Organizational Transformation and Scientific Change: The Impact of Institutional Restructuring on Universities and Intellectual Innovation

ORGANIZATIONAL

TRANSFORMATION AND

SCIENTIFIC CHANGE THE

IMPACT OF INSTITUTIONAL

RESTRUCTURING ON

UNIVERSITIES AND

INTELLECTUAL INNOVATION

RESEARCH IN THE SOCIOLOGY OF

ORGANIZATIONS

Series Editor Michael Lounsbury

Recent Volumes

Volume 25 The Sociology of Entrepreneurship

Volume 26 Studying Difference between Organizations Comparative

Approaches to Organizational Research

Volume 27 Institutions and Ideology

Volume 28 Stanfordrsquos Organization Theory Renaissance 19702000

Volume 29 Technology and Organization Essays in Honour of Joan Woodward

Volume 30A Markets on Trial The Economic Sociology of the US Financial

Crisis Part A

Volume 30B Markets on Trial The Economic Sociology of the US Financial

Crisis Part B

Volume 31 Categories in Markets Origins and Evolution

Volume 32 Philosophy and Organization Theory

Volume 33 Communities and Organizations

Volume 34 Rethinking Power in Organizations Institutions and Markets

Volume 35 Reinventing Hierarchy and Bureaucracy From the Bureau to

Network Organizations

Volume 36 The Garbage Can Model of Organizational Choice Looking

Forward at Forty

Volume 37 Managing lsquoHuman Resourcesrsquo by Exploiting and Exploring

Peoplersquos Potentials

Volume 38 Configurational Theory and Methods in Organizational Research

Volume 39A Institutional Logics in Action Part A

Volume 39B Institutional Logics in Action Part B

Volume 40 Contemporary Perspectives on Organizational Social Networks

Volume 41 Religion and Organization Theory

RESEARCH IN THE SOCIOLOGY OF ORGANIZATIONS

VOLUME 42

ORGANIZATIONALTRANSFORMATION ANDSCIENTIFIC CHANGE

THE IMPACT OFINSTITUTIONAL

RESTRUCTURING ONUNIVERSITIES ANDINTELLECTUALINNOVATION

EDITED BY

RICHARD WHITLEYUniversity of Manchester United Kingdom

JOCHEN GLASERTU Berlin Germany

United Kingdom North America Japan

India Malaysia China

Emerald Group Publishing Limited

Howard House Wagon Lane Bingley BD16 1WA UK

First edition 2014

Copyright r 2014 Emerald Group Publishing Limited

Reprints and permission service

Contact permissionsemeraldinsightcom

No part of this book may be reproduced stored in a retrieval system transmitted in

any form or by any means electronic mechanical photocopying recording or

otherwise without either the prior written permission of the publisher or a licence

permitting restricted copying issued in the UK by The Copyright Licensing Agency

and in the USA by The Copyright Clearance Center Any opinions expressed in the

chapters are those of the authors Whilst Emerald makes every effort to ensure the

quality and accuracy of its content Emerald makes no representation implied or

otherwise as to the chaptersrsquo suitability and application and disclaims any warranties

express or implied to their use

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN 978-1-78350-684-2

ISSN 0733-558X (Series)

Certificate Number 1985ISO 14001

ISOQAR certified Management Systemawarded to Emerald for adherence to Environmental standard ISO 140012004

CONTENTS

LIST OF CONTRIBUTORS ix

ADVISORY BOARD xi

EDITORSrsquo INTRODUCTIONRichard Whitley and Jochen Glaser 1

INSTITUTIONAL CHANGE AND THE TRANSFORMATIONOF UNIVERSITIES AS STRATEGIC ACTORS

THE IMPACT OF INSTITUTIONAL REFORMS ON THENATURE OF UNIVERSITIES AS ORGANISATIONS

Richard Whitley and Jochen Glaser 19

EMPOWERMENT OF FRENCH UNIVERSITIES BYFUNDING AND EVALUATION AGENCIES

Christine Musselin 51

FUNDING INDIVIDUALS CHANGINGORGANISATIONS THE IMPACT OF THE ERC ONUNIVERSITIES

Jakob Edler Daniela Frischer Michaela Glanz andMichael Stampfer

WHERE HAVE ALL THE SCIENTISTS GONEBUILDING RESEARCH PROFILES AT DUTCHUNIVERSITIES AND ITS CONSEQUENCES FORRESEARCH

Grit Laudel and Elke Weyer 111

MERGER MANIA IN SCIENCE ORGANIZATIONALRESTRUCTURING AND PATTERNS OFCOOPERATION IN AN ACADEMICRESEARCH CENTRE

Julien Barrier 141

CONTRADICTORY CONSEQUENCES OF INSTITUTIONALCHANGES ON INTELLECTUAL INNOVATION IN

THE PUBLIC SCIENCES

INSTITUTIONAL CONDITIONS AND CHANGINGRESEARCH PRACTICES IN SWITZERLAND

Martin Benninghoff Raphael Ramuz Adriana Gorga andDietmar Braun

COLD ATOMS HOT RESEARCH HIGH RISKSHIGH REWARDS IN FIVE DIFFERENT AUTHORITYSTRUCTURES

Grit Laudel Eric Lettkemann Raphael RamuzLinda Wedlin and Richard Woolley

HIGHLY ADAPTABLE BUT NOT INVULNERABLENECESSARY AND FACILITATING CONDITIONSFOR RESEARCH IN EVOLUTIONARYDEVELOPMENTAL BIOLOGY

Grit Laudel Martin Benninghoff Eric Lettkemann andElias Hakansson

PATH DEPENDENCE AND POLICY STEERING INTHE SOCIAL SCIENCES THE VARIED IMPACT OFINTERNATIONAL LARGE SCALE STUDENTASSESSMENT ON THE EDUCATIONAL SCIENCES INFOUR EUROPEAN COUNTRIES

Jochen Glaser Enno Aljets Adriana Gorga Tina HedmoElias Hakansson and Grit Laudel

vi CONTENTS

WHERE TO GO FOR A CHANGE THE IMPACT OFAUTHORITY STRUCTURES IN UNIVERSITIES ANDPUBLIC RESEARCH INSTITUTES ON CHANGES OFRESEARCH PRACTICES

Jochen Glaser Enno Aljets Eric Lettkemann andGrit Laudel

COMPUTER CORPUS LINGUISTICS ANINNOVATION IN THE HUMANITIES

Lars Engwall Enno Aljets Tina Hedmo andRaphael Ramuz

HOW DO INSTITUTIONAL CHANGES AFFECTSCIENTIFIC INNOVATIONS THE EFFECTS OFSHIFTS IN AUTHORITY RELATIONSHIPSPROTECTED SPACE AND FLEXIBILITY

Richard Whitley 367

viiContents

LIST OF CONTRIBUTORS

Enno Aljets Department of Sociology University ofBremen Germany

Julien Barrier Triangle - CNRS research unit EcoleNormale Superieure de Lyon France

Martin Benninghoff Institut drsquoEtudes Politiques etInternationales University of LausanneSwitzerland

Dietmar Braun Institut drsquoEtudes Politiques etInternationales University of LausanneSwitzerland

Jakob Edler Manchester Business School University ofManchester United Kingdom

Lars Engwall Department of Business Studies UppsalaUniversity Sweden

Daniela Frischer Faculty of Education McGill UniversityMontreal Canada

Michaela Glanz Art | Research | Service Academy of FineArts Vienna Austria

Jochen Glaser Center for Technology and Society TUBerlin Germany

Adriana Gorga Institut drsquoEtudes Politiques etInternationales University of LausanneSwitzerland

Tina Hedmo Department of Business Studies UppsalaUniversity Sweden

Elias Hakansson Department of Economic History UppsalaUniversity Sweden

Grit Laudel Center for Higher Education PolicyStudies University of Twente EnschedeGermany

Eric Lettkemann Department of Sociology TU BerlinGermany

Christine Musselin Sciences Po and CNRS Paris France

Raphael Ramuz Institut drsquoEtudes Politiques etInternationales University of LausanneSwitzerland

Michael Stampfer Vienna Science and Technology FundVienna Austria

Linda Wedlin Department of Business Studies UppsalaUniversity Sweden

Elke Weyer Center for Higher Education PolicyStudies University of Twente EnschedeThe Netherlands

Richard Whitley Manchester Business School University ofManchester United Kingdom

Richard Woolley Institute for Innovation and KnowledgeManagement Research PolytechnicUniversity of Valencia Spain

x LIST OF CONTRIBUTORS

ADVISORY BOARD

SERIES EDITORS

Michael LounsburyAssociate Dean of Research

Thornton A Graham ChairUniversity of Alberta School of Business and National

Institute for Nanotechnology Alberta Canada

ADVISORY BOARD MEMBERS

Howard E AldrichUniversity of North CarolinaUSA

Stephen R BarleyStanford University USA

Nicole BiggartUniversity of California atDavis USA

Elisabeth S ClemensUniversity of Chicago USA

Jeannette ColyvasNorthwestern University

Barbara CzarniawskaGoteborg University Sweden

Gerald F DavisUniversity of Michigan USA

Marie-Laure DjelicESSEC Business SchoolFrance

Frank R DobbinHarvard University USA

Royston GreenwoodUniversity of Alberta Canada

Mauro GuillenThe Wharton School University ofPennsylvania USA

Paul M HirschNorthwestern University USA

Brayden KingNorthwestern University

Renate MeyerVienna University of Economicsand Business AdministrationAustria

Mark MizruchiUniversity of Michigan USA

Walter W PowellStanford University USA

Hayagreeva RaoStanford University USA

Marc SchneibergReed College

W Richard ScottStanford University USA

Haridimos TsoukasALBA Greece

xii ADVISORY BOARD

EDITORSrsquo INTRODUCTION

Richard Whitley and Jochen Glaser

Since World War II the funding and governance of higher education andscientific research systems have undergone a number of major changes inmost industrialised capitalist societies In particular the state and otherfunding agencies have become more proactive in seeking to steer the direc-tion of academic research universities are being encouraged to be moreaccountable and strategic in their behaviour and the commercialisation ofresearch has taken new forms (Whitley Glaser amp Engwall 2010 Ziman2000) These changes represent significant shifts in the organisation of thesciences and have altered the nature of universities as strategic actors inmany countries including measures that increase their autonomy from thestate and strengthen their internal managerial governance (KruckenKosmutzky amp Torka 2007 Paradeise Reale Bleiklie amp Ferlie 2009Whitley 2012) In transforming the organisational contexts in whichresearch is carried out these reforms have had substantial consequences forthe dynamics of scientific change

These changes have been accompanied by a rapid increase in the typesand numbers of formal organisations involved in the production coordina-tion and evaluation of published scientific knowledge Research organisa-tions themselves have not only grown in numbers but are also more variedin their structures and missions (eg Merz amp Biniok 2010) and haveformed associations that organise lobbying or collaboration The wide-spread transition from recurrent to project funding for scientific research

Organizational Transformation and Scientific Change The Impact of Institutional

Restructuring on Universities and Intellectual Innovation

Research in the Sociology of Organizations Volume 42 115

Copyright r 2014 by Emerald Group Publishing Limited

All rights of reproduction in any form reserved

ISSN 0733-558Xdoi101108S0733-558X20140000042001

has increased the importance of funding agencies which channel moneyfrom the state or from industry to science (Benninghoff amp Braun 2010Braun 1998) Science policy making has additionally become based on alarge number of advisory bodies intermediary and lobby organisations(Van der Meulen amp Rip 1998) and the institutionalisation of formal eva-luations of research performance has given rise to agencies that oversee orconduct these (Whitley amp Glaser 2007) Researchers also make use ofcommercial organisations especially publishers (Thompson 2005) and pro-viders of research equipment materials and services (Kleinman 2003pp 9395) Finally professional organisations of academics have becomemore active in both organising collaboration and lobbying for the interestsof researchers (Schofer 1998 Simpson 2002)

With these developments control over two key processes crucial forknowledge production the allocation of resources and of reputations has become increasingly channelled through formal organisations so thatboth the strategic direction of research goals and changes in technical prac-tices depend more and more on organisational processes As a conse-quence access to and influence over formal organisations become crucialfor gaining authority over the selection of research goals and the use that ismade of research results as in the central role of advisors to funding agen-cies in the allocation of scarce resources for projects

These changes in funding and governance have been investigated byscholars in science studies higher education studies and governance from avariety of perspectives Analyses of changing sciencesociety relationshipshave focused on the emergence of a new social contract between scienceand society (Guston amp Keniston 1994) the usefulness of research for eco-nomic innovations (Berman 2012 Geiger amp Sa 2005) and the levelling outof funding in what has been termed lsquosteady state sciencersquo (Ziman 1994)Studies of university systems have been mostly concerned with the changesin their formal governance structures particularly the implementation ofsome new public management ideas (eg Amaral Lynn Meek amp MarheimLarsen 2003 Braun amp Merrien 1999 Krucken et al 2007 Marginson ampConsidine 2000 Paradeise et al 2009 Whitley amp Glaser 2007)

Specific attention has also been paid to authoritative agencies such aspublic and private funding councils and to new governance instrumentssuch as national systems of research evaluation (Martin amp Whitley 2010Whitley amp Glaser 2007) Additionally the roles of research councils asintermediary organisations between the state and scientists have enjoyedconsiderable attention (Braun 1998 Rip 1994 Van der Meulen 2003)Some analyses have used principal-agent theory for analysing relationships

2 RICHARD WHITLEY AND JOCHEN GLASER

between science policy funding councils and researchers (Braun amp Guston2003 Guston 1996 Van der Meulen 1998) while others have adoptedneo-institutionalist frameworks for the study of the international diffusionof governance structures of higher education (Drori Meyer Ramirez ampSchofer 2003) and for analysing the evolving lsquoactorhoodrsquo of universities(Krucken amp Meier 2006 Weingart amp Maasen 2007) or governance-oriented approaches for analysing changes in state-university relationshipsand university management (eg Schimank 2005 Whitley 2008 2012)

Although these approaches to institutional change in higher educationand the sciences have contributed important insights most studies havefocused on formal shifts in governance and the introduction of new instru-ments for state steering rather than undertaking comparative analyses ofhow such changes are affecting the nature of universities as organisationsand their behaviour as strategic actors in differently organised highereducation systems Additionally we still lack a theoretical synthesis of theprocesses through which the numerous changes in governance and organi-sation are affecting the conduct and content of research and there havebeen very few studies that have dealt empirically with the consequences ofgovernance and authority changes on the development of different kinds ofintellectual innovations in different sciences thus integrating the analysisof macro-level institutional change with more micro-analyses of shifts inresearch strategies and careers

It is also doubtful whether we can truly understand the nature of organi-sational change in public science systems without systematically taking intoaccount the specific nature of scientific work There is a paradox hereAlthough students of the organisational impact of governance changes onresearch practices know that this varies greatly between fields this varia-tion is rarely explicitly included in studies of organisational and scientificchange Achieving an understanding of the differential impact of organisa-tional transformations on scientific change across fields of research requireslsquobringing work back inrsquo the analysis (Barley amp Kunda 2001) In their pleato bring work back into organisational analysis Barley and Kunda notedtwo troublesome strategies in the conceptualisations of post bureaucraticorganisations namely conceptual inversion the contrasting of traditionalforms of organising with an allegedly new form and environmentalism accounting for new organisational forms and practices solely as responsesto changing environments (ibid pp 7779) Both strategies have in com-mon that they distance themselves from the detailed analysis of work prac-tices and rest more on the persuasive power of metaphor than on empiricalevidence (idem)

3Editorsrsquo Introduction

Both conceptual inversion and environmentalism can be identified inorganisational sociology and higher education approaches to the changinggovernance of research Overcoming these problems requires as Barleyand Kunda put it the blurring of conceptual boundaries between organisa-tional theory and other disciplines Applied to our topic it means aninterdisciplinary collaboration between organisational sociology and thesociology of science that can be understood as the sociology of scientificwork This collaboration poses an interesting conceptual challenge becausethe sociology of science would maintain that scientific research is conductedin organisations but by communities Thus bringing work back into thestudy of formal organisations in science requires re-thinking the relation-ship between two different types of social orders and the role they playin the production of scientific knowledge Current organisational sociologyrecognises the importance of communities for organisational processes(Marquis Lounsbury amp Greenwood 2011a) However the attempt tointegrate communities into organisational theory by conceptualising themas lsquoinstitutional ordersrsquo and lsquotypes of organisingrsquo (Marquis Lounsbury ampGreenwood 2011b) risks obscuring fundamental differences between theways in which these social forms operate differences that are highlightedby Knorr-Cetinarsquos (1999) description of lsquoepistemic culturesrsquo

The contributions to this volume address the interaction between organi-sational transformation and scientific change and the complex interplaybetween social forms in these processes at two different levels The firstpart of the book focuses on organisational transformation and asks howthis transformation is shaped by and shapes in turn the nature of scientificwork In their contribution on the impact of institutional reforms on thenature of universities as organisations Whitley and Glaser address aproblem that is not only implicit to much of the higher education policydiscourse but also reverberates in higher education and organisationresearch Do the higher education reforms currently under way in manycountries lead to universities resembling firms in their authority structures

The authors start from the observation that there is no such thing aslsquothersquo firm or lsquothersquo university Instead firms vary considerably in the author-ity exercised by owners and managers that they are granted by other socialactors and in their need to share this authority with employees if theywant to build distinctive organisational capabilities Universities in nationalhigher education systems also vary considerably in their external andinternal authority sharing and different types of universities need to bedistinguished in any analysis of authority structures Underlying thesedifferences however is the basic delegation of authority over the goals

and conduct of research and teaching that universities necessarily grant toacademics This delegation results from the uncertainties inherent to theproduction of published scientific knowledge and to the fact that the pri-mary social context for the formulation of goals workflow interdependen-cies and the evaluation of results is the scientific community

The complex relationships between organisations and scientific commu-nities are highlighted by Musselinrsquos analysis of the role of evaluations in thegovernance of French universities The higher education reforms in Francehave granted the university management considerable formal authorityover the internal affairs of universities This authority however is neces-sary but not sufficient for university management to influence decision-making about research and teaching activities It needed to be combinedwith a second element of higher education reforms namely evaluationsbased on peer reviews if the new authority of the university leadership wasto become effective Musselinrsquos case studies in three French universitiesdemonstrate how these evaluations boosted the authority of universitymanagers by legitimising actions intended to achieve positive evaluationsBeing based on peer reviews these evaluations are difficult for researchersto dismiss By strengthening the authority of university management theyamplify the effects of governance reforms

The study by Edler et al shows that much of the effect of organisationalchange on science results from its impact on individual members of otherorganisations rather than through inter-organisational interactions Byselectively funding the research of individuals and groups within organisa-tions a funding council (in their study the European Research CouncilERC) may simultaneously affect a research organisationrsquos reputationexternal funding and internal funding structure workforce and profile ofwork At the same time funding councils affect the plans and careers ofindividual researchers for which their current organisation might or mightnot be a resource

The responses by organisations and within them are similarly complexThey depend on the organisationrsquos interests and action capabilities lsquoIn-betweenrsquo organisations in terms of their collective coordination andresource capabilities showed the most changes in their internal strategieswhile powerful universities didnrsquot need to respond to the changes in theirenvironment produced by a single research council and powerless oneswere unable to respond Despite such strategic responses universities hadlittle control over the changes in research profiles resulting from significantgrants which was strongest in universities whose overall research perfor-mance was weak because acquiring one strong research group in such an

organisation meant significant shifts in their research profiles as well as bal-ances of resources and reputation In some cases organisations respondedto this lack of control and to the very low success rates by attempting tofilter grant applications for quality as understood by current elites poten-tially undermining the very intentions of the ERC to support transforma-tive research

Laudel and Weyer use a case study of the changing Dutch science systemto illustrate the impact of homogeneous institutional environments forresearch on scientific communities Dutch universities are not only expectedto develop distinct research profiles but to incorporate the statersquos thematicpriorities in these profiles The Dutch new public management reformshave provided universities with the capabilities to implement such or otherpriorities by importing or removing whole groups or departments Fundingorganisations face similar expectations and additionally a pressure towardsfunding lsquousefulrsquo research These combined expectations leave few niches forresearch that is basic but does not yield highly visible publications easilyAs a result collaboration patterns within universities might be destroyedand quasi-market failures emerge that make fields disappear at the nationallevel

In his reconstruction of a merger between French laboratories Barrierinvestigates responses to a recent trend in the French science systemDuring the last two decades the CNRS has been concentrating its fundingin larger units which creates a strong institutionalised expectation tobecome bigger Rationales of better visibility to the CNRS grant fundingagencies and commercial actors as well as science and regional policy actorswere invoked in conflicts over the merger of small laboratories to form alarger unit Barrier shows how organisational structures are used byscientists as a resource in their struggle to legitimise their research and tosecure funding The very same structures were expected to influence pat-terns of research collaboration and preferred strategies within organisa-tional boundaries The lack of collaboration between groups prior to themerger contributed to tensions between key actors that endangered thewhole process Although these difficulties were overcome the impact ofthe new structures on patterns of internal collaboration was only significantin a few very specific cases which depended on the content of research

The second part of the book presents results of a comparative project onthe impact of changing authority relations in the public sciences onconditions supporting the development of different kinds of scientific inno-vations The effects of governance funding and authority shifts on oppor-tunities for researchers to develop innovations in their fields were

compared across countries and sciences The countries included Germanyas a country in which governance reforms have been implemented relativelylate the Netherlands as a small country which was one of the firstEuropean states to begin introducing new public management changes andSweden and Switzerland which are also small and have introduced suchreforms to a more limited extent and later than the Netherlands

The comparison of innovations was intended to cover a wide range ofepistemic properties of fields and changes in research practices This wasachieved by selecting innovations in the physical sciences life sciencessocial sciences and humanities The development of these innovationsneeded to be sensitive to authority relations Since the main channelsthrough which authority is exercised are the allocation of resources andreputation innovations were selected whose development required signifi-cant resources posed reputational risks due to conflicting opinions in thescientific community about them or both An additional requirement wasthat each innovation had to be present in each of the four countries

Observing all criteria led to the selection of four innovations Bose-Einstein condensation (BEC) of cold atom gases was first realised experi-mentally in 1995 by two US research groups It took other research groupsmore than two years to replicate the first experiments In the late 1990s and2000s research about and with Bose-Einstein condensates became arapidly growing research field Evolutionary developmental biology (EDB)is an interdisciplinary perspective that grew out of the observation that evo-lutionary theory was insufficient to explain new experimental results in the1970s Since then evolutionary and developmental biologists have con-ducted a wide range of theoretical and comparative experimental researchintended to understand the interaction between the development of organ-isms and the evolution of species Comparative international large-scalestudent assessment (ILSA) is a politically motivated innovation in educa-tional science Although these comparative assessments originally were lit-tle more than a service to state education policymakers it had an impacton national educational science communities which were tasked with car-rying out the assessments and linked their research to this task Finallycomputerised corpus linguistics (CCL) the building and analysis of largedigital databases of words phrases and sentences became an innovationin the four countries at different points in time since the 1960s dependingon national research traditions in linguistics and language studies

The investigation of the development of four innovations in four coun-tries provided rich empirical material which is used for comparisons indifferent dimensions An important analytical tool for linking the content

of research to institutional and organisational conditions in many of thecomparisons is the concept of protected space This term refers to the per-iod of time in which scientists have discretion over the use of neededresources including their own efforts to pursue particular problems andapproaches before having to produce publishable and collectively valuedresults It incorporates authority over the choice and formulation of whichtopics to study how to do so and how to obtain and manage resourcesAt the macro-level of national science systems opportunities to developinnovations vary depending on the level or lsquoamountrsquo of protected spaceafforded to researchers and on its distribution between qualified researchersin a particular organisation or scientific community Protected space thusbecomes an important dimension in which public science systems can becompared (see Whitley in this volume)

At the micro-level the definition of protected space is the same but theanalytical emphasis is on the opportunities of researchers to build protectedspace for their research in different national science systems organisationsand scientific communities These opportunities depend on their organisa-tional positions success with funding agencies and priorities and prefer-ences of organisations and scientific communities Individual protectedspaces vary in the time horizons for which they afford discretion toresearchers and in the amount of resources over which this discretion canbe exercised (see the contributions by Laudel et al Engwall et al and thecomparison of organisations by Glaser et al in this volume see alsoGlaser Laudel amp Lettkemann 2014)

The applicability of protected space at both system and individual levelsprovides a means of integrating individual knowledge production processeswith the organisational conditions for these processes and the conditionsprovided by national science systems It also enables the analytical integra-tion of the influence of scientific communities and formal organisations onthe development of innovations

A complementary approach is to compare the costs of changing researchpractices This perspective was applied by Benninghoff et al to examinethe development of the four innovations in one country SwitzerlandPartly owing to its size the Swiss science system had not developed allfour research traditions from which these innovations emerged While itbelonged to the leading countries in EDB due to its strong biosciencesthe study of BEC ILSAs and CCL was not taken up by Swiss scientists asquickly because the leaders of the respective fields had incompatibleresearch priorities The authors found that the relative affluence of theSwiss public science system was a key factor supporting the development of

these innovations over time This condition had the most impact in thesciences because Swiss universities were able to support the above-averagecosts of the infrastructures required for these innovations and were able toimport innovations as was the case with BEC for which a leading Germanresearcher was recruited Finally the good funding of researchers at theseuniversities enabled at what the authors call partial switches that is thebeginning of new lines of research alongside traditional established onesThe comparison of the four innovations in Switzerland demonstrates thatnew public management reforms do not always restrict the diversity ofresearch if introduced in the context of sufficient funding and autonomyof universities

In their paper on the development of BEC Laudel et al trace the devel-opment of this innovation in five countries including Spain in addition tothe four countries that were part of the original comparative study For thefirst 10 years after its initial realisation BEC required an amount ofresources that significantly exceeds the average of the field The time it tookto achieve experimental success was difficult to predict and could extend toseveral times the common three-year term of a project This requirement ofhigh levels of protected space made it an excellent exemplar for the impactof changing authority relations on the opportunities for researchers todevelop this innovation In all five countries opportunities to develop inno-vations that needed considerable levels of protected spaces were limited toprofessors because only they combined permanent positions with thecontrol of the infrastructure that formed part of the experimental setup

For this to happen at all it was secondly necessary that the country hada research tradition of studying cold atoms which existed in Germany andthe Netherlands but not in the other countries A third necessary conditionwas access to grant funding which in the mid- to late 1990s required scien-tific communities to allocate large grants to Bose-Einstein experimentsalthough the majority opinion still was that realising BEC was either notpossible or after it had been achieved not interesting any moreFinally the comparison showed that funding for basic research has becomeprecarious in some countries which is why the research with Bose-Einsteincondensates is growing quite unevenly in the five countries investigated

EDB differs in interesting ways from BEC Laudel et al show thatbecause it is based on a change of approach that can be applied to a rangeof experimental research in evolutionary and developmental biology theepistemic pathways to the development of this innovation are various andthe protected space required by the transitions ranges from zero to morethan a million euros in the resource dimension and has time horizons that

are similar to those required by BEC As a consequence the governancesystems of the four countries provided different niches in which differentkinds of transitions to EDB could be accomplished Minimum require-ments at the international level were the acceptance by established scientificcommunities that certain standard experiments could not be conductedwithout the organisms studied by evolutionary developmental biologistsAt the national level minimum requirements included support for basicresearch topics that were not currently in fashion support that eroded inthe Netherlands Finally the considerable level of protected space neededfor undertaking comparative experimental research in EDB involved theability of research organisations to provide above-average investmentsin infrastructure in the form of breeding facilities for the new species ofinterest Swiss universities and German non-university research instituteswere best equipped to provide these investments In Sweden large grantsprovided a functional equivalent for a few researchers

The innovation in the social sciences comparative large-scale studentassessments exemplifies the importance of institutionalised research tradi-tions for the national development of an innovation While the researchtraditions in the natural sciences were internationalised and nationaldevelopments depended on the presence or absence of these research tradi-tions in each country the educational sciences are predominantly nationalresearch fields that have developed largely independently of each other AsGlaser et al show the impact of the international initiative to compare stu-dent achievements could take any one of four forms (a) accepted as a validactivity (b) considered to be falling behind the state of the art in thenational scientific community (c) seen as largely irrelevant or (d) used asan opportunity to establish the new field depending on the research tradi-tions and priorities in each of the four countries In all four though stateinterest guaranteed that the data for the international comparison wouldbe collected but the nature of the data made using them for educationalresearch difficult because there was little intellectual interest in internation-ally comparative studies and research traditions of educational sciencerequired longitudinal rather than cross sectional data This is why interna-tional comparative student assessments had a significant impact only inGermany where they were used by those who conducted them to establishquantitative studies as a legitimate enterprise of educational science

A different kind of comparison is employed by Glaser et al in theirstudy of how different types of research organisation provided opportu-nities for developing innovations The authors exploit the fact that some ofthese innovations were developed in both universities and state-funded

public research institutes in Germany and the Netherlands By comparingthe development of BEC EDB and ILSA in universities and publicresearch institutes the authors are able to identify five characteristic pat-terns of authority sharing that rendered research situations comparableacross types of organisations career positions and countries They showthat although the highly autonomous positions at German public researchinstitutes may provide the best local conditions for changing research prac-tices universities are essential for the global development of any innovationbecause only the university system provides the necessary epistemic andinstitutional diversity for simultaneously developing an innovation andbuilding a career with it At the same time the analysis demonstrates thesuperiority of discretionary block funding in situations in which researcherswant to respondent quickly to developments not yet accepted by the major-ity of their community

The reconstruction of an innovation in the humanities CCL by Engwallet al highlights the importance of disciplinary structures and authorityrelations in scientific communities as well as the challenges posed by aninnovation that introduces epistemic practices that deviate from some ofthe traditions of the field Engwall et al characterise language studies asweakly integrated and name language barriers approaches empiricalobjects (written vs spoken language modern vs historical language use)and aims (understanding language use vs support of education) as differ-ences that contribute to the fragmentation CCL and the concomitantapproach to the empirical study of languages were at odds with the theninfluential Chomskyan attempt to integrate language studies around theanalysis of generative grammars In spite of this friction the diversity oflanguage studies and their institutionalisation in separate departmentsacross Europe made it possible for CCL to develop in a variety of ways lar-gely unhampered by such opposition National traditions of research andcross-national contacts with those scholars who advanced CCL were thekey influences on how CCL developed Both early and more recentdevelopments also illustrate a feature of CCL that linked it to organisa-tional and national authority relations Its resource intensive nature parti-cularly for constructing large corpora created some problems becauseuniversities and national funding agencies had to arrange funding for thebuilding of corpora as expensive infrastructure a concept that is relativelynew to the humanities This problem was originally solved through grantsin Sweden through a state-funded institute in Germany and through com-bined grant and university funding in the Netherlands The difficulties ofobtaining infrastructure funding from research councils in the Netherlands

and Switzerland indicate that the changes in epistemic practices were stillnot fully recognised by authoritative agencies

In the final paper of this part Whitley integrates the comparative studiesby providing an account of how changing authority relations in the publicsciences affect the opportunities for the development of scientific innovationsHe does this in two steps In the first step he outlines the concepts of pro-tected space and flexibility of standards governing resources and reputationsas intermediate variables linking changes in authority relations to the neces-sary conditions for the development of different kinds of scientific innova-tions Considering the impact of the major changes in authority relationsoutlined at the beginning of this introduction on these two intermediate vari-ables emphasises their divergent and sometimes contradictory consequencesfor scientistsrsquo opportunities for developing innovations Owing to varyingstrengths and speed of changes in authority relations in different publicscience systems the influence of changes in authority relations varies consid-erably between differently organised public science systems This impactfurthermore varies depending on the content of the innovation and the fieldsin which they occur

In the second step he suggests how the different patterns of develop-ment of the four innovations investigated in the comparative casestudies can be understood in terms of the relationships between thestructure of authority relations and levels of protected space and flexibilitytaking account of the epistemic contexts in which they occur Variationsin the means of accessing protected space and other resources fordifferent researchers in each public science system help to explain contrast-ing innovation patterns These connections have a number of implicationsfor the consequences of more recent funding reductions and governancechanges

Taken together the contributions to this volume reposition researchorganisations and their environments in the approaches of organisationalsociology and higher education research Bringing work back in the studyof organisational transformation in public science systems amounts to twomajor changes First it requires looking beyond the organisations in whichresearch is conducted and research policies are constructed All theseorganisations are overlaid by scientific communities and their authoritystructures which compete with organisational authority structures for thecontrol of research and teaching the allocation of resources and the eva-luation of results At the same time the local shaping of research andteaching by organisations contributes to an aggregate effect of publicscience systems on the research conducted by scientific communities

The interaction between organisations and communities constrains theauthority over research exercised in both social contexts

Second scientific work itself has properties that modify the impact ofinstitutional environments and organisational structures Research and byextension much teaching are inherently uncertain with regard to the nat-ure of problems formulated workflow interdependencies and the meaningsof their outcomes Beyond these global properties that limit the level oforganisational control of work beyond operating groups these processeshave specific epistemic properties that make their susceptibility to organi-sational control and governance specific to fields and even to individualresearch processes or innovations Bringing work back in the analysis ofscientific change is then necessary for ascertaining how specific institu-tional and organisational conditions contribute to specific trajectories ofscientific change

ACKNOWLEDGEMENTS

We would like to thank Steven Casper Ben Martin Renate MayntzCatherine Paradeise Peter Wagner and Peter Weingart who helped bring-ing this volume to fruition by reviewing the papers We are particularlyindebted to Renate Mayntz who participated in our author workshop andhelped to significantly improve several papers We would like to acknowl-edge the help of David Bree Martin Durrell Uwe Schimank JurgenEnders Mary McGee Wood Sarah Moore and Mike Lounsbury

REFERENCES

Amaral A Lynn Meek V amp Marheim Larsen I (Eds) (2003) The higher education man-

agerial revolution Dordrecht Kluwer

Barley S R amp Kunda G (2001) Bringing work back in Organization Science 12(1) 7695

Benninghoff M amp Braun D (2010) Research funding authority relations and scientific pro-

duction in Switzerland In R Whitley J Glaser amp L Engwall (Eds) Reconfiguring

knowledge production Changing authority relationships in the sciences and their conse-

quences for intellectual innovation (pp 81110) Oxford Oxford University Press

Berman E P (2011) Creating the market university How academic science became an eco-

nomic engine Princeton Princeton University Press

Braun D (1998) The role of funding agencies in the cognitive development of science

Research Policy 27(8) 807821

Braun D amp Guston D H (2003) Principal-agent theory and research policy An introduc-

tion Science and Public Policy 30 302308

Braun D amp Merrien F-X (Eds) (1999) Towards a new model of governance for universities

A comparative view London Jessica Kingsley Publishers

Drori G S Meyer J W Ramirez F O amp Schofer E (2003) Science in the modern world

polity Institutionalization and globalization Stanford Stanford University Press

Geiger R Meyer J W Ramirez F O amp Schofer E (2005) Beyond technology transfer

US state policies to harness university research for economic development Minerva 43

Glaser J Laudel G amp Lettkemann E (2014) Hidden in plain sight The impact of

generic governance on the emergence of research fields Forthcoming In M Merz

P Sormani amp P Biniok (Eds) The local configuration of new research fields On regio-

nal and national diversity yearbook sociology of science Dordrecht Springer

Guston D H (1996) Principal-agent theory and the structure of science policy Science and

Public Policy 23 229240

Guston D H amp Keniston K (Eds) (1994) The fragile contract University science and the

federal government Cambridge MA The MIT Press

Kleinman D L (2003) Impure cultures University biology and the world of commerce

University biology at the millennium Madison WI University of Wisconsin Press

Knorr-Cetina K (1999) Epistemic cultures How the sciences make knowledge Cambridge

MA Harvard University Press

Krucken G Kosmutzky A amp Torka M (Eds) (2007) Towards a multiversity Universities

between global trends and national traditions Bielefeld transcript Verlag

Krucken G amp Meier F (2006) Turning the university into an organizational actor

In G S Drori J W Meyer amp H Hwang (Eds) Globalization and organization

World society and organizational change (pp 241257) Oxford Oxford University

Marginson S amp Considine M (2000) The enterprise university Power governance and rein-

vention in Australia Cambridge Cambridge University Press

Marquis C Lounsbury M amp Greenwood R (Eds) (2011a) Communities and organiza-

tions Research in the sociology of organizations Bingley Emerald Group

Marquis C Lounsbury M amp Greenwood R (2011b) Introduction Community as an

institutional order and a type of organizing In C Marquis M Lounsbury amp

R Greenwood (Eds) Communities and organizations Research in the sociology of orga-

nizations (Vol 33 pp ixxxvii) Bingley Emerald Group

Martin B amp Whitley R (2010) The UK Research Assessment Exercise A case of regulatory

capture In R Whitley J Glaser amp L Engwall (Eds) Reconfiguring knowledge produc-

tion Changing authority relationships in the sciences and their consequences for intellec-

tual innovation (pp 5180) Oxford Oxford University Press

Merz M amp Biniok P (2010) How technological platforms reconfigure science-industry

relations The case of micro- and nanotechnology Minerva 48 105124

Paradeise C Reale E Bleiklie I amp Ferlie E (Eds) (2009) University governance Western

Europe comparative perspectives Dordrecht Springer

Rip A (1994) The republic of science in the 1990s Higher Education 28 332

Schimank U (2005) lsquoNew public managementrsquo and the academic profession Reflections on

the German situation Minerva 43 361376

Schofer E (1998) Science associations in the international sphere 18751990 The realization

of science and the scientization of society In J Boli amp G M Thomas (Eds) World

polity formation since 1875 (pp 249266) Stanford Stanford University Press

Simpson I H (2002) Life course patterns of national associations International Sociology

17(2) 285303

Thompson J B (2005) Books in the digital age The transformation of academic and higher

education publishing in Britain and the United States Cambridge Polity Press

Van der Meulen B (1998) Science policies as principal-agent games Institutionalization and

path dependency in the relation between government and science Research Policy 27

397414

Van der Meulen B (2003) New roles and strategies of a research council Intermediation of

the principal-agent relationship Science and Public Policy 30 323336

Van der Meulen B amp Rip A (1998) Mediation in the Dutch science system Research

Policy 27(8) 757769

Weingart P amp Maasen S (2007) Elite through rankings-The emergence of the enterprising

university In R Whitley amp J Glaser (Eds) The changing governance of the sciences

The advent of research evaluation systems Sociology of the sciences yearbook (Vol 26

pp 75100) Dordrecht Springer

Whitley R (2008) Universities and strategic actors Limitations and variations In L Engwall amp

D Weaire (Eds) The university in the market (pp 23 37) London Portland Press

Whitley R (2012) Transforming universities National conditions of their varied organisa-

tional actorhood Minerva 50 493510

Whitley R amp Glaser J (Eds) (2007) The changing governance of the sciences The advent of

research evaluation systems Sociology of the sciences yearbook (Vol 26) Dordrecht

Springer

Whitley R Glaser J amp Engwall L (Eds) (2010) Reconfiguring knowledge production

Changing authority relationships on the sciences and their consequences for intellectual

innovation Oxford Oxford University Press

Ziman J (1994) Prometheus bound Science in a dynamic steady state Cambridge Cambridge

University Press

Ziman J (2000) Real science What it is and what it means Cambridge Cambridge University

INSTITUTIONAL CHANGE AND

THE TRANSFORMATION

OF UNIVERSITIES AS

STRATEGIC ACTORS

THE IMPACT OF INSTITUTIONAL

REFORMS ON THE NATURE OF

UNIVERSITIES AS

ORGANISATIONS

ABSTRACT

Recent reforms to higher education systems in many OECD countrieshave focused on making universities more effective organisations in com-peting for resources and reputations This has often involved increasingtheir internal cohesion and external autonomy from the state to makethem more similar to private companies However pre-reform universi-ties differed so greatly in their governance and capabilities that theimpact of institutional changes has varied considerably between threeideal types Hollow State-chartered and Autarkic Furthermore thecombination of (a) the inherent uncertainty of scientific research under-taken for publication (b) limited managerial control over work pro-cesses and reputations and (c) the contradictory effects of some fundingand governance changes has greatly restricted the ability of universities

ISSN 0733-558Xdoi101108S0733-558X20140000042000

to function as authoritatively integrated organisations capable of devel-oping distinctive competitive competences

Keywords Institutional reforms universities ideal types hollowstate-chartered autarkic uncertainty managerial ignorance organi-sational actorhood

INTRODUCTION

Over the past three decades or so many OECD countries have introducedsubstantial reforms to the governance of research universities (ParadeiseReale Bleiklie amp Ferlie 2009) These changes usually involved the devel-opment of new relationships with the state and other extra-mural agenciesas well as shifts in authority relations within academic institutions some-times as part of more general reforms of statesrsquo administrative structuresunder the generic title of lsquoNew Public Managementrsquo or NPM (de BoerEnders amp Leistye 2007 Ferlie Musselin amp Andresani 2009 Schimank2005) Against a common background of reduced public funding of aca-demic research and teaching or at least a substantial reduction in thegrowth rates of such support states have increasingly sought to recon-struct universities as quasi-independent collective agents on whom they canrely to realise national education and research goals Often drawing onorganisational models from the private corporate sector much of therhetoric behind these changes has suggested that they should become moresimilar to companies competing in product capital and labour marketsand lsquoentrepreneurialrsquo (Clark 1998 Marginson amp Considine 2000) Indeedthe corporate model as understood in the United States during much of thefirst half of the 20th century was widely imitated as a formal structure bymany colleges and universities in that country and is being urged upon EUmember countries in the current century (Slaughter amp Cantwell 2012Thelin 2004)

In much of continental Europe and Japan where universities have beenorganised around separate discipline-based institutes that combineresearch and teaching under the authority of full professors this involvestransforming them from arenas in which the heads of these institutes com-peted for control over resources into more internally cohesive and exter-nally autonomous organisations (de Boer et al 2007 Musselin 2007)From being rather fragmented organisations in which their administrators

had limited responsibility for or control over the research and teachingactivities their members carried out universities in these countries areincreasingly supposed to have been transformed into organisationally sepa-rate collectivities capable of developing and implementing independentstrategies in pseudo-markets

In more departmentalised academic systems mostly in Anglophonesocieties such as the United Kingdom Australia and New Zealand similarreforms have tended to combine substantial reductions in the level of publicsupport with an intensification of competition between universities forresources and reputations as well as governance shifts that enhanced theauthority of central university managers and boards and the institutionali-sation of a wide array of quality assurance and performance monitoringsystems (Dobbins Knill amp Vogtle 2011 Marginson amp Considine 2000Martin amp Whitley 2010) Many of these changes have been intended totransform predominantly collegiate organisations into more hierarchicallymanaged ones whose senior staff could be held accountable for their per-formance in a comparable way to private companies being disciplined bycompetitive markets

As a number of national case studies have shown however few if anyof such attempts have resulted in universities developing organisationallydistinctive knowledge and capabilities that might enable them to functioneffectively as strategic actors in a comparable manner to private companies(Huisman 2009 Paradeise Reale amp Goastellec 2009) While this could beattributed to inconsistencies in the ways that these policies have been imple-mented in many states andor to time lags between institutional change andorganisational responses it also suggests that there may be some funda-mental difficulties with a corporate model being applied to universities

In particular such models tend to ignore the limited ability of adminis-trators and managers to organise and control scientific research and teach-ing activities that are highly uncertain and more influenced by externalscientific communities than by organisational hierarchies (Musselin 2007Whitley 2008) Since they are poorly qualified to judge the merits of parti-cular research goals and approaches and the results they produce man-agers tend to rely on the decisions and preferences of those conductingresearch and advanced teaching activities about which strategies to pursuewith particular knowledge and skills and how to do so They therefore areunable to coordinate and control them in organisationally specific waysthat could generate distinctive competitive advantages

It is also worth noting that in any case there is no single model of effec-tive corporate governance for universities to imitate but rather a wide

21The Impact of Institutional Reforms on the Nature of Universities

variety of corporate forms and business strategies that have dominated dif-ferent market economies and industries in different historical periods asrecently highlighted by the extensive comparative capitalisms literature(see eg Crouch 2005 Roy 1997 Schmidt 2002 Whitley 1999 2007a)Thus even if they could develop some characteristics of competitive firmsit remains unclear which kind of enterprise would be most appropriate orlikely to emerge in different circumstances

Furthermore since national higher education systems and the natureof universities in them have varied so much in the 20th century wewould expect any set of institutional reforms to have varied consequenceshowever similar they may seem initially Just as overtly similar deregula-tory reforms to financial services markets in Britain and Japan havehad quite different effects because they were introduced into differentlyorganised market economies dominated by different kinds of interestgroups (Kushida amp Shimizu 2013 Laurence 2001) so too New PublicManagement influenced reforms are likely to have varied outcomes acrossdifferent higher education systems (de Boer et al 2007 Schimank 2005)

In this article we explore these points in more detail by considering howthe most common reforms are affecting key features of universities asresearch and teaching organisations and their likely impact on the kinds oforganisations that might emerge in different societies Following a sum-mary of the key features of firms as strategic actors in competitive marketsand how these vary between institutional contexts we consider the extentto which universities in different kinds of higher education systems havediffered significantly in these respects in terms of three distinct ideal typesNext we discuss how the major changes in the governance of universitiesthat have taken place in many OECD countries seem to be affecting firsttheir strategic autonomy and second their management of research andteaching activities and consequently their establishment as coherent dis-tinct and authoritatively integrated organisations in different higher educa-tion systems

THE NATURE OF FIRMS IN MARKET ECONOMIES

Private companies are commonly seen as the critical unit of economicaction in capitalist societies because they are the key collective entitythrough which private property rights holders and their delegated agentscoordinate economic activities to generate and appropriate added value

It is the combination of private ownership of the rights to an incomestream from such activities with decentralised decision making by indivi-dual strategic actors and the authoritative integration and direction ofhuman and material resources through organisational routines and proce-dures that makes firms crucial economic agents in such political economies(Penrose 1959 Richardson 1998)

The ability to direct employees to undertake specific tasks through dele-gated authority from private property rightsrsquo holders is central to the orga-nisational development of distinctive competitive competences in capitalistsocieties not least because the flexibility provided by employment agree-ments enables managers to organise economic activities in different waysand to change these to suit altered circumstances (Richardson 1998) Thisflexibility facilitates the management of increasingly complex and uncertainactivities particularly innovation As Lazonick (1991) Lazonick andWest (1998) and others have suggested the planned coordination of a spe-cialised division of labour has enabled firms to build distinctive organi-sational capabilities for developing process and product innovations on acontinuing basis

However it is important to recognise that the powers duties and socio-economic functions of private companies vary considerably betweennational jurisdictions institutional contexts and over time (Milthaupt2003 Morck 2007 Roy 1997) In particular the extent to which legallyconstituted corporate entities combine unified authority limited liabilityinvestor ownership and easily transferred private property rights in a singleorganisational form is highly variable across market economies and issubject to substantial change reflecting the contested nature of the domi-nant corporate form and major institutional differences between stateand regions (DiMaggio 2001 Dobbin 1994 Gourevitch amp Shinn 2005Goyer 2011)

This variety of leading firmsrsquo governance direction and capabilitiesacross capitalist societies suggests that there is no single model of competi-tive companies that can be taken to represent the ideal organisational formfor modernised universities to follow Rather there are many differentkinds of dominant firm types that have become established and changed indifferent institutional environments These vary on a number of dimensionsthat can be summarised under two broad headings (Whitley 2010) Firstthose dealing with issues of ownership control and direction that distin-guish between the different kinds of groups and interests dominating strate-gic decision making and their impact on leading firmsrsquo priorities oftenreferred to as corporate governance (Aguilera amp Jackson 2010) Second

those dealing with the organisation and control of work activities in waysthat add significant value to inputs and generate organisation-specificknowledge and capabilities to provide competitive advantages (Kogut ampZander 1992 Metcalfe amp James 2000 Teece Pisano amp Shuen 2000)

Firm Strategic Autonomy and Governance

Considering first how the governance of leading firms can differ betweensocieties it is important to recognise that governance involves a wide rangeof interests and issues including lsquothe whole set of legal cultural and institu-tional arrangements that determine what publicly traded corporations cando who controls them how that control is exercised and how the risksand returns from the activities they undertake are allocatedrsquo (Blair 1995p 3) In particular any analysis of how the governance of firms varies andchanges has to consider how dominant interest groups and institutionsinfluence the construction direction and development of major companiesin different jurisdictions in such ways that their dominant goals and eco-nomic consequences vary This clearly involves issues of ownership andcontrol but it also includes the impact of various kinds of employees sup-pliers customers competitors and other business associates on firm man-agement and behaviour

While in principle all private companies in decentralised market econo-mies are separate decision making agents able to decide which human andmaterial resources they will use to compete in particular markets and howthey will do so in practice the extent of their independence from lsquoexternalrsquogroups and interests can vary considerably across differently organisedpolitical economies The strategic autonomy of those running firms fromoutside shareholders business partners and employee groups is muchgreater in societies where ownership is fragmented amongst portfolioholders and most business transactions including hiring and firing are con-ducted on an armrsquos length basis than in those where share ownership ismore concentrated capital is provided more by lsquomainrsquo banks than throughcapital markets trade associations are powerful and labour organisationsare significant

In the former case what might be termed isolated hierarchies are ableto pursue the interests of the firm as understood by its managers in rela-tive isolation from other organisations and investors subject to capitalmarket pressures to achieve particular levels of profitability and growth

This can enable them to implement radical changes in business strategythrough rapid shifts in employment and asset trading on large and liquidcapital markets as well as facilitating the growth of project-based firmsin high technology industries (Casper 2007 Whitley 2006)

In contrast in the latter kind of lsquoalliance capitalismrsquo (Gerlach 1992)dominant firms tend to be more integrated into a range of commitmentsand alliances with major shareholders sometimes as members of businessgroups business partners banks industry associations and employee groupsWhile such authority sharing with major allies can encourage long-terminvestments in training and building organisation-specific competences inparticular sectors it may also limit the speed with which firms can enter newmarkets and develop radically different technologies

Organising Work and Developing Organisational Capabilities

As many discussions have emphasised the authoritative direction and inte-gration of economic activities is a key feature of firms Hamilton andFeenstra (1995 p 56) for instance claim that firms and economic organi-sations in general are lsquoabove all authoritative organisations that structurerelationships according to established rules of conductrsquo in which partici-pants recognise that they are bound to the authoritative norms of the orga-nisation and there are effective means to enforce collective rules Howevermajor differences exist in how owners and managers use their authority tointegrate and direct economic activities and so develop distinctive kinds ofcompetitive capabilities in different firms

In particular they differ in how much different groups of employeescontribute to organisational problem solving and improving performanceas many comparisons of German Japanese and US companies have indi-cated (Aoki 2001 Cole amp Whittaker 2006 Soskice 1999) Such contribu-tions depend on both the ways in which authority over task performance isshared between management and employees and on employees being com-mitted to developing firmsrsquo organisational capabilities sometimes at theexpense of improving their own individual skills

Two important ways in which owners and managers can elicit high levelsof employee involvement in joint problem solving activities and commit-ment to the improvement of firmsrsquo collective capacity to deal with complexissues are (a) to share substantial amounts of authority with them and(b) to provide long-term organisational careers Authority sharing here

involves property rights holders and their agents delegating considerablediscretion over task performance and sometimes task organisation toskilled employees and encouraging them to contribute to product and pro-cess improvements It varies in the degree of such delegation that is theamount of discretion exercised by subordinates over task performance andproblem solving and its scope that is the range of activities and decisionsover which discretion is exercised

While these aspects are often positively correlated it is clearly possiblefor managers to delegate considerable levels of discretion over specific nar-rowly defined tasks without extending it to more general features of workorganisation and involvement in problem solving On the whole the greateris the degree and scope of such authority sharing the more firms should beable to integrate different kinds of activities and types of knowledge indealing with complex problems and to develop new routines and knowl-edge as in some Danish and Japanese companies (Aoki 1994 Koike1994 Kristensen 1996 Kristensen et al 2011)

The extent of such authority sharing and nature of employee contribu-tions to problem solving are affected by the complexity transparency andpredictability of key tasks (Mintzberg 1979 pp 348379 Simpson 1985)as reflected in many discussions of the professions whose authority overtask performance has been linked to the monopoly of access to knowledgeand lsquoindeterminationrsquo in professional knowledge (Child amp Fulk 1982pp 159162 see also Freidson 1984) In general the more indeterminateare the connections between work processes and outcomes and the moreuncertain is the value of task outcomes the more authority skilled staff canexert over how tasks are performed and organised

Long-term commitment to a firmrsquos success through contributing to itsspecific knowledge and capabilities is additionally encouraged by offeringorganisational careers for those who demonstrably make such contribu-tions on a continuing basis By tying personal futures to the growth ofthe employing organisation and making credible commitments to main-tain employment across the business cycle such careers intensifyemployee commitment to the improvement of collective competenceseven if that limits their visibility on external labour markets Where onthe other hand employment is seen as being vulnerable to market andtechnological shifts as well as to changes in ownership skilled workerswill be more concerned to improve their position on external labourmarkets by enhancing their own personal knowledge reputation andskills than on sharing knowledge and opportunities with short-termcolleagues

VARIETIES OF PRE-REFORM UNIVERSITIES AND

CHARACTERISTICS OF FIRMS

From this brief account of the major variable characteristics of firms inmarket economies it is clear that most 20th century universities differed agreat deal from the stereotypical large firm able to make strategic choicesand control work processes They additionally did not have to competewith each other for resources and customers by selling their services onopen markets and have not had to develop distinctive competitive advan-tages through the development of organisation-specific collective capabil-ities They have though varied considerably in the extent to which theyconstructed separate organisational identities and have been able to exertsome independence from the state between different national higher educa-tion systems in OECD countries (Clark 1983 Coleman 1999 Rothblatt ampWittrock 1993)

These differences can be summarised in similar terms to those used todescribe types of leading firms particularly the extent of senior managersrsquostrategic autonomy and the degree to which they are able to exert unifiedauthority over the organisation and conduct of work activities to generatedistinctive collective competences to identify distinct types of pre-reformuniversities At least three ideal types can readily be distinguished thatreflect the major contrasts between higher education systems and theirnational contexts in the 20th century before major institutional changeswere introduced Hollow State-chartered and Autarkic (Whitley 20082012) Their key characteristics are listed in Table 1 and will now be brieflydiscussed

University Strategic Autonomy and Governance

Hollow universities were typically parts of the state administrationAuthority over key decisions and responsibilities was concentrated in thehands of state officials with varying degrees of delegation of authority overappointments and promotions to national disciplinary elites and seniorprofessors in each university Both academic and non-academic staffremained state employees and university administrators had little or noability to alter national terms and conditions of employment

Universities approximating to this ideal type have been significant inmany countries in continental Europe as well as parts of East Asia in

both lsquoGermanicrsquo and lsquoNapoleonicrsquo higher education systems (Dobbinset al 2011 Regini 2011) These lacked or only were able to use to avery limited extent both of the constitutive features of private companiesas key economic actors strategic autonomy and organisational capabilities

Table 1 Governance and Capabilities of Three Ideal Types of Pre-Reform Universities in OECD Countries

Organisational

Governance and

Capabilities

Hollow State-Chartered Autarkic

Autonomy of senior managers from state ministries concerning

Internal

structures

Very limited Medium High

Programmes Very limited Medium High

Student

admission and

examinations

Student numbers Very limited Limited High

Authority over resource (re-)allocation organisation structures and programmes of

State High Medium Low

Senior

management

Low Medium High

Senior academics High Medium Medium

Authority of senior

academic staff over

personnel and

academic decisions

Medium to high

often shared with

national disciplinary

elites and state

ministries

High although senior

managers can exercise

considerable influence

when they cannot agree

Medium to high with

senior managers able to

exercise considerable

influence

Authority of senior

managers over

academic careers

Very limited Limited Medium restricted by

competition for

scientific excellence and

academicsrsquo ability to

obtain external funding

Managerial ability

to develop

distinctive

organisation-

specific capabilities

None Low Low

Most of their financial and administrative decisions were made by stateagencies As Graham and Diamond (1997 p 23) point out

hellip permanent civil service officials on campus hellip answered to the central ministry for

finances and major policies on academic programmes In exchange for full funding the

universities served the governmentrsquos definition of the public interest

Senior university managers were typically elected from amongst the pro-fessors for limited periods or appointed directly by the state and expectedto resume their academic careers after completing their turn in office Theyhad little or no strategic independence and very limited freedom to shiftresources between activities subject areas and services let alone to estab-lish new areas of research and teaching or to close existing ones

The dominant organisational unit for research and teaching activities inthese kinds of university was the Chair-based Institute that combined intel-lectual leadership with administrative control of resources and careersunder the direct management of the professor Collaboration and coordina-tion of these activities depended on these professors being willing to worktogether rather than the exercise of organisational authority and so univer-sities were more a relatively amorphous aggregation of discipline-basedinstitutes than coherent integrated organisations

State-chartered universities in contrast had more distinct organisationalidentities separate from state ministries and employed both academic andnon-academic staff on terms that they were able to influence more directlyThe powers and responsibilities of senior managers as collective agentswere enshrined in formal charters from the state that granted them theright to select students decide the content and organisation of teachingprogrammes and examinations and award degrees Usually though thestate retained some control over student numbers through its funding ofplaces and was often the main supporter of academic research bothdirectly and through research foundations

In many cases it was also involved in the establishment and closure ofsubject departments In the United Kingdom for example the UniversityGrants Committee which was the main channel through which state sup-port was allocated to individual universities before the 1980s periodicallycarried out subject reviews that could result in staff and facilities beingmoved between universities (Martin amp Whitley 2010) and the state as themain employer of graduates usually had to agree to the creation of newmedical schools as well as frequently intervening in the provision of teachertraining courses

In many of these kinds of university their charters granted considerableauthority over the provision of degree programmes promoting of staff and

restructuring of departments to academic Senates and their committeessomewhat restricting the ability of senior managers to act independently oftheir senior colleagues Here the role of governing councils and boardswhich usually had a majority of non-academic members in determiningstrategic priorities was limited and vice-chancellors functioned less likechief executives and more like the primus inter pares model of chairmen ina predominantly collegiate culture

While not as hollow as most in continental Europe then state-chartereduniversities and their managers have been quite constrained in their strate-gic choices by state regulations and financial controls on the one handand by academic senates national union agreements and scientific elites onthe other hand Compared to most large companies their strategic auton-omy was limited particularly in countries where central state agencies pro-vided most of their income

In contrast autarkic universities have been much more autonomousfrom state agencies as independent organisations governed by boards oftrustees or university councils without requiring state approval to awardqualifications In contrast to both hollow and state-chartered universitiesautarkic universities have lsquofour essential freedomsrsquo that were enunciated bySupreme Court Justice Felix Frankfurter in a 1957 case (Thelin 2004p 343) lsquothe freedom to determine who may teach what may be taughthow it should be taught and who may be admitted to studyrsquo Howeverthey often depend on accreditation agencies to legitimate their degrees andof course had to attract fee-paying students and private donations to fundtheir activities

In the case of private US universities and more recently many publicones boards of trustees exercised considerable influence over strategicchoices especially the selection of the president who (usually) had muchgreater power over academic staff and the future of the university than hiscounterpart in Europe (Geiger 1986) As Trow (1993 p 292) puts it

In the case of the US hellip almost without exception our colleges and universities have

been created by a group of laymen who selected a president to actually direct and man-

age the day-to-day life of the institution And this founding body and its successorshellip

together with their chosen agent (the president) have occupied the leading positions of

authority and power in American colleges and universities and still do

The lsquocorporate modelrsquo of university governance in which trustees typi-cally saw themselves functioning as the board of directors delegating execu-tive powers to the president in an analogous fashion to the growing UScorporation in the first half of the 20th century became widespread in the

inter war period which is perhaps not too surprising given that they tendedto be senior executives themselves (Slaughter amp Cantwell 2012 Thelin2004 p 238)

Turning next to consider how pre-reform universities differed in their inter-nal management of work activities and development of organisation-specific competences it is worth emphasising the limited extent to whichmanagers and trustees in any country have been able to construct competi-tive advantages in research and teaching through their coordination andcontrol of academic work This is for two major reasons First scientificresearch undertaken to contribute to collective intellectual goals throughpublication of results is inherently highly uncertain and the significance ofresearch results is often contested Researchers themselves often do notknow how their work succeeds or fails and typically are unsure what out-comes will eventuate in any precise sense Indeed in many sciences specify-ing the nature of the problem and potential ways of dealing with it areweakly codified activities that cannot readily be reduced to routine problemsolving procedures and cannot easily be assessed as worthwhile or compe-tent by outsiders Equally the meaning and significance of research resultsare often difficult to decide by specialists let alone by non-professionalsand can be subject to later reinterpretation and re-evaluation by differentgroups as many studies of scientific controversies have shown (see egHallam 1973 Pinch 1986)

Second research universities share authority over research goals andperformance evaluation with scientific communities In competing forscientific prestige and resources based on success in contributing to scienti-fic objectives they typically defer to the authority of national and interna-tional scientific elites in determining research priorities and the standardsby which academics are judged (Glaser 2010) While the extent of suchinternational reputational control of research goals and evaluation stan-dards varies between fields and over time (Whitley 2000) universities com-peting for scientific renown on the basis of their employeesrsquo intellectualcontributions have to accept the collective judgements of competentresearchers throughout the world concerning the nature of important pro-blems and the significance of results

For these reasons the scope and degree of authority sharing differ fromthose in most firms including those whose key tasks have the closest

resemblance to public research organisations namely professional servicefirms private RampD organisations and project-based firms The scope ofauthority sharing is wider because managers have no choice but to shareauthority over goal formulation workflow interdependencies and utilisa-tion of results with employees and their scientific communities Thedegree of authority sharing is also higher because in many cases very fewemployees within the organisation have a monopoly of the necessaryknowledge and skills for defining tasks and developing approaches to sol-ving them Consequently the ability of managers to plan projects and orga-nise the division of scientific labour to achieve organisational goals in acomparable manner to RampD managers in private companies is extremelylimited In this respect universities are qualitatively distinct from most pri-vate companies and indeed many formal organisations

The authority of scientific communities over research goals and perfor-mance additionally restricts universitiesrsquo ability to restructure researchteams and generate new kinds of skills by combining staff in novel waysSince disciplinary elites and gatekeepers collectively determine the nature ofcompetent contributions and research skills universities cannotreconfigure project groups around new kinds of problems in isolation fromsuch actors if they wish to compete for scientific reputations and prestige

As a consequence the high levels of managerial authority sharing withacademics in research and teaching common in universities are largelyunintended and do not necessarily greatly enhance managersrsquo abilities tointegrate different kinds of activities and types of knowledge in a similarmanner to that in many firms On the contrary the nature of researchundertaken for publication as the central organisational task leads to diver-gent and incompatible capabilities between different fields rather than theirintegration into organisation-specific competences The uncertainties inher-ent to such scientific research means that authority over the specification ofgoals is delegated to employees who in turn formulate them with referenceto different scientific communities Quite unlike most firms then managersin universities share authority over their core processes with a great numberof divergent and highly diffuse social contexts

For the same reason long-term organisational careers may not createmuch additional employee commitment to developing organisation-specificknowledge and capabilities in each university Since promotion and salaryincreases often depend on mobility between universities or at least onbeing offered posts elsewhere loyalty to individual organisations tends tobe much less than to onersquos institute and its professor as well as to the scien-tific community at large Thus few if any academics are likely to spend

much energy on contributing to the development of university-specific cap-abilities particularly if that is at the expense of their making significantcontributions to intellectual goals

Autarkic universities have had greater powers to alter the range and nat-ure of research and teaching programmes in response to changes in demandand resource provision than did hollow and state-chartered ones often byattracting new staff and investing in new facilities They have also beenable to encourage increased involvement with private companies and thepursuit of research commercialisation through changing incentives andrestructuring employment conditions as Terman did at Stanford (Adams2009) However their opportunities too remain limited While their man-agers can in principle invest in or abandon broadly and somewhat impre-cisely defined fields of research and teaching they have to yield authorityover exactly what is done and how it is done in these fields to the academicsthey recruit

It is worth noting here that state-chartered and autarkic universities havebeen able in principle to employ a wide range of non-academic staff provid-ing services to students acting on behalf of the university administration infund raising public relations activities and other functions under theauthority of senior managers These roles could form the basis for develop-ing distinctive organisation-specific competences that enable particular uni-versities to attract large numbers of students and external resources Themore a competitive market for such inputs develops separately from thestate the more important such competences are likely to become

The limited ability of strategic managers of universities to coordinateand control research and teaching activities highlights the contingent con-nection between high levels of organisational autonomy and managerialcontrol of work activities While variations in these two aspects of organi-sational actorhood are often correlated in private companies though by nomeans always as the examples of professional service firms and manyproject-based firms illustrate (Whitley 2006) this is usually not the case inuniversities

THE IMPACT OF INSTITUTIONAL REFORMS ON

UNIVERSITY AUTONOMY

Turning next to consider the varied ways in which recent funding and gov-ernance reforms are likely to affect different kinds of universities as

organisations and potential strategic actors we now discuss how the keychanges that are taking place in many though not all OECD statesseem to be influencing universitiesrsquo strategic autonomy Essentially thereforms combine moves to make universities more organisationally distinctand well bounded as competing organisations on the one hand withincreasingly direct attempts to steer research and teaching activitiestowards public policy objectives and intensify the monitoring of academicperformance in terms of these sometimes contradictory goals on theother hand

While the more autarkic universities in the 20th century United Stateshave not been so subject to intentional institutional reforms they havenevertheless been changed by developments in their institutional environ-ments The growth of the federal research university in the United Statesand increasing involvement of previously teaching focused organisations inthe national competition for intellectual prestige and extra-mural grants inthe 1950s and 1960s began to restrict the unilateral exercise of presidentialauthority at least in the elite institutions receiving the bulk of such grants(Thelin 2004 pp 271290) From being largely subordinate staff to theexecutive hierarchy successful researchers became more influential as indi-vidual lsquogrant swingersrsquo (Greenberg 1966) who were courted by ambitiouspresidents and increasingly able to command high salaries and supportfacilities As universities became more concerned to compete for scientificreputations on the basis of their employeesrsquo contribution to knowledgethough and researchers were more able to raise project money from exter-nal sources such as state research foundations the ability of managers tocontrol academicsrsquo behaviour has declined

As well as academics becoming more significant in the decisions ofresearch universities in the United States the growing role of national gov-ernment support in helping students gain access to higher education andincreasing state regulation of universitiesrsquo personnel and other policies havebegun to affect their autonomy While then university managers andBoards of Trustees or Regents in the United States remain remarkablyautonomous in their ability to make strategic choices and set priorities forlsquotheirrsquo organisations when compared to those in most OECD countriesthey have become more circumscribed by the interests and influence of pro-fessors and federal politicalbureaucratic organisations as well as by thevariety of accrediting agencies

In the context of institutional reforms in many other OECD countriesmoves to increase the lsquoagencyrsquo of universities included reforms to fundingpractices such as a widespread shift from allocating public funds to

specific activities and roles through line budgeting to providing them asblock grants transferring academic and non-academic staff from civil ser-vice status to university employment and enhancing the authority of uni-versity senior managers and supervisory boards at the expense of theprofessoriate Attempts to steer research more directly towards publicpolicy goals have been based on the establishment and strengthening ofresearch evaluation systems (Whitley amp Glaser 2007) the formalisationof various contracts between governments and universities (Dobbinset al 2011 Meier amp Schimank 2010) the steering of public researchcouncils towards socio-economic purposes and the encouragement ofresearch commercialisation

Probably the most widespread shift in the environment of universitiesacross the OECD world since the 1970s has been the decline in the level ofpublic funding of university research and teaching per student relative tothat enjoyed in the postwar decades While this may not have directlyaffected universitiesrsquo strategic autonomy and capabilities in all countries ithas been a major feature of changing state-science relationships in manyand created the opportunity for politicians and bureaucrats to undertakesignificant restructuring of higher education systems as well of course asencouraging managers to seek alternative sources of income more activelythan before (Ballarino amp Perotti 2011)

Where such reductions have been substantial they can be expected toreinforce the consequences of associated changes in universitiesrsquo governanceand organisation particularly when these include the transfer of staff frompublic employment by the state to being employees of individual universitiesAs they become formally responsible for personnel decisions and meetingthe wage bill universities under financial pressure have to act as collectiveagencies and take a more active role than when they were primarily adminis-trative agents of the state Where on the other hand cuts in financial sup-port from the state have not been so marked the impact of many NPMreforms has not been so significant as perhaps is the case in Switzerland asdiscussed by Benninghoff et al in their article in this volume

One of the major changes associated with reductions in public financialsupport in many countries has been an increase in competitive relationshipsbetween universities as employment organisations whether for lsquoexcellencersquo(Weingart amp Maasen 2007) or for contributing to state public policy goalsand their concomitant development of separate collective identities as com-peting quasi-corporate entities While this shift is especially marked insocieties where universities were predominantly hollow organisations aspublic administrative agencies under direct ministerial control the

intensification of competition for resources and reputations has alsobecome a significant factor enhancing organisational identities in countrieswhere universities were already separately chartered organisations thatdirectly employed academic staff

Devolving responsibility to cope with resource scarcity and initiatingcompetitive relationships between universities do of course presume thatuniversities are separate organisations able to control some resources them-selves and make their own decisions about what activities to undertake andhow to do so While the extent of such delegation of resource control and(re-)allocation has often been rather less in practice than some governmentofficials have claimed (see eg Wright amp Williams Oslashrberg 2009) manystates in Continental Europe and Japan have granted university managersgreater discretion over resource utilisation by providing public funding as ablock grant

In theory this could mean that managers are able to shift resourcesbetween programmes and fields as strategic choices alter and circumstanceschange However as vice-chancellors in state-chartered universities havefound this is often easier said than done especially in the short term notleast because states often retain control over student numbers and theirassociated public funding as well as considerable influence over the creationand closing of departments in different fields (Estermann Nokkala ampSteinel 2011) Similarly making academics employees of universities mightalso be thought to enhance the strategic autonomy and powers of managersin hollow universities However since most of their salary costs continue tobe met by government ministries and conditions of service tend to remainquite standardised and set by state agencies managers have had little or noability to establish organisationally specific employment policies and sofar have evinced little interest in doing so in many countries

Combining such increases in managerial discretion with negotiatedlsquocontractsrsquo between ministries and universities whereby the lattersrsquo strate-gic choices are agreed with varying degrees of precision and equalityof influence for several years in advance (Dobbins et al 2011) mightbe seen as equivalent to the owners of a holding company setting targetsfor divisional managers of subsidiaries who are then granted enoughoperational autonomy to take responsibility for meeting them Howevergovernments rarely delegate full operational control of all aspects ofresearch and teaching and university managers have rarely had enoughindependence and authority to deliver agreed objectives in practice(Meier amp Schimank 2010) Furthermore most public policy goals foruniversities are not reducible to a single metric such as return on capital

employed or profitability and many have divergent if not contradictoryimplications

In most OECD countries cutbacks in the public funding of academicresearch and teaching and intensified competition between universities havebeen accompanied by increased state interest in monitoring their performanceIn some this has taken the form of lsquostrongrsquo research evaluation systems suchas the United Kingdomrsquos Research Assessment Exercise (Martin amp Whitley2010 Whitley 2007b) which directly link peer assessments of research out-puts to levels of financial support while others have established more indir-ect connections between various indicators of scientific and teaching successand funding decisions (Whitley amp Glaser 2007) While varying considerablyin the extent to which they are detailed intrusive and have significant finan-cial consequences all such evaluations presume that states are able to assessthe collective achievements of universities as distinctly bounded and mana-ged organisations against a variety of objectives and standards Anyincreased autonomy granted to universities is thus balanced by the institutio-nalisation of instruments and procedures for steering their activities towardspublic policy purposes

Such steering of academic research priorities by state agencies has ofcourse become much more feasible as the ability of universities to supportall the costs of scientific research has declined and scientists have had toobtain resources from public foundations and research councils to conducttheir research Over recent decades an increasing number of governmentshave developed formal science and technology policies that established stra-tegic goals for public agencies supporting university science and oriented at least in principle their allocation of resources In many state-chartereduniversities the combination of funding cuts strong evaluation systemsand steering of research support has probably reduced their strategicautonomy as they have to compete more intensively for more limited statesupport

As well as such dependence on public research councils increasing the fea-sibility of state steering of universitiesrsquo research goals it also increases theauthority of scientific elites or at least those chosen to advise these coun-cils on the merits of proposals over the direction of academic prioritiesInsofar as universities wish to establish distinctive research goals as strategicpriorities then they are constrained by both research funding agenciesrsquochoices and elite preferences in different disciplines

Another aspect of changing university identities and activities concernsthe increasing commercialisation of research results and capabilities espe-cially through the establishment of new firms with university support and

active researcher participation and intensified commitment to obtaininguniversity patents In principle the more universities are able to attractextra-mural funding from commercialisation activities the more their man-agers should be able to exercise strategic autonomy which is presumablywhy many of them have been active in seeking legitimacy for such develop-ments and establishing administrative units for exploiting new knowledgeproduced by their staff especially in the United States (Berman 2012)This reflects of course their search for new funding sources as state sup-port declines but also their wish to gain revenues that increased their dis-cretion over resource allocation which traditional academic consultingactivities rarely did

However the extent to which any increase in extra-mural revenues fromproject support and commercialisation efforts does in fact create substan-tially greater organisational independence for universities depends on theamounts thus raised and the terms on which they are made availableWhere project-based funding is accompanied by significant declines in statesupport for universities any competitive success is unlikely to enhance theirautonomy since such income including overhead costs are usually closelytied to specific research activities and is rarely available to university man-agers to use at their discretion Such limitation of their autonomy is evenmore restrictive if the state actively steers the allocation of public researchgrants towards specific purposes as in Australia (Glaser amp Laudel 2007)

Furthermore as block grants decline as a proportion of total universityincome and university dependence on extra-mural project grants growsmanagersrsquo influence over research priorities and ability to support particu-lar programmes of work declines compared to that exercised by individualresearch teams and funding councils The more important scientific reputa-tions become for the perceived success of universities and hence for thepublic performance of their managers and the more they are tied to theprovision of external research support that is largely controlled by principalinvestigators the less authority university leaders are able to exerciseauthority over academicsrsquo choices and how they contribute to universitysuccess

Similar points apply to research commercialisation revenues Most uni-versity services intended to increase these depend on the goodwill and com-mitment of researchers to notify them of possible opportunities andsupport their efforts Typically any resultant income streams have had tobe shared with research teams and their parent departments so that theamounts of discretionary resources accruing to the university as a wholehave tended to be less than is often expected (Colyvas amp Powell 2006

Krimsky 1999 Stuart amp Ding 2006) As long as universities are concernedto compete for scientific reputations based on the contributions of theiremployees to collective intellectual goals as determined by scientific elitesrather than by private interests or university managers it is difficult to seehow universities as employers could monopolise the control of commercia-lisation revenues or insist on making academicsrsquo salaries partly dependenton success in such ventures

Overall then funding and governance changes in many OECD coun-tries have probably encouraged universities to establish and reinforce dis-tinctive organisational identities as competing collectivities able to exercisesome discretion over their strategic priorities and allocation of resourcesThis has been particularly noticeable where they have previously been moresimilar to the hollow organisational ideal type However the extent of suchdiscretion remains quite severely limited in many European countries bycontinued state oversight of universitiesrsquo profiles restrictions on ministriesrsquoability to provide resources to fulfil strategic choices and political restric-tions on student numbers fee levels including zero and the ways thatuniversity qualifications are connected to access to elite labour marketsFurthermore universitiesrsquo competitive strategies in the pseudo-markets cre-ated by various state initiatives have been quite strongly constrained by themechanisms established to implement public policy goals and monitor uni-versity performance especially in countries where funding has been severelyreduced

This has been particularly the case in many of the Anglophone societieswhere state-chartered universities have previously exercised greater auton-omy from the government than those in much of mainland Europe andJapan Strong research and teaching evaluation systems state steering offunding council priorities and strict targeting of student numbers and asso-ciated public support amongst other developments have probably reduceduniversity managersrsquo strategic autonomy from the state in such countries

UNIVERSITY ORGANISATIONAL CAPABILITIES

Turning next to consider how these reforms and changes to fundingarrangements are affecting the ability of universities as organisations todevelop distinctive strategic capabilities many were expected to enhancemanagersrsquo ability to develop distinctive organisational goals and to

organise research and teaching activities accordingly In particular thecombination of increased competition for resources and reputations grow-ing managerial influence over the terms and conditions governing employ-ment relations and intensified efforts by the state and other agencies toassess their collective performance and contributions to public policy objec-tives has increased the formal authority of university managers as legiti-mate controllers of resources and priorities (Estermann et al 2011 Regini2011)

In both hollow and state-chartered universities traditionally collegiatepatterns of decision making especially about academic and personnel mat-ters are being replaced by a more top-down managerial exercise of author-ity as universities have been forced to compete more overtly with eachother and become what some have seen as lsquoentrepreneurialrsquo organisations(Clark 1998 Laudel amp Weyer this volume) The role of academic senatesin approving let alone contributing to strategic choices has become con-siderably weakened in countries like Australia (Glaser amp Laudel 2007Marginson amp Considine 2000) while senior administrative posts havebecome more obviously full time and highly paid positions Demands forgreater accountability and performance monitoring on the part of the statehave encouraged many managers to institutionalise formal procedures formeasuring the success of their employees in meeting research and teachingtargets that are summarised as key performance indicators (Glaser LangeLaudel amp Schimank 2010)

Although the combination of state policies focused on making universi-ties more effective as organisations has enhanced the authority of thoserepresenting the interests of each university and formally responsible for itsactivities managersrsquo ability to organise and control research and teachingactivities in distinctive ways to create organisation-specific competencesremains quite limited Since the lsquobaseline authority sharingrsquo in publicresearch organisations reflects properties of the key performance processesrather than managerial strategies it changed far less then could be expectedfrom the growth of managersrsquo formal authority Even in Anglophone socie-ties where university presidents and vice-chancellors have been able toboost their own salaries and perquisites and establish administrative hierar-chies mimicking those of corporate bureaucracies the management anddirection of core academic tasks remain substantially in the hands andbrains of small groups of scientists owing primary allegiance to their intel-lectual communities rather than to their current employer In this respectof course universities are not that dissimilar to many professional serviceorganisations whose central work processes are designed executed and

certified by qualified professionals and coordinated on the basis of theirstandardised skills as summarised by Mintzberg (1983) in his account ofprofessional bureaucracies

However universities differ from such professional service organisationsin their commitment to create new knowledge educate students in this newknowledge and in the much greater intrinsic technical uncertainty of theircore activities The commitment to contributing intellectual innovations inthe public sciences from incremental puzzle solving to radical reshaping ofcurrent orthodoxies means that the problems to which academics applytheir professional skills the skills themselves and the results of applyingtheir professional skills to particular problems are much lesspredictable and standardised than would be expected in most professionalservice organisations or craft-based firms (Whitley 2006)

Since the choice of problems to investigate formulation of approachesand evaluation of results is primarily decided by academics and their scien-tific communities these activities require much more continuous negotia-tion and immediate adjustment with academic colleagues throughout theworld typically through formal communication media workshops andconferences and informal communication than is the case in most profes-sional services As a result not only do university managers have to sharesubstantive control over most work processes with professional employeesthey additionally surrender the coordination and integration of researchactivities and their outputs to extra-organisational communities and theirintellectual elites (Whitley 2011)

The embeddedness of research and much advanced teaching in extra-organisational contexts that provide the dominant frames of reference foracademicsrsquo work and the fluid and constantly shifting nature of theseframes of reference render some of the tools for controlling professionalwork ineffective Thus the formalisation and standardisation of skills(Freidson 1984 pp 1318) is limited here as scientists constantly reinventthem in conducting their research Supervision by colleagues is also largelyineffectual due to different specialisations in most cases and occupationalcontrol by clients (Child amp Fulk 1982 pp 167175 Freidson 1984pp 78 Simpson 1985 pp 425426) is difficult because in many casesthe only lsquoclientsrsquo are other members of the scientific communitiesPerformance measurements based on citation counts and similar indicatorscan be interpreted as an attempt to exploit scientific communities in thisparticular role

While then managers may be able to influence the selection of scientificstaff and rely on peer judgements of the merits of their work in making

promotion decisions and allocating discretionary resources once universi-ties become direct employers of academics their ability to control thechoice of work goals the allocation of professional labour time and skillsand the evaluation of task performance is inherently limited and subsidiaryto continuing peer group assessments and research activities Essentiallytheir construction of organisation-specific competences based on the collec-tive commitment of expert staff to organisational objectives and the devel-opment of organisation-specific knowledge remains highly uncertainindirect and reactive to othersrsquo judgements

The transition from recurrent to project-based funding that can beobserved in all countries has further and ambivalent effects on the authoritysharing in universities Since peer review of competing applications forresearch funding remains the dominant means of allocating increasinglyscarce resources researchers have become even more dependent on theircommunities As a result it has become more difficult for most researchersto investigate problems that are regarded by senior colleagues as technicallyinfeasible as some physicists considered the experimental realisation ofBose-Einstein condensates (Laudel et al this volume) This means thatresearchersrsquo authority over their work is reduced in favour of scientific com-munities which thereby assume even more authority vis-a-vis universitiesAt the same time researchers who acquire grants have often the authority toindependently control them which gives them an additional base of author-ity and further reduces the authority of managers who control the shrinkingrecurrent funding

This is exemplified by the impact of major research grants from theEuropean Research Council on some universities as discussed by Edleret al in their article in this volume Because these are unusually largegrants often over five years that carry substantial prestige they are muchsought after by both individuals and universities The amount durationand prestige of such a grant produce a substantial influence over universi-tiesrsquo allocation of resources including posts and potentially enabling themto affect the long-term development of departments and faculties particu-larly the balance of activities and staff in different specialisms andtechnologies

In order to meet increasing performance expectations under conditionsof limited control over research processes some university managers areattempting to construct portfolios of lsquohigh performingrsquo research teams andinstitutes They select fields of research for which they provide an interna-tionally competitive infrastructure and attempt to hire leading researchersof these fields in the hope of achieving improved results as a lsquoreturn on

investmentrsquo The reshaping of research profiles of universities is similarlyoften concerned with creating lsquocritical massrsquo which is increasingly seen as aprerequisite for high performance (Schiene amp Schimank 2007 Laudel ampWeyer this volume) The resources for these exercises are reallocated fromfields deemed less promising in terms of research performance which oftenatrophy as a result as in the case of evolutionary developmental biology inthe Netherlands (Laudel et al this volume) However as the contributionby Laudel and Weyer to this volume demonstrates neither the effects ofinvestments nor the consequences of closing departments are easy to predictand plan and managersrsquo lack of detailed knowledge of the wide range ofsciences covered by most leading universities means that academic sciencecan end up being driven by short-term fashions and resource availability

CONCLUDING REMARKS

This discussion has highlighted a number of points about the impact ofrecent institutional reforms on the nature and characteristics of researchuniversities that are perhaps worth summarising in conclusion Firstdespite the popularity of the corporate model as an appropriate governancestructure for universities in the United States and more recently elsewherethere remain substantial barriers to universities becoming strategic actorscompeting in decentralised markets in a comparable manner to privatecompanies There are good reasons to believe that even with continuinghigher education reforms neither their governance and strategic autonomynor their ability to organise work and develop strategic capabilities willachieve the same levels as those common in private companies

Second although many OECD governments have delegated some opera-tional autonomy to universities and taken some steps to increase thepowers and accountability of senior managers relative to institute headsand professors this has rarely amounted to granting them substantial self-steering capabilities particularly where they were similar to hollow organi-sations (Estermann et al 2011) As long as the state remains the primaryprovider of funds for education and research in universities whetherdirectly through supporting undergraduate teaching or indirectly throughstate-backed loans and project support governments are unlikely to giveup responsibility for and influence on the operation of the higher educa-tion system the financing of students and how academic qualificationsaffect entry to elite labour markets Indeed the more it becomes widely

accepted that universities are core institutions of the so-called knowledgeeconomy the less likely states are to grant them high levels of strategicautonomy

Rather it seems more probable that third as universities do becomemore formally separate employment organisations with their own manage-rial hierarchies and governance arrangements states will increasinglyenmesh them in elaborate multifaceted performance monitoring regimesintended to ensure that they continue to contribute to public policy goalswhile at the same time competing intensively for resources and reputationsAs is happening in some Anglophone jurisdictions reduced public supportfor higher education has enabled governments to steer many state-chartered universities more towards policy priorities reducing both thediversity of their strategic choices and their ability to follow distinctiveorganisationally specific patterns of development While managers maysuperficially appear to have gained greater independence from state minis-tries tight government organisation of the rules of the competitive gamefor resources can so constrain their options as to restrict severely their stra-tegic actorhood

Fourth the externally guided nature and fluidity of research goals inher-ent uncertainty of work processes and contestable assessments of the valueof results in these organisations make it difficult if not impossible for man-agers to direct and control core activities effectively Not only do intellec-tual communities exert more authority over research and much teachingthan do employment organisations they also organise the coordination andthe integration of results with certified knowledge on a continuing basis

This also means that granting universities and their managers moreautonomy from state control on the presumption that they will be able tomeet demanding performance targets greatly overestimates their powersand capabilities Even where states actually do delegate substantial opera-tional autonomy to university managers and enhance their internal powersrelative to senior professors this in no way leads inevitably to their effec-tive control of research and teaching activities such they can realistically beheld responsible for each universitiesrsquo success or failure in achieving agreedobjectives Strategic independence need not and in universities usuallydoes not imply effective managerial determination of who carries outwhich activities in what ways that collectively contribute to overall organi-sational purposes

Thus while higher education reforms increase universitiesrsquo strategicautonomy and managerial control of university structures and resourcesuniversitiesrsquo ability to compete as cohesive organisations producing

distinctive contributions to knowledge pseudo-markets is severely limitedUniversity managers are unable to exercise formal authority over theiremployeesrsquo activities on the basis of delegated powers from owners in asimilar manner to most formally constituted private companies in marketeconomies Furthermore insofar as such reforms do actually promote stra-tegic autonomy but are unable to generate organisation-specific strategiccapabilities they appear to initiate and promote a bifurcation of the twoconstitutive elements of organisational actorhood

REFERENCES

Adam S B (2009) Follow the money Engineering at Stanford and UC Berkeley during the

rise of Silicon Valley Minerva 47 367390

Aguilera R V amp Jackson G (2010) Comparative and international corporate governance

The Academy of Management Annals 4(1) 485556

Aoki M (1994) The Japanese as a system of attributes A survey and research agenda

In M Aoki amp R P Dore (Eds) The Japanese firm the sources of competitive strength

(pp 1140) Oxford Oxford University Press

Aoki M (2001) Toward a comparative institutional analysis Cambridge MA MIT Press

Ballarino G amp Perotti L (2011) Italy Gradual changes and an uncertain autonomy

In M Regini (Ed) European universities and the challenge of the market (pp 168182)

Cheltenham Edward Elgar

Blair M (1995) Ownership and control Washington DC Brookings Institution

Casper S (2007) Creating silicon valley in Europe Oxford Oxford University Press

Child J amp Fulk J (1982) Maintenance of occupational control The case of professions

Work and Occupations 9 155192

Clark B (1983) The higher education system Academic organization in cross-national perspec-

tive Berkeley CA University of California Press

Clark B (1998) Creating entrepreneurial universities Organizational pathways of transforma-

tion Oxford Pergamon Press

Cole R amp Whittaker H (2006) Introduction In H Whittaker amp R Cole (Eds) Recovering

from success Innovation and technology management in Japan (pp 128) Oxford

Oxford University Press

Coleman S (1999) Japanese science View from the inside London Routledge

Colyvas J amp Powell W (2006) Roads to institutionalization The remaking of boundaries

between public and private science Research in Organizational Behavior 27 305353

Crouch C (2005) Capitalist diversity and change Recombinant governance and institutional

entrepreneurs Oxford Oxford University Press

de Boer H Enders J amp Leisyte L (2007) Public sector reform in Dutch higher education

The organizational transformation of the university Public Administration 85 2746

DiMaggio P (Eds) (2001) The twenty-first-century firm Princeton NJ Princeton University

Dobbin F (1994) Forging industrial policy The United States Britain and France in the rail-

way age Cambridge Cambridge University Press

Dobbins M Knill C amp Vogtle E A (2011) An analytical framework for the cross-country

comparison of higher education governance Higher Education 62 665683

Estermann T Nokkala T amp Steinel M (2011) University autonomy in Europe II The score-

card Brussels European University Association

Ferlie E Musselin C amp Andresani G (2009) The lsquosteeringrsquo of higher education systems A

public management perspective In C Paradeise E Reale I Bleiklie amp E Ferlie

(Eds) University governance Western European comparative perspectives (pp 120)

Dordrecht Springer

Freidson E (1984) The changing nature of professional control Annual Review of Sociology

10 120

Geiger R (1986) To advance knowledge The growth of American research universities

19001940 Oxford Oxford University Press

Gerlach M (1992) Alliance capitalism Berkeley CA University of California Press

Glaser J (2010) From governance to authority relations In R Whitley J Glaser amp

L Engwall (Eds) Reconfiguring knowledge production Changing authority relationships

in the sciences and their consequences for intellectual innovation (pp 357370) Oxford

Glaser J Lange S Laudel G amp Schimank U (2010) The limits of universality How field-

specific epistemic conditions affect authority relations and their consequences

In R Whitley J Glaser amp L Engwall (Eds) Reconfiguring knowledge production

Changing authority relationships in the sciences and their consequences for intellectual

innovation (pp 291324) Oxford Oxford University Press

Glaser J amp Laudel G (2007) Evaluation without evaluators The impact of funding formu-

lae on Australian university research In R Whitley amp J Glaser (Eds) The changing

governance of the sciences The advent of research evaluation systems Sociology of the

sciences yearbook (Vol 26 pp 127152) Dordrecht Springer

Gourevitch P amp Shinn J (2005) Political power and corporate control The new global poli-

tics of corporate governance Princeton NJ Princeton University Press

Goyer M (2011) Contingent capital Short-term investors and the evolution of corporate gov-

ernance in France and Germany Oxford Oxford University Press

Graham H D amp Diamond N (1997) The rise of the American research universities

Baltimore MD John Hopkins University Press

Greenberg D (1966) Grant swinger Reflections on six years of progress Science 154

14241425

Hallam A (1973) A revolution in the earth sciences From continental drift to plate tectonics

Oxford Oxford University Press

Hamilton G amp Feenstra R C (1995) Varieties of hierarchies and markets An introduction

Industrial and Corporate Change 4 5191

Huisman J (Ed) (2009) International perspectives on the governance of higher education

London Routledge

Kogut B amp Zander U (1992) Knowledge of the firm combinative capabilities and the

replication of technology Organization Science 3 383397

Koike K (1994) Learning and incentive systems in Japanese industry In M Aoki amp R Dore

(Eds) The Japanese firm Sources of competitive strength (pp 4165) Oxford Oxford

University Press

Krimsky S (1999) The profit of scientific discovery and its normative implications Chicago

Kent Law Review 75(1) 1539

Kristensen P H (1996) On the constitution of economic actors in Denmark Interacting skill

container and project coordinators In R Whitley amp P H Kristensen (Eds) The chan-

ging European firm Limits to convergence (pp 118158) London Routledge

Kristensen P H Lotz M amp Rocha R (2011) Denmark Tailoring flexicurity for changing

roles in global games In P H Kristensen amp K Lilja (Eds) Nordic capitalisms and glo-

balization New forms of economic organization and welfare institutions (pp 86140)

Kushida K amp Shimizu K (2013) Syncretism The politics of Japanrsquos financial reforms

Socio-Economic Review 11 337369

Laurence H (2001) Money rules The new politics of finance in Britain and Japan Ithaca NY

Cornell University Press

Lazonick W (1991) Business organization and the myth of the market economy Cambridge

Cambridge University Press

Lazonick W amp West J (1998) Organizational integration and competitive advantage

Explaining strategy and performance in American industry In G Dosi D J Teece amp

J Chytry (Eds) Technology organization and competitiveness (pp 247288) Oxford

Martin B amp Whitley R (2010) The UK research assessment exercise A case of regulatory

Meier F amp Schimank U (2010) Mission now possible Profile building and leadership in

German universities In R Whitley J Glaser amp L Engwall (Eds) Reconfiguring

Metcalfe J S amp James A (2000) Knowledge and capabilities A new view of the firm In

N Foss amp P Robertson (Eds) Resources technology and strategy Explorations in the

resource based perspective (pp 3152) London Routledge

Milthaupt C J (Ed) (2003) Global markets domestic institutions Corporate law and govern-

ance in a new era of cross-border deals New York NY Columbia University Press

Mintzberg H (1979) The structuring of organizations Englewood Cliffs NJ Prentice Hall

Mintzberg H (1983) Structure in fives Designing effective organizations Englewood Cliffs

NJ Prentice-Hall

Morck R (Ed) (2007) A history of corporate governance around the world Chicago IL

University of Chicago Press

Musselin C (2007) Are universities specific organizations In G Krucken A Kosmutzky amp

M Torka (Eds) Towards a multiversity Universities between global trends and national

traditions (pp 6384) Bielefeld transcript Verlag

European comparative perspectives Dordrecht Springer

Paradeise C Reale E amp Goastellec G (2009) A comparative approach to higher education

reforms in Western European countries In C Paradeise E Reale I Bleiklie amp

E Ferlie (Eds) University governance Western European comparative perspectives

(pp 197226) Dordrecht Springer

Penrose E T (1959) The theory of the growth of the firm New York NY Wiley

Pinch T (1986) Confronting nature The sociology of solar neutrino detection Dordrecht

Reidel

Regini M (Ed) (2011) European universities and the challenge of the market Cheltenham

Edward Elgar

Richardson G (1998) Some principles of economic organisation In N Foss amp B Loasby (Eds)

Economic organisation capabilities and coordination (pp 4462) London Routledge

Rothblatt S amp Wittrock B (Eds) (1993) The European and American University since 1800

Cambridge Cambridge University Press

Roy W G (1997) Socializing capital The rise of the large industrial corporation in America

Princeton N J Princeton University Press

Schiene C amp Schimank U (2007) Research evaluation as organisational development The

work of the academic advisory council in Lower Saxony (FRG) In R Whitley amp

J Glaser (Eds) The changing governance of the sciences The advent of research evalua-

tion systems (pp 171190) Dordrecht Springer

Schimank U (2005) New public management and the academic profession Reflections on

Schmidt V A (2002) The futures of European capitalism Oxford Oxford University Press

Simpson R L (1985) Social control of occupations and work Annual Review of Sociology

11 415436

Slaughter S amp Cantwell B (2012) Transatlantic moves to the market The United States

and the European union Higher Education 63 583606

Soskice D (1999) Divergent production regimes Coordinated and uncoordinated market

economies in the 1980s and 1990s In H Kitschelt P Lange G Marks amp J Stephens

(Eds) Continuity and change in contemporary capitalism (pp 101134) Cambridge

Stuart T amp Ding W (2006) When do scientists become entrepreneurs The social structural

antecedents of commercial activity in the academic life sciences American Journal of

Sociology 112 97144

Teece D Pisano G amp Shuen A (2000) Dynamic capabilities and strategic management In

G Dosi R R Nelson amp S G Winter (Eds) The nature and dynamics of organiza-

tional capabilities (pp 334362) Oxford Oxford University Press

Thelin J R (2004) A history of American higher education Baltimore MD Johns Hopkins

University Press

Trow M (1993) Comparative perspective on British and American higher education In

S Rothblatt amp B Wittrock (Eds) The European and American University since 1800

(pp 280299) Cambridge Cambridge University Press

Weingart P amp Maasen S (2007) Elite through Rankings The Emergence of the enterpris-

ing university In R Whitley amp J Glaser (Eds) The changing governance of the

sciences The advent of research evaluation systems Sociology of the sciences yearbook

(Vol 26 pp 75100) Dordrecht Springer

Whitley R (1999) Divergent capitalisms Oxford Oxford University Press

Whitley R (2000) The intellectual and social organization of the sciences (2nd ed) Oxford

Oxford University Press (first edition 1984)

Whitley R (2006) Project-based firms New organisational form or variations on a theme

Whitley R (2007a) Business systems and organisational capabilities Oxford Oxford

University Press

Whitley R (2007b) Changing governance of the public sciences In R Whitley amp J Glaser

(Eds) The changing governance of the sciences The advent of research evaluation sys-

tems Sociology of the sciences yearbook (Vol 26 pp 327) Dordrecht Springer

Whitley R (2008) Universities as strategic actors Limitations and variations In

L Engwall amp D Weaire (Eds) The university in the market (pp 2337) London

Portland Press

Whitley R (2011) Changing governance and authority relations in the public sciences

Minerva 49 359385

Springer

Wright S amp Williams Oslashrberg J (2009) Prometheus (on the) rebound freedom and the

Danish steering system In J Huisman (Ed) International perspectives on the govern-

ance of higher education (pp 6987) London Routledge

EMPOWERMENT OF FRENCH

UNIVERSITIES BY FUNDING

AND EVALUATION AGENCIES

Christine Musselin

ABSTRACT

This article questions how institutional change influences actorsrsquo behaviorwithin organizations affected by the evolution of their institutional envir-onment This issue is addressed by looking at how university leaders areempowered by the external reviews led by evaluation agencies andresearch councils and how they use these reviews as managerial tools andto make decisions It is argued that this process is complementary to thereforms in university governance and structures and amplifies theireffects because it is more legitimate favors organizational coupling andthe appropriation of new norms It draws on a study led in three Frenchuniversities in 2011

Keywords Empowerment external peer review funding and evalua-tion agencies institutional change organizational coupling universitygovernance

ISSN 0733-558Xdoi101108S0733-558X20140000042002

INTRODUCTION

Change has become a major issue for neo-institutionalists both sociologi-cal (Padget amp Powell 2012) and historical (Streeck amp Thelen 2005) buttheir reflections focus more on the transformation of fields than on howinstitutional change influences actorsrsquo behavior within the organizationsaffected by the evolution of their institutional environment This articleaddresses this lacuna in analyzing universities as examples of professionalbureaucracies Universities in many countries have undergone rather strongreforms aimed at strengthening their managerial capacities and transform-ing them into rationalized organizations (Ramirez 2006) but it remainsunclear how far and under which conditions such changes have been effec-tive rather than symbolic or hypocritical (Brunsson 1989) Building on thecase of French university academic managers I will argue that the intro-duction of managerial norms in universities faced a legitimacy deficit whichcould be overcome if they were able to use professional norms to makedecisions and I will show that these professional norms are produced bythe new state evaluation and funding agencies that construct external peer-based reviews for use as management tools by university leaders

In a paper published in 2007 Richard Whitley (2007) stressed the role ofthe academic elite in charge of assessing research or teaching activities orresearch projects in evaluation agencies or research councils Following thisperspective I looked (Musselin 2013) at the impact of this evolution forthe academic profession I concluded that instead of a dismissal of the aca-demic profession there is a reinforcement of the academic control throughthe role of these academic elite and the peer-reviews it produced But thereis also more differentiation The ldquohavesrdquo that is academics rewarded bythe elite or being part of it have resources enabling them to negotiate withtheir institution and are therefore stronger today while the ldquohave notsrdquo arein a worse situation than before as their failures in terms of professionalreview leave them without resources vis-a-vis managerial control In otherwords the role of the academic profession is stronger because peer-review-based judgment becomes more important but this can conflict with theneeds of institutionsrsquo management by increasing differentiation amongacademics

This article will elaborate on these points by looking more closely athow university leaders are empowered by the external reviews led byevaluation agencies and research councils I will argue that this process iscomplementary to the reforms in university governance and structures andamplifies their effects because it is more legitimate favors organizational

52 CHRISTINE MUSSELIN

coupling and the appropriation of new norms In other words the institu-tional empowerment of universities relies as much on professionals (aca-demics) as on managers and builds more on external pressures than oninternal reshaping of structures because the reinforced role played by exter-nal research councils and evaluation agencies is a powerful driver for inter-nal evolution

It draws on a 2011 study of three French universities1 (Musselin 2012)Students of a the master of sociology program at Sciences Po conductedinterviews in three universities One located in Paris and heavily specializedin health and sciences and two in the French regions one focused on socialsciences and humanities and one that is pluridisciplinary2 I will call themrespectively UniSciences UniSSH and UniPluri About 100 interviewswere held with central university managers (presidents vice-presidentsregistrars and their main collaborators) deans heads of department andheads of labs as well as with academics and administrative staff elected tosit in one of the three deliberative bodies (university council academiccouncil and council for training and students affairs) and are listed inTable 1 Because of the voluntary focus of this study on academic andadministrative staff directly involved in the management of their institu-tion this paper focuses on change in behaviors for this specific populationbut does not consider their effects on academic practices

I will start with a discussion of the literature on organizational changein general and in higher education studies in particular (1) I will go

Table 1 Breakdown of Interviews Held with University Managers

UniSciences UniSSH UniPluri

Academic university managers (presidential team) 4 7 5

Administrative managers (central administration) 12 5 7

Deans heads of departments or labs 15 11 16

Administrative managers in faculties 4 4 3

Members of deliberative bodies with no other managerial

responsibilitiesa2 5 2

Total 37 32 33

aSome of the interviewees of the other categories were also members of these bodies and inter-

viewed simultaneously for their function as administrative or academic leaders and elected

members of bodies

53Empowerment of French Universities

on with an overview on the main recent French reforms and payspecial attention to the creation of a national agency for the evaluation ofteaching and research (AERES Agence drsquoevaluation de la Recherche et delrsquoEnseignement Superieur) as well as a national research council (ANRAgence Nationale de la Recherche) after the 2006 act The next three sec-tions will be organized around the transformations induced by these twoagencies I will show that the assessment process that simultaneouslyinvolves the evaluation of teaching programs the evaluation of researchunits and the evaluation of the governance of universities favors morecentralization and more coordination within universities These processesare furthermore reinforced by the fact that resources and reputations arelinked to them

I will then argue that the results of this peer-review-based assessment like the allocation of grants by the ANR or the budgets linked to theperformance assessed by the AERES enabled universities to allocateresources on a less egalitarian basis and to close down or reconfigureresearch units and teaching programs They also encouraged the introduc-tion of incentives or other new devices in order to improve their resultsI will finally show that these are not only game playing and tactics inorder to address external demands the norms and expectations devel-oped by the peers sitting in the evaluation agency and the research coun-cil are appropriated by many of the interviewees even when they criticizethem this in turn facilitates the use of external reviews as a managementtool by university managers In a concluding section lessons will bedrawn from this case about the role of external bodies on the transfor-mation of universities and the complex links between organization andprofession

ORGANIZATIONAL CHANGE AND

TRANSFORMATION OF UNIVERSITY GOVERNANCE

Looking at Change in Organizations from Within

As stressed by Greenwood and Miller (2010) many recent developments inorganization theory have not only moved away from the study of organiza-tional design but also from the study of organizations from withinAnalysts primarily focused on the institutional environment of organiza-tions (or fields) and looked at how it exercises pressures for homogeneity

and pushes organizations in the same field to adopt the same forms(DiMaggio amp Powell 1983) They also developed explanations for changeby looking at the emergence and spread of new institutional logics(Thornton amp Ocasio 2008 Thornton Ocasio amp Lounsbury 2012) withina field leading to the transformation of the organizations belonging to thisfield The crucial role of institutional entrepreneurs in such processes wasfurthermore highlighted by many studies (Battilana Leca amp Boxenbaum2009 DiMaggio 1988 Greenwood amp Hinings 1996) The idea that com-peting logics may exist in a field and that actors are fighting one againstanother to impose their logic rather than another has also been discussedand led to the notion of strategic fields (Fligstein amp McAdam 2012) thatstresses the importance of power relations and skilled actors (Fligstein2001) for change to occur

While very effective in understanding how organizations developing simi-lar activities evolve (for instance how accounting firms became transnationalprofessional service firms as studied by Greenwood amp Suddaby 2006) neo-institutionalist approaches were less interested in looking at the impactor even the effectiveness of these transformations within organizationsFor some of them (Meyer amp Rowan 1977) this is legitimized by the factthat organizations follow institutional scripts working as rationalized mythsorganizations conform to the scripts but practices are not affected becausescripts and practices are loosely coupled Others like Fligstein (1985 1990)are more attentive to the power relations within organizations this authorfor instance considered the relationships between the main functions offirms and observed a shift of weight from production to marketing andmore recently to financial departments overtime but he was neverthelessmore interested in the impact of this shift in power for the conception ofcontrol dominating the field than in the concrete interplays and effects itmay have within the firms

Nevertheless more attention has been recently paid to what happenswithin organizations in a given field Greenwood and Miller (2010) forinstance explain how the evolution of the field of professional service firmslead to change in their organizational design and describe this shift as theresult of neo-institutionalist processes but also for more functionalistreasons Nevertheless they limit their analysis to this change in formalstructures and therefore take for granted the effectiveness of this organiza-tional design they do not consider that it might be diverted by the agencyof the actors working in these formal structures In contrast some authorssuggest putting more light on how actors confronted with a new organiza-tional design coming from an institutional change react to these evolutions

They developed the notion of ldquoinhabited institutionsrdquo (Hallett amp Ventresca2006) to describe how agency affects institutional transformations Insteadof loose coupling or passive acceptation of change they show processeswhere individuals are described as resisting adopting or interpreting thenew structures imposed on them

Conflicts around institutional logics the role of institutional entrepre-neurs the notion of strategic fields or the idea of inhabited institutions allbring some agency back into institutions and highlight the need to have acloser look at the concrete impact of institutional change on actors andhow they adopt and appropriate change but also adapt and react to itwithin the organizations of the concerned field

Following this perspective I would like to show how and under whatconditions institutional change affects practices and norms within theorganizations impacted by such change I will focus on the specific case ofuniversities because they experienced reforms aimed at transforming therelationships between professional and organizational powers by strength-ening the organizational capacities of higher education institutions andincreasing their control over professionals Did the development of a man-agerial logic in universities affect the way university managers work andmake decision

Change in University Governance

The huge literature about university governance converges on one point allauthors agree that universities are experiencing major changes and areurged to become more governed introduce managerial tools and practicesissued from firms reinforce their strategic and decisional capacities andinteract more closely with their stakeholders in order to take societaldemands better into account Such trends are exogenous they are imposedon universities and on academics For some authors this result from theimplementation of public policies inspired by new public managementandor network governance (Braun amp Merrien 1999 Paradeise RealeBleiklie amp Ferlie 2009) and for others from the diffusion of transnationalscripts introducing the model of a rationalized university (Ramirez 2006)all over the world

Whatever their origins and local implementations these evolutions sharein common the objective of building more autonomous universities andrely on quite similar strategies to reach this aim transferring to universitiesresponsibilities and domains of decision previously held by public

authorities reinforcing the executive power to the detriment of deliberativeand collegial bodies transforming the relations between universities andthe academic staff into an employeremployee relationship introducingstakeholders in the university governance reinforcing hierarchical powerdeveloping rules and procedures in order to manage the decision-makingprocesses and make them more effective etc (Braun amp Merrien 1999 DeBoer Enders amp Schimank 2007 Musselin 2005 Paradeise et al 2009)

As a result many of the reforms of the last decades aim at modifyingthe structures and decision-making bodies of universities the Qualityreform in Norway (Oslashstergren amp Stensaker 2007) the MUB in theNetherlands (De Boer 2002) the LRU in France the Law n240 ofDecember 30 2010 in Italy (Reale amp Primeri 2013) etc All modified thecompetencies of the university president the composition and attributionsof the deliberative bodies and increased the scope of responsibilities of uni-versity managers Building on the stimulating paper written by Brunssonand Sahlin-Andersson (2000) on the construction of public services intoorganizations one can argue that universities experienced a process of con-struction boundaries hierarchy and rationality that is were constructedinto organizations (De Boer et al 2007 Krucken amp Meier 2006 Musselin2007 Whitley 2008)

Nevertheless the impact of the reforms aimed at transforming universi-ties into organizations by modifying their structures and governance isoften less effective than expected One explanation lies in the fact that theyface and are slowed down by the organizational specificities that character-ize higher education institutions (Musselin 2007)

The loose coupling nature of teaching and research that is the possibi-lity to teach a class (or to work on a research project) without stronglycooperating with onersquos direct colleagues is a key organizational specificityof universities As a result top-down hierarchical forms of management arerather ineffective because the loosely coupled nature of academic activitiesweakens the robustness and connectiveness of the ties between the differentunits and between individuals This lack in hierarchical authority is evenstronger in countries where university presidents cannot appoint their inter-mediary managers in France deans3 heads of department and head ofresearch units are elected for two to five years As shown by StephanieMignot-Gerard (Mignot-Gerard 2006 Musselin amp Mignot-Gerard 2002)for France there is a rather strong gap between the ldquocentrerdquo (the directorsof the central administration and the presidential team) on the one handand the rest of the university (periphery) on the other The deans and theiradministrative staff do not feel (or rarely feel) solidarity with the center

and there is no relation of subordination neither between the president andthe deans nor between the registrar and the administrative staff located inthe faculties In order to face this situation presidents often built parallelstructures mobilizing individuals they trust (Barrier and Mignot-Gerard2013) and thus favor and develop top-down collegiality (Lazega ampWaterbled 2010[2009]) parallel to the formal structures but such informaland interpersonal ties lack legitimacy

A second organizational specificity lies in the fact that teaching andresearch are unclear technologies difficult to describe to prescribe and toreplicate Furthermore for such technologies the relationships betweencauses and effects are complex it is for instance hard to demonstrate thatstudents are successful because they have been taught that or that subjectand this way or another rather than because they were rightly selectedThis unclear nature of the technologies involved in teaching and researchhinders the intervention of university leaders on those activities as one doesnot precisely know how they work or how to improve them or make themmore efficient The power of decision therefore lies in the hands of each dis-cipline characterized by its own norms practices and values (Becher1989) university leaders are considered as incompetent in deciding forchange in a training program (or a research agenda)

These organizational specificities can be found in other organizationsbut they are rarely (if not never) simultaneously present From this point ofview they make universities distinctive from other organizations but alsodistinguish them among other professional bureaucracies In hospitals forinstance professionals are more tightly coupled to achieve their activitiesand technologies of production (surgery treatments anesthetics etc) areeasier to grasp (describe prescribe and reproduce) The simultaneous pre-sence of these two organizational characteristics is crucial in higher educa-tion institutions because they strongly impact on the ways universitymanagers can develop their role and exercise their leadership as they can-not use the same ways as in other productive organizations One centralproblem for university managers to exercise authority and use hierarchicalpower is therefore their lack of legitimacy even when they are elected

In such a situation it is not surprising to observe that university presi-dents rarely extensively use the decisional power they have on paper In thisvolume Richard Whitley and Jochen Glaser also stress the difficulties facedby university managers to control and direct core activities Personal char-isma is often a better source of influence than the formal prerogativesprovided by the law In the study she led on French universities Mignot-Gerard (2006) showed that the strength recognized in some university

presidents was always very personalized and first linked to their capacity indeveloping a vision and convincing their staff by repeating this vision onany occasion She furthermore showed that instead of selecting or differen-tiating these presidents would first of all enroll and integrate key academicsbecause of their own lack of legitimacy in imposing principles of selection

The increase in formal authority and decisional capacity attributed touniversity presidents by the reforms are therefore often not sufficient forthem to exercise more control on professionals and develop more manage-rial forms of steering because they lack legitimacy to do so

Building on the case of France I will now argue that they are able toovercome this lack of legitimacy when external peer-reviews are availableand can be used to support their decisions and increase their governancecapacity In the French case the concomitant introduction of an evaluationagency and a funding council and the passing of an act increasing the for-mal prerogatives of university leaders facilitated the effective use of theirnew margins of maneuver by these university leaders After a presentationof these two reforms I will show how academic and managerial powers arecombined rather than opposed in order to govern French universitiesmore strongly

FRENCH HIGHER EDUCATION REFORMS THE

LOPRI (2006) AND THE LRU (2007)

Within the last years two acts aimed at transforming the French higher edu-cation system were passed First in 2006 the LOPRI (act for research andinnovation) led among other measures to the creation of a nationalresearch council (ANR) and a national evaluation agency (AERES)Second in August 2007 three months after the election of Nicolas SarkozyValerie Pecresse the new Minister of higher education and research passedthe LRU act (Loi relative aux Libertes et Responsabilites des UniversitesFreedom and responsibility for universities act) that aimed at transformingthe governance of French universities in two ways by increasing the decisio-nal span of university presidents and by delegating new competencies toFrench universities including the management of their payroll and positionsthat were until then run by the ministry

In the newspapers and in debates the LRU is more often mentionedand discussed than the first one because the 2007 act was one of the firstdecisions made under the new presidency of Nicolas Sarkozy The

Minister Valerie Pecresse presented her law as a decisive rupture building anew area for the French higher education system But looking more closelyat this act it appears that it accelerated and extended but did not greatlymodify the direction set by previous acts (the 1968 Faure Act and the 1984Savary Act) or previous public policies (such as the introduction offour-year contracts between each university and the ministry since thebeginning of the 1990s Musselin 2004[2001]4) as the previous acts theLRU aimed at strengthening and increasing the institutional and manage-rial autonomy of French universities

The LOPRI is in many ways more controversial and disruptive Amongthe different new devices and instruments set by this law the more contestedwere and still are the ANR and the AERES

The ANR was created in 2005 (one year before the LOPRI that officiallygave it birth) The idea was to create an institution close to the GermanDFG (Deutsche Forschungsgemeinschaft) or the American NSF (NationalScience Foundation) able to selectively fund research projects that areeither blue sky research (the programmes blancs) or are selected throughthematic calls for proposals The agency covers all disciplines and is incharge of running the different calls Before the ANR this function wasspread between different institutions and among them between nationalresearch institutions such as the CNRS5 (National Center For ScientificResearch) and the Ministry for higher education and research The loss ofthis funding function was of course not welcome by research institutionsthe budgets they managed before 2005 were allocated to the ANR TheANR in fact centralized within one single institution different project-basedforms of funding that previously existed but the ANR is also more thanthis sum First of all the ANR received more resources than its differentpredecessors and thus accelerated the already on-going process towardmore project-based research Second the ANR became the main operatorof the different highly competitive calls for proposals (Labex Equipexetc) launched in the framework of the Grand Emprunt6 (httpwwwemprunt-national-2010fr) A large public fund was initiated to promotefuture investments and much of it (22 billions of 35 as a whole) was dedi-cated to research and universities The idea was to generously fund a smallnumber of excellent labs (Labex) excellent research equipment (Equipex)excellent higher education institutions (Idex) etc In order to run all theseprograms the ANR furthermore developed new procedures to make deci-sions but they all quite heavily rely on peer review

The AERES agency for evaluation was also created by the LOPRI andstarted to work at the beginning of 2007 It is also the product of the

centralization of former institutions and of the introduction of new devicesIt is in charge of evaluating every four years (five now) the activity of allresearch units all training programs and all higher education and researchinstitutions Its missions cover what was previously achieved respectivelyby the scientific councils of the national research institutions academicsappointed by the ministry as experts and the national council for the eva-luation of universities (CNE Conseil national drsquoevaluation des universites etdes etablissements) But like the ANR the AERES is not just the sum ofwhat previously existed It also introduces some new practices All evalua-tions are available on a website (which was only the case for the evaluationof universities before) and until 2011 research labs received a global graderanging from A+ to C7 Last but not least the results of these evaluationsare at least partly used as performance-based funding mechanisms

The staff of the AERES consists of administrative personnel who orga-nize the visits of the experts on campus design templates etc and of aca-demics nominated by the management of the AERES who are in charge ofselecting the experts discipline by discipline but also decide the criteria andprocesses of evaluation Some unions and associations of academics suchas SLU (Save universities) and SLR (Save research) developed strong criti-cisms of the AERES especially against the fact that the experts evaluatingthe research units are no longer designated through elections (and listsoften controlled by the unions) as was the case before when the CNRS orthe INSERM evaluated their labs for instance but nominated by theAERES itself

The creation of these two agencies and the LRU can easily be seen as aproof of the increasing influence of NPM in France after a long period ofresistance to this doctrine and its recipes until the 2000s (Bezes 2008) Theministry is supposed to focus on global strategies and delegate moreoperational tasks to some agencies or directly to more operational units(universities) that receive more autonomy to achieve their tasks More com-petition selective processes and performance-based mechanisms wereintroduced with the new formula for budget allocation called SYMPAthat was developed by the French Ministry about 20 of the budget isperformance-based (one part depending on four-year contracts and theother on automatic allocation based on the number of masters degreesawarded or the AERES grades obtained by the research labs etc)

But the impact of the new agencies is not restricted to their activities andto the missions they received from the state Attention should also be paidto their more indirect effects on the academic profession on the one handand on the management of universities on the other (Musselin 2013)

In this article I will in particular insist on the use of external peer reviewas a management tool by academic leaders Building on the study led inthree French universities in May 2011 I will identify more precisely howthey use external evaluation to manage their institution and how profes-sional and managerial powers are combined

EXTERNAL PEER REVIEW ENHANCES

CENTRALIZATION AND COORDINATION

WITHIN UNIVERSITIES

In French universities centralization of decision-making has been a majorprocess in the recent years This is in line with the objectives of the LRUthat sought to concentrate decision-making and responsibilities at the topof universities The radical increase in university budgets after they receivedthe management of their payroll8 and became responsible for their employ-ees is a further reason for centralization As in many countries that alsoincreased the managerial autonomy of their universities (KruckenBlumel amp Kloke 2013) the number of highly qualified staff grew in finan-cial services and human resources departments thus concentrating at thetop level more qualified individuals able to deal with the decisions and pro-blems raised by the new responsibilities The simultaneous implementationof new budgetary software (called SIFAC) amplified this trend The admin-istrative managers of universities used this opportunity to restrict the num-ber of staff able to enter and work with data in this software by trainingonly some of them in this new instrument They thus better control thequality of the data managed in SIFAC but also control the level of infor-mation accessible to different categories of staff

If not the cause for more centralization the ANR and the AERESfacilitated and increased this trend They first of all technically reinforcedcentralization As a matter of fact no application can be sent to the ANRwithout the approval of the institution that will run the money (the univer-sity or a national research institution if the research units is also recog-nized9 by one (or more) of these and prefers the money to be run by one ofthem) even if no approval is refused it provides the university with infor-mation it did not systematically get before (especially on those who appliedand failed) Similarly the procedures developed by the AERES are alsofavoring centralization The activity reports and projects prepared by theresearch units and training programs must be uploaded on the website of

the AERES by the university administration and thus the latter can bettercentralize and control what is sent

But the more important driver for centralization is that getting ANRgrants and being positively evaluated by the AERES have become impor-tant signals of the quality of the institution Communicating about successwith the ANR calls has become crucial and universities are more and moreconcerned with diffusing information about their success Directions forresearch have been created at the university level in order to help academicsapplying for the ANR (or the European Research Council) to pushresearch units to answer to call for proposals and to concentrate informa-tion about who answers and who does not who is successful and who isnot etc

The attention paid to the ANR even increased when it became the mainoperator of the calls for proposals (Labex Equipex etc) launched for theGrand Emprunt (public fund) The time constraints of the process but alsothe high stakes it represented for many institutions led university presidentsto play a major role in identifying the projects that will be submitted to thejuries of the ANR (Mignot-Gerard 2012) instead of letting bottom-upprocesses work for the emergence of research projects University leaderstook the lead and detected the projects to support They of course did nottake the risk to promote obscure research units and relied on those with astrong scientific reputation according to the AERES evaluation and thosewho were successful with the ANR But they chose among them which tosupport and thus by-passed the usual decision-making processes to keepit in their hands even if they built on scientific reputations to legitimatetheir choice

I alerted the president because he is a friend of mine I told him ldquoif you further support

this Labex you take a riskrdquo I reminded him that the deliberative bodies have not been

informed He told me he did but after checking I saw it was not the case It was not

So there is a lack of democracy hellip [] There was a kind of deal between the president

and the vice-president who talked together and decided one with another that the uni-

versity needs this Labex and they opted for it The choice of the Labex is the choice of

the vice-president because he rightly saw there was a scientific opportunity there It

was not discussed by the university bureau until last Monday it was not discussed at

the scientific council and a fortiori not by the university council but by contrast ()

all the labs concerned by the Labex were already involved in the discussion You see

we are walking on the headrdquo (Director of a lab UniSSH)

Reputation and resources are also crucial drivers for centralization inthe case of evaluation processes The introduction of performance-basedfunding even if quite limited in scope leads universities to improve their

assessment records both in order to get more resources and also to bemore attractive Because the three processes (assessment of research unitsassessment of training programs and assessment of university governance)are now led at the same moment for each university they are easier tointerconnect and coordination is also easier to achieve

The central administration and the academic leadership of the three uni-versities under study intervened in different ways First they were involvedin the preparation of the reports and projects labs and training programssent to the AERES At UniSSH and UniScience the presidential team clo-sely supervised the documents prepared for the training programs in orderto be sure that they would have a chance to get a good evaluation and thenbe accredited by the ministry

We knew that some universities sent rubbish to the AERES () The AERES did not

appreciate receiving rubbish () In 2006 when I saw the documents on my desk that

were to be sent to the AERES they were really like rubbish The training programs are

good and the professors are great and at the end this is what counts But a decent uni-

versity cannot send rubbish And it was rubbish They did not understand it was neces-

sary to provide figures () With the vice-president we quickly came back in August

and we worked on the documents (Central administration UniScience)

The first impact of the evaluation was anticipation Otherwise it is difficult to speak of

any concrete impact But the offer in training programs was built by anticipation We

tried to send the best possible projects We worked a lot beforehand and I am not sure

all this work would have been done without all these procedures (Member of the coun-

cil for teaching and student affairs UniSSH)

In some cases it became an opportunity to imagine more innovativeprojects and think of new collaborations But again presidential teams usedthe assessment led by the AERES to justify their position internally whileusing the result of the assessment as an argument in the discussion with theministry In the example related in the following quotation we see that thevice-president refused taking the risk of sending a last minute new versionof the project and suggested finalizing a new version to be tested with theAERES during the on-campus visit Because the new version had been wellreceived on that occasion the president was then legitimate to support itinternally and externally

In history they proposed two masters One rather traditional and another focused on

contemporary history with some history of art and social sciences It was very interest-

ing but the organization of the program was not very clear Nevertheless after a few

months of discussions they did not accept to move on that I finally said ldquook I do not

think it will work like that but letrsquos try and send itrdquo One week before the deadline set

by the AERES they came back to me and said ldquowe thought about it again and would

like to change our projectrdquo I told them ldquowait it is already in the pipeline there is no

way to change everything a week before the deadline you should have thought of it

beforerdquo (hellip) I told them ldquosend it as it is and use the time before the visit on campus of

the AERES to work on an alternative project and letrsquos speak of it againrdquo Of course the

AERES asked questions about the organization I was sure they would but we could

answer ldquoabsolutely you are completely right but we already have a new proposalrdquo

And I already sent it to the Ministry to start the discussion (Member of the council for

teaching and student affairs UniSSH)

University presidents furthermore used the opportunity of the AERESevaluation and the preparation of the reports to impose common rules to allIn the following example we see how a vice-president pushed for the samerequirement in the study of foreign languages in all programs of the univer-sity including those of the small department in law hosted by UniScience(despite its specialization in sciences and medicine)

In law the colleagues did not respect the requirements We are required to provide

information for the RNCP national registry on professional competencies All

this for them is hellip The university had also decided that we should help our

students to pass the TOEIC Getting competencies in a foreign language seems

unavoidable A professor in law told me ldquoforeign languages do not count We

should not carerdquo Aware of that the vice-president decided to nevertheless include

in their project ldquowe encourage students to pass the TOEICrdquo So my direction

entirely revised their document for the AERES This is so true that they could not

recognize their document when it came back from the AERES (Central administra-

tion UniScience)

Second some universities prepared the evaluation of the AERES byorganizing a mock assessment before the visit of the AERES UniSciencewas one of the first institutions10 assessed by the AERES in 2007 and wasin a way ldquostruck by surpriserdquo but the two others belong to later waves ofevaluation and had time to learn from the first experiences In both casesthey asked colleagues from other places to evaluate them in order to tryand improve their reports and projects beforehand At UniSSH this con-cerned teaching as well as research while in UniMulti only the researchunits were concerned by the mock exercise In each case the aim is toimprove the quality of the activity report that will be sent to the AERESbut it also makes one and each sensitive to what is expected from them andto the evaluation criteria they should be aware of

We tried to do our best to meet the requirements of the AERES We did it in two steps

We had a pre-evaluation organized by the university where they hellip brought hellip what

they called in English a laquo visiting committee raquo including a former member of the

AERES Mister X (Director of a research unit UniMulti)

Third the three universities exercised control on the answers written bythe evaluated after they received their AERES reviews and used the ldquorightto replyrdquo the AERES provides to each director of lab training programand institution these answers are published on the website of the AEREStogether with the reviews At UniSSH one of the university vice-presidentrewrote these replies in order to make them more positive than defensiveand very nicely explained what he tried to do

None of the reviews was a scandal But we intervened on the answers prepared by the

research units For instance one of the colleagues did not understand anything The

AERES was saying that this unit welcomes international visiting scholars for too

short periods of time And the colleague answers ldquowe welcome international scholars

for short periods because our location is not suitable for longer stayrdquo So I told him

ldquodo you think you will get international PhD candidates if you say that your building

is not suitable Say that you are developing a solid dynamics of international rela-

tionships and that you already collaborate with this and this country and that you

aim at welcoming post-docs and researchersrdquo He was developing a narrow answer

while it was easy to reverse the critics and to say he was about to expand the interna-

tional collaborations for his lab That is the work I have done for almost all answers

with more or less reformulation Because this is public this will be read looked at so

the answers to the AERES must become a way to attract people I told them ldquoyou

will apply for an ANR call and what will the experts do They will look at the evalua-

tion you got from the AERES So be careful and do not worsen your case in your

answer to the AERESrdquo So we somewhat subverted the process (Vice-President

UniSSH)

During the preparation of the evaluation and after it the universityleaders are therefore very active in trying to avoid bad assessments toharmonize some practices and provide the best image they can of theirinstitutions This occurs through more centralization and control overinformation this accentuates and builds on the centralization processalready favored by the LRU

EXTERNAL PEER REVIEW ALLOWS LESS

EGALITARIAN ALLOCATION OF RESOURCES

AND RESTRUCTURATIONS

Centralization and coordination of information in the preparation of theevaluation and application processes are only two of the effects of theAERES and ANR on institutional management The results of these pro-cesses are also used to make decisions within universities They provide uni-versity managers with the legitimacy they lack and become references forallocating resources and conducting restructurings In these ways the

assessments and decisions issued by AERES and the ANR are used asmanagement tools by the presidents and empower them they allow themto mobilize and put into action the increased decisional capacities offeredby the LRU This is particularly clear when one look at decisions about theresearch policy of the three universities under study

First because these results are produced by peer review they are legiti-mate Of course they might be contested and there exist many protestsagainst the AERES and the type of evaluations it produces as well asagainst the increased differentiation favored by the ANR but these pro-tests are limited enough for internal decisions to be made on the basis ofthese results at least in the three universities under study although one ofthem is known for its strong participation in social movements and demon-strations The allocation of the university operational research budget tothe different research units of the university is a good example of that

Since the LRU universities are responsible for the distribution of thisbudget They could have decided to attribute to each lab the amount itpreviously received from the ministry but in the three universities consid-ered here mechanisms of allocation have been discussed and accepted bythe scientific councils In all cases they weighted the amount of budgetaccording to the AERES grade (A+ A B or C) and developed a for-mula The former budget of the lab was multiplied by more than 1 inresearch units getting a A or a A+ and less than 1 for those with a CDifferent universities have though adopted some different strategies AtUniMulti (see below) they rewarded quite clearly the best and sanctionedthe ldquoCrdquo while at UniSSH they rather tried to help the ldquoCrdquo to improvethemselves but in the three cases and in most other places in Francethe rules of allocation of the research budget took the grade intoaccount

The research in this university is now organized in four research institutes They are

managerial instruments and are thematic-based To allocate funding the university uses

the same algorithm as the ministry and applies it to each institute taking into account

the number of academics etc (Question what are precisely the criteria) I do not know

precisely but the number of academics the number of research active individuals the

number of people with a second thesis (habilitation a diriger des recherches) are taken

into account I do not know precisely On top of that we introduced a supplementary

parameter taking into account the grade given by the AERES to each lab Last year

this multiplying parameter was of 14 if a lab got a A and 16 if they got a A+

(Director of a research lab UniMulti)

In all cases too this differentiation was pushed by the presidential staffbut was agreed upon by their respective academic council11 Their membersvoted in favor of the new principles of budget allocation

Second we observed that evaluation and research grants were also usedto justify restructuration In two of the three universities (UniScience andUniMulti) research units getting a C were reorganized some were sup-pressed and their members had to integrate new teams or develop new pro-jects At UniScience even a B could lead to a restructuration a researchgroup that received a B was ldquoaskedrdquo to join another lab despite the protestof the research group leader who felt they will lose the benefits of the goodrelationships they had built with the former lab Such difficult decisions arenot limited to research but may also concern teaching The creation of newtraining programs receiving a critical evaluation was also abandoned

Some colleagues wanted to create a professional master with a potential on the non

academic job market for historians Some colleagues thought it would be booming

But the evaluation of the AERES was not very positive We did not go further The

AERES was not positive the project was not well enough prepared to convince them

(hellip) If we couldnrsquot convince the AERES how could we convince students employers

professionals Maybe we will come back to it later on We considered it was probably

not ripe (Head of department UniSSH)

The perspective of getting highly selective grants may also justify restruc-turing As mentioned above the ANR was in charge of the differentcompetitive call for proposals (Labex Equipex etc) launched in the frame-work of the Grand Emprunt University presidents were very much involvedin the process and centralized the preparation of the answers they used thisoccasion to suggest mergers or justify reorganizations Meeting the criteriafixed by the ANR was a justification for some internal adaptations andarrangements Even if they remain marginal such interventions are new onthe French scene Furthermore in France and by contrast with the profil-ing policy imposed on German universities restructurings are not onlyinternal to universities but also involve strengthened relationships betweenuniversities grandes ecoles and local units of the national research institu-tions that are all invited to if not merge at least coordinate into jointregional structures

Last but not least the external pressures led by the evaluations and thecompetition for grants justified the development of internal tools aimed atimproving the performance of the university or its conformity with externalrequirements The notion of research active staff used12 by the AERESmade visible the low level of publications of some academics or their pre-ference for publication in rather peripheral journals In some cases this ledto avoidance strategies co-authorship with less research active academicswas encouraged while some labs created a category of ldquoassociated staffrdquofor research passive academics not to be counted as staff of their research

unit This also led some university presidents to imagine incentives aimedat transforming research passive staff into research active membersUniMulti for instance developed an internal call for proposals explicitlyaimed at research passive academics selected projects received someresearch money and benefited from a decrease in teaching duties

They developed a fellowship on a competitive basis to allocate 3000 euros to faculty

members who do not publish enough and who were offered 50 less teaching So they

had to write a research project and each lab could present its candidates (Director of

the department for research UniMulti)

Many evidences and decisions made therefore confirm the attention pro-vided by university leaders to the decisions and judgments produced by theANR and the AERES but also the use of these external reviews as manage-ment tools supporting internal decisions They legitimate the choices andorientations university leaders are formally empowered to make since theLRU passed

A RATHER LARGE APPROPRIATION OF THE

NORMS AND EXPECTATIONS DEVELOPED

BY THE EVALUATION AGENCY AND

THE RESEARCH COUNCIL

One should not too quickly conclude that peer-review-based decisionsmade by the AERES and the ANR only influence and empower the presi-dential team To put it differently they are not only used as managementtools The reason for that is that the AERES and the ANR do not onlyproduce decisions (on grants) and evaluations they are also producers ofnorms They contribute to the definition of what is a good research projectwhat is a good research unit or training program what are the journalswhere academics of this discipline should publish (Pontille amp Torny 2010)the right activities a research unit should be involved in the indicators atraining program should concentrate on how a research project should bemanaged and split into deliverables etc The ldquoreference documentrdquo pro-duced by the AERES to explain how the evaluation of research units willbe led is a typical example for the normative work produced by such agen-cies to ldquotellrdquo what is right and what is expected

This makes the audience and the reception of the activities of the ANRand of the AERES much larger than the presidential team and their influence

goes beyond the rather instrumental use of these results to make difficultdecisions The interviews with the deans the heads of departments and thedirectors of research units showed that they might be critical of the currenttrends but have integrated them into their own priorities and preoccupationsAs observed by Camerati (2014) in UK departments the critics are not (orrarely) oriented toward the promotion of scientific meritocracy or evaluationin itself but much more toward the way it is led the constraints it creates orthe time it takes But overall with a few exceptions the prevalence ofresearch over teaching and the increased attention paid to publications israther well accepted and even promoted When asked about their role asdirectors or deans they often spontaneously mention they have to take careof the scientific production This of course facilitates the use made by theuniversity leaders of the external reviews and recognized norms

I do not push my colleagues into more teaching because I want them to have time to

develop their research agenda I canrsquot impose them not to invest in teaching but I can

dissuade them and say be careful for your career Managing degrees will not help your

career (Dean UniScience)

They also develop a discourse about the inevitable transformation of theacademic world and the inevitable adaptation that their colleagues shouldgo through They for instance consider that their role is to encourage theircolleagues to publish in visible and reputed journals

You might still publish a very important paper in a third tier journal I say it again we

have amateurs We have colleagues deliberately publishing good papers in third tier

journals They publish in journals diffusing 50 exemplars and run by a friend of them I

do my best Slowly they come to respect what is asked We canrsquot ask for resources if we

do not evolve Twenty years ago publishing in a top journal was not a must Everyone

knew about what was out There was a specific form of evaluation based on the fact

that everybody knew everybody But it is over We canrsquot count on that anymore

Colleagues must say in which journals they publish (Director of lab UniSSH)

These interviewees all the more accepted the on-going transformationin university governance when their success in getting grants or goodevaluations was high On the one hand and on the basis of their goodresults they can negotiate more support from the university leaders but aresimultaneously less dependent on them because they get external resourcesthrough the ANR grants what they get from their own institutions is lesscrucial and they can also get further external support (from the Region forinstance or from firms looking for partners) thanks to their good reviewsThere are thus empowered twice vis-a-vis their university and vis-a-vistheir external environment

The more critical interviewees were those that were not positively evalu-ated andor were not able to get grants They generally developed concep-tual arguments against project-based research the evaluations led by theAERES and performance-based funding but their critical stand was notonly ideological It is also positional and linked to their increased depen-dence on the resources provided by their own university while at the sametime their institution more and more conditions the allocation of resourcesto their adaptation to the new ldquorules of the gamerdquo There is from this pointof view a strong contrast between the social sciences and the sciences atUniMulti or between the humanities and the social sciences at UniSSHmany of the former trying to resist while the latter were more prompt tocomply even when they disagree with the recent evolutions

CONCLUSION

In professional bureaucracies managers lack legitimacy to introduce man-agerial practices because professional norms prevail and resist The capacityof universities in lowering the impact of the organizational reforms they areimposed is linked to this deficit in legitimacy but also to the organizationalcharacteristics specific to higher education institutions that is teaching andresearch rely on loosely coupled ties and unclear technologies This reducesthe possibility for presidents to rely on classical ways of governance (hierar-chy rationalization of processes setting of boundaries etc)

In this article I argued that the increasing margins of maneuver used byFrench university leaders and managers are not directly (or only) linked tothe new capacities offered by the LRU act but to the legitimacy they gainby relying on the external reviews produced by evaluation and fundingagencies In other words French university presidents are not strongerbecause the LRU gave them more hierarchical power they are strongerbecause they can justify their decisions with the external peer-based reviewsproduced by the ANR and the AERES These evaluation and fundingagencies thus play an important role in empowering the management ofuniversities as their reviews are used as management tools by university lea-ders who built on them to make decisions allocate budgets restructureThey legitimate the use of the margins of maneuver and decisional capaci-ties provided by the reforms in university governance They make suchreforms more effective somewhat increasing the level of coupling and coor-dination within universities without nevertheless changing the fundamental

loose-coupled nature of academic activities and the unclearness of the tech-nologies they rely on

This relies on a complex but effective collusion between university lea-ders managing universities and an academic elite created by public authori-ties in order to achieve reviews and evaluations that these public authoritieswill use to make their own decisions (whom to fund with research moneywhich research projects should be funded what training program should befurther supported)

The deal relies on the legitimacy university leaders gain vis-a-vis theacademic staff by using the external reviews produced by this elite Thisallows the university leaders to mobilize and have an effective use of the for-mal organizational capacities they received from structural reforms It pro-motes the diffusion and appropriation of the norms and preferences set bythe academic elite This process does not only empower university managersbut also impacts the definition formalization and spread of academic norms

The combination (rather than the opposition) of academic and manage-rial powers seems therefore to be a condition for the effectiveness ofstructural reforms in the governance of universities that is for institutionalpressures for change to be coupled with change in governance practiceswithin organizations This finally sheds lights on the interplay between thefield of higher education institutions and the field of peer-review-basedagencies and suggests that such interplay between fields might be moreimportant for change than isomorphic processes within one field

1 This study was then completed with a questionnaire sent to the same kind ofsample in all French universities This survey confirmed the trends toward centrali-zation the development of indicators and the increasing importance of researchactivities and productivity (Chatelain Mignot-Gerard Musselin amp Sponem 2012)in French universities but does not tell as much on behaviors and practices linkedto the reforms under study here and will therefore not be used

2 In France most universities are not pluridisciplinary After the 1968 act thatfollowed the studentsrsquo movements French universities were reorganized and in bigcities they split into different institutions that often specialized into disciplinarydomains (Sciences for some sciences and medicine for others or Humanities andsocial scienceshellip) Cf Aust (2005)

3 Deans for instance are elected for a five-year mandate It is renewable but asin France this function is neither well recognized nor powerful it is rare to accept asecond mandate rather than to go back to usual academic work University presi-dents are elected for a four-year mandate renewable once

4 In France since the Revolution the ldquofacultiesrdquo (facultes) not the universitieswere the main pillars of the French university system until 1968 From this point ofview the French universities as we know them today are very recent and were cre-ated after 1968 This explains their rather low institutional consistency as it is onlyrecently (with the introduction of the four-year contracts at the beginning of the1990s) that the Ministry recognized universities (and not the faculties) as their maininterlocutors The LRU is reflecting this evolution toward stronger institutions Toread more about the (recent) emergence of French universities as institutions seeMusselin (2004 [2001])

5 In France national research institutions have been created in order to com-pensate for the lack of research in French universities The CNRS was the first tobe created in 1936 More than 11000 researchers in all disciplines are currentlyemployed by the CNRS

6 This is very similar to the German Exzellenzinitiative that also aimed at mas-sively funding some highly selected projects (Leibfried 2010) but there are majordifferences between the two processes Letrsquos point at only three In Germany supple-mentary budgets are allocated for a defined period of time there are only three calls(one for scientific clusters one for graduate schools and one for institutional pro-jects) universities are the main level at which answers are developed In France theGrand Emprunt initiative relies on the allocation of an endowment and the avail-able budget is derived from the interest gained from this capital there are morethan half a dozen of different calls and answers are joint projects proposed bygroups of higher education institutions

7 Since 2012 research units are evaluated along six criteria (publicationsattractiveness valorization and technology transfer governance doctoral programand project) For each of them a grade is attributed (A+ A B or C) and a globalassessment of the unit is written at the end of the report The global grade issuppressed

8 Until recently the ministry run the payroll of French universities and the lat-ter only managed an operating budget

9 In France some research units are called ldquomixedrdquo because they are simulta-neously located in universities and recognized by a national research institution(CNRS INSERM INRA etc) They receive funding and staff from both sides andmay decide whether a specific research project will be run by the university or bythe research institution to which they are associated10 France is divided into four (now five) territorial zones and the AERES evalu-

ates one zone each year Each campaign is called a wave11 In French universities the academic council (conseil scientifique) is an elected

body of 2040 members composed by 6080 of permanent staff (915 of thesebeing academics) by 1015 of representatives of the doctoral candidates and1030 members external to the university It decides on the scientific policy ofthe university and on the allocation of research budgets It is also involved in careerissues for the academic staff (promotion for instance)12 This notion of research active and its definition were often criticized by those

opposed to the AERES It was first introduced in the late 1990s by the ldquoscientificmissionrdquo a pool of academic experts appointed by the ministry before the creationof the AERES but the results were not publicly used and made visible

REFERENCES

Aust J (2005) La loi Faure une rupture avortee Effets de policy feedback et creation des

universites a Lyon Politiques et management public 23(1) 5369

Barrier J amp Mignot-Gerard S (2013) Leadership et changement dans une organisation plur-

aliste Le cas des transformations drsquoune universite In I Vandangeon-Derumez amp D

Autissier (Eds) Le changement organisationnel 10 etudes de cas commentees (pp

121140) Paris Eyrolles

Battilana J Leca B amp Boxenbaum E (2009) Agency and institutions A review of institu-

tional entrepreneurship Academy of Management Annals 3 65107

Becher T (1989) Academic tribes and territories Philadelphia PA Open University Press

Bezes P (2008) The reform of the state The French bureaucracy in the age of new public

management In A Cole P Le Gales amp J Levy (Eds) Developments in French politics

four (pp 172190) London Palgrave MacMillan

Braun D amp Merrien F-X (Eds) (1999) Governance of universities in a comparative perspec-

tive London Jessica Kingsley Publishers

Brunsson N (1989) The organization of hypocrisy Talk decisions and actions in organiza-

tions New York NY Wiley

Brunsson N amp Sahlin-Andersson K (2000) Constructing organizations The example of the

public sector reform Organization Studies 21 721746

Camerati F (2014) Les transformations du systeme de financement de la recherche dans les uni-

versites britanniques PhD in sociology Paris Sciences Po

Chatelain S Mignot-Gerard S Musselin C amp Sponem S (2012) La gouvernance des uni-

versites francaises Pouvoir evaluation et identite Paris CSO Monograph Retrieved

from httpcsoeduuploaddossiersRapport_GouvernanceUniversitesFrance_2012pdf

De Boer H (2002) On nails coffins and councils European Journal of Education 37(1)

De Boer H Enders J amp Leisyte L (2007) Public sector reform in Dutch higher

education The organizational transformation of the university Public Administration

85 2746

De Boer H Enders J amp Schimank U (2007) On the way towards new public management

The governance of university systems in England the Netherlands Austria and

Germany In D Jansen (Ed) New forms of governance in research organizations

Disciplinary approaches interfaces and integration (pp 137152) Dordrecht the

Netherlands Springer

DiMaggio P (1988) Interest and agency in institutional theory In L Zucker (Ed)

Institutional patterns and organizations (pp 322) Cambridge MA Ballinger

DiMaggio P amp Powell W (1983) The iron cage revisited Institutional isomorphism and

collective rationality in organizational fields American Sociological Review 48(1)

147160

Fligstein N (1985) The spread of the multidivisional form among large firms 19191979

American Sociological Review 50(3) 377391

Fligstein N (1990) The transformation of corporate control Cambridge Harvard University

Fligstein N (2001) Social skill and the theory of fields Sociological Theory 19(2) 105125

Fligstein N amp McAdam D (2012) A theory of fields Oxford Oxford University Press

Greenwood R amp Hinings C R (1996) Understanding radical organizational change

Bringing together the old and the new institutionalism The Academy of Management

Review 21(4) 10221054

Greenwood R amp Miller D (2010) Tackling design anew Getting back to the heart of orga-

nizational theory Academy of management perspectives November 7888

Greenwood R amp Suddaby R (2006) Institutional entrepreneurship in mature fields The big

five accounting firms Academy of Management Journal 49(1) 2748

Hallett T amp Ventresca M (2006) Inhabited institutions Social interactions and organiza-

tional forms in Gouldnerrsquos patterns of industrial bureaucracy Theory and Society

35(2) 213236

Krucken G Blumel A amp Kloke K (2013) The managerial turn in higher education On

the interplay of organizational and occupational change in German academia Minerva

51(4) 417442

Krucken G amp Meier F (2006) Turning the university into an organizational actor In

G Drori J Meyer amp H Hwang (Eds) Globalization and organization (pp 241257)

Lazega E amp Wattebled O (2010[2009]) Two definitions of collegiality and their

inter-relation The case of a Roman Catholic diocese Sociologie du travail in English

53(1) e57e77 [originally published in French in 2010 Deux definitions de la collegia-

lite et leur articulation le cas drsquoun diocese catholique Sociologie du travail 52

481502]

Leibfried S (Ed) (2010) Die Exzellenzinitiative Zwischenbilanz und Perspektiven Frankfurt

Germany Campus

Meyer J amp Rowan B (1977) Institutionalized organizations Formal structure as myth and

ceremony American Journal of Sociology 83 340363

Mignot-Gerard S (2006) Echanger et argumenter Les dimensions politiques du gouvernement

des universites francaises PhD in sociology Paris Sciences Po

Mignot-Gerard S (2012) Le gouvernement drsquoune universite face aux investissements

drsquoavenir Entre reactivite et resistances Politiques et Management Public 29(3)

519539

Mignot-Gerard S amp Musselin C (2002) More leadership for French universities but also

more divergences between the presidents and the deans In M Dewatripont

F Thys-Clement amp L Wilkin (Eds) European universities Change and convergence

(pp 123146) Bruxelles Belgium Editions de lrsquoUniversite de Bruxelles

Musselin C (2004[2001]) The long march of French universities New York NY Routledge

(first published in French by Presses de Sciences Po Paris 2001)

Musselin C (2005) Change and continuity in higher education governance Lessons

drawn from twenty years of national reforms in European countries In I Bleiklie amp

M Henkel (Eds) Governing knowledge A study of continuity and change in higher

education a festschrift in honour of Maurice Kogan (pp 6580) Dordrecht the

Musselin C (2007) Are universities specific organisations In G Krucken A Kosmutzky amp

traditions (pp 6384) Bielefeld Germany Transcript Verlag

Musselin C (2012) Liberte responsabilites centralisation monograph Paris CSO and ESEN

Retrieved from (httpcsoeduuploaddossiersRapport_LiberteResponsabiliteCentra

lisationUniversites_2012pdf) Accessed on February 13 2014

Musselin C (2013) How peer-review simultaneously empowers the academic profession and

university managers Evolution of the relationships between the state the universities

and the professiorate Research Policy 42(2) 11651173

Oslashstergren K amp Stensaker I (2007) Strategic responses to the quality reform A comparative

study of changes in Norwegian higher education Manuscript prepared for the biannual

NFF conference Bergen Norway

Padget J amp Powell W (2012) The emergence of organizations and markets Princeton NJ

Princeton University Press

Paradeise C Reale E Bleiklie I amp Ferlie E (2009) University governance Western

European comparative perspectives Dordrecht the Netherlands Springer

Pontille C amp Torny (2010) The controversial policies of journal ratings Evaluating social

sciences and humanities Research Evaluation 19(5) 347360

Ramirez F (2006) The rationalization of universities In M-L Djelic amp K Shalin-Andersson

(Eds) Transnational governance Institutional dynamics of regulation (pp 224245)

Reale E amp Primeri E (2013) Reforming universities in Italy Towards a new paradigm In

C Musselin amp P Teixeira (Eds) Reforming higher education Public policy design and

implementation Dordrecht the Netherlands Springer

Streeck W amp Thelen K (Eds) (2005) Beyond continuity Institutional change in advanced

political economies Oxford Oxford University Press

Thornton P H amp Ocasio W (2008) Institutional logics In R Greenwood C Oliver

K Sahlin amp R Sudaby (Eds) The Sage handbook of organizational institutionalism

(pp 99119) London Sage Publications

Thornton P H Ocasio W amp Lounsbury M (2012) The institutional Logics perspective a

new approach to culture structure and process New York NY Oxford University

Whitley R (2007) Changing governance of the public sciences The consequences of establish-

ing diverse research evaluation systems In R Whitley amp J Glaser (Eds) The changing

governance of the sciences The advent of research evaluation systems (pp 327)

Dordrecht the Netherlands Springer

Whitley R (2008) Constructing universities as strategic actors Limitations and variations In

L Engwall amp D Weaire (Eds) The university in the market (pp 2337) Colchester

UK Portland press limited

FUNDING INDIVIDUALS CHANGING ORGANISATIONS

THE IMPACT OF THE ERC ON

UNIVERSITIES

Jakob Edler Daniela Frischer Michaela Glanz and

Michael Stampfer

ABSTRACT

University governance is constantly challenged by changing expectationsand contexts New prestigious and well-endowed funding schemes areone possible source of pressure for change of university governance Thisarticle analyses the impact of one such scheme the grants of theEuropean Research Council (ERC) on the governance of European uni-versities After outlining a model of how this impact on universities canbe expected to occur we present the results of an exploratory study at avery early stage of the ERCrsquos existence (20102011) The empiricalanalysis is based on an investigation of 11 universities in eight countrieswhich shows that different kinds of universities are affected in varied and

ISSN 0733-558Xdoi101108S0733-558X20140000042003

often unexpected ways with particular differences arising at differentlevels within the universities

Keywords European Research Council project funding impact uni-versity strategies multi-level impact

INTRODUCTION

The European Research Council (ERC) was founded in 2007 as a science-led funding body intended to support lsquofrontier researchrsquo carried out byindividuals and their teams on the basis of an open competition acrossEurope The sole selection criterion for funding so far has been scientificexcellence Within its overall mission to improve scientific excellence andthe conditions under which it is produced within Europe the ERC alsoexplicitly wishes to lsquohelp universities and other research institutions togauge their performance and encourage them to develop better strategies toestablish themselves as more effective global playersrsquo (ERC MissionStatement)1 mainly through its procedures benchmarks and transparencyof results As the ERC schemes are an attractive income source and carrierof considerable prestige and reputation organisations have an obviousincentive to react to the new European instrument

The starting assumption of our analysis2 is that there are two distinctways in which organisations can respond to the ERC First they are likelyto react strategically to the fact that there is a new European excellencedriven instrument on offer that changes the overall incentive structures oforganisations and provides a pan-European benchmark Second once uni-versities have a grantee they are likely to adjust to the fact that they haveERC grantees endowed with prestigious and generous grants It is impor-tant though to note that universities and similar public research organisa-tions are limited in their strategic capabilities and ability to exercise centralmanagerial control over work processes in a comparable manner to manyprivate firms in market economies (Musselin 2007 Whitley 2008 Whitley ampGlaser this volume)

Against this background the objectives of this article are twofold It firstoffers a conceptualisation of how the ERC could affect universities as orga-nisations developing an impact model and operationalising its major inter-vening and independent variables Second it presents the results of aninitial exploratory application of this model analysing impact at a very

78 JAKOB EDLER ET AL

early stage of the ERC existence (20102011) We base our analysis on aqualitative research programme investigating eleven universities in eightcountries based on documentary analysis and an interview programme

Starting with the core question of how to conceptualise the impact ofthe ERC on universities we further address a number of empirical ques-tions What changes are induced due to the relative value universities attri-bute to the ERCrsquos new funding instruments and to the fact that they havegrantees What is the awareness and relative meaning of the ERC grantsfor strategic leadership How are the different levels within the organisa-tions affected What context and organisational conditions and transfor-mations mediate the effect of the ERC grants on organisations

The paper is structured in three parts The first develops the conceptualmodel the second summarises the methods used in our empirical studywhile the third presents our exploratory empirical findings structuredaccording to three different stylised types of organisations and three typesof impact The conclusion discusses these findings in terms of the maindimensions of the previously developed impact model

CONCEPTUAL MODEL

In this section we present the various elements of the conceptual model tobe used to operationalise and measure the likely impact of the ERC on uni-versities We begin with some reflections on the nature of universities asparticular kinds of organisations that have limited strategic capabilitiesespecially in their core functions of research and teaching

Universities as Strategic Actors

Universities today are facing a complex set of challenges and requirementsthat is well documented in the literature The New Public Management(NPM) influenced shifts of the past 20 years have slowly but steadily chan-ged the way many universities see themselves and are seen by others (egAghion Dewatripont Hoxby Mas-Colell amp Sapir 2008 Clark 1998Ferlie Musselin amp Andresani 2008) They are increasingly subjected todemanding performance requirements broadened functions such as lsquothirdmissionrsquo (Laredo 2007 Nedeva 2007) and efficiency pressures and arehaving to compete nationally and internationally against each other for

79Funding Individuals Changing Organisations

prestige resources and staff As a consequence many universities attemptto manage research activities that is to define and support effectiveresearch portfolios and establish structures and processes that are attrac-tive for excellent researchers of all career stages (eg Schmoch amp Schubert2010) It is this understanding of universities as organisational actors thatunderpins the ERCrsquos wish to improve the effectiveness of universities

However when analysing the impact of an external impulse such as thenew funding schemes of the ERC on the strategic action of universities weneed to understand the inherent limitations of universities as actors Theorganisational actorhood of universities can be understood lsquoto imply uni-fied central authority over the design of work processes the co-ordinationof their outputs and the development of collective capabilities for dealingwith problems adapting to change and seizing entrepreneurial opportu-nities through mobilizing the commitment of skilled staffrsquo (Whitley 2008p 26 see also Whitley amp Glaser this volume) Whitley argues that thedominant authority of scientific global communities in assessing the rele-vance and excellence of science and quality of scientists implies thatresearch strategies and their outcomes cannot be fully planned and mana-ged by university managers Moreover the frameworks for assessment andattribution of excellence and rewards are different in different areas ofknowledge production and thus extremely heterogeneous across universi-ties This leads to lsquohellip limitations on the ability of universities as employersto organize and direct particular divisions of scientific labour and integrateresulting outputs around specific organizational purposesrsquo (ibid) and thusto restrictions of centralised authority and collective actorhood As a resultthe management of research and advanced teaching generally occurs at thelevel of sub-units within the organisation

This overall assessment of strategic capabilities and limitations needs tobe differentiated for different types of organisations (Van Vught et al2005) that vary according to the overall level of autonomy granted to uni-versities by different states their established governance structures socialand scientific prestige the culturally embedded inclination of academicsand researchers to function within lsquomanagedrsquo systems and the wealth oforganisations This last the level of resources at the disposal of the organi-sational leadership considerably determines its ability to steer as largeshares of institutional funding or large endowments can be used for settinginternal incentives while on the other hand a transition to project basedfunding shifts the logic of resource acquisition to the level of research unitsand individuals All this implies that the actorhood of universities is deeplyshaped by the regulatory and political frameworks of national science

systems (Bleiklie amp Kogan 2007 Estermann amp Nokkala 2009 Geuna ampMartin 2003 Jongbloed 2009 Observatory of the European University2006 Oquist amp Benner 2012 Paradeise Reale amp Goastellec 2009Whitley 2008) With these qualifications for strategic actorhood in mindwe can now turn to the development of our impact model for the ERCgrants on universities We start with the independent variable the ERC itsobjectives in wishing to affect universities and its instruments

The Independent Variable The ERC its Mission for ResearchOrganisations and its Instruments

In its official mission statement the ERC is defined as the first Europeanfunding body set-up to support bottom-up investigator-driven frontierresearch which is intended to substantially strengthen and shape theEuropean research system There are three major ways in which it aims toaffect universities to help them lsquogaugersquo their performance as research orga-nisations to improve their strategies and to stimulate investment in sup-porting promising new researchers

Three General Dimensions of ERC ImpactThe ERC affects organisations by offering them three kinds of value mate-rial symbolic and normative3 The material value of grants can make a sig-nificant difference not only for the individual but also for the organisationand its sub-units At the time of this investigation the ERC offered twodistinct instruments to fund investigator-driven frontier research namelythe Advanced Grants scheme and the Starting Grants scheme StartingGrants aimed at boosting the career of the next generation of research lea-ders by promoting early scientific independence and required a respectivecommitment by the host institution that allowed potential grantees to inde-pendently direct their research and administer their resources AdvancedGrants were intended to support researchers who had already establishedthemselves as exceptional leaders in their fields with a track record of sig-nificant achievements over the past ten years ERC grantees receive up toh20 million for Starting Grants StG and up to h35 million forAdvanced Grants AdG both for a period up to five years Budgets inboth cases can be handled flexibly giving the grantees a powerful andindependent position Moreover ERC grants are portable between organi-sations and across countries4

On the symbolic level ERC grants have already developed high prestigeand reputation both for the grantee and the hosting organisation To get orto host ERC grants is considered a measure of excellence and academiccompetitiveness Lastly the ERC exerts normative influence on individualsepistemic communities research organisations and funding organisationsin Europe by providing a definition of risky and excellent research as wellas an understanding of the role and autonomy of (young) researcherswithin organisations all of which underpin the justification of the ERCand its instruments

Two Impact LogicsThe potential impact of the ERC and its new funding instruments on orga-nisations can be understood on two levels (1) the consequences of havinggrantees and more generally (2) the consequences of the sheer existence ofERC and its offer of European excellence grants on the competitive contextof organisations and thus their strategic positioning

The Consequences of Having Grantees First we assume that granteeswithin a university will make a difference due to the ERCrsquos material sym-bolic and normative value Universities may react at different levels inorder to create the legal financial career related or infrastructural environ-ment that is most favourable for the grantee to be successful ERC offersfunds of considerable size and at favourable conditions While within theresearch budgets of large organisations the individual grant may seem rela-tively small a number of awards actually can make a difference Evenmore when we focus on sub-units within organisations as this additionalsource of income may change their situation considerably both in a sup-porting and in a turbulence-stirring way

The Consequences of Having the ERC and its Instruments as such Secondwe assume that the very existence of the ERC and its offer of excellencegrants on a European level makes universities react even if they do not(yet) have grantees The logic here is that the ERC through its symbolic(ie high prestige coming with getting a grant) and normative value(through contribution to the definition of excellent research and position ofyoung researchers linked to the grants) changes the context and incentivestructures of organisations potentially leading to strategic reactions andorganisational change

Capturing Impact Operationalising the Dependent Variables

We can now further differentiate the dependent variables and define cate-gories of impact while sticking to the three explicit objectives of the ERCas regards research organisations As mentioned above the ERC seeks tohelp universities to improve their performance enhance their strategic cap-abilities and attract (and nurture) scientific researchers In this context

(1) improving (perceived and expected5) performance relates to quality andquantity of research output (impact level of journals setting newresearch agendas etc) to increased visibility of the university or indivi-dual units to extraordinary growth of research groups and to theincreased capability in raising further research income (high-levelgrants)

(2) enhancing strategic capability involves the development and implemen-tation of strategies to succeed in a global competition for staffresources and reputation This includes attempts to raise the profile andcoherence of organisations and to increase regional and global recogni-tion as well as strengthening or building-up of organisational capabil-ities to define goals and develop structures and processes to achievethem

(3) supporting and attracting talent means to support lsquopromisingrsquo research-ers and to increase the capability to successfully incorporate them intothe universityrsquos research environment This also translates morebroadly into the creation of contextual conditions and supportmechanisms within organisations that best suit (current or future) gran-tees in their drive to set up and implement challenging frontier researchprojects which might need new levels of autonomy and co-operation toallow for different kinds of research This furthermore includes(intended or sometimes unintended) effects on staff retention that isstaying attractive for both grantees and other already existing staff notdirectly profiting from an ERC grant

This operationalisation of the dependent variables constitutes the basisfor the empirical programme

Determinants of Change Defining Three Bundles of Intervening Variables

We now turn to the broad range of intervening variables that can influencehow the ERC can be expected to affect universities In this study these are

derived from the major dimensions of change elaborated in the compara-tive analysis on national reform programmes as regards governance of uni-versities in Europe in the book by Paradeise Reale Bleiklie et al (2009)(2009) and Paradeise Reale and Goastellec (2009) This comparative analy-sis has confirmed the overriding importance of power and competence struc-tures to adapt (autonomy of universities) endowment with both financialresources and human resources and ways in which those financial and humanresources can be allocated Consequently we focus on three intervening fac-tors in the unfolding of the ERCrsquos impact on universities autonomy leveland allocation mechanisms of funding and human resource policy

First the level of organisational autonomy This refers to the ability ofuniversities to determine their goals and procedures independently fromthe state as granted by the national framework and established in a his-torical trajectory This also includes the autonomy of sub-units withinorganisations which may limit the development of organisational capabil-ities at the centre but may also enhance their ability to (re-)act to theexternal impulse of the ERC National frameworks if granting formalautonomy do not automatically translate into actual autonomy At thesame time external pressures to develop organisational strategies are ris-ing both through regulations and increased competition (Fumasoli ampLepori 2011 Krucken Kosmutzky amp Torka 2007) The ERC as a newactor in this context adds to the opportunity environment by providingsupport for promising individuals and frontier scientific research(Nedeva amp Stampfer 2012)

The second set of variables has to do with research funding namely theoverall endowment with funds and the conditions under which these fundsare allocated to research We can see a considerable shift to conditionalprogramme funding that alters the rationale and degrees of freedom forresearch within organisations Universities need to adapt their strategiestowards higher funding income and diversification of activities and fundingsources External sources of various kinds probably have gained a higherinfluence on the distribution of remaining internal funds (eg Granberg ampJacobsson 2006) Actor strategies vary across organisations and can beseen as a kind of indicator for the intended positioning of different actorsincluding organisations and countries and their readiness to change(Braun 2011) As regards the ERC organisationsrsquo record with other fund-ing sources experiences as well as options might be responsible forimportant differences in ERC success ambition and ability to incorporatethe changes induced by ERC funding

Third the last bundle of variables is clustered around human resourceissues the people performing research This includes not only the quality ofpeople per se but also the way that universities can and actually do orga-nise recruitment and reward procedures and develop personnel strategiesto best serve their overall goals Inward mobility has increased though dif-ferently in different countries (Franzoni Scellato amp Stephan 2012) andunder certain constraints universities are becoming more and more fullyresponsible for their (scientific) staff As part of national autonomy reformsall over Europe the management of human resources at universities hasbecome one of the central tasks performed at the organisational level(Bleiklie 2011 Enders 2001 Musselin 2005) In this context the ERCadds new opportunities for younger researchers raises the autonomy ofresearchers and might allow for catalytic change

We do though need to stress that individual organisational contexts mat-ter strongly as changes in overall governance and fundinginstruments meandifferent things to different organisations Universities implement externallyinduced reforms in different ways as they are facing unequal starting condi-tions Adaptation pressures may be enforced blocked or re-directed throughvaried organisational settings endowments and existing prestige and excel-lence While path dependency may account for many differences it alsodemonstrates that universities are quite stable organisations where change israther incrementally implemented and mostly resource driven

The Model

Our resulting conceptual and analytical model of the ERCrsquos impact onresearch organisations is summarised and illustrated in Fig 1 We considerthe ERC and the material symbolic and normative value it offers as anexternal impulse acting on organisations Here we focus on two basicimpact logics namely the consequences of having grantees and the conse-quences of having ERCrsquos offer of European excellence grants as suchHowever such impact is shaped catalysed or hindered by additional inter-vening factors Here we consider three intervening aspects highly impor-tant (i) autonomy (ii) level and allocation mechanism of funding and(iii) human resource policies and practices

Following the ERCrsquos mission to improve the performance of universitiesenhance their strategic capabilities and attract (and nurture) researcherswe conceptualise impact along the following three dimensions (i) improving

(perceived) performance (ii) enhancing strategic capability of universitiesand (iii) attracting (and retaining) promising talent6 Further all the impactsare to be differentiated for the overall organisation and the organisationalsub-levels (faculty department institutehellip)

A few further specifications to this model First there is a certain inter-play of variables and dynamics Thus note that arrows in Fig 1 only indi-cate the direction of overall impact logics without illustrating the moredetailed feedback loops and relationships between variables (in a certainpoint of time and over time) Second we assume that the discussed aspectsof autonomy funding and human resource are strongly related to the over-all pre-existing capability of universities to act and to successfully competeas to academic performance and reputation as well as for additional fund-ing and outstanding talent It is thus obvious that our intervening anddependent variables do overlap to a certain extent Organisations have ahistory and history of course does matter This aspect is particularlyinteresting but also tricky in the context of ERCrsquos impact on universitiesThe Matthew principle seems strong but can also more generally beapplied to the attracting talent dimension of our model a considerablenumber of high-level people do attract more high-level people etc

Third as often with research designs we do face an attribution problemWe need to consider that universities are facing additional pressures toadapt and transform themselves While we assume that the ERC indeed canmake an additional difference in this transformation process the attributionof effects is quite difficult (as with all impact assessments) This means that

via the

allocated

grants

via the offer

of European

excellence

grants

perceived

PERFORMANCE

REPUTATION

Intervening variables Dependent variablesIndependent variable

iffere

isatio

ENDOWMENT FUNDING

country characteristicshellip organisational charact

HUMAN RESOURCES

AUTONOMY

STRATEGIC

CAPABILITY

other conditions

rial s

SUPPORTING

ATTRACTING

TALENT

Fig 1 The Basic Analytical Model

a solid understanding of the framework conditions organisations areencountering is thus crucial to finally assess the relative importance of theERC and the additional effect it has in relation to other pressures andopportunities We are aware that there might be impacts on organisationsthat are not covered in those three dependent variables and we are awarethat other developments might have effects in the same direction and ERCmight simply be one force out of many With respect to the results of theempirical investigation which is largely qualitative and based on a casestudy approach we emphasise that the ERC is one impulse for change oradaptation among many others

METHODOLOGY

The empirical sample comprises 16 organisations in 8 countries 11 univer-sities and 5 public research organisations This article however focuses onuniversities only Cases have been selected to cover a range of differentfunding environments size of organisations as well as a certain variety intheir strategic foci and actor capabilities as well as could be establishedex ante Although our focus is on universities hosting grantees to capturethe changes induced by having a grantee we included two universities with-out grants (referring to the first cohort of ERC grantees) Countries repre-sented in the sample establish a balanced mix of ambitious scientificallywell-performing countries moderately ambitious medium-performingcountries and countries with low ambitions and rather low performance(Braun 2011)7 Furthermore they show a range of governance patternsand HR frameworks that impinge on the way the ERC instruments canexert influence The different national conditions represented in these coun-tries and the role they play as intervening variable have been taken intoconsiderations in the university case studies We are aware that all cross-national university comparison needs to be seen in the light of the interven-ing national conditions and thus treat cases always as the combination ofnational and organisational framework conditions It is however beyondthe remit of this paper and its underlying research to deliver a systematicanalysis of the different national science systems and framework conditionswithin them

For our university sample in total 46 informed interviews8 have beenconducted on the basis of an interview guide building upon the operationa-lised variables as introduced above During interviews our major research

interest focused on perceptions of actors within respective organisations onhow the ERC has influencedinfluences organisations within the above-mentioned dimensions To abstract from individual opinions and to differ-entiate between formal and actual capability to act interviews have beenperformed on different levels within organisations (top intermediaryfaculty functionaldepartment administrative level) Finally we differenti-ate between four representations of lsquochangersquo (i) statements about reactionsthat actually represent real change (ie learning) (ii) responses claimingchange but rather representing window-dressing (iii) intentional non-change and (iv) change that is induced not by the ERC but through otherpressures and challenges

Analysis of empirical material was based on the methodological assump-tions of qualitative content analysis (according to Mayring 2008) allowingfor systematic analysis of empirical data which at the same time isinterpreted by taking into consideration the respective context (organisa-tional and country-specific intervening variables) Based on the previouslydefined research questions a preliminary set of analytical categories hasbeen defined In contrast to closed category-systems as proposed byMayring this category-system was refined in the course of our analysisuntil all relevant text passages (related to the impact of the ERC on organi-sations) were adequately recorded this was intended to allow for amechanism-oriented investigation of causal mechanisms (Laudel amp Glaser2010) The extraction of information relevant to answer our research ques-tions rests upon this iteratively refined category-system

We must stress that our empirical analysis is exploratory it serves thepurpose to develop and test the impact model and develop ideas for itsfuture improvement and application Moreover interviews were performedat a time too early to fully assess the performance difference the ERCmakes on universities in terms of scientific output prestige and furthergrant getting However the analysis could identify early signs of changeand gives indications that the symbolic and normative value of ERC doessignificantly matter

EMPIRICAL FINDINGS ON IMPACT PATTERNS

A Systematic Link Pre-Existing Organisational Capacity and ERC Impact

Our conceptual model assumed that the impact the ERC can have onresearch organisations is strongly moderated by a set of intervening

variables namely the organisational and country frameworks that deter-mine the organisational capabilities and positioning of the organisationOverall our exploratory empirical work has confirmed this assumption Itappears that there is a systematic connection between the pre-existing cap-abilities of organisations on the one hand and the nature scale and scope ofimpact the ERC has on the other hand

In particular it is possible to reduce complexity for the impact analysisby grouping the universities in our sample on the basis of our ex post analy-sis into three categories that vary in their resource endowment and co-ordination capabilities powerful in-between and powerless The assignmentof the organisations to one of these three categories was based on ourassessment of the three key bundles of both national and organisationalcharacteristics we consider important as intervening variables shapingERCrsquos impact that is (1) autonomyability to act as an organisation (atorganisational or sub-unit level) (2) pre-existing research fundingendowmentand (3) recruitmenthuman resources policies In our view these threedimensions altogether are assumed to define a universityrsquos overall organisa-tional capacity9 For the initial sample selection a simple strategy had beenapplied in order to cover a sufficient range of organisations with respect toorganisational types (universities non-university research organisations)coverage of fields (full coverage specialised into certain fieldsdisciplines)and number of hosted ERC grantees within the first cohorts of Startingand Advanced Grants (hosting more than 5 grantees up to 5 grantees no grantees) A first empirical analysis showed that certain kinds of impactare correlated with a specific set of manifestations of our intervening vari-ables This finding then was the basis of our definition of the stylised orga-nisational baskets and a preliminary allocation of organisations to thosebaskets This allocation has been further refined during the later stages ofthe analysis Because of the heterogeneity of the sample organisations inthe same category still differ considerably as regards individualorganisational characteristics10

More concretely for each intervening variable an ordinal scale wasdeveloped which allowed us not only to consider the statistical facts ofeach organisation (gathered by desk research) in more detail but to furtherinclude the more subjective impressions and perceptions of our intervieweeson their universityrsquos organisational capacities The scales were then com-bined into a summary scale and translated into our three categories ofintervening parameter sets as described above This categorisation has beeniteratively developed during the course of the qualitative content analysisof the collected interview material where we further decided to differentiate

with respect to essential supportive and hindering intervening parametersAs a result each of the three categories contains at least three organisationsfrom at least two different countries whereas several countries are repre-sented in more than one basket This is far from constituting a comprehen-sive sample but goes beyond a single-organisation case study approach andjustifies a certain approximation of results referring to the organisationaltypes

On this basis a typical stylised powerful organisation is well endowedand can draw on considerable funds available to perform research11

Furthermore such organisations are characterised by a high level of orga-nisational autonomy allowing them to actively respond and act to chan-ging environments Finally these organisations do have a clear and wellimplemented recruitment and human resources development strategy foryears that has already resulted into hiring of a larger pool of highly reputedresearchers by offering favourable highly competitive employment condi-tions (including salary levels) as well as incentive oriented human resourcesdevelopment activities This of course is only possible if organisations facea considerable degree of autonomy and can dispose of important researchendowments Overall those organisations regard themselves and areregarded by others as delivering excellent research

A typical stylised in-between organisation is characterised by a substantialoverall endowment for research that is however constantly under pressureIt shows some capability to realise and act upon emerging opportunities toimprove its own status but its strategic actor capability is limited (for his-torical institutional andor legal reasons) While those organisations oftenmay not have an elaborate formal recruitment or human resources develop-ment policy at the top organisational level on the level of operating unitsrecruitment can be quite systematic Furthermore within these organisa-tions we find acknowledged fields of excellence but often these are limitedto specific areas andor not yet fully visible In sum those organisationsboth have some basis for research excellence and for strategic actor capabil-ity to respond to the external lsquoshockrsquo of a new high level and highlydemanding ERC funding scheme

A typical stylised powerless organisation is characterised by a low levelof research funding available beyond that available funding is probablyearmarked for other tasks or cost categories than research activities Thissituation is accompanied by low levels of organisational autonomy andstrategic leadership further translating into a lack of capability to ade-quately handle human resource policies at the organisational levelHowever this lack of overall strategic capability to act as an organisation

does not necessarily imply that weak organisations are not able to performhigh-quality research and host top researchers in certain individual areas orpockets of excellence However these mainly individual pockets of excel-lence do not put an excellence mark on the organisation as a whole Theallocation of organisations to those categories is a supporting mechanismfor the analysis and generalisation as a result of the analysis as we havefound some very generally speaking systematic relation between thosebaskets and ERC impact12

Impact on Powerful Universities

(Perceived) Performance and VisibilityIn some of these organisations the ERC has led to a new discourse on whatconstitutes quality and how people should be assessed as well as on howgovernance structures should be set up to best support scientific qualityFurther many interviewees see ERC grants due to funding conditions andduration that are more generous than most national grants as potentiallyenhancing the quality of the research more generally The ERC is here per-ceived as a visible indicator of a catalyst for quality This is a unanimousassessment All interviewees at all levels confirmed that ERC grants areseen as a new and more telling indicator for research excellence for somethe ERC is even regarded as a means to overcome shortcomings they per-ceive with traditional organisational rankings or assessments Thus theseorganisations aim at confirming their position within the Europeanresearch landscape through the number of grants they receive Elite organi-sations with a high number of grantees strive to keep this level whereasthose with lower number of grants than comparable organisations see astrong need to catch up in this indicator and they have the means and theorganisational capabilities to do so

Interviewees consider the pre-existing quality as most important pre-condition for successful grant getting As top organisations are inter alia defined through having high reputation and adequate support structuresalready many of the interviewees do not see a special need for an additionalinvestment in ERC activities Further the difference the ERC makes isdependent on the existence of similar types of funding instruments in theorganisationrsquos home country If there has been such an instrument beforetop organisations have already been proficient in integrating these externalimpulses in their internal processes

Strategic CapabilitiesThe impact on strategic capabilities at the leadership level in powerful orga-nisations is not obvious While leadership has and exerts the capability torecognise and strategically incorporate emerging research fields and newlydeveloped research lines the major decisions and conditions with respect toresearch are set at a lower organisational level the operational effects arethus at the lower (and sometimes intermediate) level Even for elite organi-sations research strategy is often developed within constraints of fundingconditions and assessment exercises and their specific output-oriented anddiscipline-oriented criteria and it is often highly path-dependent The well-endowed and highly prestigious ERC grant allows a change of researchdirection which often necessitates investment of time and patience that isnot possible in more traditional funding environments With ERC lsquowe canbe much bolderrsquo and do things lsquowe could not do otherwisersquo An importanteffect here even in well-endowed organisations is lsquomaterialrsquo as ERC contri-butes to the funding of sub-units that are competing for institutional andproject grants This allows for more flexibility in one case even securingthe sustainability of a centre This overall impact of the ERC is especiallyimportant for top performing organisations that are growing quicklywhere a number of grants can be catalytically and accelerate the process oftransformation and profiling in this growth period

In terms of thematic bundling and profiling contrasting impacts are to beobserved On the one hand ERC grants have helped to mobilise and bene-fit from platforms with long term risky trajectories This was possiblethrough multiple grants in related areas On the other hand thematicbundling through ERC grants is structurally hampered as it becomes onlypossible through ex post bundling but not through the submission of jointgrants Thus it has limited impact on thematically integrative research stra-tegies of larger units faculties or large divisions which are planning to buildnew horizontal teams along clearly defined thematic lines

In powerful organisations the ERC is increasingly used as a benchmarkof units within the organisation and thus is a tool for internal competitionfor prestige and grants In two examples serial ERC successes haveseverely enhanced and strongly confirmed the reputation of a specific areawithin a top organisation In other cases specific care is taken that the ERCsuccess is reported not only externally but within the organisations expect-ing an improved bargaining position at higher levels and vis-a-vis otherunits

All of those aspects have already led to the common perception in sub-units of powerful top organisations that efforts should be put towards

getting those grants Some interviewees stressed that ERC grant gettingsupport within their units is more pronounced than support for national orEU Framework Programme funding Thus we see some investment in orderto be visible as ERC attracting unit in these organisations While in thesecases we did not observe fundamental changes in the way internal supportof research quality and income generation is organised they take advan-tage of existing structures and have strengthened and tailored them wherenecessary

Human Resource Policy Support Attraction and Retention of PromisingResearchersIn general terms the ERC as a means to increase attractiveness for organi-sations is perceived as very positive it fits into the general claim of eliteorganisations to recruit support and retain high performers However asregards the toolbox of human resource management most intervieweesreported no major changes The existing structures and incentives to attractand support talent are merely extended to ERC grantees as existing prac-tices are considered efficient enough to actually perform the task of gettingand lsquoservicingrsquo high potentials For these organisations the ERC providesfor an additional (symbolic and normative) currency in the market amutual signalling device for the individual researcher and for employingorganisations but the key criterion currently is and will remain output andpublications In some cases the larger grant is used to pay an additionalbonus for the researcher who succeeded in the ERC competition

In terms of actual recruitment the ERC has not yet had a significantimpact In our interviews cases of recruitments of grantees are not particu-larly common instead most grants (AdG and StG) were submitted byresearchers who have already been employed by the respective home insti-tution for some time At the unit level however there are additional effortsundertaken as some see a need to recruit ERC grantees or at least potentialERC grantees At times the ERC grant was even regarded as a sufficientquality stamp reducing the number of formal interviews in an ongoingrecruitment process In addition for some organisations the ERC repre-sents a welcomed opportunity to retain high-performing staff as by meansof ERC (and ERC money) generous prolongation of existing contracts canbe handled more easily

Even in elite organisations individual strategic leaders at the lower levelsof the university hierarchy perceive a potential downside of ERC grants Firstit puts normative pressure as to the conditions granted to high-performingindividuals as it potentially enhances the autonomy of individuals and thus

limits the autonomy of their strategic managers Leaders within top organisa-tions see a second downside caused by the fact that large ERC grants actu-ally can endanger the balance within units and groups and distort establishedlevels of cohesion This is specifically true when the grant allows for multiplerecruitments by the PI and thus leads to the development of research groupsthat is not planned for strategically and may not spill over to the unit moregenerally (leading to a lsquotwo class systemrsquo) In these cases unintended effectson retention of existing staff might be an issue however this needs to beinvestigated further Moreover in a comparable situation one organisationhas started to be restrictive with grant applications taking care of a balancedgrowth rather than an accidental growth within their portfolio Internalpre-selection procedures do play a decisive role here The above-mentionedchallenges of large grants especially if units have attracted multiple grantstrigger new efforts and set-up of structures to plan handle and support largeproposals and grants in sub-units by simplifying internal rules and proce-dures Even in organisations that have been reluctant to apply for externalfunds as not many schemes meet the criteria of their research staff arenow requested to consider ERC applications

Impact on in-between Universities

(Perceived) Performance and VisibilityAll interviewees agree on the validity of ERCrsquos quality brand that arisesfrom its symbolic and normative value ERC is expected to impact on qualityof research as it allows researchers to focus on basic research withoutfurther conditions However for in-between organisations ERCrsquos impacton the perceived performance of organisations or sub-organisational levels(divisions departments institutes labs groups) is closely linked to whetherthese are hosting ERC grantees or not Having (the right number of) ERCgrantees is perceived as bringing visibility and reputation for the respectiveorganisation or sub-division internally as well as externally (nationallyand internationally) With a growing number of ERC grantees acrossEurope the pressure on those organisations and sub-units within themincreases to succeed in order to further assure their visibility a commonlyused statement in this respect lsquoWe cannot afford not to succeed (in thenext round again)rsquo

Further the ERC allows certain units or research groups within anorganisation to grow extraordinarily or faster than comparable units with-out any grantees While more generally in-between organisations see this as

an opportunity for a more competitive research performance in the comingyears some of them are challenged by an inherent risk of uncontrolled andimbalanced growth triggered by multiple ERC successes pushing organisa-tions towards their limits as regards allocation of necessary space supportor infrastructure

Strategic CapabilitiesEvidence of ERCrsquos impact on the overall strategic capability is mixed Inour sample of in-between organisations there has been no evidence that amajor re-orientation of research strategies or completely new research prio-rities have come up due to ERC However often through ERC grantsalready ongoing policies and research priorities have been (i) made visibleor justified (ii) considerably strengthened in their research capacities and(iii) in some cases used as seed money directed into new more risky lines ofresearch within the field For the in-between organisations this chance ofincreased visibility also serves as a driver to adapt internal incentive struc-tures to motivate existing staff to apply for ERC grants Furthermore dueto the considerable size and the favourable funding conditions of ERCgrants research may be re-organised in a quite flexible and non-bureaucratic manner ERC grants are considered an opportunity to catch upwith excellence strategies and helping to raise a unitrsquos research capabilitiesbe it via new infrastructure via building-up critical mass via extraordinarygrowth of research groups via more incentive-driven overall financialendowment but also via new and different research methodologies orapproaches to be applied As regards the latter ERC money primarily inthe case of Advanced Grants also catalyses new more risky lines ofresearch (within already existing priority research areas) particularly as itcomes with a favourable time span In many cases ERC money acts as seedmoney and allows units or research groups within an organisation to growextraordinarily or faster than comparable units without any grantees

As ERC success is considered an easily measureable and valid qualitycriterion for research it is increasingly used for strategic decisions also at thetop level from the human resources domain to the internal distribution ofmoney Organisations allow faster promotion and leap-frogging of careersteps for ERC grantees or provide additional money or infrastructure forhisher research group or department This is not only handled ad hoc butin some cases integrated as additional momentum into existing evaluationstrategies for people as well as for organisational units Furthermore orga-nisations do not only integrate the actual ERC success into strategic con-siderations but also reflect on ERCrsquos funding criteria and thus the

perceived likelihood for winning an ERC grant in already established qual-ity assurance mechanisms

All in all while the strategic merit of the ERC more likely unfolds at thesub-organisational levels of organisations (faculties departments instituteslabs) the perceived material and symbolic value of the ERC has triggeredstrategic reaction at leadership level as well There is a stronger awarenessthat the ERC is a means to catch up as well as to signal this catch up morebroadly The ERC is used as a trigger for reinforcing quality strategies andinternal allocation for resources in some cases the university allocates addi-tional money to the groups or sub-units of grantees in others ERC grantsare considered a useful indicator for quality based allocation of competitiveuniversity budgets The main challenge for the university leadership lies inbalancing these conflicting strategies of sub-levels a challenge that is notyet met by all organisations

Human Resource Policy Support Attraction and Retention of PromisingResearchersIn the realm of organisationrsquos employment and HR policy the existence ofERC and of having ERC grantees triggers diverging promotion structuresERC grantees are by means of being branded as ERC-excellent oftenentitled to leap-frog certain career steps that is are promoted to positionswith more autonomy than colleagues of comparable academic age arealready tenured at an earlier stage or entitled to a full professorshipHowever the ways in which grantees are governed differs ranging frommere expansion of already existing exceptional promotion rules for com-parable honours to ERC grantees to newly creating staff categories andpositions for ERC grantees that substantially differ from pre-existing staffprofiles and tasks (eg research only positions in a teaching-heavyenvironment)

Even more so than in powerful organisations in-between organisationsthat seek to catch up are facing considerable challenges in this respect TheERC related reward system puts a lot of pressure onto existing staff struc-tures and organisational capabilities especially if multiple grantees have tobe integrated within a certain unit This at the same time challenges existingcohesion among the unitrsquos staff members This is not only a question ofpromotion policy on the formal level (including national legal regulationsand governance frameworks) but also closely related to the question ofmotivating and retaining existing staff and thus closely linked to the attrac-tion of talent dimension Empirical evidence in this group of organisationsshows that ERC related reputation and prestige will likely lead to further

internal differentiation among staff members Generally ERC grantees are given the amount of funding they bring and the reputation and autonomythey enjoy in a privileged bargaining position towards their host univer-sity Success in the ERC grant competition is therefore often related withmore freedom and flexibility in doing research in most cases they are addi-tionally released from teaching tasks and benefit from lsquoresearch onlyrsquo-posi-tions lsquoThe other side of the coinrsquo is growing evidence of internal imbalancesand pressures within organisations given that other staff members do notenjoy the kind of special treatment provided to ERC grantees The internaldiscrimination also increases pressure on (so far) non-successful staff mem-bers For in-between organisations those pressures are critical as large ERCgrants are potentially more disruptive than in top organisations that havebetter endowments and are already more accustomed to success in compar-able high-level funding schemes

Relating human resource strategy to research strategy our empirical evi-dence shows that a excellence-oriented strategy prevails over a (top-downpre-defined) research priorities strategy This means in essence that theurge to attract and retain top class researchers is the highest priority thedefinition of a research strategy follows from that Catching up in-betweenorganisations basically employ bottom-up strategies and allow strong fieldsto grow and become more important However in parallel they start todevelop the strategic capability to incorporate these into the overallresearch strategy

In terms of overall professionalisation of HR the ERC appears to be acatalyst but not the main driver for ongoing trends in terms of HR man-agement In a number of organisations we found strong previous experi-ence with individual grant schemes or prizes of comparable size andcomparable selection criteria on the national level Applying for ERCgrants or having attracted ERC grants only strengthens and extends exist-ing mechanisms Examples for this re-enforcement are internal peer com-mittees for developing proposals further in terms of content as animportant way of supporting for potential ERC grantees These peer com-mittees play an additional role in terms of assuring internal quality controlover research conducted by potential grantees

This use of internal reviewers as a filter for ERC applications can be con-sidered rather ambiguous On the one hand the ERC instrument targetsindividuals and seeks to fund risky unconventional research On the otherhand it wants to support research organisations in becoming more effectivestrategic actors and thus defines an organisational role in supporting ERCgrants and grant getting In the attempt to use the ERC strategically by

establishing supportive and quality assuring structures (and thus to enlargetheir management capability) organisations and their units could counter-act the spirit of the ERC especially as they are challenged to provide aninternal peer review procedure that is in line with and can match therequirements of the ERC procedure

Impact on Powerless Universities

(Perceived) Performance and VisibilityThe impact of ERC on powerless organisations that host grantees is two-fold On the one hand grantees are rare and therefore the overall impacton the organisation can be regarded as low On the other hand the numberof general success stories within these organisations is by definition againrather low Therefore the visibility of the ERC successes tends to be highalso when it comes to the issue of (perceived) performance and translatesinto enhanced international reputation Due to the small number of casesin these organisations the impact of ERC grants is still limited to indivi-dual sub-fields On this level however ERC grants seem to make a real dif-ference and allow institutes and labs to internationally better positionthemselves and to gain reputation more generally also for the very goodresearch being done before or beside the ERC grant

Further ERC grants bring the opportunity to establish new fields orresearch avenues In view of weak performance and organisational rigid-ities the funding volume flexibility and freedom coming with ERC grantsoften allow these organisations to bypass traditional hierarchies and estab-lished disciplinary boundaries In this respect ERC grants seem to stronglycontribute to the creation of additional pockets of excellence within weakerorganisations Impact thus can be considerable in individual places withinthese organisations due to individual ERC successes

Strategic CapabilitiesThe organisations covered in our sample face a number of equally seriousproblems limiting their strategic options These limitations are about mea-ger financial endowment and low flexibility of spending rigid recruitmentstructures problems with recruitment more generally and a lack of top-down quality policies steering and priority setting processes Thereforeimpacts on strategic capabilities appear only at the level of individualresearch groups labs or institutes In powerless organisations individualERC grantees with considerable funds flexibility and freedom are truly

outliers as they are able to change the geometry habits and perceptions sig-nificantly within their immediate surroundings

On the level of individual grants some first indications show that someeffects including new research avenues and higher degrees of flexibility interms of research organisation or administration can be traced to the fewERC grantees Grantees serve as role models and generally become highlyvisible they stand out due to the dearth and immobility reported for theirgeneral situation The fact that ERC grants are regarded as lsquoseed moneyrsquoand contribute to lsquosetting prioritiesrsquo in weak organisations suggests a lowlevel of strategic research planning at all levels

Further as some powerless organisations saw the ERC grant(s) as animportant source of income they developed considerable efforts to rushscientists into application processes The subsequent lack of success led todisillusion and emphasised existing weaknesses In powerless organisationsthe goals of strategic leaders at all levels focused on other national andEuropean instruments to build up state-of-the-art research infrastructuresIn some of our interviews the importance of the Structural Funds and itsbureaucratic burdens were even used as arguments why there was notenough time and attention available for developing stronger ERCapproaches

For powerless organisations both with and without grantees a generalobservation can be made First there is no impact on the overall organisa-tional setup The organisations investigated share (i) a high degree ofdecentralisation and (ii) lack of vigour as regards priority setting recruit-ment provision of incentives or quality policy These practices have notgone through considerable changes due to ERC as such or through itsgrantees Overall in such organisations the symbolic and material value ofthe ERC alone is by far not sufficient to induce the desired change

Human Resource Policy Support Attraction and Retention of PromisingResearchersPowerless organisations are in general not in a position to offer favourableconditions for employment and usually lack systematic recruitment policiesAs a consequence also the perception of potential future key employeesabout this working environment tends to be rather low A powerless organi-sation is no place to be while attractive organisations can choose amongmany strong applicants or successfully approach top scientists

Change can be and is being introduced on a lower level but mainlyerratically and unplanned These organisations try to attract researchersfrom their own countries who have started or already successfully

mastered their career abroad One strategy to do so is to focus on pocketsof excellence in certain labs or parts of the organisation which show somepotential to become internationally attractive at least in a few sub-fieldsBoth strategies are in use as regards ERC and ERC grants Excellentexpatriates are lured back with the possibility of a (Starting) Grant andERC grants help in building-up pockets of excellence However evidenceis blurred due to the small number of grants and the competitive nature ofthe ERC

For powerless organisations that do host grantees they provide valuablerole models within and beyond the organisation The impact on organisa-tions is twofold First they have something to offer (in case of success paidby ERC) they could not offer on their own Second ERC grants allow forthe establishment of specific areas offering different dynamics More free-dom more flexibility more focus on competition and quality The down-sides again are small numbers and the largely passive role of theorganisation The effective decision is made first by ERC then by theapplying researchers and only to a minor degree by the organisation itselfHowever the reported irritations within organisations caused by thosegrantees may trigger a broader re-thinking of the organisation of researchand the conditions to be provided to top researchers within those organisa-tions more generally

For these kinds of organisations without grantees there is no impact tobe found for them the record of not hosting any grantee merely confirmsunfavourable conditions in the specific case under review The performancegap is just too high the requirements of ERC are too demanding the ERCis too far away to send any signals that could trigger any response to beinterpreted as impact One potential future impact here however notdetected in our interviews could be that certain retention policies need tobe introduced or strengthened in reaction to pressures to retain people thatmight be tempted to go to more attractive organisations in order toincrease their chances for ERC success

DISCUSSION

The following research questions have been asked during our exploratoryempirical study What changes do we see because of the ERC what is theawareness of strategic leaders within universities and how are the differentlevels within universities affected How do organisational and country

characteristics determine the impact of ERC and how do ERC inducedchanges relate to ongoing transformations and other pressures towardschange The exploration outlined in this article has provided a set ofanswers to these questions that are summarised below structured alongthe key dimensions of our conceptual model

Material Symbolic and Normative Value

Our first important finding is that the symbolic value of the ERC is alreadyimmense as the sheer existence of the ERC has made a difference ERCgrants assign prestige and symbolise excellence of individuals and of hostorganisations There is a strong awareness of strategic leaders as to thiseffect and the ERC has already acquired a high level of credibility ERCgrantees are used as a benchmark for quality between organisations andwithin organisations By and large this impact could be found at all orga-nisational levels Although our sample is biased towards grant getting orga-nisations it is obvious that the symbolic value is relevant to allorganisations across the European Research Area This symbolic effect isstrongest for powerful and in-between organisations that claim to competefor research excellence Powerless organisations see less symbolic value inthe ERC as for them the prestige excellence gap seems far too big Itwill be interesting to see if this effect winds down over time as the numberof ERC grants and grantees grows and the threshold of what is perceivedto be excellent might shift In any case the high normative value of ERCinstruments puts pressure on the strategic actorhood of universities whohave no choice but to support staff to get those grants

The material value is of different importance to different organisationsIn general the more funding pressure universities have the more the mate-rial value drives their reactions This is especially true for individual unitswithin organisations for which a grant can make all the differenceHowever we have also seen counter-productive efforts of organisationsthat are poorly endowed with funds to push their staff in a rather unpre-pared and erratic manner to apply for ERC schemes Though for very well-endowed organisations even considerable ERC funding only accounts for arelatively small share of their overall institutional budget they at the sametime recognise and emphasise the value of the grants beyond both thematerial and the symbolic value mainly providing additional capacity atlower levels

The impact related to the normative value is ambiguous The ERCthrough its generous grant conditions and its explicit aims propagates a cer-tain model of researchersrsquo freedom and empowerment within organisationsConsequently we found all kinds of adjustments to create favourable con-ditions for applicants and grantees as universities compete for grantees andpotential grantees These adjustments involve more autonomy for (young)researchers better contractual conditions and the build-up of teams aroundthem Those adjustments are more or less far-reaching depending on thepre-existing ambitions of organisations and the national funding environ-ment as organisations in countries with grants similar to ERC grantspartly have made respective adjustments already prior to ERCrsquos existenceWhile the changes are in line with researchersrsquo expectations in many casesthe change of HR policies and the increase in autonomy has meant disrup-tion of an existing balance between research staff within units (facultiesschools labs) and actually less authority of strategic leaders who have nochoice but to offer greater autonomy to grantees At the unit level this hasled to bottom-up growth and research foci whereas depending on theorganisational capability and ambition this may either be enhancing orendangering systematic research planning13 This observation is related tothe tension for the strategic leadership of universities between quality andprestige on the one hand (ie getting a grant and providing the conditionsfor its successful completion) and the aspiration of strategic control ofresearch profiles and staff equity on the other hand

Impact on Performance Strategic Capability and HR

Finally across all impact dimensions our findings show that the ERC is acatalyst in most cases reinforcing certain transformations rather than trig-gering them When the ERC actually triggers strategic organisational reac-tions it is on the lower organisational levels which are more often locallyisolated and opportunity driven rather than systematically organised andspreading across the organisation

Generally it is too early to assess any impact on performance yetHowever all indications and the assessment of interviewees point towardsan effect on performance as the ERC helps to create conditions for excel-lent researcher that are conducive to producing excellent results that will bemeasured in output and impact indicators in the future The competitionbetween organisations and units in organisations to get grants will contri-bute to enhance further the likelihood of excellent research being produced

The impact on strategic capability is mixed In many cases the impact isweak on the highest organisational level The ERC does not fundamentallychange the limitations of top-level leadership to manage research asresearch direction and field specific performance criteria cannot be influ-enced very strongly Further the ERC as such does not offer bundling ofresearch projects or lines and thus horizontal management to define newcooperative areas of research in organisations is not supported by theERC However particularly in the case of several grantees in one unitbundling happens and some organisations are able to strategically incorpo-rate this into overall research strategies

At the unit level within organisations the ERC grant has been used tomake strategic adjustments towards more risky and daring researchHowever in many cases especially in powerless organisations (radical)changes in research portfolios have been un-coordinated and entirelybottom-up whereby the offer of re-direction challenges top-down manage-ment of research Strategic capabilities are most influenced in the organisa-tions that are already committed to defining an excellence-based strategyand where there is still room for learning and improvement In these casesorganisational willingness and capacity to learn between units and betweenunits and the centre can be found For more powerless organisations ERCgrants create pockets of excellence (and disruption) that are not fosteredorganisationally

The impact on attracting and managing researchers is not straightfor-ward In powerful organisations incentive structures and support mechan-isms are often already in place and the ERC does not make muchdifference In in-between or powerless organisations the ERC grant canmake a big difference in terms of professionalisation of support structureand in terms of autonomy of and reward for staff Where there is not astrong culture of competition-based allocation of reward and autonomy tostaff yet the ERC can be perceived as a challenge to HR management dis-rupting established and often cherished procedures and team balance

The Importance of the Intervening Organisational andCountry Conditions

Our empirical analysis so far has demonstrated the importance of pre-existing organisational capabilities and prevalent conditions of national fund-ing and science systems We have found a connection between the existingstrengths and development of organisational capabilities (at various levels)

on the one hand and impact of the ERC on the other hand We havebundled the various conditions within organisations into three types thatshow sufficient similarities as to the intervening variables that we havedefined For the overall organisation (ie aggregated organisational impacts)we can summarise the impact of the ERC in the simplified stylised Fig 2

This figure is an over-simplification but it is meant to highlight thataggregate effects on organisations (ie the combination of lower and higherlevel effects) in general differ for our three types For different reasonsboth powerful and powerless organisations are less affected by the creationand operation of the ERC than in-between organisations All in all power-ful organisations already show ambitions structures and processes that areconducive to getting ERC grants and despite some adaptations on lowerlevels learning and adaptations in these organisations remain limited Forthe powerless organisations the performance gap and the lack of support-ing structures are prohibitive for a broad ERC impact to unfold Strongimpact on sub-unit level is not translated into organisational lessons at ahigher level By and large within in-between organisations which oftenhave developed a conscious strategy to improve as organisations there is agreater likelihood that an impact on lower levels leads to organisationallearning at higher organisational levels Furthermore as leaders in thoseorganisations recognise the leverage of the ERC for re-positioning theirorganisation on the European landscape the ERC functions as a strongcatalyst to set up new structures and processes

Obviously as the analysis has shown this relation has to be further qua-lified for individual sub-units Here the individual grant unfolds a strongermaterial symbolic and normative value and the management of research

Overall impact onorganisations

powerfull organisationspowerless organisations

Fig 2 Simplified Illustration of Role of Pre-Existing Organisational Conditions

on Overall Impact

is less distant from the rationalities of the epistemic community allocatingthe grant in the first place What matters here for the impact on the organi-sation as such in the medium and long term it seems is the extent to whichchanges at lower levels are transmitted to the upper level or to other unitsand furthermore incorporated into the overall organisational strategy

Regarding the intervening organisational and country variables it mightappear that organisational features are often more important than countrycharacteristics in that there are strong sub-units or organisations in countrieswith a somewhat weaker overall funding arena We have experienced a num-ber of cases where individual sub-units or outlier organisations follow differ-ent rules and mechanisms and thus employ different policies The evidence isstill mixed however and further application of the approach is neededIt seems that powerful organisations can flourish also in less supportivelandscapes and the uphill struggle for powerless and even in-between orga-nisations in a weak funding landscape is a bigger one However a certainminimum financial endowment and a minimum degree of legal freedom toact have to be in place Further we concede that this claim is based on astrong simplification In reality organisational development history and cul-ture legal framework and other factors are deeply entangled (and the ERCcould become one factor over time to untie some of the historical knots)

CONCLUSION THE IRONY OF ORGANISATIONAL

ACTORHOOD AND ERC SUCCESS

Universities have to search for money that is not bound to specific review-based activities if they are to function as independent organisations asthose are beyond their own reach to steer The ERC while providing repu-tation and prestige as well as money does not add to the organisationalauthority over resources It often strengthens the position of excellentresearchers and binds the granted money to a specific activity in someinstances even leading to further claims of the grantee This binds resourceseven further or leads to the establishment of groups that would not havebeen built within the strategic planning of the organisation or even of theunit otherwise There is thus an inherent tension an irony the more theprestige of the ERC is sought the less autonomy organisations actuallymight have The normative pressure (ie ways to organise research andgrant freedom to the grantees even the younger ones) rather appears to

limit and not to enhance organisational actor capability and the additionalresources granted bind other resources and trajectories are built that werenot planned This could potentially be used in a re-active adaptive strategybut essentially limits the degrees of freedom of the organisation as a whole

On the other hand some organisations have started to interfere withERC grantees and applicants For the good of the organisation they try tomaximise the success thus putting pressure on their staff and designingsupporting structures that are entirely re-active and at worst foil the veryidea of risky frontier research Internal peer reviews may be set up in orderto assure quality of ERC applications however these may not always becapable to assess risky ERC-type approaches Further the organisationalinterference exemplifies the tension between the organisational interest andthe individual logic of the ERC

1 The official mission statement of the ERC stimulated our interest in how theERC might be affecting universities httperceuropaeumission

2 We are indebted for manifold feedback and suggestions for improvementsfrom colleagues in the EURECIA project led by Maria Nedeva and the projectrsquosAdvisory Committee (see wwweurecia-ercorg for a list of those colleagues and thefinal report Nedeva et al 2012)

3 We owe this idea of different ways the ERC exerts impact to discussions withMaria Nedeva Jochen Glaser and Dietmar Braun from the EURECIA project Theelaboration on those three ways of impact and thus any misconception is ours

4 For more details see for example ERC Grant Schemes Guide forApplicants for the Advanced Grant Call 2011 Guide for ERC Grant Holders (Part IPart II) all available at httperceuropaeudocument-library

5 The performance of organisations in terms of research output cannot yet bedetermined a proxy here will be the perception and expectation of performance andits justification in interviews

6 We acknowledge that there might be further impacts spillover effects in otherareas or unintended impacts but focus in our research on the three target dimen-sions defined by the ERCrsquos mission

7 Braunrsquos classification helps to understand the funding situation of universi-ties not their human resources policies or general governance structures Braunuses a combination of proxies that is research expenditures as a percentage ofGDP scientific articles per million population number of researchers per 1000employees and ERC success has been used (Braun 2011) While we fully agree onthe first three indicators we are a bit more sceptical about the use of ERC successrates in this context However his classification of countries to these categories washelpful for our purpose

8 We are grateful to our colleagues in the EURECIA project for supporting uswith a considerable number of interviews and related material for the cases in oursample

9 Note that our assignment of universities to the three baskets is based on amix of qualitative and quantitative variables including both country and individualorganisational characteristics which we gathered by means of upfront deskresearch literature review secondary data and our interview analysis Followingour iteratively developed research approach the baskets thus establish an integralpart of our results We are aware that there are a range of other classificationsdeveloped meanwhile most recently by Paradeise and Thoenig (2013) who use atypology based on a two by two matrix of excellence and reputation and thus deriveat four different types10 Further the allocation is based on an aggregate assessment for example

assigning University X to the top category can mean that it was relatively highranked for all three intervening variables but not necessary top ranked in all ofthem11 In this first approach we did not distinguish the different financing models

that is we did not systematically distinguish between organisations that rely moreon grants and those that can heavily or exclusively draw on institutional funding Inthe actual analysis however we did find differences and report on them12 While the baskets have worked as simplifying device in communicating broad

lines of impact the borders between the baskets are blurred This has to do with thevariety of capabilities within an organisation at the level of sub-units Thus it didnot prevent analysis of idiosyncrasies and special cases13 Note that within our empirical sample and at this early stage the type of

ERC grant (StG or AdG) made no systematic difference yet However there aresome first empirical indications that the different types of grants address differentdimensions of the researcher organisation nexus in terms of both the symbolicand normative values for example StG funding allows for new degrees of freedomfor young people that cause organisational turbulences that are not caused by AdG(as grantees most often are well established and integrated before the ERC grant)However this has to be further explored in the course of future investigation

ACKNOWLEDGEMENTS

This paper has its origins in the EURECIA project funded by the ERC asa CSA action within FP7rsquos IDEAS Programme (Grant No 229286)EURECIA set out to develop and apply a novel conceptual frameworkand methodology to investigate measure attribute and assess the impactand outcomes of the ERC and its funding schemes We are indebted formanifold feedback and suggestions for improvements to the leader of thatproject Maria Nedeva to the project team and the projectrsquos Advisory

Committee (see wwweurecia-ercorg for a list of those colleagues and thefinal report Nedeva et al 2012)

REFERENCES

Aghion P Dewatripont M Hoxby C Mas-Colell A amp Sapir A (2008) Higher aspira-

tions An agenda for reforming European universities Brussels Belgium Breughel

Blueprint Series

Bleiklie I (2011) New public management network governance and the university as a chan-

ging professional organization In T Christensen amp P Laeliggreid (Eds) Ashgate research

companion to new public management (pp 161176) Aldershot UK Ashgate

Bleiklie I amp Kogan M (2007) Organization and governance of universities Higher

Education Policy 20 477493

Braun D (2011) ERC impact on funding systems Reporting and synthesis deliverable D 93

for the EURECIA project Brussels Belgium Manchester

Clark B R (1998) Creating entrepreneurial universities Organizational pathways to transfor-

mation Oxford IAU PressPergamon

Enders J (Ed) (2001) Academic staff in Europe Changing contexts and conditions Westport

CT Greenwood Press

Estermann T amp Nokkala T (2009) University autonomy in Europe I exploratory study

Brussels Belgium European University Association Publications

Ferlie E Musselin C amp Andresani G (2008) The steering of higher education systems A

public management perspective Higher Education 56 325348

Franzoni C Scellato G amp Stephan P (2012) Foreign born scientists Mobility patterns for

sixteen countries National bureau of economic research Working Paper 18067

Retrieved from wwwnberorgpapersw18067 Accessed on February 15 2014

Fumasoli T amp Lepori B (2011) Patterns of strategies in Swiss higher education institutions

Higher Education 61(2) 157178

Geuna A amp Martin B (2003) University research evaluation and funding An international

Comparison Minerva 41 277304

Granberg A amp Jacobsson S (2006) Myths or reality A scrutiny of dominant beliefs in the

Swedish science policy debate Science and Public Policy 33(5) 321340

Jongbloed B (2009) Higher education funding systems An overview covering five European jur-

isdictions and the Canadian province Ontario Report prepared for the Hong Kong

University Grants Committee Enschede The Netherlands

between global trends and national traditions Bielefeld Germany Transcript Verlag

Laredo P (2007) Revisiting the third mission of universities Toward a renewed categoriza-

tion of university activities Higher Education Policy 20(4) 441456

Laudel G amp Glaser J (2010) Experteninterviews und qualitative Inhaltsanalyse Wiesbaden

Germany Springer Fachmedien

Mayring P (2008) Qualitative Inhaltsanalyse Grundlagen und Techniken Basel Switzerland

Beltz Verlag

Musselin C (2005) Le marche des universitaires France Allemagne Etats-Unis Paris

France Presses Sciences-Po

Nedeva M (2007) New tricks and old dogs The lsquothird missionrsquo and the re-production of the

university In D Epstein R Boden R Deem F Rizvi amp S Wright (Eds)

Geographies of knowledge geometries of power Framing the future of higher education

(pp 85103) World Yearbook of Education 2008 New York NY Routledge

Nedeva M amp Stampfer M (2012) From lsquoscience in Europersquo to lsquoEuropean sciencersquo Science

336(6084) 982983

Nedeva M et al (2012) Understanding and assessing the impact and outcomes of the ERC and

its funding schemes (EURECIA) Final Synthesis Report Brussels Belgium

Manchester UK Retrieved from httperceuropaeusitesdefaultfilesdocumentfile

eurecia_final_synthesis_reportpdf Accessed on February 15 2014

OEU (Observatory of the European University) (2006) Methodological guide Prime OEU

Guide Lugano Retrieved from wwwenid-europeorgPRIMEdocumentsOEU_guide

pdf Accessed on February 15 2014

Oquist G amp Benner M (2012) Fostering breakthrough research A comparative study

Stockholm Sweden Akademirapport Kungl Vetenskapsakademien

European comparative perspectives Dordrecht The Netherlands Springer

Paradeise C Reale E amp Goastellec G (2009) A comparative approach to higher educa-

tion reforms in Western European countries In C Paradeise E Reale I Bleiklie amp

(pp 197226) Dordrecht The Netherlands Springer

Paradeise C amp Thoenig J C (2013) Academic institutions in search of quality Local orders

and global standards Organization Studies 34(2) 189218

Schmoch U amp Schubert T (2010) Strategic steering of research by new public management

in German universities A looming statescience conflict Research Evaluation 19(3)

209216

Van Vught F et al (2005) Institutional profiles Towards a typology of higher education insti-

tutions in Europe Enschede The Netherlands University of Twente

D Weaire (Eds) The university in the market (pp 2337) London Portland Press

WHERE HAVE ALL THE

SCIENTISTS GONE BUILDING

RESEARCH PROFILES AT DUTCH

UNIVERSITIES AND ITS

CONSEQUENCES FOR RESEARCH

Grit Laudel and Elke Weyer

ABSTRACT

This article investigates the links between universitiesrsquo opportunities toshape their research profiles the changing state interest concerning theseprofiles and the impact of profile building on research at university andfield levels While the authority of the Dutch state over research profilesof Dutch universities has increased university management has consider-able operational authority over the inclusion of new research fields andremoval of existing research fields Since all universities have begun tofollow the same external signals prescribing applied research researchthat has easy access to external funding and research in fields prioritised

ISSN 0733-558Xdoi101108S0733-558X20140000042004

by the state a lsquoquasi-market failurersquo may emerge as is demonstrated forevolutionary developmental biology and Bose-Einstein condensation

Keywords New public management university research profilesresearch fields scientific communities the Netherlands

INTRODUCTION

The new public management (NPM) reforms have been studied by highereducation researchers mainly with regard to their impact on universitieswith a focus on the lattersrsquo autonomy and action capabilities (Clark 1998De Boer Enders amp Leisyte 2007 Marginson amp Considine 2000) Theimpact of these reforms on the conduct and content of research has enjoyedfar less attention with empirical studies focusing on the micro-level that ison researchers and research groups (Glaser amp Laudel 2007 Glaser LangeLaudel amp Schimank 2010 Jansen 2010 Leisyte 2007) Owing to the meth-odological difficulties of causally linking micro-level changes in researchpractices to macro-changes of research fields in a country such micro-levelimpact studies can only speculate about the impact of NPM reforms on theresearch of a country In this article we contribute to closing this gap byaddressing possible macro-level effects of an aspect of NPM that hasnrsquotenjoyed much attention so far namely the building of research profiles byuniversities and its impact on research fields at the national level

These effects are of both theoretical and political interest Theoreticallystudying the effects of changes in organisations on research fields reversesthe perspective that has dominated organisational sociology so far Therecent interest of organisation studies in communities has treated them pri-marily as a context of organisational activities for example as the domi-nant context (OrsquoMahony amp Lakhani 2011) or as a resource fororganisations (Wenger amp Snyder 2000) The example of open source soft-ware production has been used in an attempt to establish communities as adistinct organisational form of production (Seidel amp Stewart 2011) Ourinvestigation takes changes in organisations as the independent variableand asks how changes in organisations in the aggregate can affect commu-nities that are distributed across many organisations

There is also the theoretical question of what universities can actually dowith their research Investigations of the nature of the university as an orga-nisation and of properties of research processes suggest that a universityrsquos

112 GRIT LAUDEL AND ELKE WEYER

capabilities to shape its research can be increased by NPM reforms but stillhave fundamental limitations NPM reforms are supposed to turn universi-ties into corporate actors with full strategic action capability which in-cludes an increased capability to steer their core processes of research andteaching At the same time there are theoretical reasons which make thesuccess of these steering attempts rather unlikely the lsquolack of transparencyrsquoof the universityrsquos core processes (Musselin 2007) the uncertainties inher-ent to research which make scientific communities the lsquocontrol centrersquo ofresearch (Whitley 2008) the inability of universities to define researchtasks to control work flow interdependencies and to control quality anduse of results caused by properties of the researchersrsquo work (Glaser 2012)Building research profiles seems to be as far as a university can go in thesteering of its research (Von Stuckrad amp Glaser 2012 Whitley amp Glaserthis volume) which makes this process a theoretical limit of NPM reforms

Politically the possible effects of NPM reforms on research fields havebeen discussed with regard to threats to the diversity of research Theseeffects have been shown to exist in British economics (Lee 2009) and havebeen hypothesised for the interdisciplinary field of innovation studies(Rafols Leydesdorff OrsquoHare Nightingale amp Stirling 2012) In Germanythe conference of university rectors warned against a possible disappearanceof lsquosmall subjectsrsquo particularly in the humanities due to similar processes ofprofile building in German universities (Hochschulenrektorenkonferenz2007) In the Netherlands the Dutch Academy of Science pointed out thatuniversitiesrsquo responses to profile building could result in blank spots in theresearch portfolio on the national level with modern languages probablybeing the earliest victims (KNAW 2013)

Effects of profile building on research fields can be expected from theinteraction of two distinct trends in NPM reforms namely the pressure bygovernments on their universities to develop distinct teaching and researchprofiles and the strengthening of hierarchical structures in universitieswhich are supposed to enable that profile building (Meier amp Schimank2010 Whitley amp Glaser this volume) These trends represent a departurefrom past practices which of course always included decisions that shapedthe profile of a university Investment decisions decisions on the denomina-tion of chairs and recruitment decisions have consequences for the fieldsrepresented in a university and the topics addressed within these fieldsSimilarly the decisions of professors on their research topics contribute toshaping the research profile of a university However these decisions havenot always been part of a lsquomaster planrsquo They were made locally and ad

113Where Have All the Scientists Gone

hoc and included a lsquolist of disconnected choices made by individual profes-sors with respect to their theoretical and methodological interestsrsquo or otherforms of voluntary bottom-up coordination (Meier amp Schimank 2010p 214) University profiles emerged as an aggregate effect

The creation of research profiles as an intentional strategy of the univer-sity to lsquoset collectively binding rules for the membersrsquo is a relatively newdevelopment triggered by NPM reforms (Meier amp Schimank 2010p 213) Many universities are now able to manage their research and teach-ing profiles by allocating resources between competing components ofteaching and research activities which are treated as investment portfolios(Whitley 2008 p 26) The creation or modification of profiles may stilloccur for a variety of reasons including responses to external expectationsinternal initiatives by academics who want to extend their resource baseand financial considerations that are unrelated to interests concerningresearch content The new intentional building of profiles that has beenadded to this mix is usually directed towards lsquoenhancingrsquo the profile whichincludes a reduction of topics and a redistribution of resources betweenfields with the aim of concentrating resources These processes have beenhardly investigated and their effects are poorly understood

The aim of our article is to explore the links between organisationalsituations of universities and faculties their profile-building strategies andeffects of such strategies on research at universities as well as research fieldson the national level The Netherlands provide an excellent laboratory forthe analysis of such consequences because their relatively small size makesnational fields very sensitive to decisions at individual universities Thissituation enables the causal attribution of changes of research fields toprofile-building activities of universities

Our argument draws together findings from several studies of the Dutchuniversity system and will be developed in the following steps We will firstdescribe the evolution of state expectations concerning the profile buildingof Dutch universities and demonstrate that these expectations have not onlygrown in strength but have been substantially transformed from the expecta-tion that universities may develop any profile to the expectation of specificcontent namely topics that contribute to specific political goals (third sec-tion) We then analyse profile-building activities and their effects at Dutchuniversities and develop an empirical typology of strategies of profile build-ing (fourth section) Moving to the field level we demonstrate consequencesof these decisions for three fields including two that are now developingcontrary to international trends (fifth section) As a conclusion we discussthe field-level effects of profile building as a quasi-market failure of a system

in which universities must use profile building to compete for money (sixthsection)

THEORETICAL FRAMEWORK AND METHODOLOGY

The literature is rather opaque in providing definitions of research profilesand profile building that can be operationalised For the purposes of thisstudy we define the research profile of an organisation or organisationalunit as the distribution of research effort across a spectrum of researchtopics and their epistemic characteristics Thus a research profile dependson the research topics addressed by researchers of the organisation and theresources allocated to research on these topics Both aspects of a profilecan vary relatively independently of each other and are changed both unin-tentionally and increasingly by actions that intentionally aim at creating aspecific profile (profile building) Profile building at universities can includea wide range of activities from simply renaming research or defining newheadings for old research to significant investments in research centres fortopics not previously present at the university or the closing down ofresearch on topics that are no longer deemed useful for a universityrsquos pro-file Research profiles can be described by their content (research themes)their diversity (number and breadth of themes) and various epistemic char-acteristics of the research including its basic versus applied nature interdis-ciplinarity and collaborations

So far research profile building by universities has been investigated pri-marily in the German context Meier and Schimank (2010) investigated theprofile building at universities in the German state of Lower Saxony whichwas initiated by an evaluation of the university research of that state Theyfound that traditional bottom-up processes of profile building had becomesupplemented by top-down processes which were enabled by the increasedaction capabilities of German university leadership Owing to the still earlystages of the profile-building activities they observed little could be saidabout effects on the research profiles of universities Probably for the samereason Langersquos observation of attempted top-down profile building byfaculties at one German university showed nearly no effects (Lange 2007pp 164165) Germanyrsquos excellence initiative has also been consideredbecause one of its aims is to stimulate the building of distinct research pro-files (Leibfried 2010 Meier amp Schimank 2010 p 233) Again little can besaid about the effects because there is no conclusive evidence yet The sameholds for a recent study of the building of research profiles at Finnish

universities (Pietila 2014) for a descriptive account of Dutch and Germanprofile-building activities (Klumpp De Boer amp Vossensteyn 2013) and fora bibliometric investigation of a Dutch government initiative termed lsquofocusen massarsquo (focus and mass) by Van den Besselaar and Horlings (2010)

While the knowledge about effects of profile-building activities is stillvery limited the studies quoted above suggest that a variety of considera-tions may inform decisions on research profiles and the profiles emergingfrom these decisions These include at least external expectations towards auniversityrsquos profile strategies formed at different levels of the universityfinancial interests and perceived necessities and interests as well as capabil-ities of researchers Thus the building and shaping of research profilestakes place in a complicated actor constellation involving external actors aswell as actors at several levels of the university hierarchy

Our analysis of decision processes concerning university research profilesuses the authority relations perspective (Whitley 2010) which is a promis-ing way to conduct an integrated assessment of changes in public sciencesystems With this approach we can systematically analyse how authorita-tive agencies (the state research organisations organisational elites exter-nal funding agencies national as well as international scientific elites andthe researchers themselves) exercise authority over specific matters ofgovernance in our case research goals Its basic assumptions are (a) thatthe changes that public science systems go through have implications forthe relationships between actors and the way that they are able to realisetheir interests and (b) that authority relations as regards the selection ofresearch goals are the main channel through which changes in the knowl-edge production system are effectuated (Whitley 2010) We consider theauthority perspective to be particularly useful for our purpose because itenables the integration of multiple overlaying governance processesthrough which a set of actors contributed to the rise and fall of scientificfields in universities

We focus on the situations of universities in which profile-building stra-tegies are developed their decision processes on research profiles and theimpact of these decisions on research at the university as well as on fields atthe national level Concerning the situations of universities the expectationsby external stakeholders (mainly the state) and their authority and the uni-versityrsquos financial situation are of particular importance Decision processeson research profiles vary in their aims as well as the actors involved andtheir relative authority Concerning the outcomes we pay particular atten-tion to the actual changes of the profile including secondary effects acrossthe university Concerning the national level we discuss aggregate effects

which can be described as unanticipated consequences for the nationalresearch profile of the country

Our analysis is based on case studies in Dutch universities which albeitconducted for different purposes all included decision processes about uni-versity research profiles or the consequences of such decisions The empiricaldata stem from three research projects The first project seeks to find outwhether the organisational transformation of Dutch and English universitiescan be explained as transition between organisational archetypes of authoritystructures (Weyer 2014) The second project studied the impact of changingauthority relations in four European countries on conditions for intellectualinnovations in the sciences social sciences and humanities (RHESI see theintroduction to this volume) Finally a comparative project on the impact ofnational career systems on opportunities for researchers to change lines ofresearch also contributed data on authority structures in Dutch universitiesand their impact on university research profiles (Laudel 2012)

Case studies in all these projects were based on documentary analysis andsemi-structured interviews (more detail on the methodology of the investiga-tion of evolutionary developmental biology and Bose-Einstein condensationcan be found in the articles by Laudel et al this volume) Documentarysources include university newspapers internet pages minutes of universitydecision-making bodies as well as policy documents Interviews were con-ducted with members of the university top management mid-level managersheads of department and academics at different levels Interviews withresearchers were prepared by reading research-related documents (web sitespopular descriptions of research publications project proposals CVs) andindividual-level structural bibliometrics (Glaser amp Laudel 2009) The inter-views were analysed with qualitative content analysis (Glaser amp Laudel2013) Analyses for the purposes of this article focused on authority relationsconcerning university profiles and consequences of profile building for thegrowth and shrinking of fields at universities

THE EVOLUTION OF DUTCH STATE EXPECTATIONS

CONCERNING UNIVERSITY RESEARCH PROFILES

First Phase Changes in Profiles as Unintended Effects

State expectations of academic research started to rise at the end of the1970s The Dutch government wanted public research to be to a certainextent (nationally) programmed and it wanted to lsquoenhance the quality the

practical value and the accountability of researchrsquo (Blume amp Spaapen1988 p 26) This is why research evaluations were introduced in 1983

The government asked universities to submit so-called lsquoresearch pro-grammesrsquo that is coherent research plans for (initially) five years and ofgroups of at least five researchers in order to receive funding for thisresearch Part of the funding of university research was made conditionalon the approval of the research programmes by peer review committees(Blume amp Spaapen 1988 p 11 Van der Meulen 2007) Although the ori-ginal idea was to conduct ex-post evaluations of the funded programmeslater on in order to link further funding to their success this never hap-pened in practice

While this exercise had some consequences for the research profile ofuniversities they were not intended and still fall under the category oflsquounintended consequences of local decisionsrsquo described by Meier andSchimank (2010) Changes in research profiles might have occurred in uni-versities and at the national level for three reasons

The necessity to provide five-year research programmes for groups of atleast five researchers may have induced local collaborations

The success of programmes in being approved for funding was influ-enced by the fit between a disciplinersquos epistemic practices and the needto have long-term plans for larger groups

The varying success rates of programmes from different disciplines between 36 and 100 in the sample investigated by Blume andSpaapen (1988 p 21) effectively led to a redistribution of fundingbetween disciplines

None of these effects whose strength is difficult to assess due to a lackof data were intended or planned by either the state or universities Thewhole evaluation exercise was still a bottom-up process in which researchgroups chose their own research areas

The first attempts of the state to initiate profile building by specialisationand concentration in universities occurred in 1983 (lsquotaakverdeling en con-centratiersquo division of labour and specialisation) and 1986 (lsquoSelectieveKrimp en Groeirsquo selective shrinkage and growth) These policies led to theclosure of some and amalgamation of other small university departments(Blume amp Spaapen 1988 p 6) The aim of the policies was to increase effi-ciency in higher education (Westerheijden De Boer amp Enders 2009p 109) The state did not prioritise any themes

Since these first initiatives Dutch universities constantly changed thestructures and thematic labels in which research was conducted The

resulting structures are lsquovirtualrsquo aggregations of research (lsquoresearch schoolsrsquolsquoresearch institutesrsquo lsquoresearch programmesrsquo etc) that overlay the more per-manent faculties (although mergers of faculties occurred as well) Facultiesare largely considered to be the units responsible for teaching which is whythe various research units create a matrix structure of the Dutch universityOur interviews confirm that until the 2000s the impact of these structuralchanges on the content of university research remained limited (see alsoLeisyte Enders amp De Boer 2010)

Second Phase Political Pressure on Profile Building

It was only in the early 2000s that the Dutch government took concretemeasures to bring about the lsquorightrsquo content in university research It intro-duced its lsquofocus and massrsquo policy The argument is that the Netherlands istoo small to excel in all research areas Therefore resources should be con-centrated in areas where the Netherlands are excellent and which areimportant from a socioeconomic point of view (Van den Besselaar ampHorlings 2010 p 13) The concentration should be supported by the foun-dation of multi-disciplinary networks and consortia (ie build lsquocriticalmassrsquo)

To support its focus and mass policy the government initiated two paral-lel measures first public higher education institutions should develop dis-tinct research (and teaching) profiles Second the government establishedlsquotop sectorsrsquo in which research efforts should be concentrated and researchcollaboration with industry be stimulated Taking the lsquostrongrsquo researchareas of the Dutch economy as a point of departure public and private sta-keholders identified nine top sectors in the following areas Agro-FoodHorticulture and Propagating Stock High-Tech Materials and SystemsEnergy Logistics Creative Industry Life Sciences Chemicals and Water

About h15 billion was earmarked for the nine top sectors mostly byreallocating existing budgets of various ministries that manage an innova-tion budget (Technopolis 2011 pp 7 11) Particularly noteworthy in thisrespect is the decision made by the government to spend roughly half ofthe budget of the Dutch Research Council NWO on research in the top sec-tors This re-allocation raised concerns about worsening conditions for fun-damental research as voiced by the Dutch Academy of Science lsquoin essencethere is room for fundamental research in the economic top sectors butwhether that room will be made available remains to be seenrsquo (KNAW2013 p 13)

Third Phase Using Performance Agreements for Prescribing Content

In 2010 a governmental advisory committee (the Commission Veermanled by the former minister of agriculture Cees Veerman) pointed out theneed for profile building in Dutch higher education By picking up thefocus and mass argument the Commission reiterated that universities couldnot excel at everything and should therefore try to concentrate on theirstrengths in order to remain competitive with other knowledge economiesThe Ministry of Education Culture and Science followed the recommenda-tions of the Veerman report by including the demand for profile building inteaching and research in the Strategic Agenda for Higher EducationResearch and Science (September 2011) Although universities were for-mally free to decide whether to participate in the nine top sectors or nottheir active involvement was nevertheless clearly expected by thegovernment

The process of profile building in Dutch higher education was givenshape in the so-called performance agreements that were agreed uponbetween the Secretary of State and all universities In the framework ofthese performance agreements universities committed themselves to thedevelopment of unique teaching profiles and research profiles

Participation in the performance agreements was voluntary but hadfinancial consequences To ensure that universities would take the call forprofile building seriously the government earmarked h105 million (5 ofthe total higher education budget) as performance-based funding In addi-tion h 38 million (2) were to be distributed among those universities hav-ing handed in the best plans as evaluated by the ministry By means ofcoupling at least some part of the funding to specific performance expecta-tions the government reserved itself the right to withdraw this money againif universities should not reach the targets agreed upon in the performanceagreements

The evolution of profile-building activities in the Netherlands demon-strates interesting shifts in authority relations which in some respects con-tradict the central tenet of NPM reforms namely increasing the autonomyof universities with respect to formulating their own goals The originalintention of Dutch higher education policy back in the 1980s was that uni-versities should build profiles that is should concentrate their resources inselected areas which were supposed to differ between universitiesHowever at this time there were no political expectations concerning thecontent of these priority areas at universities These expectations whichincreasingly took the form of financially supported political directions

emerged in the early 2000s Clear expectations concerning the content ofprofiles were formulated and reinforced first by financial incentives in exter-nal funding and later by performance agreements and linked universityblock funding Although the Dutch publicly funded universities are in prin-ciple free to lsquoignorersquo the call for profile building their increasing financialdifficulties create a strong pressure to take the financial rewards It is henceworth investigating how universities respond to these calls when decidingabout the prioritisation of research activities

PROFILE BUILDING AT DUTCH UNIVERSITIES

Strategies of research profile building include all decision processes at theuniversity that are aimed at changing the distribution of a universityrsquosresearch effort across the spectrum of research fields The two outcomes ofprofile-building activities include a description of research topics and anallocation of resources to topics

Devolved Profile Building at University 1

The first case of profile building emerged as a response to the Dutch gov-ernmentrsquos call for creating focus and mass in university research and tocuts in university block funding The process was initiated by the univer-sityrsquos executive board1 as part of the universityrsquos strategic planning Thestrategic plan included the introduction of clear-cut criteria for the develop-ment of new lines of research and a critical review of the current researchportfolio which should be modified if necessary These aims should beachieved by the creation of faculty wide lsquofocus areasrsquo in all faculties Theprocess was supervised by the senior management team (consisting of theexecutive board and the deans) and was supported by start-up funding(conditional on the faculties providing plans for their profile-building pro-cess) In addition several budget tranches were distributed among thefaculties during the subsequent three years to support the process of profilebuilding

The selection of lsquofocus areasrsquo was largely a faculty affair The executiveboard had formulated some general expectations (eg the stipulation ofexcellent research lines) which were discussed with faculty management

The number and content of focus areas was at the discretion of the facul-ties each of which applied its own strategies

Faculty One Further DevolutionOne of the faculties the faculty of the humanities further devolved theprofile-building process by delegating the choice of focus areas to the lea-ders of the existing lsquoresearch programmesrsquo that is the current thematicunits of research The research programme leaders discussed the profilebuilding with the researchers in their programmes After this consultationprocess the research programme leaders decided that the facultiesrsquo researchprogrammes were to become the new focus areas The only actual changewas that within the research programmes (now focus areas) sub-focishould be formulated of which some tended to be slightly narrower in theirfocus than had previously been the case These narrower descriptions didnot affect the research itself Content-wise the facultyrsquos research portfolioremained by and large the same

This left the task of distributing the funds that were allocated for theprofile-building process The faculty management required the researchprogramme leaders to submit budget plans in which they explained howthey intended to use the funds The faculty management thereupon spreadthe amount of profile-building funds almost evenly across all research pro-grammes (now lsquofocus areasrsquo)

Faculty Two Participative Decision-Making at the Faculty LevelAs part of the same university-wide profile-building process anotherfaculty a science faculty applied different strategies All decisionsabout the facultyrsquos new research profile were made by the faculty manage-ment albeit in consultation with the universityrsquos executive board and seniorfaculty staff (research group leaders heads of departments etc)

While the research profile of the faculty was not changed thematicallythe profile building and the money made available for it were used selec-tively In a first phase the decision was made to strengthen the position oftwo of the facultyrsquos research areas which became the new lsquofocirsquo of thefaculty One area was a recent lsquoacquisitionrsquo a group that had moved fromanother faculty and had brought some lsquoprofile-building moneyrsquo with it Theother one was a highly successful research group in need of investmentSince the group leader had offers from other universities the faculty waskeen to provide the funds in order to make him stay The money offeredfor the profile-building process provided a good opportunity to integratethe first group and to strengthen the other which happened by channelling

all the funds available for profile building to these areas Although not allstaff members were satisfied with the faculty management decision to usethe funds selectively the faculty management could convince most of themnot least by promising that the regular distribution of funds would be leftuntouched

Bottom-Up Generation of Options and Top-DownConfirmation at University 2

In University 2 too the profile-building process was initiated by the uni-versityrsquos top management which asked the faculties to search for agreater concentration in their research portfolios In particular it wasargued that if the university wanted to maintain a strong position inter-nationally it would have to channel its resources into a number of clus-ters and disinvest from some areas The selection of focus areas wasconducted as a centrally led process in which the faculties were expectedto cooperate with each other in the creation of inter-faculty focus areasThe Executive Board did not provide financial support for the selectionprocess itself but promised financial rewards once the final selection offocus areas had been made

The selection process began as a bottom-up process The deans askedthe leaders of the research institutes (the current organisational units forresearch at university 2) to organise workshops among their research-activeacademics at which focus areas were to be suggested On the basis of theseinitial ideas researchers were expected to collaborate with their colleaguesfrom other faculties in the development of multi-disciplinary proposals thathad the potential to become part of the Universityrsquos new research profileThe proposals were collected by the deans and submitted as a proposal offocus areas to the Executive Board All proposed areas were multi-disciplinary and envisaged research collaboration among a considerablenumber of (senior) researchers from at least two faculties The majority oftopics were application-oriented and focused on issues of socioeconomicrelevance Some also addressed more fundamental questions In contrast touniversity 1 the choice of focus areas was selective in the sense that not allresearchers of a faculty were involved in them their own research activitieshence coexisted with the new focus areas

Having received the proposals the executive board asked an internalcommittee of active and emeritus professors from all faculties to assessthese plans Thereafter the list was presented to an external committee that

provided an assessment of chances and risks of each proposed focus areaNeither committeersquos recommendations led to changes of the proposalwhich was finally approved by the executive board

To support cross-faculty collaborations in the new focus areas theExecutive Board provided a start-up budget for each focus area that wastopped up by the Faculty Researchers of the focus areas were then askedto double this original investment from external grants The resulting bud-get could be used for the acquisition of additional personnel and researchequipment

Closing Research Groups at a Faculty of University 3

A faculty at university 3 responded to a growing budget deficit which hasbeen accumulating since the late 1990s The faculty responded by a decade-long process of restructuring and profile building with the aim of having lessand less costly research The dean appointed an internal committee whichproposed a plan for the restructuring of biological research at the facultyThe committee suggested an organisational structure that was supposed tolead to a more homogenous research profile to enable the creation of lsquocriti-cal massrsquo and to ensure the efficient use of infrastructure Since some groupsdid not contribute to the new profile it was suggested that they should beclosed Among the areas to be closed was a subfield in biology

The original plans were opposed by the facultyrsquos science policy commit-tee especially with regard to the closure of the biology group As a com-promise another group was closed and members of the biology groupwere asked to retire early This cost-saving strategy is possible at Dutchuniversities and is applied occasionally Dutch researchers go into earlyretirement receive pensions and continue to work at the university Thereis an obvious financial advantage for the university and little disadvantagefor the academics if they are close to retirement anyway If a faculty is infinancial difficulties the existence of such an opportunity creates a strongpressure on those who can take it If they do not agree there is the possibi-lity that some other unit or some of their colleagues are made redundant

The senior staff of the biology group agreed to help their departmentwith its financial problems However their early retirement turned out tobe only a temporary solution for keeping the original profile When thefaculty ran into financial problems again the department and faculty feltunable to honour their previous agreements The head of the biologydepartment strongly backed by the dean of faculty and the financial

officer imposed the original plans For this purpose the faculty created acommittee with lsquooutside expertsrsquo who again had the task of recommendinga profile and a structure for the department

Several factors indicate that the restructuring was triggered off by finan-cial problems and that it served the purpose of solving them rather thancreating a specific profile Thus an important criterion for including orexcluding research topics in the second round of profile building was theamount of external funding acquired by the different research groupsResearch performance on the other hand did not seem to be a selectioncriterion During the last research assessment the international reviewcommittee had positively assessed the biology grouprsquos research and hadrecommended to keep this group These recommendations were disre-garded in the decision-making process

Removing a Discipline from University 4

This section is based on an extensive description of a profile-building pro-cess published after the closure of Utrecht Universityrsquos institute for astron-omy by its last director (Keller 2012) The account was triangulated byother sources (blogs of researchers press releases of the university) andadditional inquiries

The decision on the facultyrsquos profile was a response to a budget deficitWhen the universityrsquos faculty of science was created by merging severalscience departments some of them brought substantial budget deficits intothe new faculty The department of astronomy and physics did not but wasnow part of a faculty that started with financial problems

Apparently unconvinced by the facultyrsquos own plans for overcoming itsproblems the executive board of the University of Utrecht commissionedtwo lsquooutside expertsrsquo to provide external advice on how to solve the pro-blems of the faculty Their recommendations deviated from the facultyrsquosown perceptions and plans The dean of the faculty resigned and aninterim dean was appointed for seven months The interim dean appointeda committee (whose membership was kept secret) to advise him on the dif-ferent institutes of the faculty This committee advised the dean thatastronomy should remain in Utrecht but that the previously planned addi-tional investments for the institute should not take place The instituteachieved an agreement with the interim and the prospective dean accord-ing to which the existence of the institute of astronomy was secured Theinstitute for astronomy had received a very good assessment in its last

external evaluation which was confirmed by a public internal ad hocevaluation The results of an additional ad hoc evaluation commissionedby the dean were kept secret Keller quotes the deanrsquos answer to his criti-cism of the secret nature of this evaluation

Yes as a scientist I respect the wish of scientists to be evaluated according to a well-

defined objective quantitative procedure In contrast most of the decisions that I will

have to take concerning which units of the faculty we will discontinue do not strictly

depend on such a procedure (Email of the Dean of the Faculty of Science at Utrecht

university to C U Keller 2012 p 4)

For reasons that are unknown to Keller (he suspects an intervention bythe universityrsquos executive board) the previous agreement was overturnedand the dean of the faculty decided to discontinue astronomy Accordingto Keller the dean provided only very vague reasons related to the contentor quality of research and teaching The institute was split into three partsthat were moved to three other Dutch universities

The loss of a strong university institute had consequences for local colla-borations which were epitomised by a large and prestigious theoretical phy-sics group leaving the University of Utrecht after the institute forastronomy was closed The group leader explicitly referred to the closure ofastronomy as a reason for her move

My institute does not operate in an isolated environment the elimination of other excel-

lent groups in the department of Physics and Astronomy affects us considerably

(Digitaal Universiteitsblad 4 April 2012 our translation)

The researchers themselves (here the physicists) exercised authority overtheir research goals in that they moved away from unfavourable researchconditions produced by the university through its profile-building activitiesThe reduction of theoretical physics in quantity and quality within theorganisation was an unintended effect of these activities

Authority Relations and Strategies of Profile Building

A comparison of the processes of profile building shows a pattern inauthority relations concerning research profiles In none of the cases didresearchers lose the authority over their research goals Some of them lostthe opportunity to continue their research at their university due to beingforced to leave or voluntarily leaving in response to worsening conditionsThis is however different from having other authoritative agencies

intervening in the formulation of research goals or tying the provision ofresources to specific goals

While profile building does not lsquoreach throughrsquo to the formulation ofresearch goals it decides on the presence of research areas and their accessto resources which are important conditions for the formulation ofresearch goals Three distinct combinations of thematic structuring andresource allocations occurred in the cases presented above

Purely Thematic Re-ContextualisationThis strategy consists of developing a new description of research activitieswithout changing them For example links between research fields can beemphasised by subsuming them under new headings or links to externalexpectations can be emphasised by describing existing research in a newway This strategy can be applied to create a seemingly new profile becauseresearch is a multivalent process and thus can be linked to a number of dif-ferent descriptions without being misrepresented For example a universitycan pretend to give priority to some lsquonewrsquo topics without actually changingthe allocation of resources This strategy was applied by the humanitiesfaculty in university 1

Thematic and Financial PrioritisationThis strategy links new descriptions of research to changing the relativeweight of topics in the new profile which is supported by a matching allo-cation of resources The latter can happen either by mobilising additionalresources or by taking away resources from non-priority areas and allocat-ing them to the new priority topics It implies that only some researchthemes are included into the new profile The changes in the allocation ofresources do not lead however to the appearance of new or the disappear-ance of existing research The university keeps all research efforts butchanges their prospects of development This strategy was applied by thescience faculty in university 1 and by university 2

(Positive or Negative) Selection of TopicsIn contrast to the previous strategy of prioritisation which can be seen asshifting emphasis this strategy leads to a sharp change of profile It includesthe import of new research or the removal of existing research In the firstcase additional resources or resources taken away from non-priority areasare used to create positions and infrastructure for a new area which isusually linked to some existing priority area An example of this strategyhas been observed by Lucas (2006 pp 100104) in a British university

Noticing that all biology departments that were rated highly in the researchassessment exercise had biotechnology in their profile the observed depart-ment decided to import biotechnology in its research profile The oppositestrategy negative selection means entirely removing topics for example bytaking away all vacant positions or by negotiating the redundancy of staffNegative selection is usually motivated by the necessity to free resourceseither for investing them elsewhere or for reducing budget deficits It wasapplied by a faculty in university 3 and by university 4

The occurrence of these strategies in our cases is linked to specific pat-terns in the distribution and exercise of authority summarised in Table 1which we will now further discuss Authority over research profiles that isconcerning decisions about what research areas should be at the universityand how strong these areas should be was held by the university executiveboard in all cases When the modification of whole university profiles wasdecided upon this authority had to be shared with researchers becauseexisting research competences and priorities limit the options a universityhas While researchers and research groups always have a spectrum ofpossible directions for future research this spectrum is limited and for-mulating topics outside the spectrum risks loss of performance This iswhy researchers have authority over the generation of options for a pro-file In some cases authority over options had to be shared with managersat several levels of the hierarchy However the options were generatedby researchers and higher-level managers only reserved the right toselect among them (a right which was not exercised in the cases analysedby us)

The situation was different however when research areas had to beremoved for financial reasons Again these processes have very similarauthority structures The university management shares authority withinternal and external expert committees rather than researchers Veto rightsare exercised if the higher-level management is not satisfied with the out-comes of processes

THE IMPACT OF PROFILE BUILDING ON

RESEARCH FIELDS

In the previous section we discussed characteristic strategies and financialinstruments for profile building at Dutch universities and some of the con-sequences for research in these universities We now turn to the

Table 1 Authority Relations in Different Strategies of Profile Building

University Strategy Authority over Options

for Profile

Authority over

Selection of Options

Authority over

Allocation of Funds

Veto Positions Changes in Profiles

1 Thematic

re-contextualisation

Researchers research

programme leaders

Research programme

leaders

Faculty board Faculty board

(not used)

Virtually none

1 Thematic and financial

prioritisation

Faculty management

shared with university

executive board and

senior staff

Faculty management Faculty management Faculty

management (not

Growth of two

research areas

prioritisation

institute leaders

University executive

board shared with

expert committees

board shared with

funding agencies and

national elites

University

executive board

(not used)

Growth of selected

research areas

3 Negative selection of

topics

Dean shared with

internal committee

Dean shared with

science policy

committee

Dean shared with

science policy

committee

Dean head of

Department (used

see next row)

see next row

topics

Dean shared with

internal committee

Head of department

shared with dean

Head of department

shared with dean

none Removal of several

research areas

topics

board shared with

external experts

Dean and university

executive board

Dean and university

executive board

Dean (used) Removal of one

discipline loss

of one field

AlltheScien

tistsGone

consequences of these strategies at the national level of fields of researchIn each case the strategies described in the previous section have beenapplied by all universities that host groups from a field thereby creatingaggregate effects at the field level

The Extinction of Dutch Evolutionary Developmental Biology

Evolutionary developmental biology (evo-devo) is an interdisciplinaryresearch field that combines evolutionary biology and developmental biologyIt investigates the evolution of developmental processes aiming for amechanistic understanding of phenotypic change (Sommer 2009 p 416)It can be traced back to the end of the 1970s when it became more andmore obvious that neo-Darwinian theory was unable to account for all theempirical findings of evolutionary biology was fuelled by discoveries ofgenes (HOX genes) regulating embryonic development in the 1980s andgot further impulses through the development of molecular technologytechniques (Muller 2007 Raff 2000) Researchers who want to start evo-devo research may come from a variety of fields mainly from evolutionarybiology and developmental biology By now it is an accepted field world-wide still strongly rooted in its original disciplines but with its own confer-ences and journals and emerging career tracks (for an extendedrepresentation of the evo-devo case see Laudel et al this volume) It isfundamental research that is difficult to link to medical or agriculturalapplications It can be cost intensive if comparative experimental studiesare undertaken and will often fare low on common citation-based indica-tors due to its newness and interdisciplinary nature

Evo-devo was originally deemed an underdeveloped area in theNetherlands by the scientific community which led to the creation of achair for evo-devo at one university This chair was created at the initiativeof a bioscience faculty and was co-funded by the research council and theuniversity Shortly thereafter however evo-devo research suffered fromnegative selection of related fields at several universities and prioritisationof other areas at other universities

At the university at which the chair was created a subsequent longphase of reorganisations eventually led to the closure of an evolutionarybiology institute The top management of the university had put pressureon the faculty to solve their budget problems An lsquoexpert committeersquo wasrecruited for preparing the decision It considered the evolutionary biol-ogy grouprsquos potential to acquire external funding insufficient despite the

high quality of the research which was confirmed by a previous regularevaluation The decision to close evolutionary biology had an effect thatnone of the profile builders anticipated the largest and internationallyhighly recognised Dutch evo-devo group left the university shortly afterThe prolonged process of decision-making about the fate of researchgroups had led to an atmosphere that researchers considered as unfavour-able Two leading evo-devo researchers went abroad Although movesbetween universities usually have several reasons including the prospectsof the new post (as was the case with those two researchers) a generalpattern emerges in this case of simultaneous migration Several tenuredsenior researchers moved to posts elsewhere which can be traced back tothe reorganisation

What I have seen now already a couple of times happening in various places where Irsquove

worked people get a bit sort of nervous if theyrsquore getting an opportunity to [establish]

a good research group elsewhere they leave Two or three good people left in that

A similar fate was met by one of the first Dutch evo-devo groups whohad just started to move into the direction of evo-devo The group leader(a professor) had secured an external grant and had several PhD studentsand postdocs working in this area In the late 1990s the faculty had toreduce its number of permanent positions due to financial problems As aresult the professor was forced to retire The faculty appointed a newchair in a priority area which would enable large-scale collaborationswith others PhD students and postdocs whose professors retired or wereobliged to retire went abroad or left science As a result the original evo-devo research lines that had only just started to gain momentumstopped

In addition to universities there is one public research institute (theHubrecht Institute funded by the Dutch Academy of Science) that is apotential host for evo-devo research This institute had a long tradition indevelopmental biology it is fully autonomous and the directions of itsresearch are decided by the institutersquos director This director a member ofthe national elite of biological research was not interested in and even hos-tile to evo-devo research Owing to his specific interest the institute hasshifted its research away from developmental biology and has recruitedresearch groups in other areas of biology The second reason for this shiftis the strong dependence of the institute on external funding and fundingagenciesrsquo thematic expectations (see below) As a consequence almost noevo-devo research is conducted in this institute

Apart from effects that can be traced back to universitiesrsquo and the publicresearch institutersquos profile-building activities other conditions had a stronginfluence on the development of the field most notably external fundingopportunities In recent years the most important funding agency for fun-damental biological research in the Netherlands has shifted its fundingmore and more into an applied direction and earmarked thematic pro-grams Success rates for small investigator grants have gone down from30 to 20 in the biology area This led to significant increase in the compe-tition for grants between evo-devo researchers and well established funda-mental biology groups

There is this orientation towards fundamental research in Arabidopsis There excellent

groups exist And they are very competitive in these grants for example Vidi or Vrije

Competitie But with evo-devo I felt we never had much of a chance

Several researchers had difficulties in pursuing their evo-devo researchbecause they could not get grants for it The unfavourable climate contrib-uted to career decisions to leave the Netherlands and to do evo-devoresearch abroad

Different from other countries who have large well-known evo-devolaboratories (eg the United Kingdom Germany Switzerland) research inevo-devo in the Netherlands exists today as scattered activities of indivi-duals What the researchers themselves refer to as lsquohardcorersquo experimentalevo-devo research hardly exists

So I would say I actually donrsquot know if evo-devo has any significant presence in the

Netherlands any more [hellip] I think evo-devo doesnrsquot really have a presence in the

Netherlands unless there may be some plants people that do evo-devo type research

[hellip] Some of its work may be classifiable as evo-devo But I wouldnrsquot know (Dutch

evo-devo researcher currently working abroad)

The combined actions of authorities state expectations funding agen-ciesrsquo priorities managerial decisions of research organisations all wentinto the same direction and thus drove evo-devo in the Netherlands closeto extinction

Chilling out Bose-Einstein Condensation

Bose-Einstein condensate (BEC) is a state of matter that occurs whengases of atoms or subatomic particles are cooled to near absolute zerodegree Kelvin (see Laudel et al this volume on BEC) In 1995 the firstBEC was experimentally produced by two US groups This achievement

has first triggered attempts at replication After it became apparent thatBEC can be used for a wide range of fundamental research several sub-fields of physics BEC research grew rapidly Today more than 100experimental groups and a multitude of theoretical groups worldwidework on BEC

The Netherlands had a long tradition in atomic and laser physics andwere particularly well positioned to move in this new direction One Dutchresearch group belonged to the pioneers of BEC research and made impor-tant contributions on the route of experimentally realising a BEC for thefirst time Other groups had already incorporated laser cooling techniquesin their research However the successful production of BECs and interest-ing research with them came rather slowly In the mid-2000s four groupshad succeeded in producing a BEC and established experimental BEC astheir main line of research

When universities began to build research profiles atomic and molecularoptics (AMO) physicists in general and BEC researchers in particularfound themselves in a situation where they did not fit into a priority areaAlthough there was no immediate threat for the existing groups notbelonging to a priority area also meant that AMO physics and BECresearch is not a growth area New professors are more likely to beappointed in the priority areas of the universities

And now the focus is shifting to use these lasers for life science and biomedical optics

So we have some new professors in biomedical optics Thatrsquos where most of the money

I And who made this shift

That is on the highest level of the university At the highest level of university is decided

that the natural sciences should focus their research more on the medical center (thatrsquos

the big thing over there) Actually that does not hold so much for our research So we

are a bit in a different situation But if we keep up the quality we survive

I But why actually It sounded as if the university didnrsquot really hinder your research

but didnrsquot also promote it too much

No Now in recent years universities are struggling And they need to have their profile

And [our university] focuses on medical research

I Yes But you have your lasers and you get your money from [the funding agency]

FOM How can they make you uneasy

They donrsquot throw us out yet But in chemistry now they made plans hellip chemistry had

to make budget cuts And they keep the people who are in the life sciences So if there

is at any point these political decisions are coming that there must be budget cuts made

the [university] would choose to keep the life sciences

But the problem is that BEC research is not the focus of this department [hellip] We are in

another reorganisation at the moment And it is much easier if you can say okay but

Irsquom in the focus

Beside research profile-building activities of universities profile-buildingdecisions of the most important non-university institute in this area hadalso influenced BEC research Strongly backed up by the national physicscommunity the agency funding the institute decided to move researchtowards biophysics Biophysics was considered a neglected field and neces-sary to catch up with international developments Consequently it wasdecided not to invest into BEC research despite the fact that the instituteprovided an ideal infrastructure and hosted the most successful BECresearcher

As a result of these trends the dynamics of BEC research in theNetherlands markedly differs from the international trend While the fieldstill grows worldwide new professorships are devoted to BEC in Germanyand countries like Switzerland initiate research with BEC (see Laudel et althis volume) Dutch BEC research is stagnating rather than growing andits situation may become precarious

CONCLUSIONS CAUSES AND CONSEQUENCES OF

PROFILE BUILDING

Profile building at Dutch universities emerged at the intersection of threemajor shifts in the governance of research over the last decades Scienceand higher education policy considers that the Dutch state is unable tofund research across all fields and topics and has introduced policies tofocus research funding on some areas while by implication abandoningothers This policy includes the demand that universities should not all dothe same research which is implemented rather forcefully The requirementfor focus and (critical) mass is accompanied by directions what to focus onand where to create mass The policy shift to more and stronger demandsfor lsquousefulrsquo research has made the state define research areas that promisecontributions to the solution of societal problems These areas have beenset up as targets for university profiles and have been made attractive bythe redistribution of public funding

Thus the history of profile building of Dutch universities is a history ofuniversities and academics losing authority over research topics Universitieslose authority to the state and researchers lose authority to both the state

and their universities While the latter process is in line with the NPM ideol-ogy which promotes increasing action capabilities of universities bystrengthening hierarchical steering the former is not Quite contrary to thepolitical announcements of increasing the autonomy of universities Dutchuniversities have lost autonomy over their research profiles Higher educa-tion reforms appear to have merely changed the channel though whichauthority over universities is exercised which makes the newly grantedautonomy contingent on state interests (see also Enders De Boer amp Weyer2013 on lsquoregulatory autonomyrsquo)

The shift in authority relations concerning the formulation of researchtopics that has been produced by Dutch NPM reforms can be described asfollows Academics at Dutch universities retain the authority over their for-mulation of research goals although they lost some of their authority con-cerning the profile building Their authority is epistemic authority that isauthority that is based on their exclusive access to the knowledge that isnecessary to design research topics and goals (Whitley amp Glaser thisvolume) This is why the university leadership which by now has the ulti-mate formal authority to shape research profiles devolves any processes inwhich a profile needs to be constructed from existing research Only if bud-getary considerations are dominant and consequences for research seemless important university management uses formal authority to overridethe epistemic authority of researchers usually by using outside experts as asource of information on the research and as a source that is supposed tolegitimise their decisions

The consequences of these two profile-building processes are manifoldOnly some of them are anticipated and even fewer are taken into accountin the two processes of profile building Within universities at least someresearchers in fields that are given a lower priority feel unwelcome andthreatened Even if a prioritisation strategy lets them stay with the same orslightly reduced funding they know that their position will become precar-ious whenever budget problems occur Prioritisation strategies andincreased selectivity of resource allocation in the university create a ten-dency to crowd out researchers in non-priority areas These researchersbegin to leave the university whenever the opportunity arises

These shifts and the removal of topics usually sever collaborative tieswithin the university They thus affect the removed or devalued topicrsquos(previous) environment in ways that are difficult to understand and to pre-dict This was the case with the astronomy institute at Utrecht Universitywhose disappearance discouraged theoretical physicists It may also affectthe general climate for research as we have seen for the field of evo-devo

where the closure of one institute from a closely related field led evo-devoresearchers to leave the university The case of BEC is quite similar becausethe development in the Netherlands ground to a halt at a time where BECsbecome a research method that is used in an increasing number of subfieldsof physics

The most severe consequences of profile building occur when a field dis-appears entirely This kind of quasi-market failure has been discussed inrelation to research evaluation systems (Glaser 2007) The creation ofresearch profiles by Dutch universities occurs in a similar situationThrough creating their profiles the universities compete for money boththe additional block funding that is conditional on having the lsquorightrsquo profileand research council funding that is increasingly shaped by research priori-ties set by the Dutch state The signals the universities have to follow arethe same for all universities and include priority areas of research set by thestate the state expectation of utility and the necessity of maximising grantincome which is increasingly shaped by the same state expectations Wehave shown that the management of Dutch universities has the formalauthority to significantly strengthen or to remove research areas from theirprofile and that university and faculty leaders do exercise this authority Ifall universities follow the same signals when shaping their profiles they arelikely to positively and negatively select the same research topics whichleads to the growth of some fields and the disappearance of others at thenational level This has almost happened with Dutch evo-devo researchand may well happen with BEC in the future

Thus the current authority relations in Dutch higher education create apropensity for quasi-market failure because they combine thematic priori-ties prescribed by the state a powerful demand for profile building of uni-versities financial problems of many universities and authority structuresat universities that enable far-reaching changes of profiles The warningabout a possible quasi-market failure in Germany namely the disappear-ance of the so-called lsquosmall disciplinesrsquo (HRK 2007) indicates another pos-sible reason for such failure namely the ubiquitous belief that the creationof lsquocritical massrsquo of research activities in an area creates better research(Schiene amp Schimank 2007) Such a quasi-market failure would have takena very long time to occur due to the much lower authority of universitymanagement in Germany which can only intervene when professorshipsbecome vacant However the example indicates a possible systematic side-effect of NPM Transforming the governance of universities to sendingstrong signals via quasi-markets creates opportunities for quasi-marketfailures

The argument that the small size and financial situation of the Dutchhigher education system require the concentration of resources does soundconvincing and the current processes will very likely lead to a strengthen-ing of research areas science policy wants strengthened The loss of univer-sity autonomy in the process and the shrinking or disappearance ofresearch fields are however inevitable side-effects of these processes whoselong-term impact on the Dutch science system is difficult to predict

1 The Executive Board is the highest internal governing and administrative bodyin higher education institutions It consists of three persons including the RectorMagnificus All members are appointed for a 4-year term by the UniversityrsquosSupervisory Board after receiving the views of the University Council It assumesthe ultimate responsibility for the policies of the university including the appoint-ment of deans directors and professors The Board is legally entitled to initiate newprogrammes and related research institutes and is responsible for managing the uni-versityrsquos finances and other management tasks The division of labour between themembers of the Executive Board differs from university to university

ACKNOWLEDGEMENTS

We thank Jochen Glaser for numerous discussions and ideas about mergingour two empirical studies into one article We would also like to thank ananonymous reviewer for his or her critical reading of the manuscript Thiswork was supported by two project grants from the NetherlandsOrganisation for Scientific Research (NWO) and a travel grant from theEuropean Science Foundation (ESF)

REFERENCES

Blume S amp Spaapen J (1988) External assessment and ldquoconditional financingrdquo of research

in Dutch universities Minerva 26 130

Clark B R (1998) Creating entrepreneurial universities Organizational pathways of transfor-

mation New York NY Pergamon

De Boer H Enders J amp Leisyte L (2007) Public sector reform in Dutch higher education

Enders J De Boer H amp Weyer E (2013) Regulatory autonomy and performance The

reform of higher education re-visited Higher Education 65 523

Glaser J (2007) The social orders of research evaluation systems In R Whitley amp J Glaser

tems (pp 245266) Dordrecht the Netherlands Springer

Glaser J (2012) Are universities professional organisations Paper presented at the sub-theme

17 Organizing science The increasingly formal structuring of academic research 28th

EGOS Colloquium 2012 Helsinki Finland

specific epistemic conditions affect authority relations and their consequences In

R Whitley J Glaser amp L Engwall (Eds) Reconfiguring knowledge production

Glaser J amp Laudel G (2009) Identifying individual research trails In B Larsen amp J Leta

(Eds) Proceedings of the 12th international society for scientometrics and informetrics

(ISSIrsquo09) (pp 841845) Rio de Janeiro Brazil

Glaser J amp Laudel G (2013) Life with and without coding Two methods for early-stage

data analysis in qualitative research aiming at causal explanations [96 paragraphs]

Forum Qualitative Sozialforschung Forum Qualitative Social Research 14 Retrieved

from httpnbn-resolvingdeurnnbnde0114-fqs130254 Accessed on February 12

HRK (Hochschulenrektorenkonferenz) (2007) Die Zukunft der kleinen Facher Potenziale -

Herausforderungen - Perspektiven Bonn Germany Hochschulenrektorenkonferenz

Jansen D (Ed) (2010) Governance and performance in the German public research sector

Disciplinary differences Dordrecht the Netherlands Springer

Keller C U (2012 April 25) Sterrekundig Instituut Utrecht The last years 370 years of

astronomy in Utrecht Proceedings of a conference held at Hotel Leeuwenhorst

Noordwijkerhout The Netherlands San Francisco CA Astronomical Society of the

Pacific Retrieved from httparxivorgabs12084095 Accessed on February 12 2014

Klumpp M De Boer H amp Vossensteyn H (2013) Comparing national policies on institu-

tional profiling in Germany and the Netherlands Comparative education published

online Retrived from httpwwwtandfonlinecomdoipdf101080030500682013

834558 Accessed on April 5 2014

KNAW (De Koninklijke Nederlandse Akademie van Wetenschappen) (2013) Effecten van

universitaire profilering en topsectorenbeleid op de wetenschap in Nederland Een eerste

kritische Reflectie Amsterdam the Netherlands KNAW Retrieved from httpsknaw

nlnlactueelpublicatieseffecten-van-universitaire-profilering-en-topsectorenbeleid-op-de-

wetenschap-in-nederland Accessed on February 16 2014

Lange S (2007) The basic state of research in Germany Conditions of knowledge production

pre-evaluation In R Whitley amp J Glaser (Eds) The changing governance of the

sciences (pp 153170) Dordrecht the Netherlands Springer

Laudel G (2012) Doing something new in the Netherlands The impact of research organisa-

tions and funding agencies on the start of new research lines Paper presented at the sub-

theme 17 Organizing science The increasingly formal structuring of academic research

28th EGOS Colloquium 2012 Helsinki Finland

Lee F (2009) A history of heterodox economics London Routledge

Leibfried S (Ed) (2010) Die Exellenzintiative - Zwischenbilanz und Perspektiven Frankfurt

Germany Campus Verlag

Leisyte L (2007) University governance and academic research Case studies of research units in

Dutch and English universities Enschede the Netherlands CHEPS University of Twente

Leisyte L Enders J amp De Boer H (2010) Mediating problem choice Academic research-

ersrsquo responses to changes in their institutional environment In R Whitley J Glaser amp

Lucas L (2006) The research game in academic life Maidenhead UK SRHEOpen

University Press

vention in Australia Cambridge UK Cambridge University Press

Muller G B (2007) Six memos for Evo-Devo In M D Laubichler amp J Maienschein (Eds)

Embryology to evo-devo A history of developmental evolution (pp 499524)

Cambridge MA MIT Press

OrsquoMahony S amp Lakhani K R (2011) Organizations in the shadow of communities In C

Marquis M Lounsbury amp R Greenwood (Eds) Communities and organizations

Research in the Sociology of Organizations (Vol 33 pp 336) Bingley UK Emerald

Pietila M (2014) The many faces of research profiling Academic leadersrsquo conceptions of

research steering Higher Education 67 303316

Raff R A (2000) Evo-devo the evolution of a new discipline Nature Reviews Genetics 1

Rafols I Leydesdorff L OrsquoHare A Nightingale P amp Stirling A (2012) How journal

rankings can suppress interdisciplinary research A comparison between innovation stu-

dies and business amp management Research Policy 41 1262-1282

work of the Academic Advisory Council in lower Saxony (FRG) In R Whitley amp

tion systems (pp 171190) Dordrecht the Netherlands Springer

Seidel M-D L amp Stewart K J (2011) An initial description of the C-form In C Marquis

M Lounsbury amp R Greenwood (Eds) Communities and organizations Research in

the Sociology of Organizations (Vol 33 pp 3772) Bingley UK Emerald Group

Sommer R J (2009) The future of evo-devo Model systems and evolutionary theory Nature

Reviews Genetics 10 416422

Technopolis (2011) Mini country report The Netherlands Retrieved from httpeceuropa

euenterprisepoliciesinnovationfilescountryreportsnetherlands_enpdf Accessed on

February 12 2014

Van den Besselaar P amp Horlings E (2010) Focus en massa in het wetenschappelijk onder-

zoek de Nederlandse onderzoeksportfolio in internationaal perspectief Den Haag the

Netherlands Rathenau Instituut

Van der Meulen B (2007) Interfering governance and emerging centres of control University

research evaluation in the Netherlands In R Whitley amp J Glaser (Eds) The changing

Von Stuckrad T amp Glaser J (2012) Es ist hochstens eine Kollegenschelte moglich aber die

bringt nichts - Kontingente und strukturelle Handlungsbeschrankungen der intrauniver-

sitaren Forschungsgovernance In U Wilkesmann amp C J Schmid (Eds) Hochschule

als organisation (pp 223243) Wiesbaden Germany VS-Verlag

Wenger E C amp Snyder W M (2000) Communities of practice The organizational frontier

Harvard Business Review 78 139145

Westerheijden D F De Boer H amp Enders J (2009) Netherlands An lsquoEchternachrsquo proces-

sion in different directions Oscillating steps towards reform In C Paradeise E Reale

I Bleiklie amp E Ferlie (Eds) University governance Western Europe comparative per-

spectives (pp 103125) Dordrecht the Netherlands Springer

Weyer E (2014) From loose to tight management Seeking evidence for archetype change in

Dutch and English higher education PhD thesis University of Twente CHEPS

Enschede the Netherlands

Whitley R (2008) Universities as strategic actors Limitations and variations In L Engwall amp

Whitley R (2010) Reconfiguring the public sciences The impact of governance changes on

authority and innovation in public science systems In R Whitley J Glaser amp

MERGER MANIA IN SCIENCE

ORGANIZATIONAL

RESTRUCTURING AND PATTERNS

OF COOPERATION IN AN

ACADEMIC RESEARCH CENTRE

Julien Barrier

ABSTRACT

While the formal structures of universities may predominantly reflect cere-monial rather than functional purposes attempts at changing them areusually a fertile ground for academic conflicts Taking this apparent contra-diction as a starting point the aim of this article is to explore the intriguingrole of formal structures in academic settings Drawing on a case study ofa merger and organizational restructuring process in an academic researchcentre it shows how symbolic responses to institutional pressures may haveactual consequences on research practices beyond myth and ceremony

Keywords Decoupling and recoupling processes inhabited institu-tions perspective mergers in higher education organization ofacademic research scientific cooperation

ISSN 0733-558Xdoi101108S0733-558X20140000042005

INTRODUCTION

There is an enduring intriguing puzzle in academic institutions On the onehand the design of their formal structures often seems to serve ceremonialrather than functional purposes predominantly reflecting conformity toinstitutionalized myths that is widespread taken-for-granted beliefs aboutrational models of organization (Meyer amp Rowan 1977) Universities areoften cited as a prime example of decoupling between structures and coreactivities For instance the rise of technology transfer offices in Germanuniversities has been described as a symbolic response to institutional pres-sures having little if any impact on actual university-industry relation-ships (Krucken 2003) On the other hand as pointed out by Musselin(2007) if university formal structures were all about myth and ceremonywhy do attempts at changing them usually prove so difficult and why arethey such a fertile ground for academic politics

Taking this apparent contradiction as a starting point the aim of thisarticle is to consider how and in what circumstances the formal organiza-tion of research units could affect patterns of collaboration and coordina-tion drawing on a longitudinal case study of the merger and restructuringof a French university research centre in electronic engineering Followingrecent developments in organizational institutionalism promoting a moreactor-centred approach (Hallett amp Ventresca 2006) this empirical setting istaken as a case for exploring the micro-level processes involved in thedesign adoption and transformation of structures in academic organiza-tions while not neglecting their relation to broader institutional pressures

Indeed concentration and mergers of research units are arguably notonly a significant phenomenon in the academic field in Europe but are alsoa key research site for analysing the interplay between institutional pres-sures and local work arrangements Like the lsquomerger maniarsquo in the hospitalsector in the United States (Kitchener 2002) the surge of policies advocat-ing the merger and concentration of research units exemplifies the rise ofnew powerful rational myths in the academic field For more than a decadenow policy makers in various European countries including France(Vinck amp Zarama 2007) have called for a concentration of scientificresources and personnel in lsquocritical sizersquo institutes or research centres inorder to boost their performance and outputs1 a trend Bonaccorsi andDaraio (2005) refer to as a lsquopolicy mythrsquo given the ambiguous nature ofthe relation between organizational size and scientific output

The rest of the article is structured as follows After outlining the analy-tical approach and the empirical materials of this article background data

142 JULIEN BARRIER

on the French higher education and research system are provided Thenresults from the case study are presented and discussed Although formalstructures are decoupled from core work patterns (eg cooperation pat-terns) to a certain extent our results show how reshaping structures maynonetheless have real consequences by triggering an unexpected chain ofreactions

BEYOND MYTH AND CEREMONY HOW

FORMAL STRUCTURES MAY MATTER IN

ACADEMIC SETTINGS

In order to present the theoretical background of the article this sectionoutlines three analytical issues that informed the formulation of ourresearch question and the analysis of empirical data

First this article considers the organizational formalization of academicsettings as a key yet under-researched issue Since the late 1970s most scho-lars in the field of Science and Technology and Society (STS) have beenmore concerned with bench practices or micro-processes of network forma-tion than with organizations and institutions (Frickel amp Moore 2006) atendency most evident in early laboratory studies (eg Latour amp Woolgar1979) This is not to say that science studies have ignored altogether organi-zational issues showing for instance how scientists mobilize organizationalresources to support their research agendas or the emergence of a disciplineBut STS scholars have showed little interest for the massive processes oforganizational rationalization that are so crucial to institutionalists whostress the dramatic expansion of formal structures in the field of higher edu-cation and research over the last decades (Drori Meyer Ramirez ampSchofer 2006 Krucken amp Meier 2006) As a result while the core activitiesof scientists have been increasingly dependent on their involvement intomultiple intersecting organizational forms (Barrier 2011) the consequencesof this trend at the micro-level are rarely addressed

Second although the increasing formal organization of academic activ-ities may often result in window dressing and symbolic compliance oneshould be cautious not to overestimate the extent of decoupling betweenstructures and practices The notion of decoupling originally put forward byMeyer and Rowan (1977) pointed to the gaps between the adoptionof structures reflecting conformity to institutional myths and everyday prac-tices in organizations in order to question the then-dominant assumption

143Merger Mania in Science

that formal structures were always effective means of coordination and con-trol2 Yet the original meaning of the notion has somehow suffered from itsvery success as decoupling is often assumed to be the default response oforganizations in general and academic institutions in particular to insti-tutional pressures Indeed Bromley and Powell (2012) state that much ofthe recent literature referring to decoupling has unfortunately equated adap-tation to institutional pressures with mere window dressing Resisting thisinterpretation they further note that lsquoeven ceremonial adoption can uninten-tionally trigger a chain of reactions that have real organizational effectsrsquo(Bromley amp Powell 2012 p 3) Thus instead of assuming that decouplingoccurs invariably it is arguably more fruitful to untangle the complex pro-cesses that mediate edit and shape the translation of institutionalized mythsinto practices a theme central to the lsquoinhabited institutionsrsquo perspective inorganizational institutionalism (Hallett amp Ventresca 2006)

Studies in this vein make at least two relevant contributions to ourresearch questions First they favour longitudinal approaches and focus oninteractions and the construction of meaning at the micro-level in order toidentify the complex indirect processes translating institutional pressuresinto core activities how they are lsquorecoupledrsquo (Hallett 2010 Sauder ampEspeland 2009) Second they draw attention to the internal diversity ofresponses to institutional pressures Binder (2007) shows that departmentswithin the same organization are structured along different institutionallogics which in turn shape the way they creatively adapt to externalpressures

Third a key issue is to capture how the design of formal structures mayshape and be shaped by intra-organizational dynamics Although aca-demic research units may exhibit significant variation in their organizationdepending on institutional context size and discipline they can arguablybe described as loosely coupled systems (Weick 1976) This is especiallyrelevant in the French context where they usually come as a confederationof research groups of different sizes more or less connected to each other(Louvel 2010) Having said that the question remains to understand whatnonetheless binds researchers together A first answer is to look at patternsof interdependence among researchers In the experimental sciencesresearch groups are usually based on a division of labour with functionalinterdependence stemming from the specialization of their members onvarious epistemic tasks (Shinn 1988) Yet this form of interdependenceis mostly significant at the research group level which is usually thebasic organized unit of knowledge production it is far less significant atthe institute or research centre level In addition beyond epistemic

144 JULIEN BARRIER

interdependence one may identify resource interdependence resulting fromthe need to manage of critical resources such as posts funding or researchfacilities meaning that research groups and centres alike tend to be animportant locus for the formation of local rules governing the allocation andmanagement of resources However looking at functional or task interde-pendence might not be sufficient

In a paper on the role of formal structures in universities departmentsbasic research units institutes schools and so on Musselin (1990) arguesthat they strikingly differ from similar structures in hierarchical mechanis-tic bureaucratic organizations In universities she argues the contours ofstructures are not as much based on patterns of interdependence asthey reflect the formation of defensive territories Basically they enablemembers to draw boundaries to assert their autonomy vis-a-vis othergroups (eg members of the chemistry department would not have a sayabout what is going on in the physics department) which favours theemergence of in-group identity and facilitates the development of localrules of cooperation The existence of a structure does not mean thatcooperation will ensue but it supports the maintenance of relationshipsthat would be fragile otherwise Consequently to understand how institu-tional pressures transform formal structures and affect work floorpractices one must not only examine how they relate to functional interde-pendence but also how they may destabilize or reconstruct existing defen-sive territories

DATA AND METHODS

This article is based on a longitudinal case study of a large research centrethat resulted from the merger of two research units This organization willbe referred to as the Electronic Engineering Institute (EEI a pseudonymaiming to protect the anonymity of interviewees) The study was part of alarger project on the changing conditions for academic research in Franceover the 19802007 period (Barrier 2011) It focused on the field ofacademic research in electronic engineering and micronanotechnologywhich was selected as an appropriate case for analysing the reconfigurationof relations between government academic science and industry The pri-mary aim of this project was to understand how changes in fundingregimes especially the rise of project funding affected the organizationof research and the professional autonomy of academic researchers

Empirically the design of the project was mainly based on the longitudi-nal study of two large research centres including the EEI These two orga-nizations were conceived as focal research sites that allowed for thecollection of rich data on several dimensions including funding patternsbut also the internal management and organization of the research centresas well as their relations with their institutional environment Data werealso collected to position these cases in a broader national context

In addition to the realization of a total of 128 semi-directive interviews(37 of which were conducted for the EEI case study) with academics mem-bers of university governance boards research administrators policymakers and industrial researchers this project combined archival researchquantitative data on funding as well as bibliometric data This designallowed for a strong contextualization of local observations and the trian-gulation of data It also allowed for the exploration of emerging themesFor instance when fieldwork was carried on at EEI it was decided to col-lect additional data on the merger process since preliminary interviews pre-sented the merger as a central issue Subsequently interviews with researchadministrators at the national level included questions about mergersFinally in order to get a more precise view of intra-organizational coopera-tion within EEI publication data were collected from the Web of scienceWhile cooperation does not necessarily result in joint publication (Laudel2002) co-authorship patterns can yield further insights to complement andinterpret interview data with researchers MS Access was then used to sortout raw data and identify co-authorship networks Then the softwarePajek (De Nooy Mrvar amp Batagelj 2005) was used to create examine anddraw up these networks

BACKGROUND

Positioning the EEI in the French Public Research System

A lsquojoint research unitrsquo between a public research organization the CentreNational pour la Recherche Scientifique (National Centre for ScientificResearch CNRS) and three higher education institutions the EEI reflectsa distinctive feature of the French science system (Louvel 2010 ThevesLepori amp Laredo 2007) In this context the organization of academicscience as joint research units combining public research organizations withuniversities is the rule rather than the exception

146 JULIEN BARRIER

The EEI receives funding from both the CNRS and higher educationinstitutions and includes both CNRS and higher education personnel Witha minority of CNRS researchers the staffing of EEI reflects a general trendin fields such as computer science or electronic engineering where the cen-tre of gravity of research units leans towards university professors ratherthan CNRS researchers This relates to the demographics of these disci-plines in France in which the growth of research activities has been primar-ily driven by a rising number of students in the 1990s leading to a rapidincrease in the number of university positions not matched by posts inpublic research organizations (Laredo amp Mustar 2002) (Table 1)

Without going into too much detail it is important to note that theaffiliation of a research unit to the CNRS depends on an evaluation per-formed by the Comite National3 This permanent committee is in charge ofthe evaluation of CNRS research units including their creation and disso-lution and decisions about the affiliation of new units to CNRS While

Table 1 The Electronic Engineering Institute in 2006

Main Research Directions Microelectronics Electromagnetism Telecommunication

Devices and Systems Signal and Image Processing

Staff composition (rounded

numbers)

CNRS researchers 5

Professors 60

Technical and administrative staff 25

PhD students 100

Localization and

Institutional affiliations

MidCity a large city with a student population of about

UniMedium medium size multidisciplinary university

TechSchool public engineering school

EngineeringSchool private engineering school (since 2004)

Operating budget (without

staff salaries)

Core funding 16

Project funding from national European or international

agencies 45

Funding by local and regional government 17

Funding from industry 22

criteria of scientific excellence and performance are central in the evalua-tions performed by the Comite National it also pays significant attention tothe organizational capacities displayed by research centres (FixariMoisdon amp Pallez 2009) A central script in the definition of research unitsas organizations is that they are expected to represent lsquomore than the meresum of their partsrsquo by promoting the sharing of facilities or resources aswell as encouraging scientific synergies among their members (eg thedevelopment of research projects or themes involving different researchgroups) Observations of debates at the Comite National show that mem-bers often encourage restructuring in research units to foster more cohesionor to reduce internal conflicts (Fixari et al 2009) In sum CNRS evalua-tion processes rely on an institutionalized organizational template basedon the notion that research units are (or should be) cohesive integratedorganizations4

The affiliation of research units with the CNRS gives access to materialand symbolic resources In addition to providing extra funding in the formof CNRS block grants it allows units to host CNRS researchers whounlike their university colleagues are full-time researchers without manda-tory teaching duties5 In more symbolic terms an affiliation to CNRS istraditionally considered as a strong distinctive quality signal in the Frenchsystem prompting university research units to compete for the lsquoCNRSlabelrsquo

lsquoMerger Maniarsquo in the Academic Field The Growing Concentration ofResearch Units in Electronic Engineering

Since the late 1990s it has become common place among French academicleaders and research administrators to criticize the lsquoorganizational disper-sionrsquo of research structures and the overall lack of lsquolegibilityrsquo of the Frenchresearch system Thus the concentration of research units into larger orga-nizations has been strongly encouraged The main arguments for this policywere not only to develop economies of scale in the use of research facilitiesand administrative resources but also to develop lsquocritical sizersquo research col-lectivities and boost the national and international lsquovisibilityrsquo of researchcentres This evolution concerns all disciplines and the field of electronicengineering is no exception Data on CNRS units in electronic engineeringconfirms a growing concentration of researchers into larger and larger insti-tutes a trend that accelerated during the 20042012 period (Table 2)6

148 JULIEN BARRIER

The analysis of original data and secondary sources suggests that thislsquomerger maniarsquo has not only been fuelled by policy ideas circulating at thenational level but also by the interactions between three set of actorsCNRS universities and local government First although the position ofCNRS regarding the concentration of research centres has varied overtime the global trend has leaned towards the creation of large researchinstitutes While interviews with the CNRS leadership and members of theComite National indicated some reluctance with the merger movementadvocated by policy makers at the ministry level they were nonethelessseen as a lever to promote rationalization in the use of experimentalfacilities and to promote restructuring initiatives in order to enable theemergence of new research themes or groups

Second in the early 1990s university leaders have fostered the creationof new organizational layers with the merger of research units or thecreation of clusters Given the limited institutional autonomy of Frenchuniversities the creation of new structures indirectly allowed the leadershipto define priority areas in the allocation of resources (Barrier amp Mignot-Gerard 2013) Predisposed to jump onto the lsquomerger bandwagonrsquo univer-sity leaders thus contributed to its advent

Third during the same period local governments especially regionalauthorities have become more and more involved in the funding of highereducation and research taking up a major role in financing the construc-tion of buildings and research infrastructures (Aust amp Crespy 2009) Inaddition they have sought to foster coordination among institutionslocated in the same metropolitan area Such initiatives have come to beviewed in increasingly positive terms as the notion of lsquoinnovation clusterrsquohas become more and more popular among policy makers As a conse-quence local governments have favoured mergers between research units

Table 2 Distribution of Research Units Affiliated to CNRS by Size(Number of Staff Members) in the Field of Electronic and Electrical

Engineering 19842012

lt25 2550 51100 101150 151300 gt300

1984 378 378 178 22 44 0

1994 206 294 206 147 88 59

2004 70 233 326 209 70 93

2012 67 100 200 100 300 233

UNFOLDING THE CREATION OF THE EEI

The aim of this section is to describe how two research units TelecomLaband MicroLab merged to form the EEI in 2002 Because TelecomLab wasthe main driver of the merger process and since its former researchers com-pose the bulk of the institute created in 2002 the focus on the narrativewill be mainly centred on TelecomLab

How the Idea of a Merger Came About Aligning Local Arrangements withRising Institutional Pressures

The Early Days of TelecomLab An Organizational Umbrella LooselyAffected by Institutional Pressures (19831994)From its creation in the 1980s up to the 1990s the organization ofTelecomLab can be best described as loosely coupled first it was weaklyintegrated internally second institutional pressures to conform withCNRS recommendations were largely disconnected from work practices

TelecomLab was created in 1983 as a joint research unit between CNRSUniMedium and TechSchool in a context where UniMedium encouragedfaculty members to create formal research units and seek an affiliation toCNRS Within this context the creation of TelecomLab resulted from arather opportunistic strategy to fit into university priorities and acquire theCNRS label It was little more than an organizational umbrella for researchgroups that were independent from each other Although all groups wereinvolved in telecommunications-related research each group was essentiallyconceived as an independent entity For most of the 1980s and the 1990srelations between research groups were often tense whether verging oninterpersonal competition or mutual ignorance

Each research group had been built around a single man All of them were strong char-

acters And the relationships between those guys were not easy (hellip) We didnrsquot know

well people from other groups and we often discovered what the others were doing

only when we met them at conferences Sometimes we were also in competition when

applying for funding That was really uncomfortable (Professor)

The examination of TelecomLab board minutes confirms that thissituation of minimal cooperation that is limited to the maintenance ofan administrative umbrella persisted well into the 1990s For example aprofessor showed irritation with the suggestion that the board should meetmore often he argued that the number of meetings should be proportional

150 JULIEN BARRIER

to the money that was allocated by the lab thereby it was not necessaryto meet frequently7 Likewise while the 1996 activity report (coveringthe 19921995 period) rhetorically stresses the scientific cohesion acrossresearch groups the examination of publication outputs reveals thatthere was only one paper authored by members of different researchgroups and it was not a scientific article but a paper published in a pro-fessional outlet to present research themes at TelecomLab

Regarding external pressures during most of the 1980s up to the mid-1990s affiliation with the CNRS was viewed as a matter of prestige ratherthan a strategic concern CNRS was perceived as a distant parent organiza-tion which contributed little to the funding of research activities especiallywhen compared to contracts with industry and specialized telecom anddefense-related government labs The level of third party funding was higheven according to engineering sciences standards In addition TelecomLabdid not have any CNRS researchers and PhD candidates from TelecomLabrepeatedly failed to get recruited at CNRS thus the growth and renewal ofstaff depended on recruitments at UniMedium and TechSchool which inturn depended mainly on teaching needs In other words direct resource-dependence on CNRS was weak

That is not to say that affiliation to CNRS was deemed unimportantsince it provided symbolic status and some material resources Whilethe scientific performance of TelecomLab and the overall relevance of itsresearch activities were praised in evaluation reports window dressing wasdeveloped to exhibit compliance with the managerial and proceduralcriteria used in Comite National evaluations For instance activity reportswere carefully written to stress complementarities and cohesion acrossresearch groups although they were thin or non-existent Likewise aresearcher explained that he used not to disclose the full amount of hisindustrial contracts fearing that the lab would be considered too focusedon applied research according to CNRS standards

Aligning TelecomLab with Environmental Pressures The Crystallization of aMerger Project (19941998)At the beginning of the 1990s TelecomLab was involved in a project invol-ving the creation of a lsquoMicrowave and Radarrsquo Cluster at MidCity consist-ing in the relocation of TelecomLab and other research units in a newbuilding This project was essentially driven by opportunistic motives put-ting an appealing organizational umbrella on a set of heterogeneous activ-ities to secure resources But it also had side effects as it promptedTelecomLab members to develop new interpretations about their

environment with the idea that further coordination might be developed atthe local level including other research units on campus

By 19941995 these considerations started to echo concerns about theposition of TelecomLab in a changing environment Paul a senior profes-sor and the head of a research group based at TechSchool became thechampion of a new project for TelecomLab First he argued that lsquosome-thingrsquo had to be done for TelecomLab to become lsquostrategicrsquo for CNRSwhich had started to encourage the creation of large research units Therehad been concerns over the level of CNRS support at TelecomLab sincethe early 1990s In comparison to other research units also affiliated toCNRS TelecomLab members estimated that they did not fare well Risingstudent enrolment had fuelled the creation of new faculty positions atUniMedium and TechSchool leading to the growth of TelecomLabHowever TelecomLab was still not getting any researcher positions fromCNRS This issue was considered all the more problematic as CNRS waspromoting the allocation of research positions to many research centresoutside of the Parisian area which could have benefited TelecomLabSecond Paulrsquos discourse echoed changes in the funding environment asfaculty members thought that the lsquovisibilityrsquo of TelecomLab had to beimproved A large share of project funding at TelecomLab came from thedefence and telecom sectors In 19951996 two sets of reforms in thesesectors were perceived as a threat to TelecomLab activities DGA the mainpublic agency supporting defence-related RampD made sharp spending cutsRampD in the telecom sector was also in trouble with the privatization ofFrance Telecom and the subsequent reform of CNET the government labfor telecom research These changes fuelled the idea that competition forfunding would soon become harsher and that it was necessary to defendthe collective reputation of TelecomLab From this point of view it wasalso believed crucial to maximize chances to maintain the CNRS label

The CNRS label provides an official recognition to industrial partners Mentioning

your CNRS affiliation when you apply for a grant for example DGA grants also adds

weight to your application They wonrsquot give you more money but they will pay more

attention to your proposal (Professor)

Third Paul built on these concerns to advocate a merger withMicroLab a small CNRSUniMedium unit specialized in microelectronicsSince 19931994 UniMedium had begun to encourage closer connectionsbetween the two labs in order to promote the pooling of experimentalequipment Paul pushed the idea further by proposing a merger to createan institute that would be considered as a major regional hub in electronics

152 JULIEN BARRIER

and telecommunications Paulrsquos views were shared by colleagues atMicroLab who inferred from the observation of lsquosuccessfulrsquo research unitsthat lsquoundersizedrsquo laboratories would be more and more disadvantaged inthe competition for academic positions and funding and may eventuallylose the CNRS label8 They had in mind the creation of another researchcentre the Microelectronics and Physics Institute9 an example of regionalconcentration of research capacities they wanted to follow

We saw what was going on at the national level There was the Microelectronics and

Physics Institute They had gone through a merger they were getting positions money

from the Ministry of research industry local governments and so on And we were

just bystanders waiting on the side of the road By 19951996 we started to realize that

we had to jump in if we wanted to still exist within five years (Senior professor)

This line of argument dramatizing the consequences of inaction wascentral in Paulrsquos discourse In a 1998 presentation he argued that it waslsquothe last chance for MidCity to become a recognized place in electronicsable to compare with other established clusters in Francersquo In sum the mer-ger would provide for a new organizational umbrella that could boost thevisibility of TelecomLab activities and foster synergies between researchers

Some researchers disapproved of Paul as a person casting him as tooambitious But his commitment to develop a new project for TelecomLab inthe absence of competing alternatives put him in a forceful position to defendthe idea of a merger In 1999 members of TelecomLab governing boardvoted unanimously for the merger and for the choice of Paul as the new direc-tor of the research unit that would result from the merger the Electronicsand Engineering Institute The official creation of EEI was planned for theautumn of 2000 However over a few months rising conflicts about thedirection of the merger compromised the completion of the project

Designing the New Institute Solving Old Problems ReactivatingDormant Conflicts (19992002)

Seizing the Creation of a New Organizational Umbrella to RestructureCooperation among ResearchersThe merger proposal had been mainly motivated by the prospect of anincrease in external legitimacy But the very rhetoric of local synergies thatwas put forward to draw external support started to turn into an organiza-tional reality When discussions about the actual implementation of themerger started at fall 1999 a shift occurred as debates moved towards the

definition of the governance and organization of EEI By the same tokenthe merger became increasingly understood as an opportunity for restruc-turing and developing a new culture of cooperation thus potentiallybringing the future institute in closer alignment with CNRS organizationalscripts

First Paul pushed forward a reconfiguration of the boundaries ofresearch groups to foster more interaction between researchers working onsimilar research topics While MicroLab was merely to be included as oneresearch group former TelecomLab members were to be combined into twolarge groups that cut across the lines of previous groups Second proposalswere made to pool resources the use of experimental equipment acrossresearch groups would be rationalized and a share of external fundingwould be redistributed across research groups under certain conditions(eg if a research group did not spend all the money from a contract therest would be allocated to a pool that could be used to support otherresearch groups) Third it was argued that research group leaders shouldlimit or avoid internal competition for external resources in particularwhen applying for project funding This issue was explicitly related to craft-ing and maintaining a positive perception of the research centre

Recently answering a call for proposals different members of EEI who are now

regrouped into the Systems research group have submitted competing proposals with

different industrial partners It is important that funding applications should be first

disclosed to the heads of research groups in order to avoid this lsquocacophonyrsquo that is

harmful to the lsquobrand imagersquo we all want for our research unit (EEI governing board

September 1999)

This last point illustrates how the projection of a new image started toentail more than mere window dressing However the more the mergermaterialized into new internal organizational arrangements the more itcame to be seen as a threat to the prevailing order that governed interac-tions within TelecomLab It triggered various conflicts among senior mem-bers who were unwilling to commit themselves to the full adoption of theproposed changes

When Internal Conflicts Outweigh External Pressures Breaking up the NewInstituteConflicts about the merger grew so intense that by spring 2000 the projectwas about to be abandoned Growing concerns about the internal reorgani-zation had been heightened by Paulrsquos proposals about the choice of a dep-uty director and the overall direction of the merger he argued that EEI

154 JULIEN BARRIER

should grow bigger by absorbing more research groups Because Paulsought the support of TechSchool leadership to back his propositions hisactions were soon interpreted in terms of a lsquohostile takeoverrsquo byTechSchool over the future governance of EEI

These suggestions reactivated a series of dormant conflicts among seniorstaff Although these conflicts primarily revolved around Paulrsquos leadershipand the rivalry between TechSchool and UniMedium they also involvedcomplex long time interpersonal quarrels within TelecomLab As a resultit was proposed to break up EEI into two research units Mirroring the for-mation of interpersonal coalitions the boundaries of these two units didnot follow exactly the MicroLabTelecomLab divide nor the boundaries ofTechSchool and UniMedium but cut across research themes and institu-tional affiliation of faculty members

The Institute Resurrected Realigning Internal and External ExpectationsThe merger process could have stopped here However it was soon revivedIn the autumn of 2000 the response of Comite National was clear the exis-tence of two research units was declared unacceptable It threatened to sus-pend the CNRS label if the two units did not eventually merge supportingthe lsquocreation of a single research unit gathering all key actors in Electronicsin MidCityrsquo10

This decision was perceived by several researchers as an electroshock torevive the project Although CNRS did not initiate the merger proposal itacted as an lsquoinstitutional ratchetrsquo preventing the abortion of the process Inaddition to the dire prospect of losing the CNRS label it was feared thatthe process had gone too far to be aborted it would send a very negativesignal about the capacity of the local academic electronics community tocarry out collective projects in the future In other words given the growinglegitimacy of mergers and the expectations raised by the EEI merger project both internally and externally the process had triggered a dynamic ofcommitment among actors (Becker 1960) making the prospect of an abor-tion of the project increasingly costly This interpretation is further sup-ported by the fact that the initiative to revive the merger came from a loosecoalition of junior members from UniMedium and TechSchool who fearedthat their future would be compromised by the inability of senior membersto reach an agreement11

There was a long history of conflicts and we came to see it as a deadlock We were the

young generation and we thought that our future was being hijacked that we needed a

large structure with an international visibility It was time for us to express ourselves to

say what we wanted to do even if it meant to drive out those who were in power

positions (Professor)

For them reviving the project was the only viable alternative While aca-demics can decide to leave a research unit to join another laboratory therewas no other large research unit in electronic engineering at MidCity atbest only a few could have lsquoabandoned the shiprsquo The initiative of juniorswas described as a lsquocouprsquo against seniors as the main protagonists pre-viously involved in the merger including Paul were deliberately putaside of the process

Discussions among juniors rapidly converged on the need to find a newdirector among faculty members who had kept some distance from internalrivalries Joseph a senior professor who had been long involved in theadministration of the electronics and physics teaching department atUniMedium was found to be a consensual person who could pacify thesituation He officially became the director of EEI in 2002 The same yearthe Institute was re-affiliated to CNRS with a lsquoprobation clausersquo that is itdepended on the capacity of the new institute to solve internal conflicts anddemonstrate its scientific relevance before being fully affiliated to CNRSin 2004

EXTERNAL AND INTRA-ORGANIZATIONAL

IMPLICATIONS OF THE MERGER

This section is devoted to the consequences of the merger It focuses ontwo issues how the EEI came to be seen as a nodal organization at thelocal level and how the merger affected intra-organizational dynamicswithin EEI

From Rhetoric to Reality and Reciprocally The Construction of aNew Collective Identity

Crafting an External Image Turning the Institute as a Nodal Organizationat the Local LevelThe new director of EEI Joseph was instrumental in turning the creationof EEI into a powerful rhetorical vehicle to draw support from localstakeholders Presenting EEI as a potential nodal organization at the locallevel enabled him to mobilize support which contributed to turning the

156 JULIEN BARRIER

institute into a key local actor In other words while the merger partlyresulted from external pressures to comply with an organizational templatethe legitimacy drawn from the merger was strategically reinvested to leverfurther support

During the first years of his mandate Joseph spent most of his time pay-ing visits to local stakeholders involved in the ICT sector (eg businessassociations territorial development agencies local government) Whileindividual connections existed before Joseph was the first to engage in asystematic approach to the creation and maintenance of a network of rela-tions with local stakeholders Lab council minutes over the 20022006period frequently mention the participation of EEI to local and interna-tional tradeshows in the field of electronics and telecommunications

The mere size of EEI and the absence of internal conflict allowed Josephto fashion himself as the main spokesperson for the local electronic engineer-ing academic community The lsquosuccess storyrsquo of the creation of EEIresonated with narratives circulating about the need to promote more coor-dination at the local and regional level among actors in the ICT sectorindeed local officials were becoming concerned with the decline of the sector(two corporate RampD centres closed down at MidCity in 2002 and 2003)Articles in local newspapers presented the creation of EEI as a positive movetowards overcoming industrial challenges in the ICT sector12 and the insti-tute came to be perceived as a key organization by local stakeholders

Before the creation of EEI there was no large institute in the field of electronics here

simply because people were pulling dirty tricks on each other hellip For us it is very

important to develop synergies to build projects across universities engineering

schools public research organizations hellip We found the restructuring of EEI to be very

positive it brings unity and strength (Local government administrative officer in

charge of higher education innovation and research)

EEI started to be often cited as a prime example to illustrate the aca-demic research potential of MidCity becoming a showcase for visits bybusiness or government representatives

However the increasing connections between EEI and local stakeholdersdid not have only symbolic aspects First external stakeholders saw thedirector of EEI as a contact point with the academic community in the ICTfield Local government administrative officers for instance explained thatit was crucial for them to have a limited number of trusted interlocutorswithin local academic institution in order to delegate to them the coordina-tion of proposals for research and higher education projects For exampleMidCity local government solicited EEI to support the development of

new specialized teaching programmes in line with local industrial develop-ment projects they supported In turn the growing renown of EEI and itsinvolvement with local stakeholders meant that the research centre hadbecome a flagship operation for UniMedium and TechSchool Thegrowing importance of EEI also prompted a large research group fromEngineerSchool another institution at MidCity to seek affiliation withEEI in 2004 which contributed to the growth of the Institute

Thus EEI was able to draw support from a wide array of actors Thisproved critical as illustrated by the creation of a lsquocompetitiveness clusterrsquoat MidCity in the field of ICT in 20042005 in the context of a newnational policy initiative13 Since competitiveness clusters would serve aschannels for RampD funding their creation was a matter of fierce competi-tion among firms academic institutions as well as across industrial sectorsand scientific disciplines competing for the inclusion of their goals andpriorities into the definition of clusters at the local level (Younes 2012)

Initially the perimeter and the orientations of the MidCity ICT clusterhad been defined by a small group of representatives from industryRepresentatives from higher education and research institutions were onlyconsulted in a second phase When this phase started EEI was not consid-ered a fully relevant partner since the RampD agenda developed by industrialrepresentatives emphasized research themes in image processing andcomputer science Over a few months with the help of local governmentrepresentatives and EEIrsquos parent academic institutions (UniMediumTechSchool and EngineerSchool) which were statutorily associated tothe creation of the cluster Joseph defended the idea that EEI should be arelevant partner for the cluster

ComputerLab [a large and well established research institute in computer science at

MidCity] had been obviously included in the cluster but my job was to make sure that

EEI would not be forgotten I had to go to a lot of meetings (hellip) It was necessary to

show that we could fit into the agenda (hellip) We had to identify every bit of research

[within EEI] that could be relevant for them as it was dispersed in several research

groups

These efforts proved successful since EEI eventually became a memberof the ICT cluster and was involved in several RampD projects supported bythe cluster Although it is difficult to find hard evidence that fully supportsthis interpretation it is very likely that without both the legitimacy and theinternal coordination capacity lsquospeaking with one voicersquo that resultedfrom the creation of EEI the inclusion of the research unit in the ICT clus-ter would have been compromised

158 JULIEN BARRIER

Building a New Organizational SagaThe creation of EEI set up the foundations for a new organizational saga(Clark 1972) The process that led to the creation of the institute wasreframed in positive terms stressing the ability of researchers to overcomedivergences for the lsquogreater goodrsquo Moreover because junior professors hadplayed a major role in the resurrection of the merger project marginalizingolder faculty members interviewees estimated that the merger had pro-moted a more collegial approach to the organization of research groupsand decision making In addition it was believed that the mergerbrought individual benefits to lab members Some interviewees expresseddoubts about any actual change being brought by the creation of a newstructure it was just an administrative umbrella Yet most of them especially junior faculty members were convinced that the merger pro-vided both an increased reputation and an organizational interface thatfacilitated the funnelling of resources to EEI research groups

We have more visibility Thatrsquos important to attract prospective students hellip Large cor-

porations are hearing stuff about EEI They think telecommunications Thatrsquos

EEI hellip CNRS thinks telecommunications Well thatrsquos MidCity (Professor)

Organizational Restructuring and Changes in Patterns of CooperationThe Mixed Effects of the Merger

The merger also had an impact on the research conducted by EEI particu-larly the internal pattern of cooperation Initially one of the goals of themerger was also to promote an internal reorganization of research groupsin order to foster more cooperation between researchers and more coher-ence across research themes A first reorganization in 1999 defined threelarge research groups Given the internal conflicts of 2000 this structuringwas rather virtual But the same structure remained in place when EEI wasofficially established in 2002 Later in 2004 research group boundarieswere rearranged as one of them was split in two to form a new group (theImage research group) in the same year a team of researchers from anotherinstitution at MidCity EngineerSchool joined EEI to form a fifthresearch group the Communications research group

Appeasing Tensions Developing a New Culture of CooperationIn addition to being involved in the mobilization of external supportJoseph managed to reduce tensions and to promote more cohesion withinEEI Efforts were made to limit internal competition to develop more

coordination and to pool resources For example instead of being directlyredistributed to research groups 40 of core funding was reserved bydefault for actions in the lsquocollective interestrsquo of EEI members (eg support-ing external communication) New rules were also emerging to define prio-rities in the allocation of PhD bursaries and to decide on investments inshared equipment While interviewees made it clear that these forms ofinternal coordination were emerging and not fully stabilized they viewedthem as a notable achievement given the conflicts which had plaguedTelecomLab

Although these changes mitigated conflicts and the intensity of directcompetition across research groups the impact of the merger on patternsof internal scientific cooperation was mixed On the one hand publicationoutputs do not indicate significant change in the proportion of scientificpapers co-authored by members of different research groups suggestingthat inter-group cooperation remained extremely limited14 On the otherhand while the co-authorship of papers is a good indicator of fruitful tightscientific cooperation interview data suggests that the merger facilitatedlsquolooserrsquo forms of cooperation Given the epistemic distance betweenresearch groups operating in cognate yet distinct fields cooperationmainly took the form of advice and exchanges around ideas devices analy-tical models or equipment

We could not really function like [laboratory X] where they reorganize teams on a pro-

ject by project basis Itrsquos simpler for them they are all computer scientists so you can

have many combinations Here itrsquos different we are quite heterogeneous Irsquom a remote

sensing specialist I know as much as you do about microelectronics hellip Of course if I

have a propagation problem Irsquom not going to read all the literature Irsquoll go to see the

specialists from the other group (Professor)

Another form of cooperation was the joint supervision of master-levelstudents and research interns allowing sharing expertise and testing therelevance of emergent ideas In addition researchers also reportedinstances of cooperation in large collective research projects especiallywith industry The basic rationale for the promotion of internal coopera-tion on projects was to minimize the coordination costs involved in largeprojects with multiple partners This was deemed especially relevant forindustrial projects since they usually require more integration andfinalization

People from the Systems group collaborate with our group [Telecom devices] We can

develop an antenna and all the system around it Then they can tell their prospective

industrial partners look we can do a lot here No need to go looking somewhere

160 JULIEN BARRIER

else hellip You can develop different basic blocks in the same place thatrsquos easier for an

industrial partner to assemble and coordinate all the blocks afterwards (hellip) We have

been able to do this for project X for instance (Professor)

Again while they did not necessarily resulted in tight scientific coopera-tion (ie researchers involved in different work packages in a large projectmay adjust and learn from each other without developing joint resultsworthy of publication) the development of such projects represented astrong departure from the former situation In sum while it is absolutelyclear that the restructuring could not force cooperation upon researchers especially given the diverse scientific profile of EEI it is important tonote that it fostered attempts at cooperation and thus enabled potentialcooperation to occur The emergence of a collective external identity inaddition to the regulation of internal competition had increased the incen-tives for cooperation

Cooperation within Research GroupsThe restructuring created some changes in cooperation patterns withinresearch groups albeit with different outcomes from one group to anotherThree different configurations emerged from interviews and bibliometricdata on co-authorship links between group members as shown in Fig 1The two periods presented in Fig 1 (19992003 and 20042008) corre-spond to two successive formal reorganizations of the institute first in 1999with the creation of three research groups and second in 2004 with a limitedrestructuring of research groups To draw co-authorship maps the list ofresearch group members for the two periods was retrieved and then matchedwith ISI data These graphs could be compared to long-exposure photo-graphs capturing the sedimentation of co-authorship links in two differenttime periods Since networks also evolve within each period the differencesbetween periods should not be interpreted in terms of sharp change In addi-tion Fig 2 put group co-authorship networks into the larger context of EEIduring the same period

In a first configuration illustrated by the lsquoImage research grouprsquo theconstruction of a group was mainly driven by symbolic purposes and hadalmost no impact on intra-group cooperation The group consisted of themerger of two small research teams one working on remote sensing andthe other on image processing which had different orientations andresearch themes The creation of this lsquoumbrella structurersquo intended to bal-ance the size of research groups in order to present image processing as akey topic for EEI to external audiences

2004-2008

5 1015

Image Research Group

Microelectronics Research Group

Telecom Devices Research Group

1999-2003

Ellipses Former boundariesof TelecomLab groups

2004-2008

1999-2003 2004-2008

Fig 1 The Evolution of Co-Authorship Networks in Three Research Groups

Note Nodes represent researchers who authored scientific papers only permanent

faculty members are taken into account excluding PhD students post-docs and lab

technicians Links represent co-authorship relations their thickness represent the

relative intensity of links (ie number of co-authored papers) To facilitate

comparison within a group across the two periods an identification number has

been attributed to each researcher In addition two colours have been used for

nodes to distinguish between researchers who are active in both periods (black

nodes) and researchers who are active in only one period (white nodes) For

instance white nodes appearing only in the 19992003 period represent researchers

who have either left the group (eg retirement promotion) before the following

period or who did not publish any paper in the following period

People working on image processing were really under the critical size [to be a

research group] and CNRS made remarks about it Well from a certain angle it

might be said that my team works on image So we left the Systems group to create

the Image Research group Thatrsquos also because the Systems group was too big and

we needed to balance the size of research groups But we donrsquot have the same defini-

tion of image as the other part of the Image research group (hellip) Wersquore just two

teams that get along with each other We actually exchanged some algorithms but

Fig 2 The Evolution of Co-Authorship Patterns in the Electronic Engineering

Institute

that was mainly on a case by case basis We do not have any cooperation on a joint

project (Professor)

In this case the formal boundaries of the group do not correspond to aresearch collective Research is actually performed by small informal teamsof 23 group members who collaborate on a regular basis But these smal-ler teams within the group are independent since their cognitive activitiesare not based on a complex division of labour (mainly design modellingand data processing with few experimental tasks)

In a second configuration illustrated by the lsquoMicroelectronics researchgrouprsquo the creation of a group merely turned an integrated research col-lective into a formal structure No need was felt to restructure MicroLabafter its inclusion in EEI Indeed MicroLab used to function on a highdivision of labour between its members This situation resulted from aheavy emphasis on complex experimental work going hand in hand with aspecialization of members on different aspects of the research process Inaddition to a vertical division of labour where senior researchers were incharge of looking for funding for the whole group in order to allowjuniors to focus on experimental work there was a significant horizontalspecialization among junior members as illustrated by the following inter-view quote

Professors X and Y really focus on bringing the money in going to meetings they have

a managerial role hellip As junior professors we are also involved in administrative work

but we are more orientated towards experimental work hellip And we also work for other

group members so to speak Irsquom in charge of running the materials processing equip-

ment and Z is preparing thin layers for the whole group (Junior professor)

The tight interdependence among researchers in this group is furtherindicated by co-authorship patterns which suggest a strong integration atthe group level with a relatively stable configuration over time

In a third configuration illustrated by the lsquoTelecom devices researchGrouprsquo the restructuring had a stronger impact on cooperation As thisgroup included researchers coming from three different groups atTelecomLab it initially had a high potential for conflict But under the lea-dership of younger academics the restructuring was precisely seized as anopportunity to foster the sharing of resources and to limit intra-groupcompetition

We want to avoid the old rubbish situation hellip We have defined 5 main research

topics hellip People can work on one or two research themes But they canrsquot work on all 5

We try to avoid that hellip And therersquos one person in charge of each of them who has to

channel all the relevant information to the others hellip Just a week ago I had three

164 JULIEN BARRIER

contacts with people from industry I transmitted one of them to X [a colleague in

charge of another research theme] And my colleague sent me a proposal for an indus-

trial collaboration that did not fit into his research theme (Professor)

Areas of specialization still reflected divisions among former groups to acertain extent but the restructuring also allowed for the emergence ofshared themes Although change was facilitated by the marginalization ofthe former heads of research groups one of them retired and the twoothers progressively disinvested themselves from research at EEI intervie-wees explained that they still put much effort lsquoto keep the situation into con-trolrsquo The limitation of internal competition was a necessary albeit notsufficient condition for cooperation to develop between group membersThe merger and the subsequent restructuring was used as a lever to exploitpotential cognitive complementarities between researchers as suggestedboth by interview data and the visualization of co-authorship patterns

I supervise half of my PhD students with X There is also a PhD starting soon in coop-

eration with Y hellip For a long time I kept the research line Z for myself This is new

(Professor)

The evolution of co-authorship patterns confirms that researchers thatpreviously had few or no contacts begun to develop closer collaborationBut one may note that they significantly differ from those of the highlyintegrated Microlectronics group This relates to epistemic and materialfactors Knowledge production in this group does not depend as much onheavy tightly interdependent experimental task It combines diverseactivities (eg fabrication of relatively simple devices experimental charac-terization modelling and simulation) The elementary structure of thesetasks and the nature of the functional interdependencies they create was unaffected by the restructuring But it enabled the exploration of newcombinations of topics and expertise

DISCUSSION AND CONCLUSION

In analysing this case of merger and restructuring in an organized researchunit this article shows how the formal organization of research activities ischanging in relation to the increasing diffusion and significance of newinstitutional scripts in the academic field In this perspective in line withrecent efforts to revisit the concept of decoupling in institutional theory(Hallett amp Ventresca 2006) it argues against a tendency in the literature to

overemphasize the extent of symbolic compliance in organizations15

Instead our intent was to examine how institutional myths are translatedinto local arrangements and practices Three main contributions can beoutlined from our case study

First of all this article sheds further light on an important trend in theorganization of public research in Europe that is the concentrations andmergers of research units to form lsquocritical sizersquo organizations (Bonaccorsi ampDaraio 2005) Departing from quantitative studies of the relations betweenorganizational size and scientific performance (eg Horta amp Lacy 2011) itused a case study to understand how this institutional myth is edited andtranslated at the local level This case confirms the growing legitimacy ofan emerging organizational template in the French research system thatis the large regional-based research institute but it also reveals two keyfeatures of the enactment and translation of this myth First our case illus-trates the performative dimension of the lsquocritical sizersquo myth as the mergerunfolded like a self-fulfilling prophecy The merger was perceived as a solu-tion to assert the position of electronic engineering research at MidCitywhich in turn fuelled actions intending to capitalize on the merger to estab-lish EEI as a key organization at the local level Although the merger wasinitially a defensive response to changing environmental pressures it waslater instrumentally mobilized as a rhetorical vehicle to boost the renownof the research unit in order to consolidate external networks A secondfeature concerns how the issue of organizational size was framed by actorsWhile economies of scale are usually salient in discourses promotingmergers in academia in our case actors reframed the merger as an organi-zational expansion turning it into a way to extend and diversify sources ofexternal support

In addition because notions of lsquocritical sizersquo put forward by policymakers are ambiguous our data suggests that the increasing legitimacy ofEEI did not so much derive from its conformity to a static well definedorganizational template than it stemmed from the process of aggregationand expansion In other words lsquobeing bigrsquo may matter less than lsquogettingbiggerrsquo This interpretation supported by the recent collection of follow-upinformation about EEI which has continued to grow since fieldwork wascarried out by integrating research teams from other academic institutions is consistent with the idea that institutional scripts of rationalization haveincreasingly caused reform to become endemic in organizations (Bromley ampPowell 2012)

Second our empirical findings may contribute to further exploration ofthe interplay between institutional myths organizational structures and

166 JULIEN BARRIER

practices in the scientific field issues that are not frequently addressedtogether Concerning the link between institutional pressures and organiza-tional structures a first interesting aspect of our case is that change withinEEI did not take the form of a perfect compliance to external pressuresnor was it mere window dressing Environmental pressures operatedthrough indirect processes catalysing rather than triggering change bycontributing to destabilize existing local arrangements in a chain reactionInterestingly while mergers are generally understood as a way to reinforcerelations between merged entities the merger had less impact on the linksbetween MicroLab and TelecomLab than within TelecomLab The mergerproposal threatened established defensive territories (eg the fragile equili-brium in relations between research groups at TelecomLab) a situationexacerbated by divergences about the direction of the merger which in turnprovoked a conflict among seniors almost suspending the merger processThis disequilibrium enabled junior academics to foment a lsquocouprsquo againstseniors leading to their subsequent marginalization and eventuallybrought the organization into closer alignment with the institutionalscripts diffused by CNRS (ie research centres as cohesive integratedorganizations)

Regarding the impact of organizational restructuring on research activ-ities two main results can be outlined On the one hand it appears ratherlimited in terms of changes in scientific output This certainly relates to thefact that the restructuring was unconnected to major epistemic change atthe field level in contrast with cases of thorough departmental overhaulsaccompanying the emergence of a new discipline or specialty where localreorganizations are interwoven with deep multi-level changes in the struc-ture and epistemology of a field (eg Wilson amp Lancelot 2007) In ourcase this dimension was absent the restructuring being primarily driven bya combination of non-discipline-specific institutional pressures and con-cerns over the coordination of research at the local level On the otherhand however especially in comparison with the Vinck and Zarama(2007) study of a merger between two research units in the field of micro-electronics in France our empirical evidence indicates that the merger andrestructuring did have some effects on research practices albeit mainly interms of coordination and regulation of internal competition16 Then ourresults suggest that while organizational restructuring may not affect thebasic patterns of scientific work (eg nature of core tasks types of func-tional interdependencies among researchers) structures are only partiallydecoupled from core activities Their impact is indirect because they areused by academics to define lsquodefensive territoriesrsquo (Musselin 1990) they

may inhibit or enable potentially relevant cooperation They do not asmuch bring coordination by supervision as they provide spaces forexchange learning and mutual adjustment to occur This last point sug-gests that it would be worthwhile to investigate the effects of organizationalrestructuring over longer periods of time to understand to what extent(and in what conditions) mere membership to the same structure mightfavour the exploitation of epistemic complementarities

Third our results complement and extend Musselinrsquos propositions aboutthe properties of formal structures in universities (Musselin 1990 2007)Although Musselin mainly discusses the relation between structures andfunctional interdependence our case suggests that formal structures mayactually produce another form of interdependence among their constitu-ents namely lsquointerdependence of fatersquo (Lewin 1948) In contrast with taskor functional interdependence it refers to the idea that the fate of a groupmember is inseparable from the fate of the group as a whole just like thepassengers of a raft after a shipwreck In our case the creation of a newinstitute initially intended to provide researchers with a new organizationalfacade to draw further support and legitimacy But the virtual existence ofEEI as a project paradoxically began to display potentially real (and nega-tive) consequences for researchers when it was nearly aborted Because theproject had raised external expectations about the evolution of electronicengineering research at MidCity as a whole its abortion would have hadcollective reputational costs for this group thus making local academicsinterdependent in the face of adversity17

In other words in this particular context the formalization of a newstructure had created a latent community of fate While interdependence offate may be particularly salient when a group faces collective threats ornegative sanctions it is also significant when actors benefit from the collec-tive reputation of the structure they belong to This is well illustrated bythe fact that the increasingly positive external perception of EEI as a keyactor in ICT at the regional level was seen by members as a factor enhan-cing their individual reputations further consolidating their sense of mem-bership to EEI This property of formal structures is particularly visible inthe case of research centres in the French system which are a salientdeeply institutionalized level of aggregation in the evaluation of scientificperformance Indeed in addition to establishing a strong sense of member-ship one can hypothesize that a key precondition for a structure to pro-duce a community of fate is the fact that they are subjected to collectivesanctions positive or negative Then this property of formal structures

168 JULIEN BARRIER

may also be significant in other academic systems granted that the struc-ture in question be it a department an institute and so on is taken asan object of external scrutiny evaluation or steering

Coming back to the initial puzzle of this article the ambiguous roleof formal structures in academic settings this case study suggests thatthe impact of formal reorganization on research practices depends on thealignment of external and intra-organizational conditions While it is clearthat impact is likely to be all the more significant as change in structuresresonate with wider environmental pressures it ultimately depends onexternal legitimacy being tightly linked to the mobilization of materialresources In other words when the symbolic value of structures is nota key asset or at least not believed or anticipated to be a key asset byactors in the competition for material resources changes in structuresare likely to go unheeded In addition although it is also clear that thealignment between formal structures and epistemic patterns of coopera-tion is a crucial factor a result consistent with Vinck and Zarama(2007) the impact of reorganization on practices does not depend asmuch on an lsquoobjectiversquo fit between these two elements as it depends onactors using or not structures as a way to strategically pursue theirepistemic interests

1 For example the French General accounting office described the fact that80 of academic research units counted less than 25 members as a lsquohandicap forinternational competitionrsquo (Cour des Comptes 2005 p 126)

2 Since the legitimacy and ultimate survival of an organization depends on theadoption of institutional scripts they might decouple their practices from structuresin order to protect the efficiency of their core activities

3 Its full name is Comite National de la Recherche Scientifique but to avoidconfusion with CNRS it is generally simply referred to as Comite NationalMembership of the Comite National consists of both CNRS and university person-nel who are appointed every 4 years (23 of the members are elected and 13 areappointed by the CNRS) It is divided into about 40 sections this number hasvaried over time and successive reorganizations corresponding to scientificspecialties For instance the EEI is affiliated to a section covering electronic andelectrical engineering photonics as well as micro and nanotechnology

4 Our examination of evaluation reports in different research units and probesinto CNRS central archives (especially 040074 STIC boxes 16 and 17 060107 INGbox 4 and 060111 ING box 16) indicate that the application of these organizational

criteria is consistent both over time and across research units in the field of electro-nic engineering

5 Legally they are employees of the CNRS at the national level affected tolaboratories at the local level Upon acceptance by CNRS researchers can decide totransfer from one research centre to another

6 Data were compiled from different CNRS surveys and directories To iden-tify research centres operating in the field of electronic and electrical engineeringwe relied on their affiliation to the relevant section of Comite National

7 In another instance board members complained that their colleagues did notshare information about their discussions with prospective industrial partners

8 MicroLab had been seeking to merge with another unit Actually a part ofMicroLab members were not located at MidCIty but in another smaller city By1994 this part of the lab decided to leave MicroLab to join another research unitHence MicroLab was considered undersized by CNRS A former proposal tomerge with a laboratory in physics had been rejected by CNRS for being not con-vincing in terms of scientific complementarities

9 Created in 1992 this CNRS research unit was the first to result from anambitious merger involving three different laboratories in the field of electronicengineering10 Report CNRS archives 040085 STIC box 811 Because there were no other existing research units at MidCity in electronic

engineering12 With titles like lsquoA major actor in the field of telecommunications TechSchool

and UniMedium unites to create a new research institutersquo (local newspaper article2002) and lsquoElectronics seeking a new lease of life 180 scientists prepare the futureof industrial innovationrsquo (local newspaper article 2003)13 In brief competitiveness clusters are territorial-based structures gathering

firms and academic institutions in order to foster the development of specializedlocal innovation networks in targeted areas (eg aeronautics in the Toulouse area)giving access to special funding for RampD projects Their creation depended on abottom-up competitive process a national call for proposals invited local actors tolsquodefine their goals their relevant territory and the actions they wanted to taketogether in order to reach those goalsrsquo (Younes 2012 p 341)14 From 1999 to 2003 6 out of 247 publications (24) were co-authored by

members of two different research groups From 2004 to 2008 the figure was 7 outof 349 publications (18) See also Fig 215 Our argument should not be mistaken as the trivial notion that institutional

pressures can actually have an impact on practices But given the narrow under-standing of the notion of decoupling that has become prevalent in the literature itis relevant to insist on that issue to explore further the significance of the diffusionof scripts for organizational practices16 In the case they investigated Vinck and Zarama (2007) concluded that the

merger did not have any significant impact on scientific activities and barelyaffected preexisting collaborative patterns17 This holds true because there was no alternative for most academics as they

could not easily lsquoabandon the shiprsquo For instance transferring to another researchcentre was not an easy option because MicroLab and TelecomLab were the twomain centres at MidCity

170 JULIEN BARRIER

ACKNOWLEDGEMENTS

This article draws data from a research project funded by the EuropeanCommission (PRIME Network of excellence) I also acknowledge additionalsupport from ENS Lyon IFE and CNRS (Chair on Higher Education ampResearch Policy) and I thank Pierre Barrier and Barbara Bovy for their parti-cipation to data collection Participants to the lsquoOrganizing Sciencersquo Subthemeat 2012 EGOS conference provided helpful comments on a preliminary ver-sion of this article Many thanks to Jochen Glaser Grit Laudel UweSchimank Richard Whitley and an anonymous reviewer for their construc-tive comments on earlier versions

REFERENCES

Aust J amp Crespy C (2009) Les collectivites locales face a lrsquoenseignement superieur et a la

recherche Pouvoirs locaux 10 6472

Barrier J (2011) Following scientists following the money Project funding professional

autonomy and changing work patterns in academic research Sociologie du travail

53(4) 515536

Barrier J amp Mignot-Gerard S (2013) Leadership et changement dans une organisation

pluraliste In I Vandangeon-Derumez amp D Autissier (Eds) Le changement organisa-

tionnel Etudes de cas commentees (pp 121139) Paris France EyrollesEditions

drsquoOrganisation

Becker H (1960) Notes on the concept of commitment American Journal of Sociology 66(1)

Binder A (2007) For love and money Organizationsrsquo creative responses to multiple environ-

mental logics Theory and Society 36(6) 547571

Bonaccorsi A amp Daraio C (2005) Exploring size and agglomeration effects on public

research productivity Scientometrics 63(1) 87120

Bromley P amp Powell W (2012) From smoke and mirrors to walking the talk Decoupling in

the contemporary world Academy of Management Annals 6(1) 483530

Clark B R (1972) The organizational saga in higher education Administrative Science

Quarterly 17(2) 178184

Cour des Comptes (2005) La gestion de la recherche dans les universites Report to the

President of the Republic Paris France

De Nooy W Mrvar A amp Batagelj V (2005) Exploratory social network analysis with

Pajek Cambridge Cambridge University Press

Drori G Meyer J Ramirez F amp Schofer E (2006) Science in the modern world polity

Institutionalization and globalization Stanford CA Stanford University Press

Fixari D Moisdon J-C amp Pallez F (2009) Lrsquoevaluation des chercheurs en questions

(19922009) Paris France Presses des Mines

Frickel S amp Moore K (2006) Prospects and challenges for a new political sociology

of science In S Frickel amp K Moore (Eds) The new political sociology of science

(pp 333) Madison WI The University of Wisconsin Press

Hallett T (2010) The myth incarnate Recoupling processes turmoil and inhabited institu-

tions in an urban elementary school American Sociological Review 75(1) 5274

35(2) 213236

Horta H amp Lacy A (2011) How does size matter for science Science and Public Policy

38(6) 449460

Kitchener M (2002) Mobilizing the logic of managerialism in professional fields The case of

academic health centre mergers Organization Studies 23(3) 391420

Krucken G (2003) Learning the lsquonew new thingrsquo On the role of path dependency in univer-

sity structures Higher Education 46(3) 315339

G Drori J Meyer amp H Hwang (Eds) Globalization and organization World society

and organizational change (pp 241257) Oxford Oxford University Press

Laredo P amp Mustar P (2002) Innovation and research policy in France (19802000) or the

disappearance of the Colbertist State Research Policy 31(1) 5572

Latour B amp Woolgar S (1979) Laboratory life The social construction of scientific facts

Beverly Hills CA Sage Publications

Laudel G (2002) What do we measure by co-authorships Research Evaluation 11(1) 315

Lewin K (1948) Resolving social conflicts Selected papers on group dynamics New York

NY Harper

Louvel S (2010) Changing authority relations within French academic research units since

the 1960s From patronage to partnership In R Whitley J Glaser amp L Engwall

(Eds) Reconfiguring knowledge production Changing authority relationships in the

sciences and their consequences for intellectual innovation (pp 184210) Oxford

ceremony American Journal of Sociology 83(2) 340363

Musselin C (1990) Structures formelles et capacites drsquointegration des universites francaises

et allemandes Revue Francaise de Sociologie 31(3) 439461

Sauder M amp Espeland W N (2009) The discipline of rankings Tight coupling and organi-

zational change American Sociological Review 74(1) 6382

Shinn T (1988) Hierarchies des chercheurs et formes des recherches Actes de la Recherche en

Sciences Sociales 74 222

Theves J Lepori B amp Laredo P (2007) Changing patterns of public research funding in

France Science and Public Policy 34(6) 389399

Vinck D amp Zarama G (2007) La fusion de laboratoires Processus de gestion et constitu-

tion drsquoune entite pertinente de lrsquoactivite scientifique Revue drsquoanthropologie des connais-

sances 1(2) 276296

Weick K E (1976) Educational organizations as loosely coupled systems Administrative

Science Quarterly 21(1) 119

Wilson D amp Lancelot G (2007) Making way for molecular biology Studies in History and

Philosophy of Science Part C 39 93108

Younes D (2012) Choosing the industry of an industrial cluster in a globalizing city Journal

of Change Management 12(3) 339353

172 JULIEN BARRIER

CONTRADICTORY

CONSEQUENCES OF

INSTITUTIONAL CHANGES ON

INTELLECTUAL INNOVATION IN

THE PUBLIC SCIENCES

INSTITUTIONAL CONDITIONS

AND CHANGING RESEARCH

PRACTICES IN SWITZERLAND

Martin Benninghoff Raphael Ramuz

Adriana Gorga and Dietmar Braun

ABSTRACT

This article analyses in what way Swiss academic institutions have hada favourable or unfavourable influence on changing research practicesby following developments in four scientific areas Bose-EinsteinCondensates Evolutionary Developmental Biology Large-Scale Assess-ments in education research and Computerised Corpus Linguistics Basedon empirical evidence we argue that overall a number of institutionalconditions have had a positive influence on the decisions of scientists todare a switch to a new scientific field One finds however also differ-ences in the working of these institutional conditions leading to quickeror slower developments of the four selected scientific areas

Keywords Public science systems Switzerland research practicesscientific innovations universities research councils

ISSN 0733-558Xdoi101108S0733-558X20140000042006

INTRODUCTION

This article contributes to the continuing discussion about the institutionalconditions of knowledge development (Bonaccorsi 2007 BonaccorsiDaraio amp Geuna 2010 Braun 2014 Whitley 2000 2003 WhitleyGlaser amp Engwall 2010) by analysing one particular research systemSwitzerland and assessing how Swiss academic institutions have had afavourable or unfavourable influence on intellectual innovations byfollowing the growth of research in four scientific areas Bose-EinsteinCondensates (BEC) Evolutionary Developmental Biology (ED) Large-Scale Assessments (LSA) in education research and Computerised CorpusLinguistics (CCL) In comparison to other contributions in this editedvolume where the institutional context varies while the dependent variablethe change in research practices is constant that is only one scientific fieldis analysed we look into the relationship of only one institutional contextwith a variety of scientific fields in which a change in research practices hastaken place

While the former research design enables the identification of the institu-tional conditions that are favourable or unfavourable for one scientificfield our approach is aimed at clarifying whether one and the same institu-tional context matters in the same way for changes in research practices indifferent scientific fields A strong variation in outcomes that is in thecapacity for new scientific fields to grow would suggest that the influenceof the institutional context differs between kinds of scientific field It mightbe positive for some scientific fields but not for others which helps us tounderstand the particular needs of institutional support for different scien-tific fields An overall strong capacity to spread in contrast would meanthat the institutional context has a similar positive influence throughoutscientific fields An overall weak capacity to change research practices indi-cates an institutional context with deficient support structures

Although the capacity to generalise on the base of this design remainslimited because only four case studies are considered they do represent themajor branches of the sciences physical sciences (BEC) life sciences(ED) social sciences (LSA) and humanities (CCL) Such contrasted casescan give at least an insight into the differences between change in researchpractices across the sciences and the impact of favourable or unfavourableinstitutional conditions Further research is needed to corroborate ourfindings

We proceed in the following way The next section presents conceptualconsiderations that have guided our empirical research The nature of

176 MARTIN BENNINGHOFF ET AL

scientific growth and diffusion will be defined and operationalised andwhat is needed to change lsquoresearch practicesrsquo discussed The different needsof scientific fields and the institutional conditions that can influence theirgrowth will be presented We then provide a short overview of the generalinstitutional conditions in Switzerland before we discuss the four case stu-dies The discussion describes the growth of the four cases in Switzerlandand explores the extent to which Swiss institutions have had a generallysupportive or constraining influence on all four of the developing scientificfields or rather had different effects across them In the conclusions wecome back to our research question To what extent do different kinds ofscientific institutions affect the expansion of various scientific fields

CONCEPTUAL CONSIDERATIONS

Scientific growth and diffusion are for us part of the process of lsquoscientificinnovationrsquo that is the production and diffusion of new knowledgecomponents This signifies the decision of scientists or researchers tochange their research practice1 or in other words to lsquoswitchrsquo and start towork in a yet relatively unexplored cognitive domain Original findingsare usually the starting point of the process of scientific innovation Suchfindings can only become scientific innovation if they are acknowledgedand become accepted by other scientists This happens if these otherscientists start to explore further the implications of these findings A newscientific field is created if a significant number of scientists are drawninto this exploration process and change their research practices in orderto be able to do so Institutionalisation processes in form of the creationof regular research networks associations journals and chairs in universi-ties follow

We are interested in the decision of scientists to change their researchpractice as the decisive building block for the establishment of a new scien-tific field If lsquopioneersrsquo the scientists that have made the discovery do notfind lsquofollowersrsquo the original findings will become forgotten and part of thehistory of science When therefore do scientists decide to change researchpractices In order to answer this question we use an economic interpreta-tion of scientific behaviour (Braun 2012a Kitcher 1995 Mirowski amp Sent2002 Shi 2001) Such a view presupposes that scientists take decisions in this case to enter or not to enter a new scientific field by using infor-mation about the costs and benefits involved A switch to a new field must

177Institutional Conditions and Changing Research

be lsquoattractiversquo meaning it should clearly result in more benefits than thescientist has at the moment and costs to switch should be low

Decision-Making Variables of Scientists

There are a number of variables that enter into the decision processes ofscientists considering entering new areas Only some of them are subjectto direct manipulation by institutional means in universities and fundingagencies Both the search for scientific reputation and for more lsquoeffective-nessrsquo that is better results in terms of knowledge advancement are rela-tively independent of external factors They are either decided by processestaking place in the scientific community or are a result of the cognitiveadvances that can be made once a switch has taken place Scientists canexpect some degree of better effectiveness but this is based on cognitiveevaluations of the possibilities of the new scientific field

Other variables can be manipulated or influenced

(1) Material resources have a high status in decisions of scientists Theresearcher will for example assess whether she has reasonable chancesof getting or maintaining an attractive employment when switching toa new area and she will ask whether there are sufficient chances to findresearch money research personnel and research infrastructure in thenew area If for example funding agencies do not fund the new areathe costs to find other research money will rise and the attractiveness ofthe new field will decline

(2) The time at disposition for research is another decision variableExploring new areas often need more time than continuing to work inestablished areas If research organisations or funding agencies do notprovide lsquoprotected timersquo for such longer-term investments a switchmight become unattractive Another time aspect is the amount of timeavailable for doing research compared to other activities of scientistslike teaching and administration The change of research practicesneeds a considerable amount of time that must be invested in researchInstitutional conditions define to some extent the distribution of timeof research available for scientists They can grant more or less time forresearch in different areas

(3) Academic support refers to the possibility of institutionalising thenew field within the scientific community New fields can engenderunfriendly behaviour of fellow scientists working in established fields

either by subverting publications by not granting access to scientificassociations or by resisting to the institutionalisation of the new fieldwithin the university Switchers who meet such resistance have highercosts to carry in order to overcome such resistance and many may bedeterred to switch at all when they realise that the new field may raisesuch academic resistance New interdisciplinary fields are particularlysubject to such resistance and will in general raise higher switchingcosts Academic support can however also be benevolent and favour-able for scientist The new area may be seen as a fruitful complementto existing research topics in an institute or department The lsquoswitcherrsquomay then find backing and support to build up the new research areawhich lowers switching costs

Material resources time and academic support are therefore we sug-gest the major conditions that enter into the decision-making of scientistswhen thinking about switching research practices

Types of Scientists

Switching costs are not always the same for all scientists Apart from thefact that scientific fields may raise different cost levels for examplebecause of high infrastructural needs for research equipment as in the caseof ED and BEC one can distinguish between two types of scientists inthis respect

One is the distinction between those scientists who are heavily involvedin the lsquobuildingrsquo of the new research area and those who are prepared towork in the new area and simply use what the lsquobuildersrsquo have developedThis distinction becomes most appropriate in the case of CCL and LSAwhere the creation of the lsquoresearch toolrsquo that is databanks is the maininitial contribution to be made Obviously this needs considerable timeinvestments and material resources The same is true for BEC in the case ofwhich scientists search for methods to lsquofreezersquo the movement of atomsUsers in all three cases profit from the work of builders and develop scienti-fic knowledge by using the research tools prepared Investments to be madeare accordingly much lower It needs of course both types of scientists tospread the field throughout the scientific community

The second distinction of types of scientists with regard to switchingconcerns the strategies scientists can employ in order to change researchpractices Time investment is the important variable here Scientists may

either abandon completely their former research practices (and invest alltheir energy and time in the new field) or they may use only part of theirtime while maintaining also part of their research activities in existingresearch areas One could call the first type of scientists lsquofull switchersrsquo andthe second type lsquopartial switchersrsquo Full switchers take considerably morerisks when switching because if they fail to succeed in the new area theywill suffer from loss of reputation while a partial switcher can to someextent fall back on the reputation she has built up in existing research areasand which has been further developed even when exploring the new fieldFull switchers need more lsquoprotected spacersquo that is institutional support interms of money and time resources than partial switchers Institutionalconditions support of public research organisations like universities andfunding agencies can influence the decision of scientists in this respectGenerous and unconditional support by these institutions makes it easierfor scientists to take the risk of full switching

Institutional Conditions

Institutional conditions vary in time and space and lead to different switch-ing costs and opportunities for scientists to change research practicesWhat kind of institutional conditions present attractive conditions forswitching Which ones are detrimental to changing research practices Weendeavour below to formulate some expectations about the relationship ofthe three decision-making variables we have formulated above materialconditions time resources and academic support and institutionalconditions

Institutional Conditions and Material ResourceslsquoRationalrsquo scientists planning to switch prefer one can assume havingabundant resources easy access to financial resources and a flexible use ofthese resources The most fundamental institutional condition in thisrespect is to what extent funding money is available in the national researchsystem if there are conditions of lsquoaffluencersquo or of lsquoscarcityrsquo (see Heinze ampMunch 2010 Ziman 2002) In periods of affluence funding money cannourish a large number of different research areas which raises the chancesof scientists to find financial support even in those areas of research thatare generally considered as being risky Having abundant funding allowsscientists to distribute risks among a large number of projects Chances to

become funded in this case is high for scientists going into new and unex-plored knowledge fields lsquoScarcityrsquo by contrast reduces the willingness offunders to take such risks Because fewer projects can be funded funderstend to support areas where results are certain which means most of thetime mainstream research with incremental knowledge progress There ismoreover a tendency to become more lsquoselectiversquo and to cut down averageresources per project Scientists have therefore higher lsquosearch costsrsquo in orderto find funding they must do with less resources and have fewer chancesto launch high-risk research projects

Research systems can however also be lsquoselectiversquo even in periods ofaffluence In this case it is a matter of belief about the best way to supportdiscoveries and innovations One narrative is that choosing the best scien-tists or projects and to supply them with abundant resources will result inhigher productivity than a distribution of funding money over a large num-ber of scientists and projects This narrative can lead to investments into aselected number of high risk and new research fields Switchers who suc-ceed in obtaining such funds can probably count on excellent material con-ditions but the chances to obtain these funds diminish The knowledge areathe scientist wants to work in must fit with those areas selected by fundersIf funders accept that the selected areas are chosen by scientists themselvesthe chance to obtain such funds rise but need considerable investments inorder to compete with other projects

A funding system with a plurality of funding sources is another institu-tional condition that might have a positive effect on the change of researchpractices It increases the chance to find the funding for research projectsbecause such a system allows for more variety in research Shrinkingfunding resources in one funding agency may be compensated by risingresources in another

Getting funds with abundant research money is one aspect To disposeof these funds in a flexible way is another The flexibility of fundingdepends on the policies of public research organisations in particular uni-versities to grant their scientists long-term and unconditional researchmoney Funding agencies usually give money for a short period of timeand demand a relatively precise outline and budget Recently there havebeen changes however with the launching of longer-term grants for exam-ple for centres and also occasionally to individual scientists (the advancedgrants of the ERC are a good example here) Such policies would contri-bute to a higher flexibility in the use of funding money which is importantfor switchers who work in areas that are new and must still be explored

Institutional Conditions and Time ResourcesThe fundamental decision to be made in research systems is to what extentfunding resources should be available only for shorter periods of time which reduces risks for funders as the invested amounts of money for oneproject remain low or for longer time periods which raises investmentcosts and needs substantial trust in the capabilities of scientists to beproductive Often it is institutional funding which is needed to supportlong-term areas of research or it may be project funding with a long-termperspective Another important condition is the availability of stableemployment positions that can be granted to scientists working in newareas One can think here of tenure-track or tenured positions The capabil-ity of funders to grant a long-term research perspective depends not onlyon large amounts of own resources they must also have the competencesto decide about such investments The change in some continentalEuropean universities to a new public management governance regime hasfor example given them greater operational autonomy which also includesthe use of financial resources This gives them the power to grant suchlong-term support Funding agencies have also recently started to changetheir funding toolbox by developing long-term grants In research systemsthat are built on scarce resources and a lsquoselective modersquo of choosingresearch projects long-term support will have difficulties to be financedSwitchers must in this case live with more uncertainty about the continuingfinancing of their research work and accept higher transaction costs insearching new funding

Another aspect of time resources is the possible substitution of researchtime by other tasks in universities like teaching or administration To havesufficient time for doing research in new areas is therefore also a matter ofhow public research organisations and funding agencies are liberatingresearchers from too many other tasks Funding agencies can develop fund-ing instruments that take this aspect into account and offer a lsquoresearchleaversquo for example Universities must have the flexibility and this isincreasingly the case to redistribute research time among its academicpopulation The stronger role of management and the strategic orientationof universities allow today to give some academic scientists more researchtime while burdening others with more teaching Such flexibility can allowlowering switching costs for scientists

Institutional Conditions and Academic IntegrationThe integration of new areas into established sciences depends to a largedegree on the scientific community When it comes however to the

institutionalisation of new areas in the form of recognised positions in uni-versities (eg professorships) institutional conditions in universities domatter The creation of new professorships research institutes or researchcentres in new cognitive domains can be influenced by the openness of theuniversity towards the introduction of new areas by the influence of stake-holder interests in universities by the existence of a policy of priority set-ting or by the authority of university management to impose solutions inuniversities These conditions can contribute to more flexibility in institu-tionalising new cognitive domains within universities

THE SWISS RESEARCH SYSTEM

We now outline the major features of the organisation of research andresearch funding in Switzerland that help to understand the general institu-tional environment switchers have to work in

According to the OECD (2006 p 8) Switzerland performs very welllsquoin terms of nearly all available indicators of science technology andinnovation often holding with a leading international positionrsquo It is nowonder that the country is generally regarded as a highly attractive systemfor doing research This attractiveness is demonstrated by the fact thatSwitzerland has a high influx of foreign researchers About 50 of profes-sors in 2012 came from a foreign country2 This is not only is a good indi-cator of attractiveness but let us also assume that Switzerland should bevery capable of lsquoimportingrsquo new and flourishing scientific areas relativelyrapidly into the country It then depends on institutional conditionsfor example academic integration or financial support whether there is aspread of the new knowledge

Research overwhelmingly takes place in ten cantonal universities and intwo federal institutes of technology all relatively small in size (Braun ampLeresche 2007)3 Increasingly universities of applied sciences set up in1996 and charged with more applied oriented research are also becomingimportant research places in the system Extra-university research institutesare rare in Switzerland which means that academic integration usuallytakes place in the institutional context of higher education institutions

Switzerland is a federal country This means that both the federalgovernment and a large number of member states the cantons financecantonal universities while it is the federal government alone that pays forthe two federal institutes of technology (Ecole Polytechnique Federale de

Lausanne (EPFL) Eidgenossische Technische Hochschule Zurich(ETHZ)) During a long process that started in 1988 and lasted until 2009all Swiss universities have acquired operational autonomy and negotiatetheir portfolios of activities in the form of contracts with the politicalauthorities The precise conditions and the degree of operational autonomycan vary substantially between universities (Braun 2012b)

The last decade has been characterised by a period of lsquoaffluencersquoResearch funds have been continually expanded by the federal parliamentand cantonal parliaments These increased funds resulted in considerablegrowth rates for university budgets (Baschung Benninghoff Goastellec ampPerellon 2009) though most notably it was the very well-funded federalinstitutes of technology and the five larger cantonal universities (GenevaZurich Basel Lausanne Bern) that profited most among the 12 highereducation institutions The federal institutes and the larger universitiesare also the main place for scientific discoveries The relative generousfunding has made it possible that even today Swiss researchers can counton substantial institutional and therefore long-term support for doingresearch Institutional funding has not been decreased though there hasbeen a clear tendency and political will to strengthen competition forfunds and financing by research grants from funding agencies (Leporiet al 2007) The binding of about 30 of federal funding to cantonal uni-versities to a number of performance criteria4 has together with a growingcompetitiveness of universities for funding and reputation resulted in pres-sure on scientists to submit projects for research funding Research grantshave in consequence risen considerably during the last 10 years (Braun2012b)

The Swiss National Science Foundation (SNSF) is the only fundingagency for basic research in the country and has the primary role in sup-porting the development of knowledge in universities Relatively generousfunding has enabled the continuation of high success rates in Switzerlandand throughout most disciplines At the end of the 1990s the SNSF devel-oped a major centre programme the so-called lsquoNational CompetenceCentres of Researchrsquo (NCCR) which allows for 12 years of financing incertain areas of research (Braun amp Benninghoff 2003) Because of the largesums of money that were linked to this funding instrument it not onlygave researchers a longer-term perspective for doing research but func-tioned also as a selective device for research within universities Universityleaders had to decide which research grants (and knowledge domains) theywanted to support

CHANGING RESEARCH PRACTICES IN FOUR

SCIENTIFIC DOMAINS IN SWITZERLAND

Research Methods

In order to follow the process of changing research practices in the fourscientific fields we conducted interviews with different type of actors in uni-versities (switchers and middle management) These interviews allowed hav-ing information related to the local context where researchers switch Theinterviews were based on the lsquolife-historyrsquo approach (Glaser amp Laudel2009) One difficulty was to identify individual researchers in the four fieldsThis was done by looking at institutional web sites and by referring to theISI Web of Science In total we conducted 70 interviews 22 for BEC 16for ED 18 for CCL 14 for LSA These interviews were recorded and fullytranscribed In order to preserve the anonymity of interviewees we do notmention any names of researchers

The Diffusion of Four Knowledge Domains in Switzerland

In this section we are interested whether Swiss researchers belonged to thelsquopioneersrsquo in the four cases or whether the knowledge area has been discov-ered elsewhere and Swiss researchers were among the followers to integratethe new field within the academic landscape of Switzerland If the latter isthe case how quick have Swiss scientists been to take up the new areas incomparison to other countries How rapid has then been the spread withinthe country that is the diffusion of the new area in different research insti-tutions and how many scientists have changed research practices to thenew field or in other words what has been the scope of the spread Havethere been differences between the four fields in this respect

Based on our empirical work it becomes clear that all four knowledgefields were diffused in Switzerland though there are differences in pacescope and paths of institutionalisation

ED In Line with International DevelopmentSwitzerland followed the pace of institutionalisation of ED elsewhereDuring the 1990s Swiss researchers participated in the discovery of thegene hox which led to the institutionalisation of experimental ED As at

the international level work on ED really started during the 2000s whenthe sequencing of different genomes had taken place The ED Swiss com-munity contributed to the further development by own original work (egthe identification of ED model animals) Therefore we can consider Swissscientists as pioneers in ED

An important factor for this role in ED was academic supportSwitzerland has a strong research tradition in molecular biology inSwitzerland since the 1930s and later also in genetics The fact thatthe different research domains that exist today in ED are implantedexactly in those universities which had developed this research tradition inmolecular biology and genetics is a strong indication of the relevance of aresearch tradition for the institutional integration of proximate fields ofknowledge

Looking at the ED switchers in Switzerland one finds three lsquofullrsquo switch-ers and eight lsquopartialrsquo switchers who develop new theories and methodsespecially by using new animal models and techniques coming from genetics(which is unusual in evolutionary biology and extremely costly if animalmodels are used) In comparison to other domains (ecology genetics popu-lation development genetics etc) ED until today represents a rather smallcommunity which is moreover not well connected internally due to theanchoring in different lsquomother disciplinesrsquo

Bose-Einstein Condensate Fast-Tracking the International Development byAttracting Foreign ResearchersSwitzerland has also an excellent tradition in physics which is above alllinked to nuclear physics and solid-state physics while atomic physics andtherefore cold atoms research has traditionally found little attention Thisexplains why BEC has been developed around 2000 through two scientistscoming from abroad (see Laudel et al this volume) While BEC of atomswas experimentally realised in Germany and Netherland and above all inthe USA during the 1990s Switzerland in this case was only a followercountry However a Swiss researcher has been a pioneer in realising BECwith polaritons which he discovered in 2006 A third method of condensa-tion BEC by magnetic systems has also started to be used in the 2000s inSwitzerland

Today researchers in three universities are working with the last twomethods They can be considered to be lsquopartial switchersrsquo that is they arenot investing all their research time to this issue while the two researchersworking on cold atoms in two different universities can be considered to be

lsquofull switchersrsquo who were able to overcome the high costs involved inGermany They were appointed in Switzerland once they had switched

One finds a further research strand in BEC treating the Bose-EinsteinCondensate from the point of view of theoretical physics (which can con-cern both methods presented above) It requires less investment mainlythe cost of computers with considerable calculation powers (and this canbe costly) It also requires good interactions with experimentalists andtime to master the specific physics of this field It is developed in fouruniversities with four scientists participating They can also be consideredas partial switchers with relative low switching costs linked only to thelearning of the methods and the use of highly performance computers forsimulation

CCL On the Fringe of the International DevelopmentAt the international level CCL is not a recent knowledge area (see Engwallet al this volume) It has been built up in different languages over manyyears with a particularly important progress in English language studiesScientists working in the Swiss science system have not been among thecorpus lsquobuildersrsquo and may therefore be considered as being lsquofollowersrsquoRecently however three projects emerged in Switzerland among them alsquodigital dictionaryrsquo for the German languages which has already beenfinished

There are only two Swiss CCL-builders who work almost exclusively inthis domain and have an interest in its theoretical and methodologicaldevelopment Both researchers did however not switch in Switzerland butin Germany Most researchers in Switzerland are lsquousersrsquo who work withexisting corpora in order to answer their various research questions Threeresearchers in three universities were identified doing this In Switzerlandone cannot speak of a Corpus Linguistics community both because of thesevery different ways to deal with CCL but also because of the linguistic frag-mentation of the country which gives room to different epistemologicaltraditions and implies more native language studies (four) than in othercountries

LSA in Educational Research Working through the BacklogFirst steps in LSA were the organisation of national surveys linked to theinternational programmes that existed Switzerland was not among thepioneers in this respect but started to participate driven by scientific andpolitical interest on the national level since the 1990s first in the so-called

TIMMS (lsquoTrends in International Mathematics and Science Studyrsquo)5 andthen later in the PISA (lsquoProgramme for International Student Assessmentrsquo)studies organised by the OECD TREE (lsquoTransitions from Educationto Employmentrsquo) has been a nationally developed programme in themid-2000s Those who wanted to participate had to coordinate the build-up of own databanks with international databanks (one full switcher) Theother group of scientists wanted to use these databanks Among thesescientists two had to switch which meant largely to learn the different epis-temological tradition of LSA in comparison to epistemological traditionsused in disciplines researchers were anchored in Two others were familiarwith this tradition and therefore only applied the data

ComparisonOur first question focuses on the main dynamics at work in the Swiss scien-tific landscape that is whether scientists working in the Swiss science sys-tems have participated in the set-up of the four areas as lsquopioneersrsquo orwhether they have been lsquofollowersrsquo who adopted changing research prac-tices developed elsewhere Comparing the four fields we have identifiedfour different paths of changing research practices In ED Swiss scientistshave been among pioneers backed up by a long research tradition in therelevant disciplines nurturing ED Switzerland keeps pace with other coun-tries here in the development of the research area even if the scope is notso important compared to other life sciences communities The spread ofBEC became possible because Switzerland was able to attract pioneersfrom another country Pioneering work has however also been done bySwiss researchers in one of the methods applied in BEC Despite someresearch projects CCL remains at the fringe of international developmentin the area Finally LSA researchers were able to keep in line with interna-tional developments in their subject area ED and LSA are the twodomains in Switzerland which were introduced early seen from an interna-tional perspective But if ED has taken a gradual diffusion path in linewith international developments LSA has experienced ups and downs inits development due to scientific and political interests By contrast CCLand BEC were introduced in Switzerland much later than in the pioneeringcountries The role played by foreign researchers has been important forboth innovations However the diffusion paths are different CCL remainson the fringes of international developments while Swiss BEC researchersare becoming internationally competitive and contribute in innovative waysto its further cognitive development

THE INFLUENCE OF INSTITUTIONAL CONDITIONS

ON CHANGING RESEARCH PRACTICES IN THE

SWISS RESEARCH SYSTEM

In the following we endeavour to take stock of the importance of the insti-tutional conditions in Switzerland for the decision of scientists to changeresearch practices

Institutional Conditions and Material and Time Resources

The scientists we interviewed have in general confirmed the high attractive-ness of the Swiss research system in terms of excellent research conditionsabove all in terms of funding grants high salaries and often generous start-up funds These conditions of lsquoaffluencersquo have led many switchers abroadto choose Switzerland as a host country These opinions are in line withexpectations we had about the Swiss research system The influx of foreignresearchers helps Switzerland to excel in areas that were initially not estab-lished BEC is an example though foreign researchers have also played acrucial role in ED and CCL

The attractiveness is linked to the stable and well-financed position ofprofessors in Switzerland both in the two federal and in the five cantonaluniversities examined Such positions allow building up new researchdomains especially if the necessary infrastructure has been included in thestart-up financing Switzerland has also not followed the internationaltrend to submit academics to constant evaluation processes and thereforeto a continuing pressure to publish This gives also more time to invest innew knowledge fields that still need further discovery The obvious attrac-tiveness of tenured positions in Switzerland is given the statements of thesubstantial number of foreign researchers we interviewed the majoradvantage of Switzerland as a research system Both BEC and CCL butalso ED have profited from these attractive conditions by drawing in for-eign researchers Professors profit from excellent working conditions andmoreover from continuing substantial lsquofreedomrsquo in the choice of theirresearch topics which makes lsquovarietyrsquo in the research system possible

Certainly one finds in Switzerland a stronger willingness of the univer-sity management compared to the past to think strategically both in termsof national and international competition and to orient the researchprogrammes within universities partly to pre-defined areas Despite such

tendencies professors are often hired without insisting on their fit to one ofthese priority areas but purely on the base of lsquoexcellencersquo A university policyopen to lsquovarietyrsquo and this is the result of appointments on the base ofexcellence leads to a patchwork of research fields and disciplines withinuniversities This does not mean that nominations of professors on the baseof strategic considerations are impossible but they seem still to be in theminority compared to nominations on the basis of excellence In the case ofBEC this principle of excellence has allowed the new research field tobecome established within the institutions in question CCL on the otherhand failed to find sufficient support in universities because humanities arerarely among the priority areas in Swiss universities The institutionalisationof ED was favoured by disciplinary traditions within universities that werecognitively near to the ED and which were not greatly altered by changes inthe programmatic orientation of the universities Like CCL LSA did notmeet favourable circumstances in terms of attention by the university man-agement but it was able to find external support that helped to installresearch units within universities

lsquoSeed moneyrsquo to set up new research projects can be important forswitchers to start up exploring new areas In this case we found examplesboth of university leaders demanding that seed money should be distribu-ted on the base of pre-selected research topics in the university and thatonly the proven excellence of the researcher should count

All this confirms that Swiss universities certainly have integrated a morestrategic-oriented policy in employment matters and in the support ofresearch but that a policy of lsquoexcellencersquo which gives scientists a strongfreedom in their choice of research topics still prevails thereby creatingopportunities for new fields to become integrated into universities

A policy to support a lsquovarietyrsquo of research topics instead of limiting vari-ety by selecting research topics on the base of pre-defined research pro-grammes is also supported by the research grant funding system Almostall interviewed scientists mentioned the attractiveness of the SNSF fundingwith its declining but still relatively high success rates that would increasethe chances to find support for research projects in many areas Grants addoften the necessary resources for professors to hire research personnel theycannot finance out of their institutional budgets These positive effects weretestified to above all by researchers in ED and BEC it was possible toobtain SNSF basic grants in new areas even if this has not been the pre-vious research field of the researcher Such generous funding has supporteda strategy of lsquopartial switchingrsquo of researchers who kept on working in theirexisting knowledge fields while being able to acquire supplementary

resources for new research projects that could be invested to explore newareas

Institutional funding and research grants seem to work in a complemen-tary way in Switzerland Institutional funds of universities have played arole as seed money for the exploration of the new domains have helped tobuild research centres or have simply created an institutional space forhosting the research tools (as in the case of LSA and ED) Research grantshave helped to find necessary resources in terms of research personnel andrunning costs of research projects They were also at the base of a strongerconcentration of resources in universities by way of the NCCR grants Thiscomplementarity opened up considerable possibilities for researchers Morecompetitiveness by increasing pressure on researchers to find externalresearch money did not mean moreover that fewer projects were fundedResearchers must undergo the selection process but the chances of beingfunded are higher than in many funding agencies in other countries This isaccording to us very important for explaining the ability to change researchpractices in Switzerland More competitiveness under these conditions doesnot work as a disincentive in this respect

The funding structure in Switzerland has influenced switching strategiesof researchers lsquoPartial switchingrsquo we found has been the dominant switch-ing strategy in all our four cases of changing research practices (Table 1)

Many researchers were able to work in the new fields because they didnot become lsquofull switchersrsquo Staying anchored within the existing researchareas turned often out to be important for getting resources in risky newareas There were positive spill-overs when researchers applied for researchgrants as has been described and often potential switchers got their

Table 1 Numbers of Switchers in Four Cases

Scientific Fields Full

Switchers

Partial

Switchers

Switchersa

Evolutionary developmental biology 3 8 11

Bose-Einstein condensate 2 18 20

Computerised corpus linguistics 0 16 16

Large-scale assessments in education

research

aTotal switchers means the total of researchers that we have interviewed and the not the total

of switchers in Switzerland Concerning the later we do not have data

employment positions in universities only in the traditional disciplinaryareas Partial switching allows moreover reducing the risks for the scien-tistrsquos reputation as it allows continuing profiting from the established repu-tation as long as the new area does not yet allow similar reputational gainsThe only disadvantage of partial switching though interviewed researchersdid not mention this is probably that time resources that can be invested inthe new areas are more limited in the case of partial switching compared tofull switching Such more limited time resources of switchers can howeverat least partially be compensated by the employment of researchpersonnel

There were only very few lsquofull switchersrsquo (5) in the four knowledge areaswe explored and among them some (2) had already switched abroadPartial switching (which can be either full professors assistant professorsor postdocs) seems therefore to be the main preferred switching strategyunder Swiss conditions of research reducing above all switching costsrelated to reputation resources and time

Especially with reference to BEC and ED one can say that the chancesto find adequate resources in the natural and life sciences both materialand in time were very high in Switzerland According to our investigationsthe situation for CCL and LSA was however somewhat different Weneed however to distinguish here between those who lsquousersquo the databasesthat are the main objects in these two fields and those who are activelybuilding these databases

Being a user in CCL or in LSA does not need full switching and impliesfew resources in both cases Users are contributing to the development ofthe field by working with the research tool and thereby expanding its recog-nition and significance But usually this does not need a particular positionat the university above all it does not need long-term positionsCognitive investments can be kept relatively low partly even very low assome users testified who were well acquainted with the methods that areused in LSA Funding can be acquired by the usual submission for fundingprojects There are therefore no major obstacles in this respect

Investing in building a corpus linguistic or a large assessment databasegenerates in general much higher switching costs in terms of material andtime resources

Large-scale corpus building for example is a longer-term project andneeds secured funding In Switzerland universities usually do not supportsuch an enterprise The SNSF has done it occasionally (a lsquodigital diction-aryrsquo for the German language was funded) but is increasingly reluctant tofund lsquoinfrastructurersquo or research tool projects which are the centre of both

CCL and LSA The only other agency which in this case could compensatethe lack of funding is the Swiss Academy of Humanities and SocialSciences It has however limited resources and must distribute this moneyamong several competing projects As interviewed switchers statedthe money distributed by the Academy does not cover all expenses linkedto such a long-term project In addition to other problems like the var-ious language regions in Switzerland finding research grants is thereforea major problem for corpus-builders in Switzerland One solution forthis would be the financing by the NCCR programme of the SNSF whichis however extremely competitive and has been designed for large networksof researchers and institutions This puts the natural and life sciences inan advantage compared to the social sciences and humanities which aremore based on small group and individual research In the beginning inthe early 2000s one funding round of NCCR projects was reserved to thelatter disciplines but since the mid-2000s all disciplines compete witheach other for NCCR-funds Social science and humanities project are itturned out clearly in the minority In addition it turned out that also uni-versities have been reluctant to support this field Corpus building remainstherefore a precarious undertaking in Switzerland in terms of materialresources

Builders in LSA have on the contrary found more SNSF support Inthis case the SNSF helped to support LSA as programme funding of theSNSF financed the first LSA projects the TIMMS data But it neededalso the support of policy-makers if LSA wanted to be successful above allbecause the LSA data in particular linked to the OECD PISA-programmeneeded data on pupils in schools on a national scale To achieve suchnation-wide coverage was more complicated in Switzerland than in manyother countries because of its decentralised federal structure It took severalyears to coordinate efforts in order to obtain such a nation-wide coveragePolitical support was also important for the financing of the surveys It wasthe Federal Statistical Office which for a couple of years took charge ofthe management of the data before an independent expert committee wasentrusted with this task

Swiss research networks in the field of the LSA are very heterogeneousand include academics and researchers who are still working in or comeoriginally from cantonal or federal offices In addition the integration ofthis research field into the universities was quite different one universityopened an academic position in this field within the Faculty of Educationwhile another one has established a separate organisational unit (evaluationinstitute)

Funding has been relatively piecemeal and fragmented The variousdatabanks are under different authorities The whole research area has devel-oped in an incremental way and needed a combination of academic entrepre-neurship support of faculties research funding by SNSF programmes andpolitical support by cantonal governments Finally Switzerland has beenable though with some time lag to international developments in this areato set up the research tools in LSA and even to develop a new research toolwith the TREE programme Two universities have been helpful with thecreation of one specialised professorship and one competence centre Thenomination of a professor designed for developing LSA has been possiblebecause the university in question had re-designed its faculty structure sincesome time and had created a number of interdisciplinary faculties Thisallowed the creation of an interdisciplinary education department whichdeveloped an interest in such a professorship

In summary to build the databanks as a research tool in these two caseswas certainly a demanding task for potential switchers with clearly higherswitching costs with regard to material resources than the ones switchers inthe case of BEC and ED had to carry

Another point of discussion that is related to resource needs is thedemand for lsquolong-term supportrsquo for switchers All four cases needed longer-term support in order to be developed In principle such support can comefrom universities funding agencies or stakeholders We observed thatswitching both lsquofullrsquo and lsquopartialrsquo did not occur when the academic posi-tion was not tenured Only being a professor seems to allow researchers totake the risk of entering in still unknown fields The obvious advantage isthe long-term contract and the resources that may be linked to the post ofa professor

Long-term support by the SNSF was for a long time difficult to attainUsually funding projects are around three years Follow-up projects arehowever possible but according to our interviews mostly in the naturalscience division of the SNSF do we find a tradition of supporting researchprojects on a more continuing base This tradition is well known amongnatural scientists who count on this support and often receive it These con-tinuing funds have been advantageous for the development of BECresearch in Switzerland On the other hand it has been much more difficultfor ED to get money for large animal facilities The SNSF has not beenparticularly helpful in this respect Universities provided some resources toset up animal facilities though

The NCCR programme is nowadays the only funding instrument thatcan be used for long-term support for a very limited number of

interdisciplinary projects mostly from the natural and the life sciencesThough none of our cases have obtained a NCCR funding we findresearchers most notably in BEC who participate in NCCRs and cancount therefore on recurrent funding for their projects on BEC up to 12years

We already discussed the difficulty for CCL to build up databases by find-ing adequate long-term funding LSA by contrast has found long-term sup-port for the continuation of the various surveys In the case of TREE theSNSF policy-makers and the host university together finance the surveyAnother university has created as already indicated a professorship forLSA studies and still another was prepared to build an evaluation centre andto coordinate the PISA study which is largely financed out of political funds

Costs Linked to Scientific Integration

We stated in the beginning that it might be difficult for new cognitive areasto find academic integration especially if it contradicts existing paradigmsof research Resistance of established disciplines and academics could beexpected to be the strongest in the case of ED because traditional evolu-tionary views in biology are contested One could expect little resistance inthe case of BEC as there was no contestation of existing paradigms Onthe contrary BEC gave answers to long-asked questions in atomic physicsInterdisciplinarity played no role in this case In the case of CCL and LSAacademic integration depended on epistemological orientations in the disci-plines switchers were originally anchored in like sociology psychology oreconomics If the epistemological tradition corresponded to the ones validin CCL or LSA there was obviously no reason except for struggles aboutthe distribution of finances to resist academic integration

With regard to ED we could not detect on the base of our interviewsovert academic resistance against research done in the ED field but therewere indications of more informal resistance above all from researchers inclassical evolutionary biology who were particularly challenged Theyattempted to prevent the employment of ED researchers The more visibleresistance was by scientific peers during the process of reviewing publica-tions in traditional journals ED switchers had a hard time in getting theirpapers accepted while own journals were still in the making and lackedvisibility

In the case of LSA epistemological confrontation did occur in one casewhere the employment of a new professor in this field caused considerable

resistance and could only be resolved after considerable time The confron-tation was decided in favour of LSA Two explanations can be stressedFirst there is a social demand for this kind of research that the facultycouldnrsquot ignore second there is also a research demand on the level ofscientific research community in education

We did not find a similar struggle in CCL Also BEC had no difficultiesto be accepted within the universities where it was practiced which is inline with our expectations

CONCLUSIONS

In the conclusions we endeavour to link the insights into the institutionalconditions of switching to the lsquospreadrsquo of the four new scientific fields inSwitzerland before we come back to our general question on the influenceof institutional conditions on changing research practices to what extentdo Swiss institutional conditions have a uniform or diverse influence onchanging research practices What are institutions of support which insti-tutions are detrimental to switching

Institutional Conditions and the Spread of the Four Scientific Fields inSwitzerland

In our case studies we noted that the four fields have followed differentlsquopathsrsquo in becoming part of the scientific community in Switzerland Whatare possible reasons

If we take ED first one can characterise the lsquospreadrsquo as gradual but rela-tively slow though in line with international developments Switzerlandbelonged here to the pioneers The obvious advantage for becoming a pio-neer has been the favourable academic environment in terms of existingresearch traditions that could nurture the interdisciplinary area of ED Thiscondition has facilitated scientists to explore the new field of ED As aninterdisciplinary area it met however also resistance in the academic com-munity when ED tried to find a space of its own within the scientific com-munity Together with the difficulties to obtain long-term support thismight explain somewhat the slow further development of the field inSwitzerland and the relatively small scope in terms of numbers of research-ers active in the area Long-term support exists in Switzerland up to a

certain point by institutional means of universities ED is however an areawith high demands on material and time resources Funding agency sup-port is therefore a necessary condition For a long time however theSNSF has not had instruments to fund with a long-term perspective Thismay have contributed to the slow pace of development

BEC developed rather quickly in Switzerland once foreign pioneersbecame professors in Swiss universities Clearly Switzerland did not haveresearchers of its own that went into this developing area It needed theimport of foreign researchers to set the spread of BEC in motion In thiscase the original discipline atomic physics was not strongly established inSwitzerland This explains among other things that no Swiss researcherswent into the field The attractiveness of the Swiss research place and a pol-icy of excellence the employment of new scientists on the base of theirreputation and not on the fit with existing research lines in the university allowed to draw in foreign researchers who belonged to the pioneers in thisfield Generous financial support helped to develop the area from then onLong-term support was given not only in the ETHZ in this case but also bythe SNSF where we find for the natural sciences a policy of continuous sup-port for once accepted projects though the overall amount of money maydecline in comparison to other branches of the sciences The spread of BECthroughout Switzerland was then rather quick and the country becamequite competitive in the field

In the field of CCL Switzerland has been following pioneering develop-ments in other countries It was rather late in taking up these developmentsand until today the country remains rather at the fringe of internationaldevelopments Possible reasons for the retardation are that Switzerland is acountry with different language regions which has made it more difficult tolink to international developments It needed the lsquoimportrsquo of some foreignresearchers in the German region to give the field an impetus Againattractive working conditions made the lsquoimportrsquo possible Though someprojects were launched the pace of building up a CCL community wasslow Linguistic fragmentation is a further reason for the limited scope ofdevelopments Another important variable in this respect has been thehigh transaction costs for researchers to find institutional supportInfrastructures like the databases in CCL have difficulties in finding fun-ders the SNSF only reluctantly pays such infrastructures Support in uni-versities has been non-existent and other funders did not dispose ofsufficient funding resources

Finally the situation for LSA was not so different from CCL but never-theless LSA found ways to become supported on the long term Swiss

researchers were not among the pioneers in the area but some were eager tojump on the starting train of LSA in education Temporary initial supportwas given by the SNSF Further support was however more difficultwhich explains the ups and downs in the development of the area The fed-eral structure made it difficult to develop the databases Financial supportwas fragmented and piecemeal political support was important but alsothe preparedness of some universities to host the LSA researchers by givinga position or accepting an institute Flexible arrangements between fundersand the universities made continuing support finally possible TodaySwitzerland can keep in line with international developments and research-ers have some security in remaining funded The difference with CCL hasbeen probably the political interest in the field while CCL depended on thesupport of universities alone and never reached a high status on the prioritylist of universities

Uniform and Diverse Influence of Institutional Conditions

The discussion on the link between institutional conditions and the spreadin the four fields discussed in this article has already given some indicationsconcerning favourable and unfavourable conditions In this section we tryto bring these insights together

It seems to us that a number of institutional conditions have in generalinfluenced positively the decision of scientists to dare a switch to a newscientific field To these conditions belong without a doubt the generalfavourable financial conditions that were present since the end of the 1990sin Switzerland These financial conditions allow offering attractive posi-tions in universities for scientists These positions are internationally com-petitive meaning that they are offering enough advantages to draw in alarge number of scientists The attractiveness as an important variable wasconfirmed throughout our interviews in the four cases When Switzerlanddoes not have own researchers to promote new areas it can without majorproblems lsquoimportrsquo pioneers from abroad and integrate them into the sys-tem This is not yet enough to spread the new field further but it is a begin-ning (eg BEC ED) These attractive positions are found in all the majoruniversities but given the financial advantages of the federal technical insti-tutes and their research profile technical natural and increasingly so lifescientists are particularly favoured in the Swiss context

The employment as a professor in Switzerland gives in general a goodbase for long-term research a condition for developing new fields

especially if the employment is linked to generous infrastructural support forresearch (eg ED) Such support is characteristic of new employments in gen-eral but again it is often more generous in the two federal technical institutes

The shift of universities to a NPM regime has contributed to more flex-ibility in the use of material and time resources of universities and hasallowed to lsquoexperimentrsquo in the case of institutionalisation of new scientificfields into universities (see above all the case of LSA) The policy of excel-lence applied to the nomination of new scientists we found in Swiss universi-ties has given the opportunity to completely new fields not yet anchored inthe universities to become integrated (ie BEC) On the other hand thisnewly won room of manoeuvre to decide on strategies and priorities couldalso have negative effects for switchers as the case of CCL testifies here thedecision of university leaders to value humanities to a lower extent thanother branches of sciences has been a hindrance for the developmentof CCL

A last common trait though not in itself an institutional condition wehave found in the four cases is that lsquopartial switchingrsquo is a frequent strategyof scientists used in all four cases It is questionable if this strategy is theoutcome of institutional conditions or of practices that scientists usuallyapply when considering switching Partial switching is a lsquorationalrsquo strategywe stated above in the sense that scientists avoid taking too much riskswhen switching to a new field We cannot judge whether partial switchingcontributes to a lower pace in the development of new fields as timeresources of switchers to be invested into the new area obviously are morereduced than in the case of full switchers Partial switching could expressthough a deficiency of the research system to create conditions of supportwhich let the scientist trust that she has sufficient time to develop the newarea and built up new reputation In other words long-term support mightbe judged as insufficient by partial switchers6 This point should be clarifiedin future research

In sum the Swiss research system certainly has some institutional condi-tions that have a general positive effect on the change of research practicesThere are however also differences in the working of these institutionalconditions if we take a closer look at the cases One can mention the pointsgiven below

We found a different potential of universities in Switzerland to promoteswitching Not discussing the difference between small and large cantonaluniversities here which has not been part of our investigation one canstate that the federal technical universities have in general more attractivefinancial conditions than the large cantonal universities This helps above

all the natural and technical sciences which are the main focus of these uni-versities The high standing of these universities in international rankingsdemonstrates also in general the potential for the development of newscientific fields

Infrastructural support is relatively well developed for natural and forlife sciences in universities Continuous support by the SNSF is higher inthe case of the natural sciences The position of the social sciences andhumanities is more problematic in this respect They are subject to highersearch costs when looking for institutional support and must often dependon several funding sources Priority setting in universities had detrimentaleffects especially in the case of CCL

We have found that the academic environment that is existing researchtraditions in universities matter Their presence allows to more easily con-vince scientists to switch in fields that are based on these research traditionsED has been an example while BEC and also CCL with no tradition inbasic supporting disciplines needed to import foreign researchers in orderto develop the new areas even if some ED switchers are also coming fromabroad

This overview demonstrates that some institutional conditions in a coun-try can have an overarching influence that have a positive effect on thechange of research practices There are however also institutional effectsthat lead to different chances for the spread of new scientific fields InSwitzerland a number of institutional conditions are conducive for switch-ing throughout the different branches of the sciences but it has becomeclear that the natural and life sciences are favoured by a number of condi-tions in comparison to our areas in the social sciences and humanities

Having said this we are aware that our study has its limits when gener-alising our findings We have treated only four cases out of a large varietyof scientific disciplines and research domains Nevertheless these caseshave been chosen as examples of the main branches of the lsquosciencesrsquo nat-ural science life science social sciences and humanities The differences wefound in the support of these areas in Switzerland indicate at least that onemight find such differences also for other fields in these branches Futureresearch is needed to confirm this Our study can in general guide suchfuture research by having pointed to the kind of institutional support andinstitutional deficiencies of relevance for changing research practices Theseinsights can be taken up in future research and brought to a test either byreiterating case studies or by developing a more quantitative large-scaledesign that could use our findings as hypotheses of investigation

1 Research practices are built on specific sets of concepts and theories and theuse of specific methods and research infrastructure in a cognitively delimitatedknowledge field2 Calculation based on data from the web-site of the Bundesamt fur Statistik

wwwbfsadminch3 The biggest university Zurich has 25000 students4 The main criteria are the number of students and the total amount of SNSF

research projects5 TIMSS was developed by the International Association for the Evaluation of

Educational Achievement (IEA) to allow participating nations to compare studentsrsquoeducational achievement across borders (see httpenwikipediaorgwikiTrends_in_International_Mathematics_and_Science_Study)6 The comparative analysis of ED in this edited volume suggests such a hypoth-

esis (see Laudel et al this volume)

REFERENCES

Baschung L Benninghoff M Goastellec G amp Perellon J (2009) Switzerland Between

cooperation and competition In C Paradeise E Reale I Bleiklie amp E Ferlie (Eds)

University governance Western European comparative perspectives (pp 153175)

Bonaccorsi A (2007) Explaining poor performance of European science Institutions versus

policies Science and Public Policy 34(5) 303316

Bonaccorsi A Daraio C amp Geuna A (2010) Universities in the new knowledge landscape

Tensions challenges change An introduction Minerva 48(1) 14

Braun D (2012a) Why do scientists migrate A diffusion model Minerva 50(4)

471491

Braun D (2012b) Die Forderung wissenschaftlicher Innovation an Schweizer Universitaten

In T Heinze amp G Krucken (Eds) Die institutionelle Erneuerungsfahigkeit der

Forschung (pp 65100) Berlin Germany Springer

Braun D (2014) University governance and scientific innovation In C Musselin amp

P Teixeira (Eds) Reforming higher education Public policy design and implementation

(pp 145173) Dordrecht the Netherlands Springer

Braun D amp Benninghoff M (2003) Policy learning in Swiss research policy The case of the

national centres of competence in research Research Policy 32(10) 18491863

Braun D amp Leresche J-P (2007) Research and technology policy in Switzerland In

U Kloti P Knoepfel H Kriesi U Linder amp Y Papadopoulos (Eds) Handbook of

Swiss politics (pp 735762) Zurich Switzerland Verlag Neue Zurcher Zeitung

(Eds) Proceedings of the 12th International Society for Scientometrics and Informetrics

Heinze T amp Munch R (2010 April 2223) Mechanismen der Erneuerungsfahigkeit

Working Paper Conference Institutionelle Erneuerungsfahigkeit der Forschung

Bamberg Germany University of Bamberg

Kitcher P (1995) The advancement of science Science without legend objectivity without illu-

sions Oxford Oxford University Press

Lepori B van den Besselaar P Dinges M Potı B Reale E Slipersaeligter S Theves J amp

van der Meulen B (2007) Comparing the evolution of national research policies

What patterns of change Science and Public Policy 34(6) 372388

Mirowski P amp Sent E-M (Eds) (2002) Science bought and sold Essays in the economics of

science Chicago IL University of Chicago Press

OECD (2006) OECD reviews of innovation policy Switzerland Paris France OECD

Shi Y (2001) The economics of scientific knowledge Cheltenham UK Edward Elgar

Whitley R (2000) The intellectual and social organization of the sciences Oxford Oxford

University Press

Whitley R (2003) Competition and pluralism in the public sciences The impact of institu-

tional frameworks on the organisation of academic science Research Policy 32(6)

10151029

innovation Oxford University Press Oxford

Ziman J (2002) The microeconomics of academic science In P Mirowski amp E-M Sent

(Eds) Science bought and sold Essays in the economics of science (pp 318340)

Chicago IL University of Chicago Press

COLD ATOMS HOT RESEARCH

HIGH RISKS HIGH REWARDS IN

FIVE DIFFERENT AUTHORITY

STRUCTURES

Grit Laudel Eric Lettkemann Raphael Ramuz

Linda Wedlin and Richard Woolley

ABSTRACT

Bose-Einstein condensation is a scientific innovation in experimental phy-sics whose realisation required considerable time and resources Its diffu-sion varied considerably between and within five countries that werecomparatively studied Differences between countries can be explainedby the variation in the national communitiesrsquo absorptive capacities whilewithin-country differences are due to the impact of authority relationson researchersrsquo opportunities to build protected space for their changeof research practices Beginning experimental research on Bose-Einstein condensation required simultaneous access to the universityinfrastructure for research and to grants The former is largely limited to

ISSN 0733-558Xdoi101108S0733-558X20140000042007

professors while the latter made researchers vulnerable to the majorityopinion and decision practices of their national scientific community

Keywords Scientific innovation emergence of fields authority rela-tions absorptive capacity experimental physics academic careers

INTRODUCTION

Anyone who wants to introduce a major scientific innovation in experimen-tal physics faces a challenge because designing and constructing a newexperimental system requires considerable time and resources In 1995physicists around the world were facing such a challenge if they wanted torespond to a major scientific breakthrough namely the first production ofa so-called Bose-Einstein condensate of cold atoms A Bose-Einstein con-densate (BEC) is a specific state of matter that occurs when gases of atomsor subatomic particles are cooled to near absolute zero (lt100 Nanokelvin)that is a state of very low energy A large fraction of the atoms collapseinto the lowest quantum state at which point quantum effects occur on amacroscopic scale The occurrence of the phenomenon was theoreticallypredicted by Bose and Einstein in 1924 The first BEC of atom gases wereproduced in 1995 by two US atomic and molecular optics (AMO) groupswhich innovatively combined several recently developed cooling technolo-gies (Cornell amp Wieman 2002 Griffin 2004 Ketterle 2002)

Meanwhile BECs have proven useful in the exploration of a wide rangeof problems in fundamental physics (particularly quantum theory) whichhas led to an explosive growth of experimental and theoretical activitiesPhysicists are using BEC as methods or try to manipulate this new state ofmatter for a wide variety of applications in the more distant future whichinclude atom lasers and quantum computers

There are two reasons why the development of BEC research is ofsociological interest First BEC developed in the national experimentalphysics communities at very different velocities For example it was takenup immediately in 1995 in the Netherlands and Germany while Spain fol-lowed eleven years later The extent to which BEC has become a significantpart of national communityrsquos research activities also varies Researchabout and with BECs has become a substantial and still growing part ofphysics research in Germany while it has disappeared from the agenda inSweden

204 GRIT LAUDEL ET AL

Second experimental BEC research remained very expensive and riskyfor several years until the early 2000s Researchers who wanted to developthe innovation had to make an explicit decision had to invest their wholeresearch capacity in this enterprise and had to control above-averageresources for above-average time horizons This is why BEC research was(and some of its strands still are) very sensitive to variations in the nationaland organisational governance of research

These two aspects are likely to be linked although in ways that arepoorly understood The aim of our article is to answer the question howscientific innovations can be developed by individual researchers on themicro-level under conditions of changing community expectations and indifferent systems of governance and research management With thisanswer we want to bridge the divide between macro-level diffusion studiesof fields and micro-level studies of individual research practices

CONCEPTUAL FRAMEWORK

The development of scientific innovations and of the fields that sometimesemerge with them has interested scholars in the sociology of science for along time After Kuhn (1962) introduced the notion of a paradigm and acorresponding scientific community sociologists became interested in theemergence of new paradigms Classical studies include those by Mullins(1972) on the phage group and the genesis of molecular biology Law(1973) on X-ray Protein Crystallography Mullins (1973) on ethnometho-dology and Edge and Mulkay (1976) on the emergence of radio astronomy(see also Chubin 1976 for a critical review of this research) From a cur-rent perspective many factors one would consider crucial for the emer-gence of a specialty are curiously absent from these studies The impact ofnational science policies and organisational conditions for research on theopportunities for the proponents of new fields to change their researchpractices is not discussed and the social conditions that are discussed(eg information exchange mobility a shared identity access to graduatestudents) were not linked to organisational or policy decisions The sameholds by and large for many of the constructivist studies that include thediffusion of new research practices (eg Cambrosio amp Keating 1995 1998Collins 1998 Fujimura 1988 1992 Pickering 1980 1995 Pinch 1980)

There are several reasons why these studies have neglected many condi-tions for changes of research practices that we would today consider essential

205Cold Atoms Hot Research

Conditions for research have changed considerably In particular tempor-ary positions and competition for grants which make much of currentresearch precarious are relatively recent developments as are the highereducation reforms that increase the power of university management inmany countries (see Whitley this volume) Furthermore the laboratorystudiesrsquo focus on the micro-level made it difficult to observe the impact ofmacrostructures such as institutions (Kleinman 1998 pp 285291 Knorr-Cetina 1995 pp 160163 Mayntz amp Schimank 1998 p 751)

Although these reasons for the neglect of macro-level conditions are lesspersuasive then two or three decades ago more recent studies of the emer-gence of new research fields still struggle with the micro-macro link Theystill seem to address either the macro-level of the diffusion of new findings(eg Fagerberg amp Verspagen 2009 Heinze Heidler Heiberger amp Riebling2013 Raasch Lee Spaeth amp Herstatt 2013) or the micro-level of innova-tors (Mody 2004) Studies on exceptional research (lsquocreative achievementsrsquolsquobreakthroughsrsquo) include organisational factors and sometimes grant fund-ing but were so far unable to establish systematic relationships between spe-cific conditions created by governance and specific kinds of achievements(Heinze Shapira Rogers amp Senker 2009 Hollingsworth 2008)

Studying the micro-macro link requires empirically investigating bothmacro-conditions and micro-level changes establishing how the former aretranslated into conditions for the latter and demonstrating how micro-levelchanges are aggregated In order to establish causal links specific macro-level conditions must be compared with regard to their impact on micro-level changes Although we limit our empirical study to this latter task(leaving the study of aggregation processes to further work) we still need aconceptual framework that solves three problems The framework mustenable a strict comparison of macro-level conditions that is of nationalresearch systems and influences exercised by scientific communities thecomparative investigation of translations of these macro-level conditionsinto conditions for individual researchers who decide to change theirresearch practices and a comparison of conditions for research that explainthe differential success of researchers who want to change their practicesWe use the concept of authority relations for the first and the concept ofprotected space for the second and third tasks

For an integrated assessment of changes in public science systems wedraw on the authority relations perspective (Whitley 2010) This focuseson how different authoritative agencies (the state research organisationsorganisational elites external funding agencies and national as well asinternational scientific elites) exercise authority over specific matters of

governance which we can specify for our purposes as research goals Itsbasic assumptions are (a) that the changes that public science systems areundergoing have implications for the relationships between actors andthe way that they are able to realise their interests and (b) that authorityrelations as regards the selection of research goals are the main channelthrough which changes in the knowledge production system are effectuated

Authority over research goals can only be exercised through a fewchannels namely the allocation of resources reputation and career oppor-tunities The relative authority of actors depends on their control of thesechannels which enables the construction of a framework for the integrationof national governance processes into comparable patterns By applying theauthority relations perspective authoritative agencies at the macro-levelbecome linked to meso-level and micro-level actors because authority rela-tions include all actors who exercise authority regardless of the level at whichthey are located

Our framework for comparing micro-level conditions and linking themto authority relations builds on the definition of scientific innovations likeBEC as research findings that affect the research practices of a large num-ber of researchers in one or more fields (ie their choices of problemsmethods or empirical objects) Changing research practices incurs costs andmay be risky in several respects because

they partly devalue the knowledge and equipment a researcher has accu-mulated working on previous topics

a researcherrsquos reputation may suffer if the change requires learning orexperimental redesigns and thereby delays opportunities to publishresults and

the new line of research may deviate from the mainstream of theresearcherrsquos community which again creates the risk of losing reputation

Variations in authority relations affect the creation or diffusion of inno-vations by providing different opportunities for researchers to bear therisks and meet the costs of changes in their research practices (GlaserLaudel amp Lettkemann 2014) We use the concept lsquoprotected spacersquo forcomparing these opportunities as they are provided in our investigatedcountries (Glaser et al 2014 Whitley this volume) We define protectedspace as the autonomous planning horizon for which a researcher can applyhis or her capabilities to a self-assigned task Two dimensions of this vari-able are important here The first dimension is the time horizon for whichthe capabilities are at the sole discretion of the researcher that is the per-iod of time in which the researcher is protected from external interventions

into his or her epistemic decisions and external decisions on the use of cap-abilities The resource dimension reflects the research capacity theresearcher controls in this time horizon (personnel over which theresearcher has authority time available for research equipment consum-ables etc)

Researchers create and extend protected space mainly by career deci-sions (the search for positions that provide protected space) and the acqui-sition of funding from various sources including their organisations Thebuilding of protected space links the decisions about research to authorityrelations Applying these concepts to our empirical analysis makes it possi-ble (a) to identify the authority relations the investigated researchers wereembedded in when building their protected space and (b) to assess thescope of protected space that is the numbers of researchers in differentcareer stages whose organisational position makes it possible to build theprotected space that is necessary for a change of research practices (Glaseret al 2014)

METHODS AND DATA

We use data from a larger comparative project that studies the impact ofchanging authority relations in four countries on conditions for intellectualinnovations (RHESI) to which we added a case study about experimentalBEC in Spain Our main focus was on research groups who attempted toproduce BEC of cold atom gases We identified these groups from publica-tions using the keywords lsquoBECrsquo or lsquoBose-Einstein Condensationrsquo frominternet searches of experimental physics groups at universities and fromlsquosnowballingrsquo that is by asking interviewees about their national commu-nities Our attempt to include groups that were prevented from conductingBEC research by authority relations despite their interest failed becausesuch cases are almost impossible to identify empirically

Table 1 provides an overview of the interviews For all countries exceptGermany whose BEC community is too large by now we interviewedresearchers from all atomic physics groups that conducted BEC research(usually the group leaders and in some cases also group members) OneDutch group leader and two German group leaders declined to beinterviewed However there is only one German group about which wehave little information because no former group members could be inter-viewed In order to get a better picture of the structure of relevant physics

Table 1 Overview of Conducted Interviews

Netherlands Germany Switzerland Sweden Spain

AMO BEC groups 5 (7 interviews) 8 2 (7 interviews) 1 (2 interviews) 1 (2 interviews)

Other BEC groups 3 (6 Interviews) Other physics groups 2 experimental

AMO physics

5 (8 interviews with

BEC theoreticians

1 from another physics

field)

3 (1 BEC theoretician

1 experimental AMO

physicist 1 other field)

Other informants 2 officers of funding

agency

1 officer of

funding agency

1 former funding

advisor to

ministry

Total number of

interviews

11 9 21 5 3

HotResea

communities we conducted additional interviews with other physicists Wealso interviewed officers of funding agencies

The interviews were conducted as semi-structured face-to-face inter-views1 A shared interview guide was used for interviews with BEC research-ers in order to ensure comparability of data Interviews with researchersconsisted of two main parts In the first part the intervieweersquos research wasdiscussed We explored the development of the intervieweersquos research sincethe PhD project with an emphasis on thematic changes and the reasons forthem In this part of the interview developments in the intervieweersquosnational and international communities were also discussed The discussionof the research content was prepared by analysing documents includinginternet sites Nobel lectures (Cornell amp Wieman 2002 Ketterle 2002Phillips 1998) and publications at various levels of sophistication frompopular science to an intervieweersquos research publications The preparationalso included a bibliometric analysis of the intervieweersquos publications forthematic links publications A visualisation of this publication network wasused to prompt narratives about the content of the research at the beginningof the interview (Glaser amp Laudel 2009a see also Laudel et al thisvolume)

In a second part of the interview conditions of research and the factorsinfluencing them were discussed Topics included the knowledge personnelequipment required to produce BECs source of material support andopportunities as well as constraints provided by the intervieweersquos academicposts The interviews lasted 60120 minutes All but one were recordedand fully transcribed

The analysis of interviews focused on the variables of the theoreticalframework The comparison of cases from the five countries is based on anassessment of the necessary protected space for early BEC experimentswhich could be derived from the interviews We then reconstructed animportant macro-level condition for all researchers who attempted tochange their research practices namely the international diffusion of BECresearch in the contexts of opinions and preferences in the internationaland national scientific communities For each attempt to begin the experi-mental production of BECs the building of protected space the authorita-tive agencies involved and the consequences of the exercise of authoritywere determined and compared Reasons for delayed and failed attempts toproduce BECs were also traced back to authority relations and interests ofthe actors involved Our comparison of changes of research practices at theresearcher level follows the distinction between supported delayed andprevented cases because this distinction emphasises the impact of authority

relations The empirical analysis is followed by a generalising discussion ofthe macro-micro link and conclusions

EPISTEMIC PROPERTIES OF EXPERIMENTAL BEC

RESEARCH AND ITS REQUIRED PROTECTED SPACE

From the first attempts until the early 2000s manufacturing BECs of atomswas an exceptionally complex risky and expensive undertaking even bystandards of the wider field in which it is located experimental low tem-perature physics Manufacturing a BEC was strategically uncertain becauseit was not clear that the effect could be experimentally produced and tech-nically uncertain because it was not clear how by what specific experimen-tal setting the effect could be produced (on these notions of strategic andtechnical uncertainty see Laudel and Glaser 2014) The strategic uncer-tainty concerned the question whether any gas of cold atoms would staygaseous at the low temperatures and relatively high densities required forBEC It was quite plausible that all atom gases except hydrogen wouldbecome liquids or solids if cooled so far This question stood anew for eachnew kind of atom for which BEC was attempted replications of suchexperiments were the only exception The technical uncertainty was highfor all researchers who tried to produce a BEC for the first time Setting upthe experimental system required a researcher to go through a longsequence of steps of adjusting and fine-tuning the equipment In manycases parts of the equipment were built to order by technical workshopsThis is why for a long time (at least until the early 2000s) building andfine-tuning a BEC experiment took several years It was always possiblethat the researcher could not solve the technical problems involved inwhich case the experiment failed Although the technology for BEC pro-duction has advanced during the last two decades setting up a BEC experi-ment for the first time remains a risky and demanding endeavour for aresearch group

These epistemic properties of BEC research correspond with a large pro-tected space in terms of resources and long and often unpredictable timehorizons Achieving BEC in atomic gases required the combination of themost advanced techniques for cooling atoms and trapping those with thelowest energy The research involves complex task-specific equipmentwhich is usually built from components by the researcher Depending onthe research prior to the move to BEC several of the more expensive

components might already exist in the laboratory The equipment for aBEC experiment could cost 100000500000 Euros depending on whatwas available in the laboratory Consumables (mostly very expensivecooling liquids) caused additional recurrent costs At least two full-timeresearchers (almost always PhD students) were needed to build and adjustthe experimental setup parallel work of more PhD students or postdocswould be an advantage to accommodate the technological uncertainty

Owing to the inherent uncertainties the time horizon of the experiment(from setting up the experimental system to publishing the results) mayextend beyond the usual three-year grant cycle The reputational riskinvolved is high because the experiments can fail entirely and because littlecan be published until the experiment is successful

INTERNATIONAL AND NATIONAL DYNAMICS

OF BEC RESEARCH

The development and diffusion of experimental BEC can roughly bedivided in four phases Attempts to produce BECs began in the 1980s aftera new cooling method (laser cooling) was developed This first phase endswith the first successes in summer 1995 For about three years afterwardsresponses to this success were mixed because the experiments were too diffi-cult to replicate and the potential of BECs for further research remainedunclear With the first successful replications and BECs of other atomsBEC research began to grow in the AMO community after 1998 Since theearly 2000s the manufacture and use of simple BECs became routinisedand BECs began to be used as a method in other areas of physics

Phase 1 First Attempts to Manufacture BECs (1980s to Summer 1995)

Since it had always been clear that the experimental realisation of BEC inatom gases depended on achieving extremely low temperatures experi-ments aimed at producing BECs seriously began only after a new coolingtechnique laser cooling had been developed Still the majority of thescientific community believed that producing BECs was impossible becausethe atom gases would turn into liquids or solids when cooled down to thetemperatures necessary for BEC Some believed hydrogen to be the onlypossible exception which is why major experimental efforts began in the

1980s when a few condensed matter physics groups tried to produce BECin spin-polarised hydrogen gas by combining several cryogenic methods Atthe beginning of the 1990s they were considered the leading experts in BECresearch However most AMO physicists doubted that a breakthroughcould be achieved in the near future A different route towards BEC waspursued by a small minority of researchers in the US who began to coolatom gases of alkalis This idea was met with even stronger scepticism thanthe hydrogen route

Bose-Einstein condensation was a vision of the past [early 1990s] and many people did

not believe in it They said this is not possible This will not work for all sorts of rea-

sons Just before the first Bose-Einstein condensate was produced in 1995 there were

predictions that it is not possible (German BEC researcher)

Of course all people who worked with cold atoms wrote in their applications about the

lsquoHoly Grailrsquo that this will at some time lead to BEC But this was so far away for

everybody that we thought that we will never achieve it Just imagine you always work

with buckets of water on a great plain and there are puddles everywhere and suddenly

somebody sets out to fill an ocean It is clear that it becomes more when I have more

buckets but it is unimaginable that it will work in the foreseeable future (German BEC

researcher)

In the early 1990s only two of the five national physics communitiesincluded in our study featured strong AMO physics communities TheGerman and Dutch communities had made major contributions to laser-cooling techniques Both national communities shared the internationalmajority opinion namely that BEC cannot be experimentally achievedNone of the German AMO groups pursued BEC at that time Researcherswho did consider attempts to produce BEC felt disheartened by the vastadvantage of the US groups in both resources and experience Since thenew method of laser cooling had created many new research opportunitiesthere was no reason to pursue the lsquoholy grailrsquo of BEC whose realisationwas doubtful anyway

In the Netherlands the strong tradition of low temperature physics and astrong AMO physics provided a supportive background for BEC researchOne of the Dutch research groups worked on BEC in atomic hydrogen TheDutch group leader belonged to the pioneers that paved the way towardsthe first BEC His major scientific achievements and being backed by thehydrogen community probably helped to continue this research despite theAMO communityrsquos doubts Another Dutch researcher became interested inBEC in the early 1990s but could not begin for lack of funding

The situation in the three other countries was quite different Switzerlandand Spain had no tradition in AMO physics and were unaffected by the

experimental development Switzerland has a strong tradition in nuclearand condensed matter physics instead while Spainrsquos experimental physicshas been dominated by nuclear and particle physics which had little episte-mic connections to AMO physics cold atoms and BEC research Swedenhas a long tradition in the field of AMO physics with many strong experi-mental groups However most research concentrated on molecular spectro-scopy building on the tradition of Nobel Prize winners Manne Siegbahn(1924) and Kai Siegbahn (1981) This narrow focus was criticised by aninternational evaluation committee in 1992

lsquoSweden for many years has been strong in Atomic Molecular and Optical Physics

almost all the work is of high quality and some is outstandingrsquo but also recommends

lsquo there should be less emphasis on traditional molecular spectroscopy and more on

newer fields of atom trapping laser cooling laser optics etcrsquo (Bradshaw et al 1992

pp 1617)

The fields whose strengthening was recommended were those on whichexperimental BEC built

Phase 2 Responses to the First Experimental Success(Summer 19951997)

In the summer of 1995 first empirical evidence of BEC was presented at aninternational physics conference at Capri Until the end of the year threeUS research groups were successful in producing a BEC from alkali gasesThis was immediately regarded as an outstanding contribution by AMOphysicists and by the wider physics community However the internationalAMO community was undecided whether these achievements marked theend of the long quest for the lsquoHoly Grailrsquo of BEC or the beginning of anew research area Would it open up opportunities for interesting new phy-sics or was it just the experimental confirmation of a theoretical predictionIn the year following the Capri conference it was unclear what directionresearch would take

Most German researchers tended to share the general beliefs of the inter-national community it was not clear whether BEC would indeed open anew research direction

Q [After the 1995 Capri conference] was the entire German community of the same

opinion

A (Laughs out loudly) It was very mixed Some [people] just went quiet and reverent

and saw that this is the Holy Grail But I would say that what you can do with it that

it opens up an entirely new parameter space and that you can make a completely new

physics was recognized only by few back then (German BEC researcher)

Secondly the problem of the competitive advantage of the US groupshad not vanished Thirdly other AMO fields promised better careerchances Against this general trend in the German AMO physics commu-nity three groups (two professors one junior group leader) began BECresearch after the first experimental success was announced at the Capriconference in mid-1995 Although the community approved of grants onlyreluctantly all three researchers could immediately begin with the first stepsof building the experiments which included simulations and invitations tothe innovators from the United States in order to learn the tricks of thetrade from them The first German BEC (and first BEC outside the UnitedStates) was achieved in 1997

In the Netherlands the researcher who had originally worked on BECin hydrogen switched to the alkali route because the technology hadbecome superior to the complicated cryogenic approach for BEC in hydro-gen Four more researchers became interested in pursuing BEC research inalkalis in this phase However only one professor and a researcher whosemove was tolerated by his professor could start in that period Tworesearchersrsquo change of research practices was delayed by authority rela-tions as we will explain in more detail in the following section

Phase 3 Growth of BEC Research within AMO Physics (Since 1998)

The year 1998 witnessed new BECs being produced in many countries Theresearch now moved beyond the replication of the original results becauseit became obvious that BECs provided many opportunities for interestingtheoretical and experimental research Since then more than a hundredresearch groups worldwide have produced BECs BECs of atoms of otherelements photons and molecules have been produced The technologiesfor producing BECs have been improved which made it possible to useBECs as tools in several other research areas The growth of BEC researchwas thus accompanied by an internal differentiation

Again the German AMO physics communityrsquos attitude towards BECparalleled that of the international community The great potential of BECexperiments became widely accepted More groups began BEC researchand the experiments of all German BEC groups moved beyond the replica-tion of original experiments in new directions Today about 15 GermanAMO physics groups work on BEC

The new attitude of the German AMO community was also reflected inthe decisions on project funding Grants for BEC research were approvedwithout problems and the delays in producing BECs were tacitly acceptedby approving grants for researchers who had already had a grant but werenot successful within the first three years

In the Netherlands the growth period of BEC in AMO physics wasbriefly reflected in a dedicated grant programme for the support of BECresearch Two further Dutch atomic physics groups began their BECresearch Different from Germany the communityrsquos attitude to BEC wasless favourable

But the situation in Holland was not very favourable [several years] after the first BEC

was observed Some people thought why are you going to do this now Other coun-

tries are ahead of us why should we do that There was one experiment in Amsterdam

where they achieved Bose Einstein condensation and some thought well perhaps this

is sufficient why do you need four groups (Dutch BEC researcher)

Different to their German counterparts the Dutch funding agencybecame impatient soon and downsized the funding programme when noBECs were produced after three years Further grant funding was concen-trated on two groups one of which already had produced a BEC in 1999The funding problems caused delays and forced one group to give up theirattempts altogether In the end three more groups achieved BEC between2004 and 2006 Currently four groups continue BEC research

From 1998 onwards BEC research began in the three other investigatedcountries too In 1998 a young Swedish researcher on a 4-year fellowshipposition began BEC research which he intended to conduct parallel toanother line of research that was also based on laser cooling In spite of theresearcherrsquos appointment as professor at a Swedish university his group didnot succeed due to funding difficulties In 2009 the group leader left thecountry without having achieved BEC and the Swedish group dissolved

The Swiss physics community caught up with BEC research by recruitinglsquoready-madersquo BEC researchers This was not completely premeditated Theuniversities recruited whom they considered as the best candidates in quan-tum physics who happened to be BEC researchers However the decisionwas likely to be influenced by the high potential of the BEC field In 2000the first AMO physics groups began to work on BEC a second group fol-lowed recently Both groups expanded rapidly with several parallel BECexperiments with at least one group belonging to the international elite

The situation in Spain was unusual in that the first BEC experiment inSpain was initiated by a researcher from a different community namely the

theoretical physics community of non-linear optics He had conductedtheoretical BEC research since the late 1990s and had made important con-tributions to the field But it was only after he became a full professor thathe considered the move towards experimental BEC and he started in2006 Taking into account his limited access to resources he wanted to pro-duce what could be considered a lsquostandard BECrsquo at that time (a BEC ofRubidium atoms) However this intention was not enthusiastically met bythe national community which doubted the value of a Spanish lsquosimplersquoBEC given how far other groups worldwide had moved ahead The com-munityrsquos prevailing attitude was that it would be better lsquovalue for moneyrsquoif he continued to produce theory papers Being fully autonomous as auniversity professor the researcher tried nevertheless but failed because hedid not receive grant funding

Phase 4 Utilisation of BEC in Other Areas of Physics(Since the Early 2000s)

With BEC increasingly proving its usefulness as an instrument for funda-mental research questions in physics and beginning to show remote applica-tion opportunities the production of BECs became interesting for otherareas of physics such as condensed matter physics In Germany severalgroups worked on BEC topics related to condensed matter physics

This diffusion of BEC in other fields of physics is most pronounced inSwitzerland where the strong condensed matter community turned to BECresearch Around 2001 a condensed matter physicist began BEC researchbecause he wanted to add a BEC research line to his research portfolioDue to the high technical uncertainty it took around seven years to achievesuccess which was immediately recognised as a major achievement by theinternational community Two more condensed matter physicists workedon BEC with completely different approaches In one case BEC was aserendipitous discovery while the other researcher is only marginally inter-ested in BEC and has not yet produced a BEC after eight years

Fig 1 summarises the diffusion dynamics of BEC research in cold atomsin the five countries we studied The Netherlands and Germany had startedearly on with BEC research Switzerland and Sweden followed relativelylate and Spain only recently started Germany has the strongest BECcommunity now with several groups belonging to the international elite TheNetherlandrsquos BEC community consists of four groups that are internationallyrecognised Switzerland has one group belonging to the international elite

and another strong group2 Several groups could not immediately start BECresearch (the delays are marked grey)

HOW VARIOUS SETS OF AUTHORITIES ALLOW

HINDER PREVENT BEC RESEARCH

Having established interests and influences of international and nationalscientific communities we now turn to a second set of macro-level condi-tions and their link to changes of research practices We begin at the micro-level of individual changes by asking how changes of research practices wereaccomplished A first comparison across all countries reveals that a largenumber of these changes were supported by the set of authority relations inwhich they were made while other attempts to produce BECs were delayedor even prevented by the exercise of authority as summarised in Table 2

The dynamics of the changes in research practices vary considerablywithin countries The shortest time between the start of BEC research andsuccessful publication is one year (for the researcher in Switzerland whohad successfully produced BECs before) the longest time is ten years (for aDutch researcher whose move to BEC was delayed by authority relations)

Fig 1 Diffusion Pattern of BEC Research in Five Countries

The delays and failed attempts certainly were also influenced by the cap-abilities of researchers which we could not reliably compare across allcases in five countries (see Glaser amp Laudel 2009b on that problem)However we can trace the reasons for delays and failures back to particu-lar patterns of authority relations which systematically differed from thoseof the supported cases

Since the variance in dynamics is bigger within countries than betweencountries we begin our analysis by comparing supported delayed andprevented cases Findings on countries will be synthesised in a subsequent step

Authority Relations Supporting BEC Research

The ideal-typical situation of a researcher whose move to BEC researchwas supported by authority relations was somebody who controlled anadaptable infrastructure and had access to external grants in a way that

Table 2 Impact of Authority Relations on Changes of ResearchPractices

Impact of Authority Relations on Changes of Research

Practices

Supported Delayed Prevented

Occurrences Germany

Netherlands

Switzerland

Germany

Netherlands

Sweden

Time from first intention to begin

of work (years)

Immediate start Germany 27

Netherlands 36

Immediate start

Time from begin of work to

achieving BEC (years)

Germany 27

Netherlands 3

Switzerland less

than 1

Germany 2

Netherlands 49

publication of results

Germany 38

Netherlands 4

Switzerland 1

Germany 38

Netherlands 610

supported the unpredictable time horizons Researchers in Germany theNetherlands and Switzerland found themselves in that situation They heldprofessorial positions which granted them discretion over the necessaryinfrastructure (laser equipment) and some personnel Building or changingthe infrastructure for BEC research was supported by resources providedby the university which were granted as start-up packages on appointmentas professor or as loyalty packages if a professor received an invitation towork elsewhere but stayed at the university Many professors already hadsubstantial parts of the necessary equipment (eg the lasers) and thus oftenneeded only relatively small amounts of additional funding of ca h100000Technical support provided by the research organisations was crucial forbuilding the experimental setups In terms of the required knowledge theAMO physics groups were either themselves familiar with laser optics andcooling technologies or they hired postdocs who had learned it in the lead-ing laboratories abroad

With the exception of one German professor whose start-up packagewas so generous that he could work without external grants for severalyears the money for additional equipment and personnel came from exter-nal grants For that (personnel) and for specific equipment additionalresources were needed This made professors dependent on their scientificcommunities whose attitudes towards BEC and grants for it changed overtime As we saw in the previous section the majority opinion in scientificcommunities was not in favour of BEC in the first two phases The Germancommunity would nevertheless provide grants albeit reluctantly so

The application was in June rsquo95 In May I believe was the [Capri] meeting In June I

sent the application out for review This review process was stopped by the DFG []

The approval of my application is well I would have to look it up but I think that it

lasted almost two years It was approved when I achieved the BEC (laughs) Maybe it

was just one or one and a half years or something like that So it was extremely tough

They posed further questions they did not answer for months and it was terrible

(German BEC researcher)

The Dutch community did not approve of any grants for BEC exceptthose for the early innovator until the late 1990s and cut funding soonagain in the early 2000s (see Glaser et al 2014 for a comparison ofGerman and Dutch decision practices on BEC grants) This attitude con-tributed to several delayed cases (see the following section)

For professors who held non-BEC grants their communitiesrsquo reluctancedid not matter due to another practice namely the complete transfer ofauthority over the use of the money to researchers once the grant wasapproved lsquoBootleggingrsquo the use of the grants for different purposes was

tacitly approved by both the German and the Dutch funding agencies Thisled to the paradoxical situation that a community did not explicitly approveof grants for BEC but let researchers use grants on BEC that were approvedfor other purposes

Another tacit practice helped adjust the grant funding to the long andoften unpredictable time horizons of BEC research Grants were usuallygiven for three (Germany) or four years (Netherlands) while producing aBEC could take much longer The scientific community responded to thisdiscrepancy by awarding new grants regardless of the experimental successof previous ones

Well I must say that we have always been supported by the Deutsche

Forschungsgemeinschaft especially with these high-risk projects So in the case of BEC

which as I said took seven years you could have said many times lsquothatrsquos itrsquo and lsquothere

will never be resultsrsquo Nevertheless we have always been successful in writing applica-

tions (German BEC researcher)

The same pattern occurred in Switzerland in the 2000s where a con-densed matter group received consecutive grants for its non-atomic BECand succeeded after seven years

The ideal-typical pattern thus consists of a professor who controls alaboratory infrastructure and some personnel has access to technical work-shops and can utilise grants for a change of research practices There weretwo deviations from this pattern which nevertheless still included authorityrelations that made the change of research practices possible

Several researchers started BEC on non-professorial positions In thesecases the necessary access to infrastructure and personnel was not givenwhich made the researcher dependent on others German and Dutchresearchers below the professorial level could start their BEC work becausetheir professors (or directors of institutes) approved and granted themaccess to their infrastructure Researchers whose professors did not grantthat access were delayed in their move (see the following section) In onecase this access was granted by a Dutch faculty which wanted to compen-sate for the lack of grants for BEC

The other deviation from the ideal-typical pattern is that of the twoSwiss researchers who produced BECs of atoms Both researchers suc-ceeded in a very short time due to prior experience and above-averagefunding They both had produced their first BEC abroad and had to re-create experimental setups they were already familiar with when theymoved to Switzerland Both researchers received considerable start-upfunding from their university departments for equipment and several PhD

and postdoctoral positions In addition one of them was permitted by hisformer lab leader to take the equipment for his previous BEC experimentwith him Both researchers extended their protected space quickly by exter-nal grants from the Swiss National Science Foundation (SNF)

The ideal-typical pattern and the deviations from it make the authorityrelations that supported the move to BEC research clearly identifiableResearchers had to have authority over infrastructure and personnel whichwas granted by their universities to professors They also needed grantswhich the scientific community was initially reluctant to give directly Thisexercise of authority concerning dedicated BEC grants was compensatedfor by the transfer of authority over the use of grants to researchers Thusthe transfer of authority over resource by both universities and scientificcommunities enabled the building of the large protected spaces

Authority Relations Delaying BEC Research

In several Dutch and German cases authority relations delayed changes ofresearch practices because researchers could not build the necessary pro-tected space Either of the two processes identified above the transfer ofauthority over infrastructure from universities to professors and the transferof authority over grants from scientific communities (through fundingagencies) to researchers could be absent or halted If this was the casesome researchers waited for authority relations to improve which causedthe delays

The most frequent reason for delays was the lack of access to infrastruc-ture Two Dutch researchers on tenured but non-professorial positions hadto postpone their BEC research in the late 1990s because their professorswere not interested in BEC research and therefore would not lsquolendrsquo theirinfrastructure for this topic This lack of opportunities to build protectedspace contributed to their fears of not being able to compete with thegroups in the US Thus they only started around five years later with BECresearch Two German researchers were also delayed by missing access toinfrastructure One of them had to wait for a professorial position whilethe other had become appointed but did not receive the start-up packagefor several years He had successfully negotiated when he was recruited butdid not receive the money due to financial difficulties of his university

Missing project funding was a reason for delays in the NetherlandsThe main funding agency for physics research Stichting voor Fundamenteel

Onderzoek der Materie (FOM) was initially reluctant to fund BECresearch In the first phase prior to the first experimental success onlyone researcher received grants for research on BEC in hydrogen Anotherresearcher who wanted to take up the idea of BEC in alkalis before thesuccess in the United States did not receive grants and had to postpone thestart of BEC research

At the end of the 1990s Dutch researchers joined forces and wrote abottom-up funding proposal dedicated to BEC research This proposal wasreviewed and had to be revised which delayed the funding programmeuntil 2000 This funding programme gave researchers who wanted to pro-duce BECs access to dedicated grants The programme was evaluated afterthree years and was stopped because no further BECs were achieved afterthe first success in 1999

A And then there was the [hellip] programme on cold atoms starting somewhere in 2000

and that lasted only three years and then it was stopped by FOM because they thought

there was not enough progress

Q In form of publications

A In the form of Bose Einstein condensates Because you have to realize it started in

2000 and then we of course promised Bose Einstein condensate here in Holland and

two years later there were still no Bose-Einstein condensates And then the funding

agency said we stop the programme (Dutch BEC researcher)

The funding agency FOM established a second funding programme forBEC but gave all funding of this programme to the group of the researcherwho already had produced a BEC and to a junior research group leader inhis department These groups were the only two whose research was nothindered by insufficient funding The other groups faced shortages Twogroups continued their BEC research but could not set up parallel experi-ments which was common in BEC research As a result both groupsrsquo suc-cesses were considerably delayed A third group had to give up BECresearch entirely (see below)

The cases of delayed success in producing BECs confirm the necessaryauthority relations derived from the cases that were not delayed Research-ers that were successful in the end were delayed because they had to waiteither for a position that gave them authority over the infrastructure andpersonnel provided by the university or for grants that provided the neces-sary complementary funding The following cases of unsuccessful experi-mental BEC demonstrate what happens if the two conditions are neversimultaneously fulfilled

Authority Relations Preventing BEC Research

A Dutch a Swedish and a Spanish group failed in their attempts toproduce a BEC Having tried for seven to eight years these groups had notproduced BECs and two abandoned experimental BEC research Whileepistemic reasons may have contributed to these failures the authority rela-tions in which these researchers worked also deviated considerably fromthose characterising successful cases The three cases have in common thatat some point the continuation of BEC research of all three groupsdepended on experimental success that is on an externally enforced timehorizon of their protected space that was too short

A Dutch group was confronted by the termination of the BEC fundingprogramme by FOM after three years (see the preceding section) The uni-versity did not compensate for the exclusion of the group from grant fund-ing and the group could not obtain enough grants to continue BECresearch by bootlegging money

A Swedish researcher on a 4-year fellowship had obtained knowledgeabout laser-cooling experiments in one of the leading US laboratories andmanaged to secure a grant from a major foundation for setting up a laser-cooling laboratory at his home university The start-up process was slowbecause neither the fellowship nor the infrastructure grants allowed for anyfunding of additional research positions The faculty did not allocate PhDpositions very likely because he was not an established researcher on apermanent position yet BEC research also was new to the physics depart-mentrsquos research agenda The researcher could later secure two PhD posi-tions for his laboratory which both were jointly funded by his departmentand by external grants Three years later in 2001 the researcher wasappointed as tenured associate professor at another Swedish universityThe recruitment came with no start-up funding and further funding forPhD students or equipment remained problematic for the group

The group felt there was a lack of support from the university for theBEC research Support further declined when a new dean of the depart-ment was appointed There was a lack of understanding for the technicaluncertainties of BEC research and disappointment about insufficient publi-cations Eventually the group leader managed to secure funds from smalllocal Swedish funding agencies to fund doctoral students for his laboratoryStill the group had only minimum resources in an environment where theycould not collaboratively use the infrastructure of others While fundingfor equipment did not seem to be a major problem the grants left littleroom for the experimental failures that were unavoidable under conditions

of high technical uncertainty Technical support by workshops which hadbeen crucial for the German and Dutch BEC groups was almost non-existent The group could not recruit experienced postdocs from otherlaboratories One PhD student travelled to one of the well-known laser-cooling labs in the United States to obtain additional knowledgeComparing the conditions in the US lab with what was available in hishome setting he said that his group had to work lsquowith duct tape andhome-made solutionsrsquo to make the experiment work BEC was neverachieved In 2009 the group leader moved abroad to take up a professor-ship and the only attempt of establishing experimental BEC research inSweden ended

In Spain a theoretician who was internationally recognised for his con-tributions to theoretical BEC research became interested in experimentalBEC which he started when he obtained a professorial position in 2006Although the Spanish physics community did not support his idea of pro-ducing a standard BEC he managed to obtain some resources for theexperimental work A PhD student on a 4-year scholarship provided by theministry was awarded through a general PhD programme and not specifi-cally dedicated to BEC research but could be oriented this way Theresearcher also received two small 2-year grants from the national govern-ment and regional government (together around h60000) The researcherfurther bootlegged funding from grants for his theoretical research to buyequipment This enabled the beginning of building the experimental setupalbeit on a shoestring budget He got some ad-hoc assistance from hiscolleagues-experimentalists in the department There was no technical sup-port from the university at all Through visits of leading European labora-tories he and his PhD student obtained the necessary knowledge about thecooling techniques of BEC experiments The visits were made possible byhis reputation for theoretical BEC work

The reputational risk of this attempt at experimental BEC was reducedsince the researcher continued his theoretical BEC research and was thusable to continually publish However the time horizon of protected spacewas limited to two years because the Ministry made funding conditional onproving experimental success

Yes we finished the MOT [magneto-optical trap] and just on time In the Ministry

they said okay if you donrsquot have the MOT by this date you will not have more money

So we were really desperate and I remember the day when I was with this experiment you know in these experiments you have lots of things to tune - and I was there and I

was very impassioned with a detector we had there quite slow 100 times more and we

saw a flash there and okay and I said to [the PhD student] have you seen that and she

said yes and we were very delicately moving and then that was when we got it And yes

it was a really happy day One week before the deadline (Spanish researcher)

The group succeeded only with the first experimental steps but not withproducing a BEC Subsequent evaluations by members of the national phy-sics community questioned the scientific significance of the standard BECFinally the national funding agency rejected further grant applications fortwo successive years

If we look at the authority relations in the three cases of failed attemptsthe contingencies of BEC research become apparent Researchers simulta-neously need access to infrastructure and grant money for a sufficiently longtime horizon which means that they depend on the university having theseresources and granting them and on the approval of their research by theirscientific community In the Netherlands the funding agency decided not toprovide grants for BEC research to particular groups anymore because thetime horizons were at odds with common expectations In Sweden thelimitations of grants were not overcome by support from the universitypartly because university funding was limited and partly because theresearcherrsquos colleagues also expected quick results In the Spanish case BECresearch appeared somewhat like a suicide mission a scientific communitywith strong misgivings funding agencies controlled by the government withlittle understanding for the protected space needed Only the flexible use offunding made it possible to start the experimental BEC work at all

MACRO-MICRO LINKS IN THE DEVELOPMENT OF

SCIENTIFIC INNOVATIONS

Having identified the processes by which international and national scienti-fic communities developed the innovation and the role of authority relationsin the building of protected space for changing research practices at theindividual level we can now return to our question about the link betweenmacro-level authority relations and individual changes of research practicesWe focus our discussion on the exercise of authority by the actors involved

Absorptive Capacity of Scientific Communities and Exercise of Authority

Scientific communities exercise authority through their scientific elite whichcontrol the grant funding that has become a necessary source of protected

space in experimental research In the case of basic research the state and(in Sweden) private foundations supplied these resources but transferredtheir authority over them almost completely to the scientific elites Theirdecisions in funding agencies which are macro-level actors bypassed allother actors including universities and their sub-units and directly shapedresearchersrsquo opportunities to build protected space This is why the commu-nitiesrsquo attitudes towards BEC research and to the use of grants for it werecrucial for micro-level changes of research practices (Table 3)

Our reconstruction of the emergence and growth of BEC research in theinternational scientific community and in five national scientific commu-nities shows that the idea of prior research creating absorptive capacitywhich has been introduced by Cohen and Levinthal (1990) to organisa-tional sociology is also applicable to scientific communities and theirresearch traditions We found four distinct situations each of which cre-ated specific responses to the innovations Germany and the Netherlandshad strong research traditions in the field in which the innovation wascreated and featured research groups that contributed to the groundworkfor the experimental success This could be expected from the largeGerman science system which is likely to host most research traditions

Table 3 Absorptive Capacity and Attitudes towards the Use of Grantsfor BEC in Five Communities

Germany Netherlands Switzerland Sweden Spain

Absorptive

capacity

High High Low (high in

condensed

matter

physics)

Low Low

Acceptance of

proposals that

contradict majority

opinion

Yes with

reluctance

Only for one

member of

international elite

observed

Not observed No

Acceptance of time

horizons exceeding

terms of grants

Yes Only for one

member of

international elite

Yes Not observed No

Transfer of

authority over use

of grants to

researchers

Yes Yes Not

observed

Limited Yes

in some form The absorptive capacity of the Dutch community was his-torically contingent The Netherlands had a strong research tradition inlow temperature physics that dates back to the 19th Century which pro-duced many important contributions to cooling techniques

The example of Sweden highlights the contingent nature of such tradi-tions Sweden featured a strong AMO physics tradition which howeverwas narrowly focused on the legacy of two Swedish Nobel laureates to theexclusion of those subfields of AMO physics that provided the absorptivecapacity for BEC research Switzerland and Spain did not even have AMOphysics communities

Consequently the German and Dutch research groups were best posi-tioned to develop the innovation and responded first by producing BECsThe Swiss response is yet another illustration of the importance of absorp-tive capacity Apart from the recruitment of two researchers who producedBECs of cold atoms several initiatives to produce BECs were developed inthe condensed matter physics community which is a strong Swiss researchtradition Having no absorptive capacity Swedish and Spanish physics hadto rely on individual activists who both can be said to have failed due tothe wider physics communityrsquos lack of understanding for the intricacies ofBEC research particularly the technical uncertainties and the resultingunpredictable time horizons

The degree to which the communityrsquos majority opinion actually mat-tered depended on the decision style of grant funding processes (Glaseret al 2014) Of the three communities for which this could be observedbecause proposals that were at odds with the majority opinion were sub-mitted the German community exhibited a pluralistic attitude and (albeitreluctantly) funded proposals the majority did not consider worth fundingThe Dutch and the Spanish community did not with the Dutch communitynot believing in the possibility of BEC in alkalis (first phase) and in thescientific merits of continuing with BEC after the first experimental success(second phase) The Spanish community did not believe in the merits of aleading theoretician venturing into experimental BECs

The communityrsquos response to the unpredictable and often long timehorizons of BEC research is a third aspect of the exercise of authority thatsignificantly affected researchersrsquo opportunities to build protected spacewith grant funding The German and Swiss communities accepted this pro-blem and responded by approving new grants despite the previous ones notbeing successful The response of the Dutch community is rather surprisingbecause AMO physicists would know the problem (and have it experiencedthemselves) However the decision was made at the level of funding

programmes with the wider physics community being included in a deci-sion about fields to which the money should go In Sweden the attitudetowards time horizons was not observed for the grant funding process butwas clearly visible in the critical attitude of the BEC researcherrsquos faculty con-cerning delayed publications In the Spanish case the impatience manifesteditself in intermediate lsquomilestonesrsquo for setting up the experiment that seem tohave generated at the ministerial as well as the community side of funding

The opportunities to build protected space were finally affected by theextent to which authority over the use of grants was transferred toresearchers once the grants were approved German Dutch and Spanishresearchers were able to overcome other obstacles set by their communitiesor by administrative delays in the grant funding process by lsquobootleggingrsquomoney from grants they received for other purposes This opportunity gavea clear advantage to lsquowealthyrsquo researchers that is to those who were suc-cessful in acquiring many grants In the Swiss case the transfer of authoritywas not observed because bootlegging was not necessary The same appliesto the Swedish case although some of the grants the BEC researcherreceived carried strong limitations (they could be used only for equipment)

Thus although scientific communities used only one channel to exerciseauthority over BEC research this channel turns out to be quite complexThe approval of grants and expectations concerning their use and outcomescarry the use of both specific authority concerning the content of research(in our case BEC research and its time horizons) and unspecific expecta-tions that include the extent to which authority over grant use is transferredto grant holders

The Translation of Authority Relations in Micro-Level Conditions forChanging Research Practices

The second essential source of protected space for researchers was thefunding provided by their research organisations (mostly universities for acomparison of universities and public research institutes see Glaser et althis volume) The authority exercised through this channel also created acomplex pattern that varied between countries The state was the ultimatesource of all resources for BEC research (a Swedish private foundation beingthe only exception) but transferred its authority over all these resources toother actors

Authority over resources for research is transferred to research councilsand universities in all five investigated countries The degree to which

authority is transferred varies and has been greatly increased with recenthigher education reforms in the Netherlands Germany Sweden andSwitzerland While the use of resources within universities was previouslyprescribed in great detail universities now have more authority over theuse of their resources

In all five countries control over the expensive infrastructure for experi-mental research was concentrated on professors This situation occurredregardless of the progress of higher education reforms Even in the Dutchsystem whose career structure and university governance was reformed inorder to reduce authority of professors the latterrsquos control over infrastruc-ture and PhD students remains This is partly due to scarcity Universitiesdo not have enough resources to provide all their academics with infra-structure and if a selective approach is necessary anywhere concentrationon the most highly reputed academics suggests itself

In university systems that are further reformed professors share theirauthority with their parent faculty which has the authority to allocateadditional resources or positions for PhD students We observed this splitauthority in the Netherlands and Sweden In Germany the authority overinfrastructure is still transferred exclusively to professors while the author-ity of the faculty was not relevant to the Swiss and Spanish cases

This made professors the gatekeepers of the BEC innovation in all fourcountries While the grant allocation process was more diverse and also pro-vided researchers below the professorial level with resources for BECresearch the second essential source for protected space the universityrsquosinfrastructure could only be utilised by professors or those whose researchprojects professors approved of This is why fellowship positions rarely pro-vided enough protected space for a change of research practices towardsBEC Even generously funded fellowships depend on pre-existing infrastruc-ture which made fellows dependent on those who control it The very fewexceptions from this pattern we observed confirm rather than challenge itThey include one decision of a Dutch faculty to provide a non-professorialresearcher with resources for infrastructure and a German researcher whohad accumulated grants that might have sufficed for a change of researchpractices (which was not tested because he was appointed professor soonthereafter)

CONCLUSIONS

The experimental manufacture of BEC constituted an extreme case becauseit required protected space that is large both in the resource and the time

dimension from the beginning of the 1990s at least until the beginning ofthe 2000s This made it a very good test case for the observation and analy-sis of authority relations Our analysis enables some general conclusionsabout the impact of authority relations on researchersrsquo opportunitiesto change their research practices for following or generating scientificinnovations

A first important conclusion concerns the role of epistemic traditions forcreating absorptive capacity in scientific communities For scientific innova-tions to be taken up and developed in a country this country not onlyneeds interested researchers but also a basis in the form of prior knowledgeand techniques While larger countries can be expected to have this basisfor most innovations its existence in smaller countries depends on histori-cal developments and path dependencies This is why for many countriesthe opportunity to immediately respond to a scientific innovation is contin-gent on its research traditions

A second conclusion concerns the impact of authority relations onresearchersrsquo opportunities to build protected space If we start from theassumption that any researcher in any position might want to change theirresearch practices in order to either create an innovation or to develop aninnovation that has been created elsewhere we can assess how authorityrelations reduce this space

If building the necessary protected space depends on access to the infra-structure provided by universities the opportunity to create or develop aninnovation is limited to those who control it that is to professors and thosewhose research is approved by professors This dependence is underminedwhere faculties have the right to allocate infrastructure However thismoves a researcherrsquos dependence from individual professors to local organi-sational elites

If building protected space additionally depends on external grants theopportunity to create or develop an innovation is thematically limited toinnovations that are considered worth pursuing by a scientific communityOnly if a scientific community transfers the authority over this particulardecision to the researchers by not making the majority opinion a bindingfoundation for the approval of grants researchers can work against themainstream It was interesting to notice that all investigated communitiesmitigate some of the consequences of their decisions by transferring far-reaching autonomy over the use of grants once they have been awarded

These dependencies prevented some researchers from changing theirresearch practices while others experienced considerable delays It is nowpossible to assess the impact of some of the larger longer-term trends ofchanging authority relations in the public sciences (see Whitley this

volume) The transition from predominantly recurrent funding to a splitsystem in which recurrent funding must be supplemented by competitivegrant funding has ambivalent effects It increases the number of researcherswho can build protected space by including staff below the professoriallevel At the same time it makes all researchers dependent on the authorityof their national and international scientific elites which can cause consid-erable delays and may prevent some innovations altogether Higher educa-tion reforms and the transfers of authority it involves do not affect thelimitation of innovation opportunities to those who control the infrastruc-ture (ie professors) but may create additional opportunities for non-professorial staff by transferring authority over resources to faculties Thefunding of fellowships and temporary research groups is not sufficient forinnovations requiring large protected space but may be beneficial for lessdemanding innovations

Finally the increasing state interest in research serving societal goals didnot play a direct role for the purely basic BEC research However asLaudel and Weyer (this volume) suggest the total protected space for basicresearch might shrink due to funding problems and the increasing incor-poration of state priorities in science policies

1 Telephone interviews were conducted with three of the researchers in theSwedish case2 The research of Swiss condensed matter physicists who work on BEC cannot

be compared to the other cases in terms of authority relations and protected spacewhich is why we didnrsquot include them in Fig 1 and Table 2

ACKNOWLEDGEMENT

We are grateful to Jochen Glaser whose suggestions have greatly improveda previous draft

REFERENCES

Bradshaw A M Cabbibo N Dalgarn A Kienle P Ramsey N Wilkins J amp Eberson

L (1992) International evaluation of Swedish research in physics Report to the research

council Stockholm Sweden Naturvetenskapliga forskningsradet (NFR)

Cambrosio A amp Keating P (1995) Exquisite specificity The monoclonal antibody revolution

New York NY Oxford University Press

Cambrosio A amp Keating P (1998) Monoclonal antibodies From local to extended net-

works In A Thakray (Ed) Private science Biotechnology and the rise of the molecular

science (pp 165181) Philadelphia PA University of Pennsylvania Press

Chubin D E (1976) The conceptualization of scientific specialties Sociological Quarterly 17

448476

Cohen W M amp Levinthal D A (1990) Absorptive capacity A new perspective on learning

and innovation Administrative Science Quarterly 35 128152

Collins H M (1998) The meaning of data Open and closed evidential cultures in the search

for gravitational waves American Journal of Sociology 104 293338

Cornell E A amp Wieman C E (2002) Nobel lecture Bose-Einstein condensation in a dilute

gas the first 70 years and some recent experiments Reviews of Modern Physics 74

875893

Edge D amp Mulkay M J (1976) Astronomy transformed The emergence of radio astronomy

in Britain New York NY Wiley

Fagerberg J amp Verspagen B (2009) Innovation studies The emerging structure of a new

scientific field Research Policy 38 218233

Fujimura J H (1988) The molecular biological bandwagon in cancer research Where social

worlds meet Social Problems 35 261283

Fujimura J H (1992) Crafting science Standardized packages boundary objects and ldquotrans-

lationrdquo In A Pickering (Ed) Science as practice and culture (pp 168211) Chicago

IL The University of Chicago Press

Glaser J amp Laudel G (2009a) Identifying individual research trails In B Larsen amp J Leta

(ISSIrsquo09) Rio de Janeiro Brazil (pp 841845)

Glaser J amp Laudel G (2009b) On interviewing ldquogoodrdquo and ldquobadrdquo experts In A Bogner

B Littig amp W Menz (Eds) Interviewing experts (pp 117138) London Palgrave

MacMillan

Glaser J Laudel G amp Lettkemann E (2014) Hidden in plain sight The impact of everyday

governance on the emergence of fields In M Merz amp P Sormani (Eds) The local con-

figuration of new research fields On regional and national diversity Sociology of the

Sciences Yearbook Dordrecht the Netherlands Springer

Griffin A (2004) The first BEC conference in Levico in 1993 Journal of Physics B Atomic

Molecular and Optical Physics 37 Retrieved from httpiopscienceioporg0953-4075

377E02 Accessed on February 15 2014

Heinze T Heidler R Heiberger R H amp Riebling J (2013) New patterns of scientific

growth How research expanded after the invention of scanning tunneling microscopy

and the discovery of buckminsterfullerenes Journal of the American Society for

Information Science and Technology 64 829843

Heinze T Shapira P Rogers J D amp Senker J M (2009) Organizational and institutional

influences on creativity in scientific research Research Policy 38 610623

Hollingsworth J R (2008) Scientific discoveries An institutionalist and path-dependent per-

spective In C Hannaway (Ed) Biomedicine in the twentieth century Practices policies

and politics (pp 317353) Bethesda MD National Institutes of Health

Ketterle W (2002) Nobel lecture When atoms behave as waves Bose-Einstein condensation

and the atom laser Reviews of Modern Physics 74 11311151

Kleinman D L (1998) Untangling context Understanding a university laboratory in the

commercial world Science Technology amp Human Values 23 285314

Knorr-Cetina K (1995) Laboratory studies The cultural approach to the study of science In

S Jasanoff G E Markle J C Petersen amp T Pinch (Eds) Handbook of science and

technology studies (pp 140166) London Sage

Kuhn T (1962) The structure of scientific revolutions Chicago IL The University of Chicago

Laudel G amp Glaser J (in press) Beyond breakthrough research Epistemic properties of

research and their consequences for research funding Research Policy Retrieved from

httpdxdoiorg101016jrespol201402006

Law J (1973) The development of specialties in science The case of X-ray protein crystallo-

graphy Science Studies 3 275303

Mayntz R amp Schimank U (1998) Linking theory and practice Introduction Research

Policy 27 747755

Mody C M (2004) How probe microscopists became nanotechnologists In D Baird

A Nordmann amp J Schummer (Eds) Discovering the nanoscale (pp 119133)

Amsterdam the Netherlands IOS Press

Mullins N C (1972) The development of a scientific specialty The phage group and the ori-

gins of molecular biology Minerva 10 5182

Mullins N C (1973) The development of specialties in social science The case of ethno-

methodology Social Studies of Science 3 245273

Phillips W D (1998) Laser cooling and trapping of neutral atoms Reviews of Modern

Physics 70 721741

Pickering A (1980) The role of interests in high-energy physics The choice between charm

and colour In K D Knorr R Krohn amp R Whitley (Eds) The social process of scien-

tific investigation (pp 107138) Dordrecht the Netherlands Reidel

Pickering A (1995) The mangle of practice Time agency and science Chicago IL The

Pinch T J (1980) Theoreticians and the production of experimental anomaly The case of

solar neutrinos In K D Knorr R Krohn amp R Whitley (Eds) The social process of

scientific investigation (pp 77106) Dordrecht the Netherlands Reidel

Raasch C Lee V Spaeth S amp Herstatt C (2013) The rise and fall of interdisciplinary

research The case of open source innovation Research Policy 42 11381151

Whitley R (2010) Reconfiguring the public sciences The impact of governance changes

on authority and innovation in public science systems In R Whitley J Glaser amp

HIGHLY ADAPTABLE BUT NOT

INVULNERABLE NECESSARY AND

FACILITATING CONDITIONS FOR

RESEARCH IN EVOLUTIONARY

DEVELOPMENTAL BIOLOGY

Grit Laudel Martin Benninghoff

Eric Lettkemann and Elias Hakansson

ABSTRACT

Evolutionary developmental biology is a highly variable scientific innova-tion because researchers can adapt their involvement in the innovation tothe opportunities provided by their environment On the basis of com-parative case studies in four countries we link epistemic properties ofresearch tasks to three types of necessary protected space and identifythe necessary and facilitating conditions for building them We foundthat the variability of research tasks made contributing to evolutionarydevelopmental biology possible under most sets of authority relationsHowever even the least demanding research depends on its acceptanceas legitimate innovation by the scientific community and of purely basic

ISSN 0733-558Xdoi101108S0733-558X20140000042008

research by state policy and research organisations The latter conditionis shown to become precarious

Keywords Scientific innovation emergence of fields authority rela-tions life sciences evolutionary developmental biology

INTRODUCTION

In this article we look at the development of a scientific innovation that inmany respects epitomises the varied nature of scientific change Differentfrom innovations that emerged from specific experiments or methodologi-cal developments (see eg Laudel et al in this volume on Bose-Einsteincondensation) evolutionary developmental biology (evo-devo) has a muchlonger trajectory with more diffuse roots It can be traced back to the endof the 1970s when it became more and more obvious that neo-Darwiniantheory was unable to account for all empirical findings of evolutionarybiology (Muller 2007) particularly the rapid changes in the forms oforganisms evident from the fossil record and the origins of traits that didnot constitute an adaptation to the environment

It became increasingly obvious that these explanatory deficits of neo-Darwinism were

due to its treatment of development as a lsquoblack boxrsquo and the consequent absence of

the generative rules that relate between genotype and phenotype (Muller 2007

pp 500501)

The discovery of genes regulating embryonic development in the 1980s(HOX genes) and advances in molecular and genomic techniques made itpossible to address specific questions by comparing the development of dif-ferent organisms which led to increased understanding of developmentalmechanisms on the molecular level (Gerson 2007 Raff 2000 Muller2008) This research recently received a further impulse by a breakthroughin the development of sequencing technologies which made the sequencingof whole genomes affordable for single research groups

These conceptual and methodological developments affect a variety offields Responses to the new ideas range from purely conceptual develop-ments to complicated experiments Research in evo-devo can be conductedin many ways and thus can adapt to specific organisational conditions andauthority relations However some features of different approaches to evo-devo make them sensitive to organisational conditions which is why theydo not thrive in all organisations and countries

Evo-devo research can be taken up by researchers from either develop-mental or evolutionary biology fields by integrating the complementaryperspective and can be either theoretical or experimental Experimentalevo-devo research is always comparative but varies in the types of organ-isms that are used and in the ways in which the empirical evidence aboutthe organisms to be compared is acquired Most of the research comparestwo or more species but some scientists compare transgenic organismsbelonging to the same species The organisms studied can be classicalmodel organisms (such as mouse Drosophila or Arabidopsis) or they canbe selected because they are best suited to answer specific evo-devo ques-tions (eg hedgehogs or snakes) The comparisons required by evo-devoresearch can also take a variety of forms Many researchers experimentwith organisms from one species and acquire information about the organ-ism(s) they use for comparisons either from the literature or by collaborat-ing with colleagues who investigate the other organisms Others experimentwith more than one organism Most of these possibilities occur in threemajor fields in which evo-devo perspectives have taken hold namely zool-ogy plant biology and palaeobiology

The resulting combinatorial complexity of intellectual transition pathsto evo-devo and their varied intellectual and resource costs for researchersmake evo-devo both a very interesting and a very challenging case forsociologists for studying the impact of authority relations on the develop-ment of innovations Evo-devo is a very interesting object of study becauseresearchers can adapt their involvement in the innovation to the opportu-nities provided by their environment They can temporarily or permanentlychoose degrees of involvement whose costs match the niches provided forthem by the various interacting organisations At the same time this flex-ibility makes it much more difficult to attribute variations in the develop-ment of evo-devo research to specific authority relations it is not only thatthey enable or prevent evo-devo research but they also enable or preventcertain kinds of evo-devo research

In this article we take up this challenge by identifying the conditionsunder which specific research problems of evo-devo biology can be formu-lated and solved This amounts to formulating a lsquopopulation ecology ofresearch tasksrsquo that in many ways resembles the population ecologyapproach in organisational sociology (Hannan amp Freeman 1977 1989) Atthe same time we follow Hodgson (2013) who resolved the argument aboutselection versus strategic adaptation that has arisen in the context of popu-lation ecology in exactly the same way as it is done by our empirical objectie by applying an evolutionary developmental perspective We combine an

237Highly Adaptable but Not Invulnerable

evolutionary approach that asks which niches enable support or preventspecific kinds of research tasks with a developmental approach that askshow researchers together with other authoritative agencies co-created thesevery niches and adapted their research tasks to them

Analysing the ecology of evo-devo research tasks involves a causal argu-ment consisting of three steps which we present after introducing the theo-retical background and the methodological approach of our research Firstwe analyse the epistemic characteristics of different kinds of evo-devoresearch and identify the kinds of protected space that scientists need forengaging in the various lines of this research Second we identify the neces-sary and facilitating conditions for building these kinds of protected spaceand the authoritative agencies controlling them This analysis enables athird step in which we ascertain how researchers could build protectedspace in different countries

THEORETICAL BACKGROUND

The emergence of fields (or scientific specialties as they were called at thattime) enjoyed much attention in the sociology of science during the late1960s and 1970s1 Inspired by Kuhn analysts attempted to identify rela-tionships between cognitive dynamics and social patterns in the early stagesof development Focusing on the link between the dynamics of ideas andpatterns of interaction the studies of emerging scientific specialties did notpay much attention to the conditions under which the research was con-ducted and the opportunities for changing research practices provided bythese conditions In 1976 Edge and Mulkay produced a synopsis of studiesof emerging specialties that identified three common features of specialtyemergence the growth of the specialties from innovations at the margins ofestablished disciplines the mobility of researchers and ready access of theproponents of the new specialties to graduate students (Edge amp Mulkay1976)

Evo-devo does not though fit these patterns because research in thisarea grows on top of rather than beside existing fields The differentiationpattern observed in the cases studied in the 1960s and 1970s was one oflsquobranchingrsquo according to which new research areas emerged around newsets of problems research technologies or empirical objects (Mulkay1975) By contrast evo-devo adds a new layer of empirical research and anew frame of reference in which experiments can be designed and data

compared This means that moves of researchers from old to emergentfields can be partial and may be reversed

The early studies of emerging fields also had a rather narrow focus inthat they did not link the observed social conditions for emergence to orga-nisational or policy decisions This might have been due to the differentsituations of scientists at the time the fields emerged particularly regardingaccess to research funds and tenured posts Most of the researchers in thesestudies were already on permanent contracts had relatively easy access toresources and were able to move between universities as opportunitiesarose Once established in tenured posts they were much less dependent onauthorities (organisational managers disciplinary elites and external agen-cies) than are their colleagues today

More recent science policy studies that look at the shaping of emergentfields (such as nanotechnology) by governance structures typically take thereverse perspective They look almost exclusively at policy measures chan-nelling resources to emerging fields and trying to influence the directions oftheir development (see Glaser et al 2014 for a critique) Scientific innova-tions that do not attract political attention eg due to lack of potential forapplications and the processes through which governance structures andprocesses affect the development of fields do not tend to be the focus ofthese analyses

Investigating the evolution of evo-devo research links the problemchoices of researchers to the niches created by contrasting governance struc-tures in different countries We consider evo-devo as a scientific innovationwhich we define as a research finding that affects research practices (choicesof problems methods or empirical objects) of a large number of researchersin one or more fields2 Changing research practices incurs costs and mayberisky eg by devaluing knowledge equipment and reputation accumulatedwith previous research Our comparative framework focuses on differencesin authority relations (Whitley 2010) as the key mediating factor connect-ing governance structures to changes of research practices and uses theconcept of protected space (Glaser et al 2014 Whitley this volume) tocompare the opportunities for researchers to change their practices underthe specific conditions created by different sets of authority relations

The state research organisations external funding agencies and scientificelites exercise authority over research choices through three main channelscontrol of resources the allocation of reputation and the provision ofcareer opportunities For these channels the relative authority of each setof actors can be assessed and a framework for the comparative analysis ofeveryday governance of research activities formed accordingly This

framework can be linked to the micro-level of changing research practicesby identifying the authority relations that conditioned how researchersattempted to move to evo-devo research

The impact of particular sets of authority relations on the opportunitiesfor changing research practices towards evo-devo can be ascertained bycomparing the level of protected space that is required for different kindsof changes with the level that can be built by researchers in different sets ofauthority relations We define protected space as the autonomous planninghorizon for which a researcher can apply his or her capabilities to a self-assigned task The two critical dimensions of this variable are the time hori-zon for which the capabilities are at the sole discretion of the researcherand the resources (including personnel over which the researcher hasauthority and the actual time available for research) Researchers createand extend protected space mainly by career decisions (the search forpositions that provide protected space) negotiations with managers of theirresearch organisation and the acquisition of funding Important dimen-sions of this last variable are the likelihood of success in receiving fundingand the speed with which such funding can be obtained

METHODS AND DATA

We use data from the larger comparative project that studied the impact ofchanging authority relations in four countries on conditions for intellectualinnovations that is summarised in the editorial introduction Throughsearches in publication databases (Web of Science Google Scholar) andInternet webpages we detected research groups who had included evo-devointo their research portfolio We also asked our interviewees which otherevo-devo researchers in their national community they were aware of InSweden Switzerland and the Netherlands most evo-devo researchers couldbe interviewed In Germany with its large evo-devo community onlyselected cases could be investigated Researchers who presented evo-devoas a research focus on their websites were selected taking into accountthe variation of disciplines (zoology and plant biology) and of types ofresearch organisations (university and public research institute) Additionalinformation was obtained by interviewing experts in the field and heads offaculties and institutes The distribution of our interviewees is summarisedin Table 1 The table also lists the situations we investigated as casesnamely situations in which researchers successfully or unsuccessfullyattempted to build protected space Not all of our interviewees undertook

these attempts some of them were postdocs whose research was determinedby the decisions of their group leaders Furthermore we consulted self-descriptions of the evo-devo community on the development of the field(eg Carroll 2006 Hall 2012 Laubichler amp Maienschein 2007 Minelli2008 and the authors quoted above)

The interviews with researchers consisted of two main parts In the firstpart the intervieweersquos research that contained the innovation was discussedin the context of the intervieweersquos research projects exploring the continu-ity and all thematic changes and reasons for them Parallel developmentsin the intervieweersquos national and international communities were alsodiscussed The discussion of the content of scientistsrsquo research required thedevelopment of the interviewersrsquo knowledge to an lsquoadvanced laypersonrsquosrsquolevel and the negotiation of a level of communication at the beginning ofthe interview (Laudel amp Glaser 2007 see also Collins amp Evans 2002 onthe level of expertise necessary for competent interaction) Therefore it wasprepared with Internet searches and publications at various levels of diffi-culty (from popular science up to an intervieweersquos publications) were usedInterview preparation also included a bibliometric analysis of the intervie-weersquos publications that enables the identification of thematically linkedpublications A visualisation of this publication network (see Fig 1) wasused to lsquostimulate the recallrsquo (Dempsey 2010) and to prompt narrativesabout the content of research (Glaser amp Laudel 2009)

In a second part of the interview conditions of research and the factorsinfluencing them were discussed Topics included the knowledge personnel

Table 1 Overview of Interviews and Cases

Germany Netherlands Sweden Switzerland

Researchers 7 group leaders

1 PhD student

11 group leaders

2 PhD students

3 group leaders

2 postdocs

7 group leaders

5 postdocs

Other informants 3a 4b

Total number of

interviews

8 16 5 16

Number of transition

situations

12 13 7 9

aOne researcher who gave background information on the evo-devo field one director of an

institute one officer of a funding agencybOne evo-devo background three heads of institutes or deans of faculties

and equipment required to conduct evo-devo research sources of materialsupport and opportunities as well as constraints provided by the intervie-weersquos organisational positions The separation of the discussions of theseconditions of research from the content of research is important because itlimits the extent to which interviewees present their own subjective theoriesabout how current funding conditions made them conduct their currentresearch The interviews lasted 60120 minutes They were recorded andfully transcribed

For the comparison we developed typologies for evo-devo practicestransition situations and protected space In order to deal analytically withthe task of identifying influences of authority relations on intellectual tran-sition patterns we reduced the complexity of these patterns in two subse-quent steps First we used the combinations of variables with the strongestinfluence on transition costs to identify ten common transition situations tostart evo-devo In a second step we allocated these situations to differentcategories on the basis of the level of protected space they required

lsquoTHE SNAKE TOOK US ABOUT THREE YEARSrsquo TRANSITIONS TO EVO-DEVO

In order to identify the impact of authority relations on transitions toevo-devo we must first establish what kind of transitions could take place

Fig 1 Example for a Research Trail of an Evo-Devo Researcher (the circles are

publications the size of the circles indicates the number of citations the lines show

thematic connections between publications)

and what protected space these transitions required The difficulty costand risks of a move to evo-devo depended on the epistemic properties ofthe move which included properties of the research task chosen by aresearcher and the researcherrsquos disciplinary background We derive the pro-tected spaces necessary for a transition in two steps First we identify theproperties of transitions to evo-devo that affected the necessary protectedspace Second we distinguish between large medium and small levels ofprotected space that the transitions required

Properties of Transitions

Four properties of transitions to evo-devo research that affected the neces-sary protected space included (a) the empirical strategy (b) researchersrsquo ori-ginal disciplinary background (c) the types of organisms used forexperiments and (d) the approach to comparisons The combined variationof these properties produces the enormous variance in evo-devo researchtasks and associated necessary protected spaces

Empirical StrategyOne of the most important distinctions for the transition to evo-devoresearch was that between experimental evo-devo research tasks and otherforms of evo-devo research (Table 2) The easiest way to engage in evo-devo research and one that was the entry ticket to evo-devo for many ofour interviewees can be described as conceptual extension Conceptualextension occurred when researchers continued their evolutionary biologyor developmental biology research line including the presentation of their

Table 2 Empirical Strategies Affecting the Protected Space Required forTransitions to Evo-Devo

Empirical Strategies Impact on Dimensions of Protected Space

Resources Time horizon

Conceptual extension None Low

Theoretical research None Low

Bioinformatics research None Low

Experimental research Depends on other properties Depends on other properties

findings to traditional audiences and used evo-devo as an additionaltheoretical framework in which they contextualised their findings Thislsquodual usersquo of experimental research for traditional and evo-devo questionsdid not require any changes in experimental strategies or designs It onlyrequired that the researchers acquire the evo-devo theoretical frameworkand concepts and look for possible comparisons of their findings in a moresystematic manner

A We just searched the literature and found that there were similar things and we dis-

cussed them We [developmental biologists] never decided to go in the evo direction in

fact But the evo dimension is always involved in what we are doing So [hellip] in presen-

tations and seminars it is often discussed but we are not approaching that directly

Q Why didnrsquot you decide to go more into the evo direction

A Because we are quite focused we have all the tools and genetic resources to go on

approaching these questions in the mouse and we have nothing to approach it in

another system We keep contact with these other groups because if they would have

tools in the system to test the hypotheses that comes from all that that would be nice

But I donrsquot think we will do it ourselves You cannot become excellent in everything

We are specialized in mouse3

Evo-devo research could also be theoretical in the sense of purely con-ceptual development ie not linked to experiments at all A third variantof non-experimental evo-devo is mathematical modelling in bioinformaticswhich is conducted either with biological data or by building more abstractmodels For a biomathematician beginning to model evo-devo processesrequires the transfer of mathematical skills to new questions

The empirical strategy chosen affected mostly the resource dimension ofprotected space All three non-experimental strategies required relativelyfew resources while the resource demand of experimental evo-devo can beconsiderable The learning time involved in the non-experimental forms ofevo-devo research did not substantially delay these research processes andthus did not require a specific time horizon of protected space Howeverresearchers who just added evo-devo perspectives to their experimentalwork could encounter difficulties when they tried to publish outside theirdisciplinary journals attempted to publish controversial evo-devo conceptsor had a strong evo-devo focus as in the following example

One thing I saw is that twice I had papers containing more evo-devo which were not

considered at all for the main journal of molecular evolution Molecular Biology and

Evolution which is the best journal of molecular evolution And two of our articles which were then published in good journals moreover have not been accepted in

this journal They thought that it was not about molecular evolution because there was

a strong developmental axis which is quite strange from my point of view

Thus adding evo-devo perspectives could affect the time horizon ofprotected space through delays in publication Overall however onlyexperimental evo-devo research required significant levels of protectedspace which depended on the three variables that are now discussed

Disciplinary Background of ResearchersAmong our interviewees who conducted dedicated evo-devo experimentsthose who came from evolutionary biology had to modify their research toa larger extent than their colleagues from developmental biology (Table 3)Evolutionary biologists who wanted to move to evo-devo research had tolearn molecular genetic techniques and embryological methods which mostdevelopmental biologists already knew and only had to adapt to the newevo-devo questions

You know normally as an evolutionary biologist we usually receive a piece of tissue in

a tube and thatrsquos all what you see from the animal [hellip] And then we also started to

learn some techniques so we have a collaborator [hellip] [in the UK] And he is a purely

developmental biologist who is also interested in Evo-Devo and he helped a lot to

learn things so I have been to his lab about three times or four to learn new techni-

ques So I went once to learn basic histology another time for In situs and another

time for skin culture So that has been a great help

You need to know how to breed this organism You must be fully familiar with it that

is you need to know how it develops Not just in molecular terms but by observing it

under the microscope So what happens during the development You need to know

its whole development Then you need to influence this development molecular geneti-

cally and you need to know these methods Thatrsquos what mainly happens in the lab

Table 3 Disciplinary Background Affecting the Protected SpaceRequired for Transitions to Evo-Devo

Disciplinary Background of Researchers

Moving to Evo-Devo

Impact on Dimensions of Protected

Evolutionary biology None Medium

Developmental biology None None

The epistemic background could have a significant effect on the amountof newly acquired knowledge required to engage in evo-devo Evolutionarybiologists needed more time to establish the methods necessary for experi-mental evo-devo and thus a longer time horizon of protected space

Type of Organisms Used in ExperimentsThe choice of organisms to compare had a much stronger impact on neces-sary protected space than the general disciplinary background of research-ers and associated research practices (Table 4) Experimental biologistsgenerally prefer to work with a few organisms about which much knowl-edge has been accumulated over the last decades and with which methodsare known to work well However these so-called model organisms arenot necessarily best suited for answering evo-devo questions not the leastbecause the evolutionary branch on which they are located constrains thechoice of organisms they can be compared to This is why it is often attrac-tive for evo-devo researchers to work with entirely different organismsHowever including new organisms in experimental research often incurshigh costs because little is known about them in the beginning whichmakes breeding them and conducting experiments with them more difficultand often more risky than the work with model organisms4 Comparedto the organisms which are well known to researchers and were chosenbecause they are easy to breed and to use the organisms best suited forevo-devo research often require more time

For some methods [the transfer] worked straight away others needed five years In

some cases we had to spend ten years to transfer each method from Caenorhabditis

[model organism] to [our evo-devo organism]

Itrsquos also this with evo-devo you have to adapt to completely different constraints in

new sorts of species So now we apply ultrasound to our creatures Here we are we

had to use ultrasound people never apply ultrasound to their mice So we do

Table 4 Type of Organisms Affecting the Protected Space Required forTransitions to Evo-Devo

Type of Organisms Used in Experiments Impact on Dimensions of Protected Space

Model organisms None None

Non-model organisms high high

ultrasound on our hedgehogs and on [another non-model organism] to see at which

stage embryos are to be able to take them Only the breeding of these animals that

took us two years to master the breeding

A And for that we had to develop these three techniques

Q Was this complicated to set up and letrsquos say expensive

A Yeah it was complicated because in fact you do it through detours Because it is a

technique that worked very well in the fly but does not in vertebrates for rather compli-

cated reasons So we had to introduce a kind of molecular scissors in [our vertebrate]

tinkering

Evo-devo researchers who did not work with classical model organismsoccasionally had difficulties when attempting to publish their results Theyfaced criticism because they did not meet the standards of work that haddeveloped for model organisms The new organisms could not be manipu-lated the same way as model organisms because the knowledge and toolshad not been developed yet As a consequence the functional tests thatwere expected in the field and particularly from reviewers could not beconducted with these organisms

I And what was the problem in the view of the [journal] reviewers

The problem that came back to us was that we lacked the ability hellip We were showing

this one gene a gene that is very famous for all sorts of reasons to be involved in the

very early development of the trait that we were looking at And the comments that

prevented it to making it to higher journals were that we were lacking what we call

functional tests in developmental biology which means the ability in gene in a develop-

ing system to see what does it do to the system when you manipulate it We could see

that it was expressed we could see that it was there before this or that we could see

that it was expressed not in lineages versus not others What we could not do was

manipulate it on the wing to ask the question okay when I mess it up what happens

Yes in fact we have difficulties for some of our evo-devo work If you go to high profile

journals they would come back and say why didnrsquot you do a functional study Why

havenrsquot you studied how the genes are actually working rather than just seeing if they

are expressed And the answer is you cannot but it is still very interesting to find out

what you can find out about these organisms [hellip] And that is not always very much

because they are very difficult animals to work on there are all sorts of technical

problems that stop you from doing this very high end technical type of things But the

really big developmental journals they want technical high end papers in general So

they question some of these things we send in because they do not work with compara-

tive things they work on model organisms They do not see the enormous challenges

that it presents You just cannot do it it is the short answer

Q In terms of publications and related to your switch to Evo-Devo was it easier or

more difficult to publish

A No much more difficult [hellip] It takes place in a more restrictive setting as it is

more multidisciplinary where we must satisfy both developmental geneticists and evo-

lutionary biologists This is a problem to publish the evo-devo [work] because journals

are either one or the other typically So often we see immediately lsquoOh yes thatrsquos a devel-

opmental geneticist [reviewer] he wants functional in every wayrsquo He does not realize that

we are working on something other than the mouse He asks infeasible things Or he

asks or he criticizes disregarding the fact that we do not work with the mouse We expect

to be told lsquowow itrsquos still fantastic what you did with your animals So it is not easyhellip

Working with new organisms had a strong impact on both the resourceand the time dimension of protected space In particular the time dimen-sion often became unpredictable due to the necessity of establishing meth-ods for the new organisms and due to delays in publishing Only a fewspecial evo-devo journals exist so far (McCain 2010) Although they havean interdisciplinary focus they have a lower reputation than the establisheddisciplinary journals This could hinder the advancement of careers andgrant applications At the same time the epistemic rewards of this strategyare likely to be higher due to the organismsrsquo suitability for evo-devo ques-tions Introducing new evo-devo-specific model organisms in a laboratoryis the lsquohigh risk high rewardrsquo strategy of evo-devo It requires a larger pro-tected space in both dimensions but is also the potentially most rewardingin terms of scientific yield5

Approach to ComparisonsAnother decision of researchers that considerably affected the protectedspace needed for transitions to evo-devo research concerned theways in which comparisons between species were conducted (Table 5)

Table 5 Approach to Comparisons Affecting the Protected SpaceRequired for Transitions to Evo-Devo

Approach to Comparisons Impact on Dimensions of

Protected Space

Comparative

experiments

Organisms from one species None None

Organisms from similar species Low Low

Organisms from dissimilar species High High

Experiments

compared to data

from the literature None None

from collaborators Low-medium Low-medium

Comparisons could either be achieved by carrying out comparativeexperiments with different organisms in the researcherrsquos laboratory or byexperimenting with just one type of organism and using externally pro-duced data for comparisons The most difficult and thus most resource-intensive and time-consuming way to conduct comparative evo-devoexperiments is working with two different organisms in one laboratoryIn most cases this approach to comparative research is more risky andcostly because researchers have to establish (introduce breed and under-stand) a second organism in their laboratories (non-evo-devo research ismostly conducted with only one organism) The advantage of compara-tive experiments is that the experiments are designed within the sameconceptual and methodological framework which guarantees the best fitof data

One thing is the feeling for the organism in biology you have to learn you have to

know the local effects It takes a while before you get to know a [certain plant species]

the things that you can do and the things that you cannot do what is normal what is

abnormal [hellip] like diseases in the greenhouse You have to know it looks like that you

have to treat like that You donrsquot ever see anything about it in the paper but if they all die

because of some disease that takes so long to learn that At some point Arabidopsis

entered on the lab floor people hated it because they were not used to it You have to

get used to it and know what you can do with it and what you should not do

There are several ways to reduce these investments One approach wasto conduct comparative experiments but to use only one species fromwhich different transgenic individuals are constructed Since only one spe-cies is used both breeding the organisms and the application of methodsare less risky and costly because of the overall similarity of the organismsinvolved Our sample included one plant biologist and one zoologist whoused this approach Another way of reducing risks and costs is producingthe comparative information internally but using species that are very simi-lar (eg two species of fish) Again breeding and using the second speciesrequires less learning because of the similarities and the experimentalmethods are more likely to be easily transferred

Costs of comparative experimental research can also be reduced wheninformation about the second organism is acquired from external sourcesrather than internally produced by experimenting with several organisms inone lab This can be done through collaboration with other evo-devoresearchers who are specialised in the other organisms An even easier wayis to include published data on the organisms or to use bioinformatics data-bases However both collaboration and reliance on literature or databasesreduce a researcherrsquos control of the experimental approach and the data

that can be used The data that are accessible this way may not fit the speci-fic evo-devo question thereby limiting the potential epistemic rewards ofthese cost-effective strategies

Q But where exactly comes the evolutionary part in

A It is when there were enough sequences from different species across the whole range

to compare them

Q How do you do this You have Arabidopsis and you have [another plant species]

A You look up the databases You have your gene that you are working with and put

it against the whole database The computer screens it compares it one by one This

one matches 80 So you compare the [other plant species] gene to the Arabidopsis

gene hellip Same evolutionary origin 80 identical

Q You used Drosophila then as a second model organism for comparison [hellip]

A Well it was more at the theoretical level because for obvious reasons there is much

much more work done in Drosophila than not only in insects but in many other organ-

isms probably So it was my comparison insect I was working on a gene called [hellip] and

most of the work has been done in Drosophila Of course this was my [hellip] reference

point at least from the literature comparison perspective

This discussion of the four major variables affecting the level of pro-tected space required for different transitions to evo-devo research demon-strated the complexity of research strategies and the large lsquopopulationrsquo ofpossible evo-devo research tasks Each of the tasks within this populationrequired specific protected spaces which in turn implied a specific fit withthe lsquoauthority landscapersquo in which researchers made decisions on tasks andapproaches In order to enable comparisons of authority relations and theirimpact on the selection of evo-devo research tasks we now reduce thecomplexity by identifying three levels of protected spaces that can be linkedto particular sets of authority relations

Three Levels of Required Protected Space

The impact of these four variables on the resources and time horizondimensions of required protected space can be integrated by distinguishingthree levels of necessary protected space which are linked to the degree towhich the specific evo-devo research deviates from established researchpractices as summarised in Table 6

1) Large protected space was needed when a move to evo-devo includedproducing the empirical evidence by comparative experiments with twoor more dissimilar species In exceptional cases experiments with onlyone organism could also require large protected space eg when a non-model organism was established by an evolutionary biologist whomoved to evo-devo research and who had to establish a molecular biologylab The other exception was a researcher who experimented with onlyone organism and acquired the data about other organisms from colla-borators He compared so many organisms that he had to build a ratherlarge network of collaborators which in the end was as time-consumingas establishing a second organism in onersquos own lab

[hellip] there were two problems to do experiments in comparative development One was

to get the embryos from all these different species It has taken me years to build up a

big network of collaborators So I can now get embryos from very rare species quite

easily [hellip]

With the [hellip] limbs we were working for seven years Collecting the species [hellip] clon-

ing the genes waiting for a new species to come in Comparative studies take a long

long long time

Researchers who kept several different species in their laboratoriesneeded breeding facilities (eg animal houses) for these different speciesincluding technical support Manipulating additional and particularlynon-model organisms in the lab also required a longer time horizonMolecular genetic tools needed to be adapted to non-model organisms

Table 6 Degrees of Protected Space Required by Different Types ofEvo-Devo Research

Degree of Protected Space (in

Resource and Time Dimensions)

Types of Transitions to Evo-Devo Research

Large Setting up comparative experiments with dissimilar species

Setting up experiments with one non-model organism

when coming from an evolutionary biology background

Organising large collaborative networks for comparisons

Medium Setting up comparative experiments with organisms from

one species or from two similar species

Small Adding evo-devo perspectives to traditional experiments

Theoretical or bioinformatics research

particular those that were unique even to the evo-devo community Dueto the lack of genetic tools the work with certain species is technicallyvery difficult and hence very time consuming Obtaining sufficientembryos from certain species required a high amount of time becausemating does not occur every year

2) Medium protected space was required for evo-devo research thatexperimented with only one species and acquired the empirical evidenceabout the other(s) through collaborations from the literature or fromdatabases of gene sequences The necessary protected space was also med-ium for evolutionary biologists who added only a small lsquodevo partrsquo to theirwork eg by investigating the function of single genes Another approachto limiting the costs of comparisons was conducting within-species com-parisons which required the applicationdevelopment of genetic tools foronly one species or working with species that were similar enough not toincur additional costs for breeding and adapting genetic techniques

Researchers who worked on only one organism nevertheless neededtime to adapt genetic tools for manipulation if this organism was a non-model organism This was even impossible in some cases which led topublication difficulties This is why the necessary protected space is med-ium in both the resource and time dimensions

3) Small protected space was required by all evo-devo research that wasnon-experimental either because it just added the evo-devo conceptualperspective to the interpretation of traditional results or because it wastheoretical or bioinformatics research The transition to evo-devo couldbe achieved by continuing onersquos work largely unchanged and addition-ally framing experimental results in an evo-devo context The protectedspace required by the experiments was necessary for the non-evo-devoresearch ie for common experimental research in the field The actualtransition to evo-devo did not require additional resources or timePublishing results could also be continued in the researcherrsquos main fieldThe necessary protected space was also low for all researchers whodeveloped theoretical concepts in evo-devo (eg certain hypotheses) orused bioinformatics data for answering evo-devo questions In bothcases costly experiments are unnecessary

THE IMPACT OF AUTHORITY RELATIONS ON PATHS

TO EVO-DEVO

With our empirical basis we could test numerous situations of evo-devotransitions (requiring low medium or high levels of protected space) in

four national constellations of authority relations This enables us to deter-mine necessary and facilitating conditions for evo-devo research and howthey are met Further we can specify who is actually able to build the neces-sary protected space in the four countries and why

Ecological Niches for Evo-Devo Research Provided by DifferentSets of Authority Relations

Having identified three types of protected space that can be linked to mostof the research tasks formulated by evo-devo researchers we can now askhow different sets of authority relations created niches for researchers inwhich these protected spaces could be built We do this by identifyingnecessary and facilitating conditions for building protected space and theauthoritative agencies involved in creating these conditions (Table 7)

The most general condition for evo-devo research to be possible at all isthat scientific communities accept the outcomes of evo-devo research ascontributions to scientific knowledge As we have seen in the discussion ofpublication problems this condition is not always fulfilled because evo-devoexperiments can have difficulties in meeting the standards set by researcherson traditional model organisms These affected researchersrsquo choices parti-cularly those of researchers who still had to advance their careers

For a transition to evo-devo to be possible in any national sciencesystem this type of research had to be deemed worth supporting by theauthoritative agencies that provide positions and resources to researchersSome properties of evo-devo research turned out to be particularlyimportant

Evo-devo research is basic research for which applications (be they med-ical or agricultural) cannot be convincingly promised

The research on which evo-devo conceptual extensions (evo-devo per-spectives) build also is basic research with no applications in sight

Evo-devo research is non-mainstream interdisciplinary research thatrisks less recognition and delayed publication of results

These properties may be at odds with political or managerial preferencesfor contributions to societal ends and with national and organisational eva-luation systems This is why the actual fit of basic research in general andevo-devo research in particular with the interests of research organisations(universities and public research institutes) is a necessary condition for allthree types of evo-devo research Therefore the three conditions are neces-sary for all evo-devo research to be undertaken are (1) evo-devo results

Table 7 Necessary and Facilitating Conditions for Building Protected Space for Evo-Devo Research

Conditions for Evo-Devo

Research

Protected Space

Large Medium Small

Resources for breeding

facilities and large groups

long time horizons

Resources for evo-devo

experimental projects sometimes

unpredictable time horizons

Resources for theoretical or

experimental research

Acceptance by scientific

communities of evo-devo results

as scientific contributions

Necessary condition Necessary condition Necessary condition

Acceptance of basic research by

the research organisation

Acceptance of evo-devo research

by the research organisation

Project funding for basic

research

Necessary condition Necessary condition Necessary for evo-devo conceptual

extension of experimental research

facilitating for other

Control of infrastructure Necessary condition Necessary condition Necessary for evo-devo conceptual

Above-average investment in

infrastructure

Necessary condition Facilitating condition Indifferent

Tenured position Necessary condition Facilitating condition Indifferent

Project funding for

interdisciplinary collaborations

Facilitating condition Facilitating condition Indifferent

LAUDELETAL

must be accepted by the scientific communities as scientific contribution(2) Basic research must be accepted by the research organisation (3) evo-devo research must be accepted by the research organisation as sum-marised in Table 7

Beyond these three necessary conditions other conditions vary in theirimplications for developing different levels of protected space Project fund-ing for basic research is a necessary condition for most experimentalresearch making theoretical evo-devo research and the research workingwith databases the only forms of evo-devo research that can be undertakenwithout such project funding Experimental research in the biosciencesadditionally requires control of infrastructure in the form of a laboratoryand the basic equipment that comes with it The crucial step of experimen-tal evo-devo research is establishing the different species as experimentalobjects This includes building the infrastructure for breeding the speciesand developing the molecular genetic techniques that are necessary tomodify the organisms In other words an evo-devo laboratory has to beestablished

This infrastructure is commonly provided by universities Providing highlevels of protected space involves above-average investments in infrastruc-ture A dedicated evo-devo-lab is large and often depends on expensivebreeding facilities for more than one species Research organisations mustbe able and willing to make these investments

The medium-sized protected spaces for evo-devo research do not dependon above-average investments However since they are linked to dedicatedevo-devo research questions additional investments might be necessary ifnew species are to be bred Such evo-devo research will be facilitated byinvestments in the additional infrastructure requirements

The long and more importantly often unpredictable time horizonsof large and medium-sized protected space for evo-devo research made atenured position important for most researchers While they were necessaryfor research requiring large protected space we did observe transitionsto evo-devo research that required medium-sized protected space byresearchers in fixed-term posts Some laboratories specifically recruitedpostdocs with evo-devo experience in order to expand this line of researchHowever the time horizon provided by fixed-term positions was not alwayssufficient One researcher postponed a more lsquohard-corersquo approach ofevo-devo because the labour-intensive developmental genetics work wouldhave extended beyond the duration of his position Finally researchersrelying on collaborations for the comparisons between species benefitedfrom project support for interdisciplinary collaborations

Who Could Build Protected Space in the Four Countries

Having identified the necessary and facilitating conditions for buildingprotected space and the authoritative agencies controlling these conditionswe can now ascertain how protected space could be built in the four coun-tries The discussion of conditions in the previous section already indicatedthat opportunities strongly depended on researchersrsquo career positionsand that there may be functionally equivalent ways of providing the otherconditions This is why the following discussion is focused on the threetypes of protected space and the ways in which conditions for buildingthese spaces in the different countries are compared

The building of large protected spaces was the most demanding taskbecause it depended on simultaneous and sometimes coordinated supportof all authoritative agencies Acceptance of evo-devo research by scienti-fic communities is a condition on the international level which existedfor all of our four countries to the same extent In the mid-90ies sup-port for basic research and particularly of the kind of basic researchrepresented by evo-devo was available and led to initiatives for evo-devo to be established in two of the smaller countries Sweden and theNetherlands The third small country Switzerland had a strong researchtradition in the relevant biosciences and has actively contributed to thedevelopment of evo-devo while the larger German science system alsohosted a strong tradition in these fields Support for evo-devo includedproject funding for the basic research on which evo-devo builds andfor evo-devo research itself In all four countries this depended onnational and international funding agencies providing project fundingfor basic research ie on the research councils6 which represented stateinterests

I had from the start the opportunity to acquire external grants And it always worked

It is not particularly easy in the evo-devo area because it is not medical applied

research It is always easier to get money for those areas because there are predefined

research priorities for which you can apply directly We do not fit in these We always

have to apply for programmes that provide complete freedom content-wise That is tra-

ditionally the DFG Individual Grants Programme This always worked if it was a good

evo-devo project (German researcher)

We found a few exceptions from the need for gaining grants In Germanstate-funded non-university research institutes and at one Swiss universityresearchers had sufficient recurrent funding to immediately start evo-devoresearch without being dependent on external grants

The fate of evo-devo in the Netherlands demonstrates however thatthe support for evo-devo research by the state research organisations andfunding agencies cannot be taken for granted anymore The state hasrecently formulated strong political expectations concerning the utility ofuniversity research and has specified these conditions by defining priorityfields Universities are expected to include these fields in their researchprofiles and are financially rewarded for doing so Funding agencies areconfronted by similar expectations and follow them

The money for evo-devo has dried up and it is really difficult to get funding unless it

has an applied aspect So I moved a lot more recently into hellip [a biomedical direction]

So I still do a bit of evo-devo [hellip] And we can use it as an assay But pure evo-devo is

kind of on the way out I think Because it is fundamental research and all the funding

is shifting well a part of funding is shifting to clearly applied research (Dutch

researcher)

Dutch universities responded to state expectations by discouraging thekind of basic research represented by evo-devo and the fields on whichit builds and increasing their demands for research that is connected toapplications Biologists in the Netherlands are increasingly expected toshow potential biomedical or agricultural applications an expectation thatusually cannot be met by evo-devo researchers Universities ceased to investin groups that cannot meet these expectations One evolutionary biologygroup was closed and several biology chairs were re-dedicated towardsmore application-oriented research

Since publications in high-impact journals were an important promotioncriteria and a requirement for grant applications one researcher movedaway from evo-devo research

I tried writing a [grant] two years ago for the evo-devo stuff And I didnrsquot get [it] And

probably next year I will write a [grant] about [hellip] development simply because we

have now two publications in the pipeline with this group and they will go to big jour-

nals and they will be cited Simply your chances are much bigger And career-wise

being on a tenure track and people looking at your H-index and whatever it is these

days you feel like you have to make those kind of choices rather than risking doing it

again and getting rejected again (Dutch researcher)

The worsening climate for basic research let several evo-devo researchersleave and taking up offers abroad As a consequence of these develop-ments evo-devo research in NL has almost disappeared (see Laudel ampWeyer this volume)

The control of infrastructure that was necessary for building largeprotected space was limited to professors or in Germany directors at

state-funded public research institutes Professors could build evo-devolabs from start-up packages they received with their appointments InSwitzerland Sweden and the Netherlands evo-devo researchers wereactively recruited with the offer of establishing an evo-devo lab Howeverhigh levels of protected space depended on above-average investments Inour sample only Swiss universities were affluent enough to support localresearch interests that required such investments (see Benninghoff et althis volume) The strong research tradition in developmental biology andin evo-devo (Swiss researchers contributed to some of the early discov-eries that set the field in motion) created enough interest in some universi-ties to make them invest in the required infrastructure In Germanyinvestments of this kind were made in state-funded public research insti-tutes rather than universities Although the biosciences are strong enoughfor interest in evo-devo to emerge in many German universities the lat-terrsquos capacity for above-average investments is limited (on the comparisonof universities and public research institutes see Glaser et al in thisvolume) The case of Sweden demonstrates that exceptionally large grantscould provide a functional equivalent to university funding in infrastruc-ture One of the Swedish funding agencies intended to promote evo-devoresearch

The VR [Research Council] had decided to push for funding for a professorship in evo-

lutionary biology at the border between traditional and modern molecular evolutionary

biology And that had then been set for some sort of competition among universities in

Sweden and it was Uppsala who had won it And [hellip] there were all sorts of different

candidates with different profiles that were interviewed But several [hellip] had probably

some sort evo-devo perspective or at least some kind of combination of historical data

with molecular data (Swedish researcher)

In this case the majority of the funding came from the grant with somematching funds contributed by the university

The long or at least unpredictable time horizons of large protected spacefor evo-devo research made tenured positions important Since these had tobe combined with the control of infrastructure most researchers who builtlarge protected spaces had professorial posts and headed evo-devo labora-tories Only one Swiss and one German researcher on temporary positionsattempted evo-devo research that required large protected space The Swissresearcher could utilise the large protected space created by the professorbut within the limits of the lab (particularly the animals established there)The German researcher worked as a group leader at a state-funded institutein which evo-devo research was firmly established and which provided

excellent facilities as well as extraordinary technical support He had discre-tion over several PhD students and divided his research into a risky evo-devo line requiring large protected space and a low-risk research line thatsecured publishable results

Medium levels of protected space could be built without above-averageinvestments in infrastructure This was possible in all four countries forresearchers who could establish an evo-devo lab and had access to projectgrants Tenured positions were an advantage due to the sometimesunpredictable time horizons which is reflected by the fact that Dutchresearchers on tenured positions below the professorial level built medium-sized protected space Researchers on fixed-term positions could buildmedium-sized protected space when working in an evo-devo lab but facedthe restrictions described above

Low levels of protected space were easiest to build In the case ofconceptual extension however it was small only for the evo-devo part ofthe research while the foundations of disciplinary experimental research inthe biosciences still had to be provided For these small protected spaces tobe built basic research had to be tolerated by universities and had to beable to attract some grant funding

The low threshold for entering some forms of evo-devo research and theopportunities to develop it gradually without having to give up previouslines of research suggest that some kind of evo-devo research is possibleeverywhere While this is true to a limited extent career expectationsof prospective evo-devo researchers affected their decisions For youngresearchers the opportunity to build a career with evo-devo research wasvery important Evo-devo research or the basic research it builds on had tobe able to attract funding and results had to be published For researchersstill having to meet criteria for tenure avoiding evo-devo research was thesafest option which was chosen by two Dutch researchers

The existence of large evo-devo labs and the many opportunities forlsquolow-threshold transitionsrsquo made the move to evo-devo research possiblefor researchers at all career stages Laboratories whose leaders were inter-ested in evo-devo without wanting to move to it themselves offered oppor-tunities for postdocs to begin evo-devo research The dedicated evo-devolaboratories also provided the necessary protected space for youngerresearchers who in many cases specifically sought employment in an evo-devo lab in order to move into that field

Under these conditions it was quite common for researchers whobecame interested in evo-devo to start with topics that required small pro-tected space (eg theoretical evo-devo or adding evo-devo perspectives to

traditional experiments) Researchers thus could explore the potential ofevo-devo ideas as well as publication opportunities without risking toomuch because their experiments still lsquocountedrsquo in the traditional contextIf things went wrong it was only the lsquoadd-onrsquo that failed while the basisof experimental research still provided publication opportunities andaccess to grants If they were successful they moved to evo-devo thatrequired medium-sized or large protected space One developmental biolo-gist had started with theoretical evo-devo work He first added to his clas-sical developmental experimental research conceptual considerations onevo-devo Then he moved to do theoretical evo-devo research using litera-ture data Finally he set up an evo-devo lab and started to do evo-devoexperiments

A I switched really from classical evo-devo to much more molecular stuff Now Irsquom

looking at the old theoretical papers I did on evo-devo and now say letrsquos re-address

this work with molecular tools

Q And classical would be just looking at the morphology

A Morphology All the literature historical aspects theoretical aspects Now it is

much more experimental molecular

An evolutionary biologist began his evo-devo research in the early1990s ie at a time when the methodological development of evo-devomethods were still at its early stages He published an article that waspurely theoretical

A [hellip] That was the argument I put forward in this article without any empirical evi-

dence whatsoever looking at the morphology [hellip] It also reveals so to speak the gap

between conceptual insight and technical ability I mean it is possible to write some-

thing like that as soon as you have the idea but to actually do something [empirically]

took time

This work required only very small protected space The transition toevo-devo benefited from collaborations which supported the researcherrsquoslearning The researcher later acquired his own evo-devo lab and eventuallybuilt large protected space for evo-devo research

The specific nature of evo-devo research as a scientific innovation thatnot only leads to new experiments but also combines and overlays tradi-tional experimental research made it possible and often easy to add evo-devo to already existing lines of research and to move to the new researchonly partially This is trivial for all cases in which traditional experimentalresults were just interpreted in evo-devo theoretical contexts but alsoapplied to those researchers who built medium-sized and large protected

spaces Only one Swiss two Dutch and one German researcher of thosethat we interviewed changed their research completely and conducted onlyevo-devo research in the end

CONCLUSIONS

If we stay in the vocabulary of our biologists evo-devo research has provento be a highly adaptable species The variability of research tasks contribut-ing to the progress of the new field made the field as a whole highlyadaptable to variations in research traditions and authority relationsIn spite of this variability however both individual strategies for buildingprotected space and the development of evo-devo at national and interna-tional levels depended on particular sets of authority relations Generalconclusions about this dependency can be drawn as follows

The most fundamental condition is the acceptance by scientific commu-nities of the legitimacy of the innovation To be a researcher means produ-cing contributions to the knowledge of onersquos scientific community whichhas exclusive authority over this definition This authority has beenincreased by changes in the governance of science because new evaluationprocedures also channel it through grant funding organisational evalua-tions and individual evaluations for recruitment and promotion purposes

An important condition at the national level is the acceptance of researchof the evo-devo type purely basic non-mainstream low-impact bynational authoritative agencies These include the state national elitesfunding agencies research organisations and organisational elites TheDutch case shows how the new public management reforms and the increas-ing incorporation of public policy goals in science policies can lead to asituation where all these authoritative agencies become lsquoalignedrsquo andresearch that does not fit state expectations be crowded out This appliesto both evo-devo research itself and the fields of evolutionary and develop-mental biology it depends upon

Third while some forms of evo-devo research might be possible evenunder the worst conditions evo-devo as a field can only progress if medium-sized and large levels of protected space can also be built Researchersusually plan some years ahead when formulating research tasks whichmeans that it must appear possible to them to make a career with evo-devowhich in turn requires moving from positions in which small protectedspaces can be built to those that enable the building of larger ones

Fourth the large protected spaces required by lsquotruersquo evo-devo research experimental comparisons of unusual organisms within one lab highlightedyet another necessary feature of authority relations Large protected spacecan only be created if authoritative agencies set priorities either becausethey want evo-devo research or because they want specific researchers andaccept that they have to create the infrastructure in order to attract theresearchers However not all research organisations appear to be able tosupport these investments German public research institutes Swiss univer-sities and a Swedish funding agency were able to create these niches as wasa lsquojoint venturersquo of a Dutch funding agency and a university The fact thatthese particular sets of authority relations all deviate from those commonlyavailable at the contemporary average European university indicates thatthe general shift from block funding for research to external project grantfunding might have overshot its target and now unduly limits the authorityof universities

Finally the fate of evo-devo research in the Netherlands points to thedanger of an international homogenisation of authority relations Thechanges in authority relations that endangered evo-devo research in theNetherlands occur in many countries If the trend towards a homogenisa-tion of state interests thematic priorities for research and intrusive shortterm governance instruments persists research of the lsquoevo-devo-typersquo mightbe crowded out in the future Scientific research appear to thrive under con-ditions of institutional diversity because research tasks will be formulatedwhere the necessary protected space can be built It doesnrsquot matter to inter-national scientific communities which country provides the conditions Ifhowever no country at all provides opportunities to build specific pro-tected spaces some kinds of research might disappear or at least sloweddown considerably

1 See Mullins (1972) on the Phage group and the genesis of molecular biologyEdge and Mulkay (1976) on the emergence of radio astronomy in BritainBen-David and Collins (1966) on Psychology Dolby (1976) on physical chemistryLaw (1976) on X-ray Protein Crystallography Mullins (1973) on ethnomethodologyFisher (19661967) on the theory of invariants Chubin (1976) published a criticalreview of the research on the emergence of scientific specialties Two anomalous fieldshave also been studied The field of lsquoN-raysrsquo died relatively quickly (Nye 1980) whilethe cold fusion lived on for quite some time (Lewenstein 1992 1995 Simon 1999)2 Thus we do not apply the concept lsquoscientific innovationrsquo to all findings that

are accepted and adopted by other researchers (as does Knorr-Cetina 1981 p 66)

but limit it to findings that have implications for a whole scientific community suchas Lynch and Jordanrsquos example of a biological innovation (the polymerase chainreaction) shows (Lynch amp Jordan 2000)3 Interview quotes in German Swedish and French were translated by us4 In some cases these difficulties motivated the selection of species similar to

those of established model organisms which made breeding and manipulating themeasier and increased the likelihood that methods could be successfully transferred5 In recent years evo-devo has been developing its own model organisms Thus

analytical toolkits become available to manipulate those organisms (Sommer2009) However most of our interviewees didnrsquot have this advantage yet when theymoved to evo-devo research Therefore the time period of a researcherrsquos transitionto evo-devo is important in terms of transition costs6 The Deutsche Forschungsgemeinschaft (DFG) in Germany The Schweizerische

Nationalfonds (SNF) in Switzerland the Nederlandse Organisatie voor WetenschappelijkOnderzoek (NWO) in the Netherlands and the Vetenskapsradet (VR) in Sweden

ACKNOWLEDGEMENTS

We would like to thank Jochen Glaser for invaluable suggestions how tocope with the unusual complexity of this comparative case study andRichard Whitley for helping us to clarify the argument and improving thereadability of this article

REFERENCES

Ben-David J amp Collins R (1966) Social factors in the origins of a new science The case of

psychology American Sociological Review 31 451465

Carroll S B (2006) Endless forms most beautiful The new science of Evo Devo and the making

of the animal kingdom New York NY Norton

448476

Collins H M amp Evans R (2002) The third wave of science studies Studies of expertise and

experience Social Studies of Science 32 235296

Dempsey N P (2010) Stimulated recall interviews in ethnography Qualitative Sociology 33

349367

Dolby R G A (1976) The case of physical chemistry In G Lemaine R Macleod

M J Mulkay amp P Weingart (Eds) Perspectives on the emergence of scientific disci-

plines (pp 6373) The Hague Mouton

Fisher C S (19661967) The death of a mathematical theory A study in the sociology of

knowledge Archive for History of Exact Sciences 3 137159

Gerson E M (2007) The juncture of evolutionary and developmental biology In M D

Laubichler amp J Maienschein (Eds) Embryology to evo-devo A history of developmental

evolution (pp 435463) Cambridge MA MIT Press

Glaser J amp Laudel G (2009) Identifying individual research trails (pp 841845) In

B Larsen amp J Leta (Eds) Proceedings of the 12th International Society for

Scientometrics and Informetrics (ISSIrsquo09) Rio de Janeiro

governance on the emergence of fields Forthcoming In M Merz amp P Sormani (Eds)

The local configuration of new research fields On Regional and national diversity

Sociology of the sciences yearbook Dordrecht Springer

Hall B K (2012) Evolutionary development biology (evo-devo) Past present and future

Evolution Education and outreach 5 184193

Hannan M T amp Freeman J (1977) The population ecology of organizations The American

Journal of Sociology 82 929964

Hannan M T amp Freeman J (1989) Organizational ecology Cambridge MA Harvard

University Press

Hodgson G M (2013) Understanding organizational evolution Toward a research Agenda

using generalized Darwinism Organization Studies 34 973992

Knorr-Cetina K (1981) The manufacture of knowledge An essay on the constructivist and con-

textual nature of science Oxford Pergamon Press

Laubichler M D amp Maienschein J (2007) From embryology to evo-devo A history of devel-

opmental evolution Cambridge MA The MIT Press

Laudel G amp Glaser J (2007) Interviewing scientists Science Technology amp Innovation

Studies 3 91111 Retrieved from httpwwwsti-studiesdeojsindexphpstiarticle

view8970 Accessed on February 12 2014

Law J (1976) Theories and methods in the sociology of science An interpretative approach

In G Lemaine R Macleod M J Mulkay amp P Weingart (Eds) Perspectives on the

emergence of scientific disciplines (pp 221231) The Hague Mouton

Lewenstein B V (1992) Cold fusion and hot history OSIRIS 7 135163

Lewenstein B V (1995) From fax to facts Communication in the Cold-Fusion Saga Social

Studies of Science 25 403436

Lynch M amp Jordan K (2000) Patents promotions and protocols Mapping and claiming

scientific territory Mind Culture and Activity 7 124146

McCain K W (2010) Core journal literatures and persistent research themes in an emerging

interdisciplinary field Exploring the literature of the evolutionary developmental biology

Journal of Informetrics 4 157165

Minelli A (Ed) (2008) Evolving pathways Key themes in evolutionary developmental biology

Mulkay M J (1975) Three models of scientific development The Sociological Review 23

509 526

Muller G B (2007) Six memos for evo-devo In M D Laubichler amp J Maienschein (Eds)

Muller G B (2008) Evo-devo as a discipline In A Minelli (Ed) Evolving pathways Key

themes in evolutionary developmental biology (pp 530) Cambridge Cambridge

University Press

Nye M J (1980) N-Rays An episode in the history and psychology of science Historical

Studies in the Physical Sciences 11 125156

Raff R A (2000) Evo-devo The evolution of a new discipline Nature Reviews Genetics 1

Simon B (1999) Undead science Making sense of cold fusion after the (Arti)fact Social

PATH DEPENDENCE AND POLICY

STEERING IN THE SOCIAL

SCIENCES THE VARIED IMPACT

OF INTERNATIONAL LARGE

SCALE STUDENT ASSESSMENT ON

THE EDUCATIONAL SCIENCES IN

FOUR EUROPEAN COUNTRIES

Jochen Glaser Enno Aljets Adriana Gorga

Tina Hedmo Elias Hakansson and Grit Laudel

ABSTRACT

The aim of this article is to explain commonalities and differences in theresponses of four national educational science communities to the sameexternal stimulus namely international comparative large scale studentassessments that offered vastly improved comparability of nationalresults from the beginning of the 1990s The comparison shows theepistemic traditions of educational research in the four countries and

ISSN 0733-558Xdoi101108S0733-558X20140000042009

properties of the data produced by the international comparative studiesto be the central explanatory factors for commonalities and differencesof responses to the new studies

Keywords PISA studies educational sciences epistemic traditionsstate interests

INTRODUCTION

While most fields of research are spread across many countries the degreeto which they are truly internationalised varies enormously At one end ofthe spectrum we find completely internationalised fields in which all quali-fied researchers regardless of their country of origin or residence addresssimilar problems in similar ways While preferences for particular themesor approaches may vary between countries these are variations within ajoint enterprise that is advanced by contributions from all these countriesThe contributions fit because the knowledge from which problems arederived and methodological standards of research are shared acrosscountries

In contrast there are many fields in which goals and approaches aremore nationally or regionally specific Applied research for instance oftenaddresses innovation needs or societal problems of only one or few coun-tries (Luukkonen Persson amp Sivertsen 1992) In other cases the languagein which research is communicated is central to the research effort and willthus separate researchers in a country or a group of countries from othersworking on similar problems In yet other cases the empirical object of theresearch is nationally specific for example for palaeontology (Reguant ampCasadella 1994) or for research on social phenomena that exist in only onecountry (Mayntz 2001 pp 2027) In the most nationally specific casesresearch problems empirical objects and communication media are allexclusively national

The educational sciences have a range of features that are close to bothof these poles They typically investigate nationally specific empiricalobjects (education systems) and are often geared towards contributions tothe improvements of these systems It comes as no surprise that educationalresearchers in each country have a strong national focus and that researchtraditions vary between countries At the same time the possibility of

268 JOCHEN GLASER ET AL

international comparative large scale student assessments (ILSA) in varioussubjects indicates that many education systems share basic features includ-ing what is taught at what age of the students it is taught how it is taughtand by what organisations it is taught Research on these phenomena canbe expected to address similar problems and to produce results that can betransferred between national contexts thereby constituting an internationalfield

Why is it then that education researchers in four European countriesresponded to the emergence and rise of international large scale studentassessments in different ways Why was the educational science communityof one country (Sweden) involved in the very early stages of these develop-ments but lost interest at a time when lsquotrue comparabilityrsquo was achievedwhile the community of another country (Netherlands) was never reallyinterested at all the community of a third country (Switzerland) only afterthe whole enterprise had been under way for quite a while and someresearchers in a fourth country (Germany) became excited mobilisedresources for educational science at a scale previously unheard of and cre-ated a new subfield of educational science that led to a split of theirnational professional association The stimulus these national communitiesresponded to was the same for all of them namely the efforts by an inter-national elite and two international organisations to compare studentachievements across a large number of countries

The aim of this article is to explain commonalities and differences in theresponses of four national educational science communities to the sameexternal stimulus namely international comparative large scale studentassessments that offered vastly improved comparability of national resultsfrom the beginning of the 1990s We link commonalities and differences inthe responses to the epistemic traditions in the educational research of thefour countries to political structures and state interests and to propertiesof the data produced by the international comparative studies

In the next section we provide a brief overview of the history leading tothe new ILSA studies of the changes that justify the term lsquonew ILSA stu-diesrsquo and of our empirical investigation We then present case studies onthe impact the participation in these assessments had in the four countriesA comparison of these cases reveals the central explanatory factors forcommonalities and differences of responses to the new ILSA studies by thefour national educational science communities Our conclusions focus onthe tension between the political demand for comparative data andattempts to make scientific use of them

269Path Dependence and Policy Steering in the Social Sciences

COMPARING THE IMPACT OF INTERNATIONAL

LARGE SCALE STUDENT ASSESSMENTS

The Emergence of lsquoNew ILSA Studiesrsquo

The basic idea of international comparative large scale student achieve-ment studies is to compare the outcomes of education in the participatingcountries by submitting students in primary and secondary education tostandardised tests of their knowledge and abilities The attempts to con-duct international comparative studies of student achievement dateback to the 1930s of the 20th century were interrupted by World War IIand were put on the agenda again by the UNESCO in the 1950s In1958 the International Educational Association (IEA) was founded by agroup of leading educational research institutions under the direction ofthe UNESCO and began to regularly conduct comparative achievementstudies of students in different subjects across nations (Grisay amp Griffin2006)

Early comparative studies of student achievements suffered frommethodological problems that limited the value of comparisons betweencountries Available statistics were limited the degree of standardisationwas insufficient to guarantee data comparability and the methodology wasnot sufficiently advanced to support quantitative comparisons It was onlyat the beginning of the 1990s that the methodological quality of such stu-dies was improved sufficiently to meet accepted standards of comparativequantitative research These improvements included Rasch Scales ItemResponse Theory and Plausible Value Technique as well as matrix samplingwhich enables the coverage of larger areas of the curriculum withoutincreasing the number of questions an individual student had to answerThese new tools significantly improved opportunities for comparative ana-lyses They were first used by the IEA at the beginning of the 1990s in thepreparation of the Third International Mathematics and Science Study(TIMSS) in which all four countries under investigation participated andare now applied in all ILSA studies

In addition to these methodological improvements the organisationof the ILSA studies underwent a change in the early 1990s The studiesbecame hierarchically organised and standardized More internationalplayers entered the scene At the end of the 1990s the OECD introduced theProgramme for International Student Assessment (PISA) studies in whichall four countries investigated here also participated (Grisay amp Griffin 2006)

As a result of these developments all four countries were confronted bythe same external stimulus from the beginning of the 1990s They wereexpected to become part of hierarchically organised standardized methodo-logically advanced comparative studies of student achievements (Table 1)These studies were run by international organisations in collaborationwith an international elite of educational researchers and were aimed atproviding data for educational policy-makers in the participating countriesThey measured student achievements in various subjects on the basis ofstrictly equivalent instruments common definitions of target populationsand standardised procedures They also included a limited set of contextvariables on different levels including variables describing the organisationof school education in each country instructional practices and studentcharacteristics

As a result even countries that had previously participated in ILSA stu-dies were confronted by a new instrument new data collection protocolsand new opportunities to conduct international and national comparisonsof student achievements This is why the changes in the early 1990s can beconceptualised as a stimulus to educational science communities The sti-mulus made possible a wide range of responses by national educational

Table 1 New ILSA Studies and Participation of the Four Countries

Time Study Netherlands Sweden Switzerland Germany

19941997 Third International Mathematics

and Science Study (TIMSS)

X X X X

2000 Programme for International

Student Assessment (PISA)

X X X X

2001 Progress in International Reading

Literacy Study (PIRLS)

2003 PISA X X X X

2003 TIMSS X X X

2006 PISA X X X X

2006 PIRLS X X X

2007 TIMSS X X X

2009 PISA X X X X

science communities and national policy-makers with regard to the designof the studies the collection of the data and the analysis of data

The Empirical Investigation

Our investigation of responses to the lsquonew ILSA studiesrsquo was part of theRestructuring of Higher Education and Scientific Innovation (RHESI) pro-ject a comparative investigation of the impact of changing authority rela-tions in the public sciences on the opportunities for scientific innovationssummarised in the introduction to this volume We conceptualised the newILSA studies as such an innovation and asked how the authority relationsin the four countries affected the changes in research practices that wererequired by the conduct of national parts of ILSA studies and enabled bythe new research opportunities the ILSA data provide

The empirical investigation was based on interviews with researcherswho conducted ILSA studies (administered the national surveys or contrib-uted to the development of the methodology) or used ILSA data for sec-ondary analyses observers of the field policy-makers and administratorsat universities (Table 2) Interviews with ILSA researchers focused on theconditions under which they moved to producing or working with ILSAdata In addition we used a rich body of written sources including discus-sions of ILSA studies in the national educational science communitiesreports biographies and websites of the organisations conducting ILSAstudies

Our analysis revealed major differences between the impact of the newILSA studies and that of the other innovations studied by the RHESIproject (see the contributions by Laudel et al and Engwall et al to thisvolume) While the change of research practices by individual researcherswas crucial for understanding the impact of authority relations on the

Table 2 Interviews Conducted in the Four Countries

Netherlands Sweden Switzerland Germany

ILSA researchers 7 5 6 11

Others 2 policy-makers 7 observers

5 university managers

2 observers

2 funders

Total 7 7 18 15

development of innovations in the other three cases the conditions for andeffects of the new ILSA studies were mostly determined at the level ofnational educational research communities and their interaction with thestate and international agencies

An important condition that modifies the impact of authority relationson changes of research practices is the epistemic structure of the field(s) inwhich an innovation occurs This epistemic structure is only partly reflectedin its authority structure particularly the degree to which a unitary elitecan exercise authority over research goals and standards for researchOur analysis therefore focused on the following questions

What were the major research traditions in the national educationalscience fields that provided connection points for the new ILSA studies

How were educational researchers of a country involved in the earlydevelopment of ILSA studies prior to the 1990s

How was educational research institutionalised in the beginning of the1990s

How was participation in the new ILSA studies organised What skills had to be learned by researchers conducting the new ILSA

studies and analysing the data Did researchers of a country take part in the development of the new

ILSA methodology Did educational researchers collect additional data in connection with

the new ILSA studies How were ILSA data used What research questions were answered

using ILSA data What was the impact of the new ILSA studies on the national educa-

tional science community

These questions are now answered by four brief case studies whose sub-sequent comparative analyses will reveal the factors shaping nationalresponses

ONE STIMULUS FOUR DIFFERENT RESPONSES

The Netherlands ILSA as Distraction

At the beginning of the 1990s Dutch educational sciences had astrong quantitative research tradition Quantitative methods had diffused

into educational sciences in the Netherlands since the 1980s and were com-mon and accepted within several subfields from the 1990s One of thesubfields of Dutch educational research school effectiveness research evenaddressed research questions very similar to those of ILSA studies TheNetherlands was the only country with strong school effectiveness researchbesides the United States and the United Kingdom (Creemers amp Scheerens1994 Ditton 2009 p 252) Research focused on influences on student per-formance at different levels of aggregation such as the individual studentthe classroom and the national education system It differed from othersubfields of educational science in its inclusion of an international com-parative perspective Already in 1968 the Dutch government founded theNational Institute for Educational Measurement (CITO) one of the firstinstitutes for educational measurement worldwide CITO has developededucational tests for monitoring and evaluation purposes

Dutch educational scientists actively participated in ILSA studies asmembers and chairs of international steering committees from the 1980sWhen Tjeerd Plomp chaired the steering committee of TIMSS and the IEA(19891999) his research institute was the national hub for the Dutchparts of these studies However Dutch researchers contributed little to themethodological research that prepared the lsquonewrsquo ILSA studies from theearly 1990s These contributions primarily came from the United StatesUnited Kingdom Germany and Sweden (Keeves amp Lietz 2011)

The participation in the lsquonewrsquo ILSA studies could build on theseresearch traditions and on a well-established and institutionalised field ofeducational science A substantial part of educational research had beenorganised in five research institutes which though part of universitiesalmost entirely depended on external research contracts with the stateAdditionally educational science was (and still is) conducted within someuniversity faculties for example psychology departments

The costs of participation were met by the state who issued calls for ten-ders to administer ILSA studies in the Netherlands Researchers apply forthe administration of these studies because they need the funding ThePISA studies are now administered by CITO which was transformed intoa commercial research organisation that also develops tests for nationalachievement studies

As a consequence data production was more or less a routine taskeven though it was time-consuming and required considerable skills Allresearchers who used ILSA data were trained in quantitative educationalscience methods which made learning of specific analytical methods unne-cessary Owing to the wide acceptance of quantitative educational research

Dutch participants in ILSA studies didnrsquot have reputational problemseither

The researchers who were involved in the data production had hardlyany influence on the content and methodology of the studies which sepa-rated it from their own research interests The main task of those who con-ducted the Dutch part of an ILSA study was to develop a samplingstrategy for their country

A You probably know about those international studies that there are only very few

margins which you can use yourself because more or less all is dictated by international

agencies So I didnrsquot have any input on what we were doing at [a certain ILSA study]

Q So you have got basically the Dutch part and had to follow the instructions

A Yes you could say that definitely1

Other attempts by Dutch educational researchers to influence the inter-national data production failed too One researcher succeeded to includean important variable (to him) in the Dutch national survey but did notachieve the same change on the international level

I never did this on the international level this is impossible Because it is an intergo-

vernmental body PISA and OECD And there I donrsquot have influence This doesnrsquot

mean that the secretariat doesnrsquot know that I do this I gave a lecture at the PISA office

[hellip] They found it nice and that was that

In the Netherlands ILSA studies did not trigger additional data collec-tion because there has always been a research tradition of quantitativecomparative school effectiveness research which continued independentlyof the ILSA studies Dutch educational researchers have conducted regularlongitudinal national studies of student achievements since 1988 Theministry was interested in the regular monitoring of student performance inprimary and secondary schools and in factors influencing performance(Kuyper amp van der Werf 2012 pp 16)

So far ILSA data have been used almost exclusively for national reportsdemanded by the state One of the reasons why ILSA data are rarely usedfor research is the scarcity of time Researchers who had main responsibil-ities in the ILSA data collection and had an intimate knowledge of the sur-vey data lacked the time for using the data because they had to administerthe national surveys while their colleagues depended on contract researchin other areas to fund their own positions Meanwhile the Dutch researchcouncil granted a larger project proposal of four research groups that pro-posed an extensive use of ILSA data

A second reason for the reluctance of Dutch educational researchers touse PISA data is their dominant interest in their national educationalsystem which makes data on other countries and comparisons with othercountries less interesting The national focus is reinforced by the interestsof the Dutch ministry of education which still is the major funder ofeducational research

A third important reason why Dutch educational researchers donrsquot useILSA data is the nature of the data themselves Researchers consider thesedata unsuitable for answering their research questions and as generallyinferior to the national data they produce independently of ILSA studiesAccording to the Dutch educational researchers interviewed the currentimportant research questions of the field require longitudinal data whichthe ILSA studies donrsquot provide (see also Dronkers amp Robert 2008 p 549)

What is the biggest downside of PISA is that it is cross-sectional Thatrsquos a main pro-

blem certainly in the Dutch context where we have much better data if we would only

analyze Dutch kids So we have the VOCL cohorts previously now the COOL data

These longitudinal data are much better potentially than PISA hellip We can answer

important questions which you canrsquot answer in PISA which is the extent to which

choices in education are based on performances or to what extent something else is hap-

pening in the choice processes and you canrsquot assess that because you need to have prior

information about their school test result and then see what happens in the choices later

on so you need to longitudinal data (Dutch educational researcher)

Those few Dutch educational researchers who used ILSA data did sobecause they were interested in international comparisons They often usedthese data in an opportunistic way in combination with national data andother international data such as the European Social Survey Only tworesearchers used ILSA data regularly One of them an educational sociolo-gist saw the data as crucial for establishing a new line of research

I realized these are very rich data for the things that I want so looking at the educa-

tional systems and how are they different between countries [hellip] basically if you look

at the comparative stratification literature for example the field that prepared me for

getting into [this] project itrsquos very important to look at the level of stratification of edu-

cational systems

Although Dutch educational research is still predominantly orientedtowards the national education system publications of research with ILSAdata were accepted by the community

Overall the impact of the lsquonewrsquo ILSA studies on the Dutch educationalscience community was rather low Data production was considered a rou-tine process by most and the few researchers who wanted utilise the ILSAdata collection for their own research failed due to the rigidity of the inter-national data collection protocol The data produced by ILSA studies are

rarely used mainly because there is little interest in research questions thatcan be answered with these data This situation might change due to recentresearch council funding for research using ILSA data

Sweden Focus on Methodology

The Swedish community of educational science has over time changed intobeing increasingly multi-disciplinary and fragmented (Hansen amp Lindblad2010) Many of the early educational researchers were educational psy-chologists Most research in educational science was and still is providedby departments of pedagogy located in faculties of social sciences at publicuniversities Historically the role of quantitative research in pedagogy hasbeen strong in Sweden However quantitative research gradually lost itssignificance in the 1970s as a more societal-oriented and qualitativeapproach in research methodology was introduced in pedagogy

There was skepticism about the more quantitatively oriented pedagogy in Sweden dur-

ing the 70- and 80-centuries Husen [a pioneer of early ILSA studies see below] was not

Prophet in his own laboratory so to speakhellip

The new approach was also reflected in policy-making in educationAs an outcome pedagogical research became increasingly separated fromquantitative methods and measurements as well as approaches inspiredby psychology In addition academic courses in statistics disappeared inthe national curriculum in pedagogy at undergraduate and postgraduatelevels More recently interest in quantitative methods has increasedagain especially in pedagogy Sweden is considered by interviewees ashaving a strong international reputation in statistics and statisticalanalysis

Swedish participation in ILSA studies is embedded in its long traditionof studying the efficiency of the Swedish national education system In1992 and 1995 two larger national studies were conducted which howeverused less advanced methodologies than the international studies

Sweden actively participated in ILSA studies from the beginning Theheyday of ILSA studies in Sweden appeared in the late 1950s and wasstrongly linked to enthusiastic pioneers such as professor Husen and theestablishment of IEA and its first international studies After a period ofregression interest increased again when international large scale studentassessments were transformed into more reliable and systematic test modelsin the late 1980s and 1990s

Participation in primary ILSA studies has never been a problem inSweden in financial or reputational terms Swedish statisticians havebeen central in the development of the ILSA methodologies both beforeand after the early 1990s For instance they have been successful withdeveloping computer programs for multivariate statistical analysis such asLISREL being known as lsquothe state of the artrsquo and used in ILSA analysis

The Swedish National Agency for Education which was established in1991 is the central authoritative agency in this field It is also responsiblefor the national implementation of the ILSA studies and funds the primarystudies on behalf of the government The administration and coordinationof the national implementations of ILSA studies are contracted out to pro-ject groups operating at public universities following a competitive applica-tion procedure based on criteria such as reputation and the internationalbackground of the university as well as the potential of the university tocreate a good research environment The project groups are coordinated byresearchers acting as national project leaders The national operations ofproject groups such as the reporting of raw data are strongly controlled bynational and international authorities Their discretion is limited to thesampling strategy (on the basis of predetermined selection criteria)

The project groups responsible for producing LSA data at the nationallevel are free to include additional questions for studying specific aspects atthe national level not being covered by the standardized assessments In2011 a national network supported by the National Research Councilwas established for contributing to re-analyses of large-scale studies Oneof its aims is to complement large-scale studies with the collection of quali-tative data

As in the other countries the main target of the international large scalestudent assessments and their results in Sweden is policy-makers followedby school principals and teachers Most of the results of primary data ana-lysis are published in form of national reports and as shorter summarieswhich are co-published by the national project groups and the NationalAgency for Education

In some cases the National Agency orders in-depth analyses of certainareas of the tests Researchers at Swedish universities studied the validity ofTIMSS and PISA surveys with a focus on the relevance of internationaltests for the Swedish curriculum and on comparisons to national testsSimilar studies were conducted in the area of reading comprehension Inthe mid-1990s the private funding agency Riksbankens Jubileumsfondfunded a project at Gothenburg University that created a moderncomputer-based public domain covering all ILSA studies conducted in

Sweden before 1995 The aim was to make old ILSA data lsquoanalysablersquo andcomparable by translating it into a format that is compatible with modernsoftware

Most of the researchers who were interested in using ILSA data movedinto the field due to their strong methodological research interest

It was mainly my interest in method that led me to my involvement in re-analysis stu-

dies of ILSA data [A colleague] led a project in which I got involved where we did two-

level analyses using structural equation modeling first developed by Bengt Muten

These methods have been regarded as extremely complex and even though we had spe-

cial competence within these areas we also ran into problems we had to focus much on

developing new computer programs

Beyond these methodological studies the size and scope of this fieldis still limited and it is mainly concentrated in the central researchenvironments of Sweden such as Gothenburg Umea and Mid SwedenUniversity Some of the researchers involved in primary ILSA studieshave also used the data for secondary analyses However the tight sche-dule for running and administering the ILSA tests often reduces the flex-ibility of researchers to use the data This is why they are often unable toexploit their good access to data and proper knowledge about the datasample for doing such studies Researchers also reported financialobstacles

It is very time consuming to prepare and analyse the data and from that also to author

papers It is hard to finance this type of research in Sweden it seems almost impossiblehellip

A second reason for the limited use of primary data in Sweden is theircomplex structure which requires highly advanced quantitative methods ofdata analysis Interviewees described this as a result of the shift from theexperimental and research inspired design of the old ILSA studies to amore evaluation-focused research agenda The design of the tests used inILSA studies and the sampling strategies make it impossible to use themethods of quantitative data analysis most quantitative pedagogicalresearchers are familiar with Some knowledge of psychometrics econo-metrics and advanced statistics is required Owing to a relapse of interest(and consequently education) in quantitative methods only few Swedisheducational researchers are currently able to conduct causal analyses withILSA data This is why both young and more established scholars whowere interested in ILSA data experienced a need for further methodologicaltraining before conducting secondary analyses In order to learn how touse ILSA data researchers participate at specific workshops being orga-nised by international organisations

More recently other academic disciplines such as economics politicalscience and statistics have started to use ILSA data (Forsberg amp Lindberg2010) Also the National Agency of Education has arranged a seminar ser-ies together with the National Research Council to encourage researchersto do in-depth analysis of test results

So far the major impact of the lsquonewrsquo ILSA studies on the Swedishcommunity is a re-awakening of the methodological interest in quantitativestudies By now many of the younger researchers involved in the pro-duction of primary data have invested time in learning how to use the stan-dardized statistical methods and tests developed by the internationalorganisations Research questions around ILSA are mostly methodologi-cal and the use of ILSA data is seen as hampered by an insufficient under-standing of its methodology which has to be overcome first Howeverresearchers are turning towards secondary analyses of ILSA data a trendthat is encouraged by policy actors and research councils

Switzerland Routine Data Collection and Slow but SteadyInstitutionalisation of Data Use

In the early 1990s educational research in Switzerland was even morestrongly fragmented than that of other countries because the multiplicityof cantonal education systems in a federal state linguistic and culturaldiversity and the co-existence of hermeneutic and empirical researchtraditions added to the traditional multi-disciplinary nature of educationresearch

Despite this fragmentation two main topics dominated Swiss educa-tional research in the 1990s the role of compulsory education and a grow-ing interest in the study of effects of educational processes (Gretler 2000)Both topics provided a fertile ground on which the participation in and useof ILSA studies could grow However the quantitative research traditionwas rather weak in Switzerland and varied with the size of cantonal officesfor educational research Cantons with larger offices conducted more edu-cational research and thus had a stronger interest in these methods thansmaller offices with fewer researchers

Prior to 1995 only some of the Swiss cantons sporadically participatedin ILSA studies (eg the canton of Geneva in 1959) which is consistentwith the cantonal authority over education The first participation ofSwitzerland as a country occurred when the country joined TIMSS in 1995

Educational research in Switzerland is institutionalized in a variety ofpublic research organisations In the early 1990s this mostly took place inCantonal Universities and in Cantonal Offices for Educational ResearchThe Universities of Teacher Education which provide teacher education asUniversities of Applied Science since 1995 are likely to become anothertype of organisation in which educational research relevant to ILSA studiesis conducted However this has not happened yet despite early localattempts and a specific funding instrument supporting cooperation betweencantonal universities and universities of applied sciences being introducedby the Swiss National Funds (SNF) which has awakened the interest ofsome researchers in ILSA studies

After the TIMSS study the political decision was made that Switzerlandshould only participate in one type of ILSA study namely PISA Fundingand management of the participation also changed The research councilfunding which was oriented towards research in education was replacedby a funding scheme for the management of the study Funding is splitbetween the confederation (60) and the Swiss Conference of CantonalMinisters of Education (40) The Swiss Federal Statistical Office wastasked with PISA data collection and management The analysis of PISAdata is conducted ad hoc by consortia of researchers which in 2008 alsotook over data collection from the federal statistical office This move wasmotivated by the restrictions the statistical office put on the access to dataand on additional analyses When forming consortia researchers alreadyhad experiences with large-scale quantitative studies They participated inOECD training workshops to acquire the specific methodological expertiserequired for PISA analyses

While the members of the consortia were mainly concerned with produ-cing data and ad hoc analyses other Swiss researchers linked additionalresearch to the Swiss national parts of the ILSA studies The researchers con-ducting TIMSS complemented the comparative tests of achievements by col-lecting and analysing video data In a project funded by the SNF researchersfilmed and analysed mathematics lessons with students who participated inTIMSS in order to understand the main determinants of educational successSNF funding also motivated researchers to link the study of transitions fromeducation to work to the PISA studies The longitudinal study lsquoTransitionsfrom Education to Employmentrsquo (TREE) was initiated by a researcher froma cantonal office for educational research in 1999 following a recommenda-tion by the OECD The PISA 2000 data collection provided a window ofopportunity for this project which followed the educational and vocationalpathways of the 6000 students who took part in the PISA study

PISA data are still rarely used for research beyond data analysis for edu-cation policy Research in the educational sciences is highly individualisedwith each researcher striving for an original approach and using their homediscipline (sociology psychology and economics) rather than educationalscience as their major frame of reference

These approaches included critical perspectives on PISA and researchquestions that did not entirely fit the international PISA methodologyand data Swiss researchers tried to avoid the hierarchically organised andstandardized studies in order to critically investigate PISA data (usually byadditional research)

To say it very shortly and maybe a little bit extreme these data TIMSS and PISA

they are good for politics but they are not very useful for teacher education and for

teachers on this level If you want to use this kind of data than you have to go far

beyond this kind of system monitoring (hellip)

A second reason is that the access to recent data has been restricted atleast initially to those who produced the data

For example there were guidelines regulating the access to data for external researchers

[researchers who were not members of the consortia conducting the primary studies]

And often it was like this lsquoPISA data will be given to researchers who are in the group

the consortiumrsquo usually six months before publication because they have to prepare

the national report lsquoAnd then for a year after the publication of data there will be no

access to the data They will have to waitrsquo So they did everything to block access to

external researchers to PISA data

The use of ILSA data is likely to increase now that the access to datais controlled by researchers and due to initiatives by several universitiesto institutionalise educational research linked to ILSA studies Until2012 however the use of ILSA data at the cantonal level was stillrestricted for researchers who were not involved in the data productionThey faced a one-year embargo before they could access the data

Several Swiss universities responded to the growing state interest byinstitutionalizing ILSA-related research more strongly In two universitiesnew research centres were created In 1999 the University of Zurich createda Competence Centre for Educational Evaluation which was transformedinto a self-funded Institute for Educational Evaluation in 2003 The direc-tor of the institute also is the national coordinator of the PISA consortiumAt the University of Bern the Research Center for Educational Economywas founded in 2001 It was asked by the national PISA steering group tostudy the influence of studentsrsquo social background on their performance inPISA tests In addition to these centres the University of Genevarsquos

department of education created a professorship for the analysis of effectsof education systems These developments point to a slow growth of theSwiss educational science community or at least of the part of the commu-nity engaged with ILSA studies

Germany ILSA as Kick-Start of a Quantitative EducationResearch Community

In the beginning of the 1990s German educational sciences were domi-nated by a hermeneutic research tradition that was focused on the under-standing of interactions between teacher and student It was assumed thatthe effects of school education exclusively depended on these interactionswhich were treated as specific to each classroom situation In this paradigmof educational science there was no room for quantitative comparativeresearch

Quantitative educational research was also close to impossible inGermany at that time because researchers could not produce appropriateempirical data Access to data (including data collection in classrooms) hadto be granted by the federal states which withdraw their initial permissionafter being disappointed by the low scientific quality of studies in the 1960sand 1970s As a consequence Germany ceased participation in interna-tional comparative studies and quantitative research came to a halt foralmost three decades

The 1980 and 1990s before TIMSS had actually been dead years for German empirical

educational research

Since there was no quantitative educational research in Germany at thebeginning of the 1990s there was no community in which reputations couldbe earned with conducting ILSA studies or quantitative educationalresearch in general There were no positions or research infrastructures atuniversities and only limited research infrastructures at public researchinstitutes There was no project funding for quantitative educationalresearch and it was impossible to build a career in educational science onthe basis of quantitative studies Only few researchers remained who hadthe knowledge required by a participation in ILSA studies Political sup-port for ILSA research was also low

This relapse occurred after some German researchers had been invol-ved in the early ILSA movement These researchers (among them theGermany-based Neville Postlethwaite) were outsiders to the German

educational science community They were based in state-funded researchinstitutes outside universities which also supported the early participationin ILSA studies prior to the denial of access to data by the federal states

The prospects of ILSA studies changed fundamentally when after manyyears of neglect political actors developed an interest in international com-parative and intra-German comparative data on school effectiveness AfterGerman reunification a lack of knowledge that could inform the govern-ance of the educational sector became apparent and education policy-makers became interested in Germany participating in ILSA studies Theygranted researchers access to schools and turned to the only organisationsthat could possibly manage this participation in ILSA studies namely thethree state-funded research institutes whose mission was educationalresearch

In the context of a shrinking discipline of educational sciences and asignificantly increased competition for state funding of public researchinstitutes due to German unification public research institutes in the early1990s were forced to look for ways to legitimise further research Since edu-cational research was the raison drsquoetre for some of these institutes theirdirectors had to accept the request to participate in ILSA studies As aconsequence some researchers in these institutes were effectively forced todrop their previous research and to move to ILSA studies

German researchers did not just administer their parts of the interna-tional studies but also actively engaged in the methodological work at theinternational level and thus began to play important roles in the interna-tional development (including the provision of data analysis services)

And simultaneously it happened that I was proposed to join the international expert

group at OECD or ACER (hellip) It was important from a German point of view to sit in

the first row and to participate in test development and to ensure the test quality to get

involved in international discussions These expert groups were relatively small about

seven or eight people who came from different countries

The move of researchers to ILSA studies was expensive and bore signifi-cant reputational risks The material costs of ILSA research were howevereasily met due to its location at state institutes (see Glaser et al thisvolume) and the political interest in these studies

German ILSA researchers managed to exploit the new political interestfor extending the collection of data beyond those required for the interna-tional comparison They enlarged national samples of ILSA studies inorder to obtain robust data for intra-country comparisons which are ofspecial interest in Germany due to the federal statesrsquo authority over

education In the context of PISA the sample was extended to 44000students which made comparisons between federal states possible and wasconsidered a breakthrough by educational researchers

LSA researchers also constructed national longitudinal studies aroundinternational LSA studies These data which trace educational attainmentof the same students in several subsequent years are generally consideredto be much more useful for educational research than the cross-sectionaldata produced by ILSA studies Similar to their Swiss colleagues Germanresearchers turned TIMSS in a longitudinal study

We did TIMSS as a longitudinal study We started one year before took a new sample

that was untouched for the international comparison (hellip) We were the only country

that added a longitudinal design from the beginning We had observation data we had

video data in comparison to Japan and the US (hellip) The international organisation of

the study was not interesting for me

The use of data emerged slowly with the accumulation of data from sub-sequent international studies The federal structure of the German educa-tional system which has 16 different educational systems under the controlof the federal states enables the replication of ILSA-type questions at thenational level

Access to the ILSA data and publication of results remain difficult forsome researchers because some of the questions about differential educa-tional attainment in Germany are considered politically sensitive andeither access to data or publication of results is hindered by the state ThelsquoPISA shockrsquo the perception that German student achievements are onaverage far worse than previously assumed created an intense politicaldebate about the causes of these results Neither all of these causes nor allthe causes of differences within Germany between education systems ofthe federal states are open to educational research at the moment Someresearchers who want to investigate specific questions in that context findthemselves barred from access to the necessary data

Parallel to the slowly growing use of ILSA data an interest in longitudi-nal data emerged More recently the interest in longitudinal data led to theinstitutionalisation of a large national longitudinal study the so-callednational educational panel study (NEPS) In the context of these develop-ments many researchers changed their research practices to the productionof LSA-style data

When German educational policy-makers decided to participate inthe lsquonewrsquo ILSA studies administering the studies bore significant reputa-tional risks for researcher It was not clear at all whether an academic career

in educational science was possible on the basis of ILSA research Therewere no positions at German universities for ILSA researchers and the hos-tility of traditional educational scientists towards ILSA studies made it seemunlikely that professorships would be given to ILSA researchers

However the German ILSA researchers were able to transform the rea-wakened political interest into political and financial support for quantita-tive empirical research beyond the immediate contribution to internationalstudies The researchers at public research institutes who moved into thefield of ILSA studies were originally not particularly interested in thisresearch but recognised the research opportunities created by the politicalinterest and the absence of competition The federal states had no depart-ments of their own that could produce independent scientific data as inputfor governance Nor were there nearly enough researchers at universitieswho would or could meet this demand

Participation in international studies and accompanying national studieslet quantitative educational studies emerge as a separate subfield of educa-tion research in Germany The large research programme which was guar-anteed through the periodic cycles of international studies and thecontinuous commitment of political actors created many well-funded posi-tions for research and training Soon researchers participating in the inter-national studies could earn reputations within a multi-disciplinary field ofeducational science sociology and psychology The states also began toinitiate the institutionalisation of quantitative empirical education researchat German universities

Owing to the specific methodological knowledge that was required forconducting ILSA studies none of the educational researchers in the herme-neutic tradition moved to quantitative studies Instead researchers frompsychology economics and sociology migrated to the new field The con-ceptual and epistemological differences between traditional and lsquonewrsquo educa-tion researchers remained and led to a split of the professional associationIn 2012 quantitative educational researchers founded their own professionalassociation the Association for Empirical Education Research

EXPLAINING COMMONALITIES AND DIFFERENCES

IN NATIONAL RESPONSES TO ILSA STUDIES

The comparison of the four cases shows that the factors most important tothe other innovations discussed in this volume access to resources and

reputational risks didnrsquot make a difference to the development of ILSAstudies in the four countries The strongest authoritative agencies whichshaped the emergence and persistence of ILSA studies in all four countrieswere the states and their interest in monitoring comparing and improvingtheir education systems and the international agencies and their expertpanels which determined what was done when it was done and how it wasdone This basic authority structure did not vary between the fourcountries

The interest of the state in participating was strong enough to securethat administering the survey and providing reports was sufficientlyfinanced and researchers were funded for each study In the Netherlandsand Switzerland the administration of surveys was occasionally outsourcedto non-research units which emphasises the routine character of this taskAccess to resources was secure and conducting the new ILSA studiesinvolved no reputational risks for Dutch Swedish or Swiss researchers

The organisation of authority relations made a difference only inGermany where the first groups to conduct ILSA studies were lsquoshelteredrsquoin public research institutes which protected them from their communityrsquosantagonism The political interest in ILSA studies led to continuous fund-ing which created positions and career opportunities and finally led to theemergence of a community

The involvement of a countryrsquos educational scientists in ILSA studiesbefore 1990 also made little difference In the three countries involved withsetting up ILSA studies (the Netherlands Sweden and to some extentGermany) this tradition lapsed prior to the new ILSA studies

Most of the commonalities and differences of responses by the four dif-ferent educational science communities can be explained by nationally spe-cific relationships between the nature of the stimulus the new ILSAstudies and the national epistemic context in which it operates (Table 3)

The four national educational science fields vary considerably intheir thematic composition and methodological research traditions whichmade the new ILSA studies arrive in different contexts Three of the fourcountries had quantitative research traditions that addressed researchquestions similar to those asked by ILSA studies Dutch Swedish andSwiss educational researchers were all interested in the effects of theireducational systems on achievement to some extent and addressed thesequestions by quantitative approaches Although quantitative researchreceded in Sweden and never was a dominant approach in Switzerland itwas legitimate in all three countries and provided some fertile ground forthe new ILSA studies Only German educational science was dominated

Table 3 Relationship between New ILSA Studies and Educational Research in the Four Countries

Epistemic traditions on

which ILSA could build

School effectiveness research Efficiency of the national

education system

Effects of

educational

processes

Strength of quantitative

research in the early 1990s

Strong Limited Limited Very weak

Involvement in developing

the ILSA methodology

No Yes No Yes

Additional data production

linked to ILSA studies

None Additional items in ILSA

surveys qualitative data

Qualitative and

longitudinal data

expansion of ILSA

samples

Qualitative and

longitudinal data

expansion of ILSA

samples

Research building on ILSA

studies

Few secondary analyses

(international comparisons)

Mainly methodological few

secondary analyses

Mainly critically

methodological few

secondary analyses

Methodological and

few secondary

analyses

Reasons for limited use of

ILSA data

Time constraints for data

producers initially little

funding for secondary

analyses own longitudinal

data fit research questions

better little interest in

international comparison

analyses insufficient

methodological knowledge of

other researchers

Limited access to

data until 2008

data donrsquot fit

research questions

Limited access to

data data donrsquot fit

research questions

Impact on educational

science community

None Strengthening of quantitative

research tradition

Strengthening of

quantitative

research tradition

Emergence of

quantitative

community

JOCHEN

GLASER

by hermeneutic approaches and was at the same time hostile to quantita-tive educational research As a consequence ILSA studies were effectivelyestablished outside the dominant educational science community

Regardless of compatible traditions in three countries and a rapidlygrowing community in the fourth the new ILSA studies donrsquot appear to bewell embedded in educational science research Researchers in Swedenadded questions to the survey and German and Swiss researchers enlargedthe sample in order to support comparisons between the educational sys-tems of federal states German and Swiss researchers complemented thefirst new ILSA study (TIMSS) by qualitative data and turned it into anational longitudinal study The research accompanying the new ILSAstudies is mainly methodological The use of ILSA data for educationalresearch beyond the delivery of policy reports has started with considerabledelays and keeps growing slowly but steadily in all four countries least soin the Netherlands

Our interviews revealed several interesting reasons why ILSA dataappear to be difficult to use The first reason is time It was clearly stated inthe Netherlands and Sweden that one can either administer the nationalsurvey for an ILSA study or use the data for research because administer-ing the national survey is a very time-consuming task that consists of muchroutine work While state funding was sufficient for data collection andanalyses for policy reports in all four countries it did not cover secondaryanalyses for scientific purposes Funding for secondary analyses emergedwith considerable delays in all four countries

This dilemma impedes secondary analyses because ILSA data are quitecomplex and difficult to understand and having produced them best pro-vides the necessary understanding This is why those who conduct thenational studies also were those best placed to use the data At the sametime there is little overlap between producers and users of the data in allfour countries

Secondary analyses of ILSA data are also made unattractive by the lackof control of data collection The content methodology and outcomes ofILSA studies are largely decided by the panels of international expertswhich plan and administer the studies for either the IEA or the OECDConsequently it doesnrsquot matter at all who administers the national studiesas long as the rules set up by the international leaders are followed Datacollection is a routine task that can equally be conducted by state bureau-cracies or education researchers This emphasis on the international com-parability of data maximises political use but ignores possible limitations toscientific use

This was an issue for Dutch researchers who tried to include variablesbut did not succeed In Sweden and Switzerland this wasnrsquot even tried andGerman educational researchers actively participated in the methods devel-opment but did not attempt to change the variables The only opportunityresearchers had was including additional questions in their national sur-veys which made international comparisons the main purpose of newILSA studies impossible As a result ILSA data are likely to slightly mis-match most research questions and compete with data produced by theresearchers themselves that is data that were controlled and fully under-stood by those who analyse them in countries with a national tradition ineducational research

The limited interest of educational science communities in internationalcomparisons is yet another reason why ILSA data are used onlyreluctantly Participation in ILSA studies has not changed the fact thateducational research has a strong national orientation which makes inter-national comparisons a marginal research theme The countries with inter-nally differentiated education systems (Switzerland and Germany) canconduct internal lsquoILSA-stylersquo comparisons within national boundariesHowever these comparisons are also mainly of political interest Beyondthese comparisons in federally organised countries few researchers appearto use data from other countries at all

Another important property of ILSA data that limits their use for edu-cational research has surfaced in all four cases ILSA data are cross-sectional rather than longitudinal which serves their political purpose ofinternational benchmarking but makes them less suitable for researchpurposes Educational researchers consider longitudinal data necessary foranswering their research questions This makes ILSA data lsquobadrsquo data asDutch researchers who have their own longitudinal data clearly stated Theattempts by German researchers to make the TIMSS data part of a longitu-dinal study and the recent initiatives of the now-established German com-munity to create national longitudinal data reflect the same problem Swisseducation researchers responded to this problem by adding a longitudinalstudy to the first PISA survey Another more indirect sign of the unsuit-ability of ILSA data for educational research is the large proportion ofILSA-related research that uses the data to answer methodological ratherthan substantial questions (the methodological research on the validityof ILSA studies in Sweden) or occurs outside the core educational sciencefor example as sociological research into inequality in the NetherlandsSwitzerland and Germany

Finally it appears that ILSA data need to exist for some time beforeapproaches to using them are developed These data do after all constitutea new empirical research object that competes with those researchersalready have established The research opportunities need to be discoveredwhich will be done by new generations of researchers After the lsquonew ILSAstudiesrsquo began in the early 1990s we finally observe a rise in ILSA data usefrom the mid-2000s onwards This indicates that educational research com-munities in all four countries including the ILSA-driven and rapidlygrowing German community needed to learn what can be done withdata that were produced first and foremost for political purposes The slowgrowth of the use of ILSA data also indicates a diffusion process It mightturn out that ILSA data have their more innovative uses not withineducational research but in neighbouring fields such as the sociology ofinequality which might in turn change the composition of the complexfield of educational research

These difficulties to link educational research to the new ILSAstudies explain their differential impact on national educational science com-munities This impact is negligible in the Netherlands because Dutch educa-tional researchers continue their nationally oriented quantitative research withtheir own data which they consider superior In Sweden and Switzerland thenew ILSA studies strengthened the quantitative research tradition that wasalready there In Germany where no tradition existed that was compatiblewith the new ILSA studies the introduction of ILSA studies due to stateinterest kick-started the growth of a quantitative educational science commu-nity State-funded research institutes and large amounts of funding tempora-rily protected new quantitative researchers from the authority of theirnational community Having grown and become firmly institutionalised thenew community now appears to be turning towards more interesting thingsthan administering and analysing the new ILSA studies

CONCLUSIONS TENSIONS BETWEEN SERVICE

DELIVERY AND ENDOGENOUS RESEARCH

DYNAMICS

The new ILSA studies that were conducted from the beginning of the1990s provided us with a quasi-experiment Four different science systemsreceived the same stimulus of exactly the same type of empirical study

The lsquointerventionrsquo of new ILSA studies was exogenous to the four nationalscientific communities in several respects They did not emerge from theresearch of any of the national communities even though the generaltopic addressed by it was established in three of them The protocol ofdata collection was non-negotiable and could only be added to by nationaleducational science communities The studies applied a new methodo-logy that enabled reliable international comparisons of student achieve-ment and educational policy-makers in the four countries were interestedin the results and made available the necessary resources for conductingthem

It is important to keep in mind that the lsquopanels of international expertsrsquodirected these internationally comparative studies in the absence of aninternational scientific community with common research priorities Whileall four countries had or have leading researchers who were members ofinternational expert groups or organisations and promoted the participa-tion in ILSA studies in their countries there is little indication in our fourcase studies of a set of shared scientific problems that guides ILSA studiesInstead these studies owe their existence to a political interest in comparingstudent achievements The international experts leading ILSA studiesappear to be an international group of members of national elites ratherthan the elite of an international community

The new ILSA studies thus were exogenous to the four national educa-tional science communities in two respects They were epistemicallydecoupled from educational research in the four countries and they weredriven by external (political) interests This situation provided us with theopportunity to identify the factors responsible for the different responsesto this stimulus in the four countries The analysis of authority relationsbetween international agencies national policy interests and scientific com-munities on the one hand and national epistemic traditions of educationalresearch on the other hand enables the following four conclusions

First the impact of any scientific innovation on a national scientificcommunity strongly depends on its compatibility with national researchtraditions (for the same point see the contributions to this volume byLaudel et al Engwall et al and Benninghoff et al) In the case of newILSA studies the impact was strong in the country that didnrsquot have compa-tible traditions (Germany) and much weaker (Sweden and Switzerland) orabsent (Netherlands) in countries that had research traditions asking com-patible questions and using compatible methods In these latter countriescertain research traditions were strengthened while in Germany a wholenew community of quantitative educational researchers emerged

This pattern is exactly the opposite of what has been observed in theother RHESI case studies It also is somewhat counter-intuitive becauseone could expect a stronger impact on fields with a smaller epistemic dis-tance to the innovation This leads us to a second conclusion namely thatthe development of an innovation crucially depends on the research oppor-tunities it offers (a point that has been made by Pickering 1980 for thechoice between theories) The new ILSA studies differ from the other inno-vations discussed in this volume in this respect because they had little tooffer to educational researchers Modifications of the data collection proto-col were strongly limited to additional questions in national contexts andenlargements of the national samples Secondary analyses of the data werelimited by time constraints for those who produced the data methodologi-cal difficulties for others and more generally by the nature of the dataCross-sectional data were of no use for educational research communitieswhose research traditions generated questions that can only be answeredwith longitudinal data

The third conclusion is therefore that political interest and funding cangenerate scientific services but are not sufficient to establish scientific inno-vations This argument has a long tradition in science studies possiblybeginning with Polanyi (1962) reinforced by empirical studies on lsquoplannedresearchrsquo in the 1970s (Van den Daele Krohn amp Weingart 1979) andapplied to the analysis of research councils as intermediary organisations(Braun 1998)

The demand for services is not inconsequential though Services pro-vided by a scientific community may strengthen lines of research that canbe meaningfully linked to the service as is illustrated by the strengtheningof quantitative research traditions in Sweden and Switzerland The Germanexample demonstrates that state interest can be utilised by researchers tofurther their own research and professional agendas and how far this utili-sation can go Germany owes the new ILSA studies the emergence of awhole new research community and the emergence of a quantitativeresearch tradition in educational science

The case of ILSA studies thus demonstrates the limited impact of exter-nal authority on the endogenous dynamics of research fields One of themore surprising findings of our research is the slow pace at which data thatare produced with immense efforts are used for educational research Thisleads us to our fourth conclusion Having defined a scientific innovation asa research finding that affects the practices of many researchers in a field itseems doubtful that the new ILSA studies constitute such an innovation atall A significant impact could be observed only in Germany where the

emergence of a quantitative educational science tradition can be consideredas such a wide-ranging change of research practices But even in Germanyone could ask whether it was a research finding that brought about thechange of practices The process can be better understood as institutionalentrepreneurs exploiting a state request for professional service for buildingtheir own research community

The use of ILSA data is nevertheless growing in all four countries whichsuggests that the scientific impact of new ILSA studies on educationalscience communities still lies ahead and that innovations in the social scienceand humanities are developed more slowly not least due to the varyingnational research traditions (see also Engwall et al in this volume on aninnovation in the humanities) The slowly but inexorably growing use ofILSA data for genuine educational research and research in adjoining fieldsas well as the growing support for this research by the scientific communitiesdemonstrates that a scientific community will respond to the emergence of anew research object (however awkward) but at its own pace

1 All quotes are from interviews with education researchers from the four coun-tries Interviews with Dutch researchers were conducted in English quotes fromSwedish Swiss and German researchers are our translations

REFERENCES

Creemers B P M amp Scheerens J (1994) Developments in the educational effectiveness

research programme International Journal of Educational Research 21 125140

Ditton H (2009) Familie und Schule Eine Bestandsaufnahme der bildungssoziologischen

Schuleffektforschung von James S Coleman bis heute In R Becker (Ed) Lehrbuch

der Bildungssoziologie (pp 239256) Wiesbaden VS Verlag fur Sozialwissenschaften

Dronkers J amp Robert P (2008) Differences in scholastic achievement of public private

government-dependent and private independent school A cross-national analysis

Educational Policy 22 541577

Forsberg E amp Lindberg V (2010) Svensk forskning om bedomning en kartlaggning The

National Research Council Report 2

Gretler A (2000) Die schweizerische Bildungsforschung der Nachkriegszeit im Spiegel ihrer

Institutionen und ihrer Themen von der Geschichte zu aktuellen Fragestellungen

Schweizerische Zeitschrift fur Bildungswissenschaften 22 111144

Grisay A amp Griffin P (2006) What are the main cross-national studies In N K Ross amp

I Jurgens-Genevois (Eds) Crossnational studies of quality of education Planning their

design and managing their impact (pp 67102) Paris UNESCO Institute for

Educational Planning

Hansen M amp Lindblad S (2010) Om forskningskommunikation och publiceringsmonster

inom utbildningsvetenskap en studie av svensk utbildningsvetenskaplig forskning vid tre

larosaten The National Research Council Report 10

Keeves J P amp Lietz P (2011) The Relationship of IEA to some developments in educa-

tional research methodology and measurement during the years from 1962 to 1992 In

C Papanastasiou T Plomp amp E C Papanastasiou (Eds) IEA 19582008 50 years

of experiences and memories (pp 217252) Nicosia Cultural Center of the Kykkos

Monastery

Kuyper H amp van der Werf G (2012) Excellente leerlingen in het voortgezet onderwijs

Schoolloopbanen risicofactoren en keuzen GION Universiteit Groningen

Luukkonen T Persson O amp Sivertsen G (1992) Understanding patterns of international

scientific collaboration Science Technology amp Human Values 17 101126

Mayntz R (2001) Die Bestimmung von Forschungsthemen in Max-Planck-Instituten

im Spannungsfeld wissenschaftlicher und ausserwissenschaftlicher Interessen Ein

Forschungsbericht Koln Max-Planck-Institut fur Gesellschaftsforschung

and colour In K D Knorr R Krohn amp R Whitley (Eds) The social process of

scientific investigation (pp 107138) Dordrecht Reidel

Polanyi M (1962) The republic of science Minerva 1 5473

Reguant S amp Casadella J (1994) English as Lingua Franca in geological scientific

publications a bibliometric analysis Scientometrics 29 335351

Van den Daele W Krohn W amp Weingart P (Eds) (1979) Geplante Forschung Vergleichende

Studien uber den Einfluszlig politischer Programme auf die Wissenschaftsentwicklung

FrankfurtMain Suhrkamp

WHERE TO GO FOR A CHANGE

THE IMPACT OF AUTHORITY

STRUCTURES IN UNIVERSITIES

AND PUBLIC RESEARCH

INSTITUTES ON CHANGES OF

RESEARCH PRACTICES

Jochen Glaser Enno Aljets Eric Lettkemann and

Grit Laudel

ABSTRACT

In this article we analyse how variations in organisational conditions forresearch affect researchersrsquo opportunities for changing individual-level orgroup-level research programmes We contrast three innovations thatwere developed in universities and public research institutes in Germanyand the Netherlands which enables comparisons both between organisa-tional settings and between properties of innovations Comparing thedevelopment of three innovations in the two types of organisationsenables the identification of links between patterns of authority sharing

ISSN 0733-558Xdoi101108S0733-558X20140000042010

at these organisations and the opportunities to develop innovations Onthis basis the distribution of opportunities to change research practicesamong researchers in the two countries can be established

Keywords Scientific innovations universities public research insti-tutes Germany Netherlands authority distribution

DO RESEARCH ORGANISATIONS MATTER

Research organisations significantly contribute to the opportunities forresearchers to change their research topics or approaches Although organi-sations can exercise only very limited control over the content of theiracademicsrsquo research (Musselin 2007 Whitley 2008 pp 3536 Whitley ampGlaser this volume) their role as interfaces between society and scientificcommunities is widely acknowledged Organisations provide salaries forresearchers and resources needed for research They communicate societyrsquosexpectations to researchers and useful knowledge as well as cultural contri-butions by researchers to society They are therefore important governanceactors in their own right mediators between policy and research and envir-onments for research When researchers decide on their future researchthey take into account their current conditions of research as well as thosethey can anticipate for the relevant future and many of these conditionsare shaped by their research organisations

Thus while organisations cannot easily make researchers change theirtopics they can make such changes more or less likely While this influencehas been investigated by a number of research strands there still is a dearthof knowledge about the influence of specific organisational settings onchanges in research across different fields of knowledge Conditionsenabling changes in research vary between and within organisations as dothe necessary conditions for such changes in different fields

A prime example of this variation is the relationship between organisa-tional funding and external project-based funding of research Since thefunding of science in most countries has shifted from recurrent to project-based competitive funding to an extent that makes nearly all empiricalresearch dependent on external grants it is difficult to compare the differ-ential impact of the two modes of funding on the change of researchprogrammes and practices Recurrent and project-based funding can becompared however when public research institutes based on recurrent

funding and universities whose researchers depend on project-based fund-ing are active in the same field

In this article we consider how variations in organisational conditionsfor research affect researchersrsquo opportunities for changing individual-levelor group-level research programmes We take advantage of the compara-tive approach of the RHESI project (see the introduction by Whitley andGlaser in this volume) which studied the development of four innovationsin four countries Three of these innovations were developed in differenttypes of organisations which enables comparisons both between organisa-tional settings and between properties of innovations

In the remainder of this article we first present our theoreticaland empirical approach We then compare the development of three inno-vations in different types of organisations and identify links betweenpatterns of authority sharing at these organisations and the opportunitiesto develop innovations Conclusions concern the distribution of opportu-nities to change research practices among researchers in the two countriesadvantages and disadvantages of competitive grant funding and thespecific roles of different types of organisations in the development of scien-tific innovations

APPROACH

Theoretical Background

The influence of organisations on the research conducted in them has beeninvestigated from several perspectives Higher education research is mainlyinterested in the changing governance of and within universities whilescience policy studies focus on the governance of research organisationsBoth fields overlap and share a focus on the increasing autonomy of univer-sities and new capabilities to allocate resources invest in research and teach-ing programmes and recruit personnel (eg Bleiklie amp Kogan 2007 Clark1998 Louvel 2010 Marginson amp Considine 2000 Sanz-Menendez ampCruz-Castro 2003) The few studies of the impact of such changes on thecontent of research emphasised disciplinary rather than organisationaldifferences between effects (Glaser Lange Laudel amp Schimank 2010Leisyte 2007) or are concerned with the consequences for research perfor-mance (Jansen 2010)

299Where to Go for a Change

Constructivist studies in the sociology of science have observed theimportance of organisational contexts for research but typically beingsingle-case studies considered only very general aspects of organisationalcontexts such as the control of organisational resources (Knorr-Cetina1981) or the reduction of complexity in hierarchical decision-making(Vaughan 1999)

There is also a research tradition that directly addresses the impact oforganisational conditions on research productivity Early studies tried tolink publication productivity to variables such as group size and groupstructures (Bland amp Ruffin 1992 Knorr amp Mittermeir 1980 McCarrey ampEdwards 1973 Pelz amp Andrews 1966) and investigated the relationshipbetween organisational control and the autonomy of research (Blau 1973Kaplan 1964 Lambright amp Teich 1981 Zabusky amp Barley 1997) Thistradition has been revived by studies that took exceptionally successfulresearch as their starting point and investigated organisational conditionsfor such research

In his study of major discoveries in the biomedical sciences for exampleHollingsworth (2008 p 231) found that lsquomajor discoveries tended to occurmore frequently in organizational contexts that were relatively small andhad high degrees of autonomy flexibility and the capacity to adapt rapidlyto the fast pace of change in the global environment of sciencersquo This wasconfirmed by Heinze Shapira Rogers and Senker (2009) who additionallyfound lsquocreative accomplishmentsrsquo to be associated with lsquosmall group sizeorganizational contexts with sufficient access to a complementary varietyof technical skills stable research sponsorship timely access to extramuralskills and resources and facilitating leadershiprsquo (ibid p 610)

This brief overview highlights the point that only a few of the many stu-dies of research organisations employ a comparative perspective on organi-sational contexts or kinds of research conducted in them Comparingorganisational contexts for research requires disentangling a complex net ofmulti-level causation In their decisions about research researchers musttake into account not only conditions produced by their organisations butalso many influences generating in the organisationrsquos environment some ofwhich lsquobypassrsquo the organisation and address researchers directly Theseinclude interests and preferences of scientific communities funding organi-sations and collaborators as well as legal frameworks and societal normsconcerning research In order to render these influences comparable and toseparate them analytically from organisational influences we apply twoframeworks one for comparing the impact of governance and one forcomparing the micro-conditions for changes of research practices

The differential impact of governance arrangements for researchers indifferent organisational environments is compared by analysing authorityrelations concerning the formulation of research goals (Glaser 2010Whitley 2010) An actorrsquos authority is defined here as institutionally shapedinfluence (Schimank forthcoming) Authority relations then encompassthe relative authority that can be exercised by members of a configurationof interdependent actors with regard to a specific subject matter (in ourcase the formulation of research goals) This notion is both more specificthan concepts of governance insofar it focuses on actors (authoritativeagencies) and uses institutional structures and processes of governance aslsquobackground informationrsquo on how authority is produced and exercised

At the same time it is more inclusive than much empirical research ongovernance because it always includes all actors who have authority con-cerning a specific decision process regardless of their inclusion in particulargovernance instruments This enables the analytical integration of allgovernance processes relevant to a researcherrsquos decisions Since bases ofauthority and the exercise of authority are limited to a few channels thecontrol of resources reputation and career opportunities the relativeauthority of actors can be assessed and a comparative framework beformed This enables comparisons which in our case means comparinghow universities and research institutes are embedded in and modifyauthority relations concerning the formulation of research goals In thesystem of authority relations research organisations are both an authorita-tive agency in their own right (an actor with specific interests) andlsquochannelsrsquo for the authority of the organisationsrsquo external stakeholders(Schimank forthcoming)

Our dependent variable is the change of research practices which weunderstand as types of actions aimed at producing new knowledge that arecharacterised by specific theoretical frameworks objects methods andobjectives With this definition we attempt to distinguish between the con-tinuous adaptation of actions to the circumstances of producing new knowl-edge (Knorr-Cetina 1981) and changes of one or more of the constitutiveelements of a research process (problems empirical objects or methods)The latter are usually more consequential because they extend beyond a sin-gle research project and create new trajectories for research

Examples of such significant and enduring changes of research practicesthat diffused in scientific communities include the diffusion of the molecu-lar biology paradigm and practices in cancer research (Fujimura 1988) thecreation of monoclonal antibodies and their subsequent widespread use inthe biosciences (Cambrosio amp Keating 1995) the adoption of a new theory

as an explanation of new particles observed in high-energy physics experi-ments (Pickering 1980) and the emergence of new specialties followingfrom the adoption of new practices (see Edge amp Mulkay 1976 for an over-view) These examples have in common that in response to new findingsresearchers changed one of the constituting features of their knowledgeproduction and headed off on new lines of research or lsquoresearch trailsrsquo(sequences of thematically linked projects see Chubin amp Connolly 1982)Since they affected the practices of many members of at least one scientificcommunity we call them scientific innovations

Changes in research practices are very difficult to compare across fieldsof research due to their idiosyncratic content However they can be com-pared in a framework that focuses on the requirements of such changesChanges in research practices incur costs and may be risky in severalrespects They incur costs because

bull they partly devalue the knowledge and equipment a researcher has accu-mulated working on previous topics

bull a researcherrsquos reputation may suffer if the change requires learning orexperimental redesigns and thereby delays opportunities to publishresults and

bull the new line of research may deviate from the mainstream of the researcherrsquoscommunity which again creates the risk of losing reputation

We use the concept lsquoprotected spacersquo for comparing the opportunitiesfor researchers to pay these costs Building on Whitley (2012 and in thisvolume) while adapting his definition for the purposes of our empiricalinvestigation we define protected space as the planning horizon for which aresearcher can autonomously apply his or her capabilities to a self-assignedtask1 Dimensions of this variable are the time horizon for which the cap-abilities are at the sole discretion of the researcher (the time horizon forwhich the researcher is protected from direct external intervention into hisor her epistemic decisions and external decisions on the amount of capabil-ities) and the resources (including personnel over which the researcher hasauthority and time for research) Protected space can be constructed byresearchers in a variety of ways which include interactions with universitiesand state-funded research institutes and other authoritative agencies Thismakes it possible to link the specific protected space required by a changeof research practices to the specific authority relations centred on differentorganisational settings

Authority relations can be compared according to the protected spaceresearchers can build within them This includes the shape of protected

space (extensions in the time and resource dimensions) and its scopeie the number and positions of researchers who can build protected spacethat is large enough to enable a change of lines of research Our investiga-tion focuses on the different shapes of protected space (their extension inthe two dimensions) that occur in two types of organisations namely uni-versities and state-funded research institutes

Empirical Approach

Our comparative approach includes conditions provided by universitiesand state-funded research institutes in two countries for three differentscientific innovations For this comparison we could utilise a sub-sampleof the cases selected for the RHESI project Opportunities for cross-country comparisons were limited by the different role of state-fundednon-university research institutes in the four national science systems andwithin those in different fields of research Among the countries studied bythe RHESI project only the German and the Dutch national sciencesystems feature state-funded public research institutes and thus enable thecomparison

These two public science systems included in our empirical analysis dif-fer considerably in both size and structure as is illustrated by Table 1 TheGerman science system is not only much larger but also features a strongstate-funded non-university research sector In the Netherlands the shareof state-funded non-university research is much smaller Both countrieshave in common their reliance of experimental research at universities onresearch council funding Most university research and some research atthe research institutes depends on external grants

The two countries feature different academic career patterns Germanyhas a chair system in which all positions below the professorial level arefixed-term positions The Dutch system has permanent positions below theprofessorial level Finally the Dutch higher education reforms since theearly 1980s have considerably changed authority structures in universitiesby transferring the authority over budgets and staff to universities andstrengthening the authority of university managers vis-a-vis academics InGermany these reforms began more than a decade later While the levelsof autonomy of universities and the authority of university managementhave increased here they are still much lower than in the Netherlands

We compare changes in research practices that occurred in threedifferent fields namely the experimental production of Bose-Einstein

condensates (BECs) in physics evolutionary developmental biology (evo-devo) and international comparative large-scale assessments of studentachievements (ILSA) in educational sciences (PISA being the best-knownexample) Their key features are summarised in Table 2

Bose-Einstein condensates are a new state of matter that occurs inatomic gases at temperatures very near to absolute zero (lt100 nanokelvin)BECs of cold atom gases were first produced in 1995 by combining severalrecently developed cooling techniques The diffusion of the experimentalproduction of BECs began two years later The production of BECs origin-ally required considerable protected space Researchers needed to learn sev-eral cooling techniques and to combine them in a new experimental

Table 1 Key Features of the Two Science Systems Whose OrganisationsWere Compared

Features of Science Systems Germany Netherlands

Size of university sector

(billion h)a118 42

Size of public non-university

research sector (billion h)a73 13

External funding Essential most important

funding agency is the Deutsche

Forschungsgemeinschaft (DFG)

Essential most important

funding agencies are the

Nederlandse Organisatie voor

Wetenschappelijk Onderzoek

(NWO) and its Stichting voor

Fundamenteel Onderzoek der

Materie (FOM) for physics

Key academic positions at

universities

Professors (tenured)

Assistants (fixed-term)

Professor Universitair Docent

Universitair Hoofddocent (all

tenured)

research institutes

Director and head of

department (tenured)

Researcher (fixed-term)

Director (tenured)

Group leaders (five-year

tenure track)

Researchers (fixed-term)

Progress of higher education

reforms

Since the mid-1990s limited

redistribution of authority

Since the early 1980s

significant redistribution of

authority

aYear 2009 sources Federal Ministry of Education and Research (2012 p 58) and Ministerie

van Onderwijs Cultuur en Wetenschappen (2012 p 25)

setting The equipment was (and still is) expensive costing about 500000Euros Building and fine-tuning the experimental system took a long time(at least one and often two years) and required the coordinated work of atleast two people During this time only few intermediate results could bepublished As a consequence researchers beginning the production ofBECs faced the reputational risk of not having significant publications fora long time if the experiments failed Thus researchers who wanted to pro-duce BECs needed control of a research group whose research they coulddirect of laboratory space in which they could set up their experiment andof significant additional resources (between half a million and one millionEuros) for at least three years

Evo-devo is a highly heterogeneous life science field which evolvedaround a set of concepts and questions that explore links between the evo-lution of a species (the subject matter of evolutionary genetics) and theembryonic development of its individuals (the subject matter of develop-mental biology) It provides a new perspective on existing data and a dis-tinct set of research questions that require comparative experimentalresearch on genetically different organisms preferably from two differentspecies This comparative approach poses a major challenge because mostresearchers traditionally work with just one species For the purposes of thiscomparison we reduce the complex pattern of changing research practices thatis described by Laudel et al in this volume to two strategies The low-risk

Table 2 Key Features of the Three Innovations Compared

Properties of

Innovations

Bose-Einstein

Condensation

Evolutionary

Developmental

Biology

International Large-Scale

Student Assessments

Resources required sim500000 euros Up to 1 million

Up to 1 million euros

Time horizon Long and

unpredictable

Long and

unpredictable

Short and predictable

Nature of risks Experimental

failure

Experimental failure No recognition of

contributions in scientific

community

German organisational

contexts

Universities and public research institutes

Dutch organisational

contexts

Universities and public research

institutes

Only universities

strategy entails conducting experiments with the methods and the species theresearcher is familiar with and adding a comparative perspective by obtaininginformation on other species from the literature through collaboration or byincluding model organisms that are already well understood The protectedspace required for this strategy is relatively small but so are the possiblereputational gains The high-risk strategy aims at establishing new species asmodel organisms and developing new methods This strategy requires largeprotected space but also promises higher reputational gains

Since ILSA cannot be considered an innovation in the Dutch contextand was conducted exclusively at universities in the Netherlands (seeGlaser et al this volume) only German researchers at universities andinstitutes can be included in the comparison Although ILSA have beenconducted for a long time by various international groups changes in theirmethodology and organisations significantly improved the comparability ofnational results and thus mark the early 1990s as an innovative period Inthis period German education policy-makers decided to renew Germanyrsquosparticipation in ILSA studies thereby initiating a process that movedquantitative studies to the core of German educational sciences This situa-tion implied a major reputational risk for those taking up ILSA Since theGerman educational science community did not consider quantitative stu-dies as a legitimate way of doing educational research it was difficult to seehow a standard academic career could be built with producing ILSA dataAt the same time ILSA studies incurred significant material costs Fundingthe large groups necessary for the German part of such a study requiredsignificant amounts of money (around one million euros)

Taking up ILSA also required being included in one of the internationalnetworks conducting the international comparative studies which in turnstrongly depended on having built a reputation and a personal network inthe field Thus researchers who wanted to be involved in the German partof an ILSA study not only had to cope with a strong unfavourable majorityopinion of their community In order to lead the German part of such astudy they needed to be well connected internationally and to be able toraise funding that was unheard of in this field

Our analysis is based on documents describing the changes of lines ofresearch at the international and national field levels interviews withresearchers who worked with the innovations or attempted to do so in uni-versities and public research institutes and background interviews withobservers of the fields officers of funding organisations and universitymanagers The distribution of these interviews is shown in Table 32

Interviews with researchers focused on the position of the epistemicswitch in the intervieweersquos research biographies and on the necessary andactual conditions for that switch Owing to the small size of the Dutchscientific communities all researchers who took up BEC and evo-devoin the Netherlands could be approached and interviewed In Germany weapproached all researchers who could be identified as having taken up BECin the 1990s all researchers who led projects producing ILSA data sincethe early 1990s and a sample of researchers who began evo-devo researchearly (in the 1980s or 1990s) Two German BEC researchers declined to beinterviewed

Table 3 Interviews Conducted in the Two Countries

Innovations Germany Netherlands

Bose-Einstein condensation 9 11

with researchers 8 7

Evolutionary developmental biology 8 16

International large-scale student assessments 15 Not included

with researchers 11

Table 4 Number of Research Situations Described in the Interviews

Innovations Research Situations

In Universities In Research Institutes

Germany Netherlands Germany Netherlands

Bose-Einstein condensation 9 6 4 1

changes of research practices 8 5 2 1

Evolutionary developmental biology 6 12 4 1

International large-scale student

assessments

14 Not

included

16 Not

included

changes of research practices 4 8

The cases that form the basis of our analysis are distinct situations inwhich our interviewees built protected space at universities or publicresearch institutes The research situations include those in which intervie-wees attempted to change their research practices by moving to one of ourthree innovations and those in which they already worked with the inno-vations and to which they referred for comparisons (Table 4) The num-ber of these situations often exceeds the number of interviewees Themobility of researchers included moves between organisations (eg from aresearch institute to a university) and moves between key academic posi-tions (eg if a German assistant or a Dutch Universitair Hoofdocentbecame professor) This is why the number of cases we could analyse islarger than the number of interviews conducted Not all attempts tochange research practices were successful The situations include research-ers in all key academic positions at both types of research organisationswhich is why we are confident that the typology of research situationsthey described is exhaustive

BEGINNING NEW LINES OF RESEARCH IN

DIFFERENT ORGANISATIONAL SETTINGS

Bose-Einstein Condensation

In spite of the differences between German and Dutch universities thesituations of university researchers who wanted to move to BEC researchin the two countries were quite similar German university professors couldconstruct the necessary protected space if they already had the basic infra-structure for producing BECs namely a laboratory that was equipped forcooling atoms or working with lasers The first two German researcherswho produced BECs were university professors who had part of the neces-sary infrastructure already in place because their previous research wasmethodologically similar to BEC Two other researchers had used theirpostdoctoral positions abroad to learn the methods for producing BECsWhen they returned to Germany and were appointed professors they usedtheir start-up packages to equip their laboratories for producing BECsFour other professors used loyalty negotiations to acquire the funding forthe new equipment3 In all cases start-up or loyalty packages boughtthe major equipment and recurrent funding for one or two positionswhich almost covered the workforce requirements for producing a BEC

However recurrent funding was never sufficient to purchase the expensiveequipment or to pay the running costs of the laboratory

Irsquove got something like one million start-up funds from my appointment negotiations

which was relatively high but the recurrent resources are negligible It is about 40000

Euros Half of it 20000 Euros is spent for the machine workshop telephone and

copies something you need anyway These 20000 Euros are a tiny fraction of what it

[running the equipment] actually costs (university professor)

The costs of actually running the equipment (which included costs ofadditional personnel) had to be covered by external funding This created acontradiction between the standard time span of grants (three years) andthe time span necessary to produce BEC which was impossible topredict and often was longer than three years However it turned out thatintermediate results were accepted by the reviewers of the DFG and otherfunding agencies This attitude guaranteed continuing access to grantswhich compensated for grant cycles that were too short for producing BEC(see Laudel et al this volume Glaser et al 2014) This willingness of fund-ing agencies to support BEC continued unchanged Nevertheless the lsquograntbusinessrsquo could delay the progress of research for months This endangeredthe position of German university professors in the international competi-tion which in some cases took the form of lsquoracesrsquo for BECs of specificatoms

Given this description of situations of German university professors itcomes as no surprise that only one researcher below the professorial levelat a university tried to switch to BEC At this time he had no discretionover recurrent funding or PhD students who were all hired and supervisedby professors The overall time horizon of his research was determined by afixed-term position of five years This small protected space increased in theresource dimension after the researcher received a large grant for juniorresearch group leaders from DFG

Then I got an Emmy Noether grant of about 350000 Euros Ultimately that was the

lionrsquos share of the material investments The whole infrastructure was provided by [my

professor] I also had a PhD student from his university budget

Although this grant bought the (low-budget) equipment moving toBEC experiments still depended on the permission and support by hisprofessor It is impossible to say whether the protected space was sufficientin the time dimension because the researcher was appointed professor atanother university before he achieved BEC as a junior research groupleader He used his start-up package from the new university to build anew experimental system for BEC

Despite the different university structures the description of theGerman situation applies to Dutch BEC experiments too Although Dutchacademics below the professorial level have tenured positions they have nocontrol of the technical infrastructure for research or access to PhDstudents Professors control the infrastructure of laboratories and shareauthority over recurrent funding and PhD students with faculty manage-ment In the case of BEC this control mattered because as in Germany allmoves to the production of BECs at universities ultimately depended ondecisions by professors

The situation of Dutch university professors also resembled that of theirGerman colleagues in that they depended on external funding whenbuilding their BEC experiments Their recurrent funding paid for the basicinfrastructure of a laboratory one or two PhD positions and technicalworkshops at the universities that could be used to custom-build experi-mental equipment Specific equipment consumables and additional person-nel had to be funded from external grants The time horizons provided bythe grants were better (four years) but the communityrsquos tolerance of delaysin outcomes was lower (see Laudel et al this volume)

The authority of professors over the resources provided by the universitybecomes apparent if we look at three academics below the professorial levelwho attempted to begin the production of BEC Two of them became inter-ested in BECs soon after the first experimental successes but could not pur-sue this interest because their professors did not approve One of them wasin a situation that exactly mirrored that of the German junior researchgroup leader quoted above He received a large grant that paid most of thecosts of BEC research but was not large enough to pay for the infrastruc-ture too Only when he was appointed on an independent tenure-trackposition and received a start-up package from his university could hefinally undertake the production of BECs A third academic could changehis research because his professor tolerated the change and granted accessto his infrastructure for these experiments Although two of the non-professorial researchers had tenured positions and thus a long-time horizonfor their research and all three had the opportunity to independently applyfor external funding their dependence on their professorrsquos infrastructurewas decisive for their opportunities to build the protected space necessaryfor the production of BECs

Directors at German public research institutes (directors of the institutesand heads of departments) had in common with university professors theircomplete freedom of choosing research topics and the long-time horizonsof their permanent positions They had a considerable advantage over their

colleagues at universities however because the recurrent funding of theinstitutes was sufficient to conduct BEC research without any additionalexternal grants While the same amount of resources can be amassed byuniversity professors from external grants the recurrent funding saves timeand lowers the risk of interruptions by failed grant applications as aresearcher who worked both at a university and public research instituteexplained

Letrsquos put it this way with regard to my financial situation I actually had the same finan-

cial resources I do not have more money now than I had for example at [university A]

I had excellent funding [at university A] which was as good as now [in the public research

institute] You can have in Germany in Europe an outstanding financial support But of

course I had to do much more for it I spent a lot more of my time on administrative

work The time I spent on writing applications I can now spend on physics

Among the three interviewed directors at public research institutes wasone who had started BEC research as PhD student and followed this lineof research over his whole career ie did not need to change his researchpractices A second director changed his research to BEC without needingto acquire external funding A third director did not change his owndirection but acted as sponsor ie enabled the change of research by aresearcher in his department This researcherrsquos protected space was stillsmall because of the missing autonomy and limited discretion overresources He had to ask for money but was supported by this director Heimmediately received some money from the directorrsquos recurrent budget foraltering the experimental system

Well more money hellip went into [the directorrsquos] main activity hellip But there was some

surplus money And that we could kind of use (or convince him to by) this and that So

I remember that at some stage I wrote a one-page letter to [the director] whether he

would have 50000 Deutsche Mark I suppose and then I could change my experiment

to a BEC experiment

Since the money available ad hoc from the directorrsquos budget was not suf-ficient the researcher had to apply for additional grants for the necessarymodifications of the experimental system and for recruiting an additionalPhD student However he could immediately begin to reconfigure theexisting equipment while he waited for the grant to be approved

The situation at the Dutch public research institute was markedly differentThe institute was funded by the funding agency for the physical sciencesFOM with the mission to explore new directions of research Its missionand themes are controlled by the Dutch physics community and funding ispartly conditional on the approval of research programmes by FOM The

director of the institute is primarily a manager whose time for research islimited The institute consists of research groups that are headed by research-ers who are appointed to five-year tenure-track positions They rarely stayat the institute for more than 10 years because they move to universityprofessorships

In 1996 the elite of the physics community which was represented inthe FOM board decided to reduce research in nuclear physics and toestablish biophysics as new research field at the institute Biophysics wasconsidered underdeveloped in the Netherlands at that time The boardexpected half of the institutersquos research to be biophysics by 2006 As a con-sequence of this thematic reorientation the director of the institute aninternationally leading BEC researcher decided to keep only a small groupfor BEC research and to orient the other groups towards biophysics

While the director could use the excellent infrastructure and basic equip-ment provided by the institute the institutersquos recurrent funding was notintended to be the sole basis for research and was not sufficient to producethe protected space for producing BECs The director and group leadershad to apply for grants with FOM the same agency that provided the insti-tutional funding for the institute

And [the institute] hellip normally the groups were relatively small maybe three or four

PhD students per staff that was the maximum On my CV it would say external grants

And I would say half was effectively hellip internal grants and half was really earned on

my merit and writing something original

Evolutionary Developmental Biology

The possibility of entering the evo-devo field with low-risk strategies letmany researchers at universities gradually move to evo-devo by stepwiseincreasing their commitment (see Laudel et al this volume) Most of themalso kept other lines of research besides evo-devo This was in principlepossible for professors as well as researchers below the professorial levelFor example two German interviewees moved to evo-devo after they werealready appointed as university professors and have mainly worked onlow-risk projects ever since Another researcher began evo-devo researchwhen he was a postdoc but decided to move to the high-risk strategy whenhe was appointed to his first professorial position The decision went handin hand with the winning of a research prize that substantially increased hisstart-up funds

German professors who wanted to conduct experimental evo-devoresearch needed start-up or loyalty packages for establishing an evo-devolaboratory and for establishing breeding facilities regardless of the riskinessof their strategies Owing to the modularity of the research problems andexperiments (which is different from the production of BECs) researcherscould partially overcome limitations of their own resources by sharingequipment and staff with neighbouring research groups The large univer-sity landscape of Germany made access to a wide range of data sourcesandor expensive equipment relatively easy However the high costsinvolved in developing the high-risk strategy of evo-devo research causedby the required breeding facilities for multiple new species appear to haveexceeded the opportunities of university start-up and loyalty packages Wefound no university professor who engaged in such a strategy (see Laudelet al this volume)

Even for low-risk experimental evo-devo research all university profes-sors were dependent on external funding sources (prizes project grantsetc) to create sufficient protected space for experimental evo-devoUniversitiesrsquo recurrent funding is usually not sufficient to pay even themaintenance of the laboratory

Well the university supported us here so to speak by paying electricity paying the tel-

ephone paying for the heating and for the paper for the copy machine (university

professor)

All German researchers below the professorial level who changed theirresearch practices started their evo-devo careers by adopting model organ-isms and tools that had already proven their usefulness in other specialtiesTheir cases indicate that starting evo-devo is possible in early career stagesBut again all PhD students and postdoctoral fellows relied more or less onthe goodwill of their supervisors or heads of laboratory who had to grantthem the use of infrastructure for their evo-devo research

Despite the shortages of recurrent funding the evo-devo field is growingfast since the late 1990s One interviewee estimates around 500 researchers inGermany who actively participate in todayrsquos evo-devo conferences The rea-son given by all interviewees for the lsquoexplosionrsquo of evo-devo approaches isthe funding policy of the DFG which still provides funding for investigator-driven basic research despite global and national trends of expecting medicalapplications from the life sciences

It is not very easy in the evo-devo field because it is not medical not applied research

It is always easier to get money for these areas because there exist pre-formulated

research programmes for which you can apply we do not fit into these We always

have to apply for programmes that provide complete freedom in terms of content

Classically this is provided only by the DFGrsquos Individual Grants Programme

In the Netherlands the opportunities for building protected space atuniversities are again similar to those in Germany Owing to the variety ofstrategies with which costs and risks of evo-devo research can be con-trolled both professors and other researchers moved to evo-devo researchat various stages (see Laudel et al this volume) After evo-devo was origin-ally considered important enough by the community to warrant the crea-tion of a chair at one university increasingly specific demands of statepolicies concerning university profiles changing criteria for grant fundingand the strong authority of Dutch university management worsened thesituation of evo-devo researchers at universities from the mid-2000s

The basic research character of evo-devo research its limited access tohighly ranked journals the resulting difficulties to attract grant fundingand its distance from the national priorities for research made evo-devoresearch unattractive for universities The problems with research fundingcan be traced back to new priorities of the Dutch national research councilNWO Grant funding from NWO has been losing its independence frompolitical expectations Today an orientation towards applications andnational priorities is increasingly important for proposals to be successfulAs a result evo-devo research has largely disappeared at Dutch universi-ties and only a few scattered researchers engaged in some kind of evo-devoresearch remain (see Laudel amp Weyer this volume)

German researchers at public research institutes could often move toevo-devo research without applying for external funding at all and wereable to apply high-risk strategies This happened in four cases reported byour interviewees Two researchers decided to move to evo-devo researchwhen they were directors at institutes while the others had already movedto evo-devo at their universities and brought their evo-devo research withthem when they were appointed directors at institutes Of the two directorswho changed their research one recruited a researcher for this purposeThis researcher was the only German interviewee who chose to enter theevo-devo field with a high-risk strategy The risk remained manageablebecause he was a member of a large research group and was backed by hisdirector The other director strongly supported the move to evo-devo bygroups at his institute In both cases no external funds were necessary

All four research institutes have an excellent infrastructure and directorscan direct the research of several staff members (technicians PhD studentsand postdocs) Researchers at public research institutes usually followedmore risky and time-consuming research lines than their colleagues at

universities The relatively large group size (at least 20 researchers) allowsdirectors to spread the risks of developing new tools andor model organ-isms among several individuals

The dilemma you are confronted with is If it [the experiment] works the person who is

responsible got a project If it doesnrsquot work you try for some time but in this case the

researcher needs a back-up project to successfully finish his PhD or to get some results

from his postdoctoral work (director research institute)

Such lsquoback-up projectsrsquo are very common in large research institutes Theyexist at universities as well but here lsquoback-up projectsrsquo depend on the success-ful acquisition of several parallel grants Another advantage of researchinstitutes is that directors and heads of laboratories are not bound to theshort-term time horizons of project grants On the basis of large and contin-ual amounts of recurrent funding they can plan high-cost research lines thatwill pay off in terms of reputation only after five or even ten years For exam-ple some new ecological approaches in evo-devo require expeditions tocollect new organisms and the establishment of field laboratories abroadThese tasks can be afforded by research institutes but not by universityprofessors who depend on the Individual Grants Programme of the DFG

In the Netherlands one public research institute is a potential host forevo-devo research Authority relations at this institute which is funded bythe Dutch Academy of Science are similar to those of Max PlanckInstitutes and some Leibniz Institutes in that the directors of the institutecan autonomously determine the institutersquos research agenda Howeversimilar to the physics institute funded by FOM the bioscience institute pro-vides only limited recurrent funding Group leaders receive funding fortechnical support and one PhD position and are expected to apply forexternal grants for their research Again the situation of group leaders issimilar to those of university professors

The Dutch institute has a strong tradition in developmental researchand was therefore a potential host for evo-devo research However thedirectors of this institute decided not to include evo-devo research in theirresearch agenda and recruited more and more groups from other researchareas Thus the basis for evo-devo research gradually vanished

International Large-Scale Student Assessments

In the early 1990s there were few professors at German universities whoconducted quantitative empirical educational research The study of theconditions leading to educational success was dominated by qualitative

approaches derived from the humanities and neglected the measuring ofcompetences (cf Fend 2010 pp 286290 Zedler amp Dobert 2009pp 2428) As a consequence the first researchers who wanted to changetheir line of research towards ILSA were also pioneers of quantitativecomparative empirical educational research in Germany It was only inresponse to the outcomes of the first German ILSA studies in the late1990s that more professorships for (quantitative) empirical educationalresearch were created a process which has continued ever sinceThese professorships provided the protected space for the change ofresearch to ILSA and secured careers in the context of reputational risksSince a professorrsquos infrastructure was limited the large investments neces-sary for ILSA research projects could only be made with external fundingwhich in the case of ILSA was provided by the state Accessing this exter-nal funding required seed-funding for preparatory activities which put uni-versity professors at a disadvantage

hellip these 80000 Euros without which you could not have done the preparatory work

within the international network All these things had to be decided and done relatively

quickly If I had had an application processes with the EU and external peer review and

BMBF and such things nothing of this would have happened hellip International things

sometimes run very fast and a decision must be made somehow within one or two

weeks Then you are either in or out

When the professor who described this problem first applied for leadinga study he solved the problem of missing seed-funding by asking colleaguesfrom a public research institute to lsquolendrsquo him the necessary money He thencould acquire the external funding for the study from the state ie by cir-cumventing the peer review of the potentially hostile educational sciencecommunity The state was willing to grant funding because it was interestedin the results of ILSA studies With these grants university professorscould expand their research capacity Thus university professors simulta-neously had protected space that was small in the resource dimension(recurrent funding) and long in its time horizon (permanent autonomousposition) and protected space created by grants that was somewhat largerin the resource dimension but limited in its time horizon

At the beginning of the diffusion of ILSA studies in Germany no uni-versity positions below the professorial level were available for work onthis topic The established educational research professors did not hireassistants for ILSA studies because they disliked this kind of research As aconsequence no protected space was available for non-professorial aca-demics and no change of lines of research was possible for them4

Given the conditions at universities it comes as no surprise that ILSAstudies were first taken up by German public research institutes Three ofthese institutes had been set up with the mission to conduct empirical edu-cational research which made them the natural place to conduct ILSA stu-dies and gave the directors at the institutes little opportunity to declinesuch an involvement even if they wanted to (see Glaser et al this volume)

The directors at public research institutes who changed their lines ofresearch benefited from a large protected space with sufficient resources andlong-time horizons They had authority over significant recurrent fundingand positions which made it possible for them to form the large consortiarequired for conducting ILSA studies in Germany The seed money for pre-paring the participation in future studies was easily available from recurrentfunding The external funding for conducting the studies could be acquiredfrom the state due to the latterrsquos interest in the results and due to the reputa-tion of the institute Regular contacts with national and international elitesare easy to maintain for a director at an institute A researcher who workedin both types of organisations highlighted the differences

I can just easily without thinking about it fly to America with two people and can get

the information from an internal meeting of the organisation that conducts the study

That costs a couple of thousand Euros and I donrsquot have to think about it A normally

funded professor could not do all this This is easy for someone who has a research

institute but close to impossible for someone with a regular university chair who can-

not do this kind of studies

For the directors institutes provided much better conditions for switch-ing than universities did for professors Especially the university professorsrsquolack of access to seed-funding which is necessary to fit into the tight inter-national time schedule of ILSA studies created a huge barrier for ILSAresearch at universities Consequently one director at a public researchinstitute declined a professorial appointment at a university He wanted tomove to that university for personal reasons but ultimately didnrsquot becausehe believed his ongoing ILSA research could only be conducted at a publicresearch institute

That was a very hot time where I thought I could not take this with me that would not

work You can do this only in [institute] with these resources

From the other perspective a university professor explains why heceased his involvement in large studies

hellip these large studies have grown to a size that you cannot handle as a normal univer-

sity professor with two or three assistants You need a lot of additional resources hellip

The protected space for staff at public research institutes was not limitedin the resource or networking dimensions Working on full-time researchpositions researchers below the level of a director had a large researchcapacity as well as opportunities to learn about the new methods and newresearch objects Although they depended on the institutersquos director theywere provided with sufficient resources because the director decided toestablish the group

And there was really no restriction of resources When I arrived there I asked [the insti-

tutersquos director] ldquoWhat about the budget for the libraryrdquo He looked surprised and said

ldquoHow What You donrsquot have a budget What you want to buy you simply buyrdquo

And then I asked ldquoWhat about student assistants How much money can I spend on

them How many thousandsrdquo ldquoWe donrsquot have budgets You get what you needrdquo

The time horizon of the protected space of researchers was limited byfixed-term positions Even researchers on permanent positions below thelevel of a head of department were expected to leave the research instituteafter the habilitation and to move to a professorial position at a university

Researchers at public institutes who were not directors could not auton-omously choose their research topics They had no discretion over recur-rent funding and few opportunities to acquire grant funding independentlyOwing to their fixed-term positions their opportunities to conduct researchthat is risky in reputational terms (delayed publications no publications atall) were very limited If they wanted to conduct ILSA studies howeverthey had access to the necessary resources and international contacts Thisis why research institutes might not have provided better conditions forsolitary changes of research practices but certainly provided better condi-tions for such changes when the new practices already were represented by adirector

PATTERNS OF AUTHORITY SHARING AND

OPPORTUNITIES TO CHANGE RESEARCH

PRACTICES IN TWO TYPES OF RESEARCH

ORGANISATIONS

The comparison of situations of researchers who wanted to change theirpractices by developing an innovation makes it possible to identify charac-teristic patterns of authority sharing which enable specific protected spaceto be built and thus create distinct opportunities to change research

practices We will do this in three steps First we describe the patterns ofauthority sharing we observed in our comparative case studies Thereafterwe link patterns of authority sharing to opportunities to build protectedspace for the change of research practices In a final step we discuss thedistribution of these patterns in the two science systems

Patterns of Authority Sharing

Five distinct patterns of authority sharing by researchers can be identifiedin terms of which authoritative agencies enabled the building of protectedspace and thus with whom researchers shared authority over their researchgoals (Table 5) First full authority of researchers over research goals wasobserved in some German public research institutes where directors didnot have to share their authority at all This pattern is created by the stategranting sufficient recurrent funding for the institutersquos research and relin-quishing its authority over the uses to which these resources are putResearchers who have full authority can build almost any protected spacethat might be necessary in their field They have full discretion over thenecessary resources and enjoy long-time horizons due to their permanentpositions at research institutes The institutes themselves are regularly eval-uated but can be assumed to exist for a long time

Second mission-bound full authority was observed at German publicresearch institutes whose directors also have full discretion over a largeamount of resources but are bound by their institutersquos mission If the state

Table 5 Patterns of Authority Sharing

Authority Pattern Authority over Research Goals Shared between

Researcher State Scientific

community

Management Senior

researchers

Full X

Mission-bound full X X

Community-shared X X

Embedded community-

shared

Enmeshed X X X X

voices expectations that fall within the scope of its mission institute direc-tors have little choice and need to organise a change of research practiceswith at least some of the institutersquos resources Mission-bound authoritymeans that researchers have to adapt to expectations of authoritativeagents who however do not intervene in the formulation of research goalsor selection of approaches Researchers who have this authority can alsobuild most protected space that might be necessary in their field

Third community-shared authority is a pattern that is characteristic forGerman university professors and group leaders at the two Dutch insti-tutes After they are appointed these academics have full discretion oftheir recurrent funding Since this funding is insufficient for experimentalresearch however they need to supplement it with grants which are con-trolled by their scientific community This makes any change of researchpractices and particularly the switch to a scientific innovation dependenton the decision practices and epistemic preferences of the community(Glaser et al 2014) We speak of shared authority if researchers need tonegotiate goals or approaches with other authoritative agencies Due tothe esoteric nature of the knowledge required for competent negotiationsauthority is likely to be shared with peers ie with researchers from thesame field

Fourth embedded community-shared authority is the authority ofresearchers who not only share their own authority over research goalswith their community but also find this authority sharing embedded withintheir organisational environment which constrains their authority Wefound two versions of this embedded community-shared authority in theNetherlands First university professors today share their authority withthe university management which although unable to influence the use ofthe recurrent funding by professors directly has discretion over the exis-tence of a professorrsquos field of research at the university and can grant orwithdraw recurrent funding As the example of Dutch evo-devo researchdemonstrates the university management (university leadership and deans)can strengthen limit or discontinue the presence of fields at the universityThe shared authority of Dutch university professors is contingent on thisparticular authority a relationship that becomes apparent only in specificcases (see Weyer amp Laudel this volume for more examples) Second asimilar embedded community-shared authority characterised the situationof the director of the Dutch physics institute who enjoyed an excellentresearch base but depended on his community for grants and found himselfembedded in a decision process of the agency that funds the institute aboutthe latterrsquos research programme

Finally enmeshed authority is a pattern encountered by all researchersbelow the professorial or director level at universities and public researchinstitutes in both countries These researchers had to share their authorityover means for building protected space with many others They had nodiscretion over infrastructure and thus depended on senior researchers inthe organisation (their professors or directors) who could grant them accessto infrastructure They had to share authority over their research goalswith the senior researchers which also embedded them in the latterrsquosauthority relations If they needed additional grants they had to shareauthority with their communities

The Impact of Authority Patterns on Opportunities toChange Research Practices

These five authority patterns are linked to different opportunities for build-ing protected space in terms of both resources and time horizons and toopportunities for changing research practices which we compare in termsof the limitations under which the authority patterns operate (Table 6)Full and mission-bound full authority patterns have in common thatresearchers have discretion over the resources necessary for virtually all

Table 6 Impact of Authority Patterns on Opportunities to BuildProtected Space and to Change Research Practices

Authority Pattern Opportunities to Build Protected Space Limitations of

ProtectionResources Time horizon

Full Available as needed Long None

Mission-bound full Available as needed Long None within bounds

of mission

Community-shared Contingent on approval of

grants

Split Delayed rarely

insufficient

Embedded

community-shared

Contingent on support by

university and approval of

grants

Split Delayed

occasionally

insufficient

Enmeshed Contingent on permission by

senior researchers and approval

of grants

Split or

fixed-term

Delayed insufficient

in most cases

protected space that might be necessary in their field and enjoy the long-time horizons of permanent positions and the tenure of their institutes Theboundaries set by missions of institutes are likely to matter only in veryrare cases

The community-shared authority pattern provides only limited opportu-nities to build protected space It is contingent on grants (the acquisition ofgrants and their size) in the resource dimension In the time dimensionprotected space is split along the two sources of resources ie between thelong-time horizons of permanent discretion over some recurrent fundingand the time horizons of grants that are necessary to supplement recurrentfunding These conditions do not necessarily constrain the opportunities tochange research practices However the building of protected space isdelayed by the grant approval process if the community and the fundingagency do not tacitly accept the use of awarded grants for research on pro-blems for which they were not awarded There are also cases in which thecommunity enforces its majority opinion that some changes of researchpractices are not worth funding as the Dutch physics community did inthe early phase of experimental BEC research In these rare cases thecommunity-shared authority pattern is insufficient

The embedded community-shared pattern enables the same building ofprotected space as the community-shared pattern but is additionally contin-gent on the acceptance of a research field by the university and mightrequire additional negotiations of the thematic scope in which protectedspace can be build It is therefore likely to be insufficient more often thanthe community-shared pattern

Enmeshed authority made changes of research practices dependent onthe acceptance of professors or directors contingent on the authority struc-tures these gatekeepers were embedded in and in most cases also contin-gent on the decisions within the scientific communities The resources thatcould be mobilised with it were contingent on both the contributions bysenior researchers and the grant decisions in the scientific communitiesTime horizons were either limited to fixed terms or split In Germany afixed-term time horizon is associated with this authority pattern becausethe fixed-term grants are combined with fixed-term positions of researchersThis also applies to some of the Dutch researchers below the professoriallevel particularly to those whose positions are funded by fellowshipsThe regular Dutch university positions below the professorial level Universitair Docent and Universitair Hoofdocent are permanent positionswhich creates split time horizons As a result of the multiple dependencelimitations to building protected space are common which in turn means

that this authority pattern only rarely supports changes of researchpractices

Distribution of Authority Patterns in the German and Dutch Science System

Having identified the ways in which the different authority patterns affectthe opportunities for changing research practices we now need to identifytheir distribution throughout the two national science systems Table 7 liststhe empirical cases we found and used to identify the authority pattern andconclusions about the distribution of the patterns

The range or scope of situations in which the full and mission-boundfull authority patterns occur is small because there are very few positions inGerman public research institutes whose incumbents enjoy such relativelyuntrammelled authority They are also not available in all fields of sciencewhich is why it is only by chance that a potential innovation can besupported by this authority pattern

The scope of community-shared authority patterns is medium inGermany where it includes all university professors and small in theNetherlands where it includes only directors of research institutes It isfurther limited by the infrequent situations in which German universityprofessors can build protected space for more radical changes of research

Table 7 Distribution of Authority Patterns in the Two Countries

Authority Pattern Observed Cases Distribution

BEC Evo-devo ILSA

Full Germany Germany Only in Germany very few

Mission-bound full Germany Only in Germany very few

Community-shared Germany

Netherlands

Germany Germany Medium in Germany

(professors) very few in the

Netherlands (directors of

institutes)

Embedded

community-shared

Netherlands Netherlands Medium in the Netherlands

Enmeshed Germany

Netherlands

Germany

Netherlands

Germany Widespread in both countries

practices namely appointment and loyalty negotiations Only in thesesituations can the inert part of their protected space the infrastructureand basic equipment that are paid for from recurrent funding be shapeddifferently to support new research

Embedded community-shared authority patterns also have a mediumscope because they occur for all Dutch university professors However thelatter have better opportunities to negotiate ad hoc funding for changes intheir infrastructure due to the increased authority of university manage-ment over resources The most common authority pattern in both countriesis that of enmeshed authority It was found with all researchers below theprofessorial level at German and Dutch universities

If we compare the distribution of authority patterns in the two countriesinteresting commonalities and differences become apparent First opportu-nities to change research practices are largely restricted to lsquoprofessors andaboversquo since building protected space in the experimental sciences dependson access to infrastructure and authority granting this access is restrictedto professors in both countries Second Dutch university reforms have notmodified the distribution of authority between professors and non-professorial academics at universities Possibly due to the overall scarcityof recurrent funding this funding is still concentrated on professors Atthe same time university reforms did modify the authority distributionbetween professors and the university management Authority over theexistence of fields at universities is now shared with the management withthe latter having a veto position in this particular decision process (seeWeyer amp Laudel this volume) This creates an embedded community-shared authority pattern that does not yet exist in Germany Third Dutchpublic non-university research institutes provide authority patterns resem-bling those of Dutch university professors rather than being an alternativeto them This is why the characteristic authority patterns that can be pro-vided by public research institutes those that make researchers fully inde-pendent from other authoritative agencies do not appear to be firmlyinstitutionalised in the Netherlands which points to a different nationalpattern of authority sharing (Glaser et al 2014)

CONCLUSIONS

Protected space enables changes of research practices because it providesautonomous discretion over resources for a specific time horizon which

also suspends reputational consequences of not meeting expectations of thecommunity for that time (for the latter see also Hackett 2005) The oppor-tunities for researchers to build such protected space depend on authoritypatterns in their research organisations We could identify five distinctauthority patterns and trace their distribution across the national Germanand Dutch science systems Our analysis leads to four conclusions

First opportunities to change research practices are very unevenly dis-tributed in the two national science systems The majority of researchersface an enmeshed authority pattern in which building protected spacedepends on too many authoritative agencies to have high changes ofsuccess In both countries professors have opportunities to build protectedspace and act as gatekeepers because they control the infrastructure atuniversities

Second the community-shared and embedded community-sharedauthority patterns which we found for professors may delay the change ofresearch practices and introduce additional risks compared to full authoritypatterns The early changes to BEC in German universities do not invali-date this conclusion because the researchers in question had already built aresource base and needed only few additional resources once the possibilityof producing BECs was demonstrated

Third the full and mission-bound full authority patterns provide asignificant advantage both with regard to immediate and unconditionalaccess to resources and protection from reputational consequences if achange of research practices contradicts the communityrsquos current majorityopinion It is quite difficult to base the decisive stages of onersquos career onresearch contradicting the mainstream (and proved impossible in the caseof ILSA at German universities) It became possible in universities onlyafter research conducted at public research institutes established the field

Thus our analysis points to the necessity of being able to build protectedspace quickly without needing the agreement of other authoritative agenciesin the early stages of the diffusion of new research practices which includesthe original scientific innovation This underlines the importance of fullauthority patterns for flexible responses to new developments on the micro-level even though the grant funding on which community-shared authoritypatterns are based may provide a more flexible national-level system

Fourth the empirical findings appear to prove the superiority of blockfunding as a condition for major changes in research practices While thismight well be the case (and further research is needed to make this pointmore firmly) the conclusion that block-funded public research institutesare a solution to the problems of grant funding would be premature Even

in Germany with its large sector of public research institutes these insti-tutes do not cover all fields and thus cannot foster innovations across thefull spectrum of scientific disciplines There are also several important rea-sons why universities will always play a key role in the diffusion of scientificinnovations even if these innovations are created in other organisationalsettings First universities still offer more career opportunities for research-ers because there are many more positions for both young researchers andprofessors Research institutes do not provide positions for researchersfrom all fields and provide significantly fewer positions for both youngresearchers and directors if there is a research institute for a field Secondthe involvement of academics at universities means that new practices maydiffuse to teaching Third academics at universities have the best access topromising PhD students Last but not least only universities can grantPhDs and thus the entrance into research careers Owing to this strongposition in academic career patterns universities are lsquoobligatory points ofpassagersquo (Latour) in the diffusion of any scientific innovation which makesthem a limiting factor if they cannot sufficiently contribute to the protectedspace for their researchers

1 The idea of lsquoprotected spacersquo has been previously used by Rip (1995 p 86) todescribe the laboratory as a space in which researchers are shielded from interference(see also Krohn amp Weyer 1994) Rip and others have also used the concept fordescribing a situation that is necessary for the early development of technologicalinnovations (Kemp Schot amp Hoogma 1998 Rip amp Schot 2002) Rip has recentlyextended the concept to lsquoprotected spaces of sciencersquo in society (Rip 2011) Our useof that concept deviates from Riprsquos in that we define it at the micro-level of indivi-dual researchers and their projects include the protection from reputational conse-quences in the scientific community introduce the time horizon for which aresearcher is protected and link it to the macro-level by asking for whom theseindividual-level protected spaces are provided2 We would like to thank Raphael Ramuz for providing access to one of the

interviews he conducted which is relevant to the German case3 German university professors receive money for initial investments in their

research and teaching infrastructure on appointment (start-up packages) and foradditional investments when they negotiate staying with their university after beingoffered a position elsewhere (loyalty packages)4 In order to make sure that this observation is not an artefact of our selection

of interviewees we searched the German LSA community for researchers thatmight have switched at a German university in the 1990s and were not professorsWe are sure that this didnrsquot happen

REFERENCES

Bland C J amp Ruffin M T (1992) Characteristics of a productive research environment

Literature review Academic Medicine 67 385397

Blau P M (1973) The organization of academic work New York NY Wiley

Chubin D E amp Connolly T (1982) Research trails and science policies In N Elias

H Martins amp R Whitley (Eds) Scientific establishments and hierarchies

(pp 293311) Dordrecht Reidel

Federal Ministry of Education and Research (2012) Bundesbericht Forschung und innovation

2012 Bonn Bundesministerium fur Bildung und Forschung

Fend H (2010) Bildungsforschung Von 1965 Bis 2008 Ein Zeitzeugenbericht Zu

Fortschritten Ruckschlagen Und Hohepunkten In C Ritzi amp U Wiegmann (Eds)

Beobachten Messen Experimentieren Beitrage Zur Geschichte Der Empirischen

PadagogikErziehungswissenschaft (pp 275303) Bad Heilbrunn Verlag Julius

Klinkhardt

Glaser J Lange S Laudel G amp Schimank U (2010) Informed authority The limited use

of research evaluation systems for managerial control in universities In R Whitley

J Glaser amp L Engwall (Eds) Reconfiguring knowledge production Changing authority

relationships in the sciences and their consequences for intellectual innovation

governance on the emergence of fields In M Merz amp P Sormani (Eds) The local

configuration of new research fields On regional and national diversity Sociology of the

Sciences Yearbook Dordrecht Springer

Hackett E J (2005) Essential tensions Identity control and risk in research Social Studies

of Science 35 787826

and politics (pp 317ndash353) Amsterdam IOS Press

Disciplinary differences Dordrecht Springer

Kaplan N (1964) Organization Will it choke or promote the growth of science In K Hill

(Ed) The management of scientists (pp 103127) Boston MA Beacon Press

Kemp R Schot J amp Hoogma R (1998) Regime shifts to sustainability through processes of

niche formation The approach of strategic niche management Technology Analysis amp

Strategic Management 10 175198

Knorr K D amp Mittermeir R (1980) Publication productivity and professional position

Cross-national evidence on the role of organizations Scientometrics 2 95120

Knorr-Cetina K D (1981) The manufacture of knowledge An essay on the constructivist and

contextual nature of science Oxford Pergamon Press

Krohn W amp Weyer J (1994) Society as a laboratory The social risks of experimental

research Science and Public Policy 21 173183

Lambright W H amp Teich A H (1981) The organizational context of scientific research In

P C Nystrom amp W H Starbuck (Eds) Handbook of organizational design volume 2

Remodeling organizations and their environments (pp 305319) Oxford Oxford

University Press

Leisyte L (2007) University governance and academic research Case studies of research units

in Dutch and English universities Enschede CHEPS University of Twente

McCarrey M W amp Edwards S A (1973) Organizational climate conditions for effective

research scientist role performance Organizational Behavior and Human Performance

9 439459

Ministerie van Onderwijs Cultuur en Wetenschappen (2012) The science system in the

Netherlands An organisational overview Retrieved from httpwwwgovernmentnl

filesdocuments-and-publicationsleaflets20120417the-science-system-in-the-nether-

landsthe-science-system-in-the-netherlands-an-organisational-overviewpdf Accessed

on Febraury 11 2014

traditions (pp 6384) Bielefeld transcript

Pelz D C amp Andrews F M (1966) Scientists in organizations Productive climates for

research and development New York NY Wiley

Rip A (1995) New combinations European Review 3 8392

Rip A (2011) Protected spaces of science Their emergence and further evolution in a chan-

ging world In M Carrier amp A Nordmann (Eds) Science in the context of application

Rip A amp Schot W (2002) Identifying loci for influencing the dynamics of technological

development In K H Sorensen amp R Williams (Eds) Shaping technology guiding pol-

icy Concepts spaces and tools (pp 155172) Cheltenham Edward Elgar

Sanz-Menendez L amp Cruz-Castro L (2003) Coping with environmental pressures Public

research organisations responses to funding crises Research Policy 32 12931308

Schimank U (forthcoming) Wandel der ldquoRegelungsstrukturenrdquo des Hochschulsystems und

die Folgen fur die wissenschaftliche Forschung zur Entwicklung der Governance-

Perspektive In A Bora A Henkel amp C Reinhardt (Eds) Wissensregulierung und

Regulierungswissen Weilerswist Velbruck

Vaughan D (1999) The role of the organization in the production of techno-scientific knowl-

edge Social Studies of Science 29 913943

Whitley R (2012) Institutional change and scientific innovations The roles of protected space

and flexibility Paper prepared for presentation to Sub-theme 17 Organizing Science

The Increasingly Formal Structuring of Academic Research at the 28th EGOS

Colloquium 57 July 2012 Helsinki

Zabusky S E amp Barley S R (1997) ldquoYou canrsquot be a stone if yoursquore cementrdquo Reevaluating

the emic identities of scientists in organizations Research in Organizational Behavior

19 361404

Zedler P amp Dobert H (2009) Erziehungswissenschaftliche Bildungsforschung In R Tippelt amp

B Schmidt (Eds) Handbuch Bildungsforschung (pp 2345) Wiesbaden VS Verlag fur

Sozialwissenschaften

COMPUTER CORPUS LINGUISTICS

AN INNOVATION IN THE

HUMANITIES

Lars Engwall Enno Aljets Tina Hedmo and

Raphael Ramuz

ABSTRACT

Computer corpus linguistics (CCL) is a scientific innovation that hasfacilitated the creation and analysis of large corpora in a systematic wayby means of computer technology since the 1950s This article providesan account of the CCL pioneers in general but particularly of those inGermany the Netherlands Sweden and Switzerland It is found thatGermany and Sweden due to more advantageous financing and weakercommunities of generativists had a faster adoption of CCL thanthe other two countries A particular late adopter among the four wasSwitzerland which did not take up CCL until foreign professors hadbeen recruited

Keywords Scientific innovation linguistics country comparisoncorpus-building case study

ISSN 0733-558Xdoi101108S0733-558X20140000042011

INTRODUCTION

Human communication in written and spoken form has since long inter-ested scholars over the world One classical method has been the collectionof examples of linguistic usage in order to analyze variations in construc-tions dialects etc In contrast to this are deductive theories that imply theconstruction of theoretical examples and testing them on informants Bothof these approaches experienced significant changes in the decades follow-ing the Second World War In terms of empirical studies the developmentof computer technology dramatically changed the potential for handlinglarge bodies of text in a more systematic way More or less at the sametime Noam Chomskyrsquos ideas for focusing linguistics research on the under-lying constructions the generative grammar had a significant impact onmany scientists particularly in the United States As a result the 1960switnessed considerable tensions between empirically and theoreticallyoriented linguists Today this is less the case as computer corpus linguis-tics (CCL) that is the creation and analysis of large corpora in a systema-tic way by means of computer technology minus has become quite widelyestablished1 This state of affairs is demonstrated by a large number ofbooks introducing the field (eg Biber Conrad amp Reppen 1998 Halliday2004 Kennedy 1998 McEnery amp Hardie 2011 Meyer 2002 Teubert ampCermakova 2007) and two fairly recent handbooks (Ludeling amp Kyto20082009 OrsquoKeeffe amp McCarthy 2010)2 Moreover in contrast to the earlyphase when academic entrepreneurs took the lead CCL has become a large-scale effort within the larger field of e-humanities3 Needless to say advancesin computer technology have been significant for these developments

The construction and analysis of computerized corpora of both writtenand spoken linguistic usage represented a considerable innovation in lan-guage studies in the sense that it involved the development and learning ofnew skills and enabled new kinds of research questions to be consideredCorpus-building though requires considerable protected space in the senseof researchers having long periods of discretion over the use of neededresources to pursue particular problems and approaches before producingpublishable and collectively valued results since it is both time-consumingand requires major financial and technical support Since the level of pro-tected space afforded to scientists varies between countries it would not besurprising if the development of CCL occurred in different ways and at dif-ferent rates in different national academic communities These differencescan be expected to reflect variations in the organization and preferences ofnational scientific elites funding opportunities and the organization of

332 LARS ENGWALL ET AL

intellectual and administrative authority in the main centers of academicresearch mostly universities In this article we therefore compare theearly development of CCL in four European countries (Germany theNetherlands Sweden and Switzerland) and suggest explanations forthe variations in the adaptation of CCL Before doing so we discuss thecontext and the development of CCL as a scientific innovation

THE CONTEXT AND THE DEVELOPMENT OF CCL

The Context of CCL

The epistemic field of language studies has by its very nature been ratherweakly integrated4 The main reason for this is the great number of lan-guages in the world as well as the common use of university languagedepartments to train teachers of the native language in schools As a resultfaculties of the humanities in many universities have had a large number ofsmall departments specializing in specific languages Traditionally the clas-sical languages Greek and Latin were strong but with the passage of timethe modern languages of English French and German grew in importanceas did the study of native languages5 While the degree of integrationbetween scholars of different languages tends to be weak within universi-ties the level of integration among language-defined international commu-nities consisting of scholars of English French German etc tends to begreater than that found in many other fields in the humanities

However it is not only language barriers that hamper integrationAnother important circumstance is a diversity of approaches to languagestudies One such division is that between philologists and literature scholarsAnother is that between written and spoken language And a third divisionis that between historical and modern language use As mentioned aboveparts of language studies have also been closely associated with educationparticularly efforts to develop educational material such as grammars dic-tionaries and vocabularies

Nevertheless efforts to increase the intellectual integration of languagestudies have been made over time In the beginning of the last century theSwiss structural linguist Ferdinand de Saussure (18571913) provided anintegrating framework based on a distinction between langue (the gram-mar) and parole (the spoken language and the written text) (Saussure1916) Although widely used and cited it did not have the same effect on

333Computer Corpus Linguistics An Innovation in the Humanities

the field as the generative grammar introduced by Noam Chomsky (b1928) in the 1950s It was first presented in Chomsky (1957) and was lateron developed in a number of works (Chomsky 1965 1972 1982 19952002)6 In these Chomsky criticized the structural approach of Saussureand introduced another pair of concepts competence (the language knowl-edge of a native speaker) and performance (the language used) BasicallyChomskyrsquos ideas imply that language studies should focus on the deepstructures of languages through a deductive approach rather than throughempirical studies

In terms of empirical studies language scholars have long used corporafor the production of dictionaries dialect atlases and grammars A veryearly example is a German frequency dictionary (Kaeding 18971898) pro-duced by the stenography expert Friedrich Wilhelm Kaeding (18431928)Other early examples are Henmon (1924) a French frequency-based word-book and the publications of the American and Canadian Committees onModern Languages (cf eg Buchanan 1931 Cheydleur 1934 Morgan1933 Vander Beke 1929) Similarly De la Court (1937) presented frequen-cies from a one million corpus of Dutch compiled before the advent of thecomputer

To a large extent these studies had educational motives that is a wishto map basic vocabularies and their frequency structures In the 1930s theyinspired the Harvard linguist professor George Kingsley Zipf (19021950)to formulate what has become known as Zipfrsquos law that is that the pro-duct of rank and frequency in a word distribution tends to be constant(Zipf 1932) In this way linguists began to cross-disciplinary borderstoward the natural sciences At the same time word frequencies alsocame to interest non-linguists such as the British statistician Udny Yule(18711951) who provided a statistical study of literary vocabulary (Yule1944 see also Simon 1955 Ijiri amp Simon 1977)

The Development of CCL

Against this background it is not surprising that language scholars wereattracted to the opportunities offered by postwar developments in informa-tion technology to construct new corpora In addition to facilitating thecounting of words these developments also made it possible to produceconcordances that is an output that put the individual words into theircontext An early contribution was made already in the early 1950s minus withthe support of the IBM Chairman Thomas J Watson Sr minus by the Italian

Jesuit Father Roberto Busa (19132011) through his studies of the Latintexts of Thomas Aquinas (cf eg Busa 1951)7

Another European pioneer was the Frenchman Bernard Quemada(b 1926) who started his work on CCL in the 1950s in Besancon8

Through a considerable faculty grant and contacts with the French compu-ter company Bull he created a laboratory for the study of French vocabu-lary9 He also managed to convince the founder of lrsquoInstitut National de laLangue Francaise (INaLF) Paul Imbs (19081987) of the advantages ofusing electronic data processing10 In this way he eventually becamethe editor of 30 volumes presenting the historical French vocabulary(Quemada 19591993) and could in 1968 defend a thesis on dictionariesof modern French (Quemada 1968) As director of INaLF (19771994) hebecame a key person for French CCL At an early stage Quemada arrangedsummer schools which together with his publications were important forthe diffusion of CCL (see further Zampolli Cignoni amp Peters 1981)

Apparently independent of these European-based researchers theRumanian born Stanford professor Alphonse Juilland (19232000) pro-duced frequency dictionaries of four Roman languages Spanish (Juilland ampChang-Rodriguez 1964) Rumanian (Juilland Edwards amp Juilland 1965)French (Juilland Brodin amp Davidovitch 1970) and Italian (Juilland ampTraversa 1973) At about the same time another Stanford professor inthe German Department Alan Pfeffer (b 19072003) compiled a spokenGerman corpus based on almost 80 hours of recordings11

On the British Isles Randolph Quirk (b 1920) was the pioneer He firstlaunched the project Survey of English Usage (SEU) at the University ofDurham (see eg Quirk 1957) for which he collected written texts as wellas spoken English In 1960 he moved to University College London andrecruited the Swede Jan Svartvik (b 1931) for the further work This inturn paved the way for Svartvikrsquos Swedish doctoral dissertation (Svartvik1966) Later on in 1975 back in Sweden as professor in Lund Svartviktook the initiative for a sister project of the London survey Survey ofSpoken English (SSE) Together the two surveys resulted in a corpus ofone million words from 100 written texts and 100 spoken texts (seeQuirk amp Svartvik 1978) which were used for the writing of grammars (seeQuirk Greenbaum Leech amp Svartvik 1985 Quirk amp Svartvik 1972Svartvik amp Sager 1971)12

Although Busa Quemada Juilland and Quirk appear to have beenthe real CCL pioneers the literature often points to the Brown corpusat Brown University in Providence RI created by Henry Kucera(19252010) and Nelson Francis (19112002) as the first major corpus

It contained around one million words that had been published in theUnited States during the year of 1961 It was analyzed and publishedin 1967 as Computational Analysis of Present-Day American English(Kucera amp Francis 1967) This corpus later on provided the basis for thepublication of the first edition of American Heritage Dictionary (Morris1969) It was also the inspiration for many followers in the field of CCLThe closest follower was the CAMET project (Computer Archive ofModern English Texts) launched in 1970 by the then Reader in English atLancaster University Geoffrey Leech (b 1936) Targeting British Englishit was collected according to the same principles as the Brown corpus inorder to facilitate comparisons Through collaboration with Norwegianscholars it eventually became the Lancaster-OsloBergen (LOB) corpus andwas completed in 197813

Forces Against and for CCL

It is thus very apparent that the development of computer technology wassignificant for the development of CCL However it is also important tonote the contemporaneous questioning of the scientific validity of theconstruction of vast databases of linguistic usage The key person in thiscontext was Noam Chomsky and his advocacy of the generative grammarapproach to language studies (Chomsky 1957 1965) He and his followersargued that language could be better and more easily studied by confront-ing native speakers with constructions of word combinations rather thanby studying vast materials of written and spoken language In this waycorpus linguistics was to a large extent challenged by generativists withChomsky (1957 p 159) claiming

Any natural corpus will be skewed Some sentences wonrsquot occur because they are

obvious others because they are false still others because they are impolite

This attitude created according to Filmore (1992 p 35) a tensionbetween ldquoarmchair linguistsrdquo (who tested theoretical constructions) andldquocorpus linguistsrdquo (who collected and analyzed vast materials of languageused) Such a tension persisted even after CCL had become more wide-spread and according to Johansson (2008 p 33) ldquothe negative view of cor-pora found in early generative linguistics persisted in many circlesrdquoLikewise Jan Svartvik (2007 p 11) has stated that

hellip there might have been moments when being named [a corpus linguist you] felt like

discovering your name on the passenger list for the Titanic

For corpus linguists the introduction of the generative grammarapproach thus indeed implied that they had to fight an uphill battle parti-cularly in the United States14 In addition to these negative attitudes theywere also at a disadvantage owing to a need for more resources technicalproblems copyright issues and fewer opportunities for the publication ofarticles in international journals This was also reinforced by critical atti-tudes from literature scholars Nelson Francis a pioneer through theBrown corpus thus once told his colleagues ldquoOne of my colleagues a spe-cialist in modern Irish literature was heard to remark that anyone whowould use a computer on good literature was nothing but a plumberrdquo(Francis 1986 p 5)

In relation to the tensions between the supporters of Chomsky and cor-pus linguists it is also important to bear in mind that linguists do not onlydeal with present-day language which permits the testing of constructionson native speakers A considerable number of linguists instead deal withhistorical material which makes the confrontation with native speakersproblematic and as a result tensions with Chomskyans less significant Aprime example of this is Father Busa and his studies of Thomas Acquinas(cf above) Likewise the former director of INaLF Robert Martin hasdenied in an interview any critical attitudes toward his corpus work onmedieval texts15

While the Chomsky approach implied challenges to CCL some com-mercial forces were working in the other direction As already mentionedthe Brown corpus became the basis for a new dictionary of AmericanEnglish and other publishers took a similar interest Oxford UniversityPress (OUP) thus collaborated with the Arts Computing Centre atWaterloo Canada to create the Oxford Dictionary of English (Johansson2008 p 35)16 This eventually led to the creation of the British NationalCorpus which is an industrialacademic consortium led by OUP andfunded by commercial partners as well as the UK government It now con-tains 100 million words17 Needless to say the development of this corpusas well as other corpora has been strongly facilitated by changes in printingtechnology of publishing houses since the 1970s leading to easy access oftexts to the content of newspaper articles books and other publications

Another source of support for CCL came from the efforts to use compu-ter technology for translation18 Thus as early as in 1962 the Associationfor Machine Translation and Computational Linguistics (AMTCL) wasfounded for ldquothe international scientific and professional society forpeople working on problems involving natural language and computationrdquowhich in 1968 took its present name the Association for Computational

Linguistics (ACL)19 This organization was followed by others like theAssociation for Literary and Linguistic Computing (1973) and theAssociation for Computing and the Humanities (1978) (Johansson 2008p 35) In the wake of the foundation of these organizations several journalshave been established among which Computational Linguistics has thehighest impact factor (see further Engwall amp Hedmo 2012a)

Conclusions

From the above we can see that while CCL developed as part of a traditionof corpora construction in natural language departments that were relativelyweakly integrated it faced an intellectual challenge from efforts to creategreater integration through the development of general language theoriesIn some universities these theories led to the creation of departments oflinguistics that focused on the identification of general characteristics ofwritten and oral communication as the key component of the scientific ana-lysis of language However the variety of natural language institutes anddepartments in many European universities and their role in training lan-guage teachers during much of the postwar period meant that generativistswere not able to dominate language studies to the extent of preventing allcorpus construction and the use of computers in doing so in all nationalpublic science systems

As an innovation in the humanities then CCL needed researchers whohad the protected space to undertake long-term projects without havingto publish their findings within short-term project cycles and were able toaccess for the time large-scale computer facilities with technical supportbut was regarded as a legitimate scientific activity in some academicunits albeit to varying degrees among countries and universities In somecases commercial support from publishers was important in providingsupplementary funding especially in the United Kingdom and the UnitedStates

THE EMERGENCE OF CCL IN FOUR EUROPEAN

COUNTRIES

Having provided a general background regarding the development of CCLwe will now turn to the analysis of the emergence of CCL in Germany the

Netherlands Sweden and Switzerland In so doing it is first important topoint out that we will focus on corpus builders since they are the originalinnovators paving the way for corpus users The basis for our analysis ofthe development of CCL in the four countries is constituted by (1) writtensources (ie publications by corpus linguists and websites) and (2) inter-views with researchers in the field and other experts in language studies Intotal we interviewed 52 people distributed across countries and CCL com-munities as summarized in Table 1

Germany

Developments in Germany are closely associated with das Institut furDeutsche Sprache (IDS) in Mannheim It was founded in 1964 by the fed-eral and provincial governments to study and document the ldquocontemporary

Table 1 Interviews Undertaken for the Study

Country English

French

German

Swedish

Others Total

France 2 2

Germany 1 1 3a 1c 6

Netherlandsb 3 1 3 1 8

Swedenc 3 1 1 6 11

Switzerland 5 3 9 5d 22

United Kingdom 2 1 3

Total 14 7 12 3 6 4 6 52

aThese three interviewees were in Corpus Linguistics and Morphology Computational

Linguistics and Theoretical Computational Linguistics and Computational LinguisticsbTwo of the interviewees started to build and use corpora during their PhD work the others

included corpora later in their research career One researcher (who had also started to work

with corpora during the PhD period) migrated to the Netherlands from abroad Figures refer

to the main language that the interviewee has been focusing on Five of the eight interviewees

also did studies of other languagescInterviewees include pioneers belonging to the first generation of CCL research as well as

more contemporary researchers in the area of CCL research building and using spoken and

written corpora in the area of general linguistics and natural languagesdThese five interviewees were two senior linguists and three deans

usage and recent history of German languagerdquo20 Several of the early cor-pora were thus the outcome of IDS initiatives Nevertheless the very firstapplication of CCL appears to be a project at the University of Saarland inSaarbrucken led by Hans Eggers (19071988) intended to analyze Germansyntax Although this corpus was originally conceived as early as 1956 ittook six years until 1962 before the research group was created and theproject could start It then took another six years to complete thework The resulting Saarbrucker Korpus (see second column in Table 2)consists of 200000 words of German texts These were obtained by a sys-tematic selection of sentences from Rowohlts Deutsche Enzyklopadie andFrankfurter Allgemeine Zeitung followed by a reduction of the materialfrom 100000 to 11000 sentences in order to make the corpus manageableThe processing of the corpus implied a cross-disciplinary collaborationfrom 1964 with Institut fur Angewandte Mathematik und Rechnungszentrumat the university The outcome was a database and publications

Table 2 Three Early Corpora in Germany

Corpus Saarbrucker Korpus Bonner Zeitungskorpus Freiburger

Korpus

Start year 1962 (1956) 1964 1968

Principal

Investigator (PI)

Hans Eggers

(19071988)

Manfred W Hellmann

(b 1936)

Hugo Steger

(19292011)

Age of PI at the

55 (49) 32 39

Language German German German

Institution University of Saarland IDS Bonn IDS Freiburg

Type of material Written texts

(encyclopedia and

newspaper)

Newspapers from FRG

and GDR

Spoken public

sources

19601974

Corpus size (million

words)

02 31 05

Funding DFG IDS DFG IDS

Output Database and

publications

Database and publications

among them a dictionary

Database and

publications

References Eggers (1964 1969) Hellmann (1984 1992) Gesprochene

Sprache (1974)

presenting the work undertaken and syntactic analyses (Eggers 1969Hellmann 1976 pp 250 251) The project was supported by DeutscheForschungsgemeinschaft (DFG)

The other two major early German corpora stemmed from initiativesundertaken by the IDS The first of these followed the creation of aresearch unit in Bonn that led to the Bonner Zeitungskorpus (third columnin Table 2) compiled by Manfred W Hellmann (b 1936) This project sup-ported by DFG compared newspaper texts from the Federal Republicof Germany (FRG die Welt) and the German Democratic Republic(GDR Neues Deutschland) In addition smaller samples were added fromthe FRG titles Bonner General-Anzeiger (1964 and 1974) and FrankfurterRundschau (1974) as well as the GDR titles Norddeutsche NeuesteNachrichten (1964 and 1974) and Der Morgen (1974) As a result thecorpus came to include 31 million words The material was processed bymeans of the concordance computer program KWIC (KeyWords inContext) and was made accessible for other users on microfiches It hasbeen used for linguistic analysis of modern German21

The next IDS corpus was the Freiburger Korpus (fourth column inTable 2) for which work started in 1968 It was led by Hugo Steger(19292011) who had just taken over the chair of German philology atthe University of Freiburg The corpus was directed toward spoken stan-dard German with the purpose of describing and analyzing its grammaticaland stylistic features For this purpose a sampling of 222 sound recordingsfrom radio television and of other types of communications was underta-ken for the period 19601974 In this way the corpus came to contain 68hours of interviews discussions entertainment lectures reporting andstories which in transcribed form corresponded to about 500000 wordsThe outcome was a database and a series of publications22

For later developments it is particularly worth mentioning theMannheimer Korpus which was launched in 1968 by Ulrich Engel (b 1928)with the support of IDS Mannheim This corpus consists of two parts onelarger (22 million words) for the period 19501967 and a smaller one(03 million words from the years 1949 1952 19601974) The first partconsists of 293 texts taken from seven novels (among them Heinrich BollsrsquoAnsichten eines Clowns and Thomas Mannrsquos Die Betrogne) the memoirs ofthe former FRG President Theodor Heuss scientific and popular scientificliterature popular fiction and articles for newspapers and magazines whilethe second part contained 52 texts for the period 19491974 with a some-what broader frame such as user manuals and textbooks23 The outcome isa database included in the IDS archives

Still another later corpus was the LIMAS Korpus Bonn It is of particu-lar interest to the present study since it again clearly demonstrates theinfluence from the pioneers in the United States Although the languagestudied in the corpus is German it used the same design as Kucera andFrancis (1967) for their Brown corpus that is selecting 2000 words from500 different sources from the years of 1970 and 197124

All in all we can conclude that the early development of CCL inGermany was strongly influenced by IDS that provided considerable pro-tected space for the CCL researchers In the decades after1970 the signifi-cance of IDS has increased even more as IDS has since merged a numberof individual CCL outputs into a huge database of different corpora25 Bythe end of 2013 the database thus contained six billion words more thanfour billion of which are open to the public Through the various IDSinitiatives the institute today plays a dominant role in German CCLThroughout the goal of the CCL activities has primarily been to answerlinguistics research questions rather than finding commercial applications

The Netherlands

The early development of CCL in the Netherlands followed internationaldevelopments although on a smaller scale most likely due to a strongerChomskyan community among Dutch linguists In the 1960s a numberof small-scale activities appear to have started at numerous places in theNetherlands One of these was the work in Amsterdam by de Werkcommissievoor Taalstatistiek which resulted in a 44000-word corpus of newspapertexts (Berckel Corstius Mokken amp Van Wijngaarden 1965) As mentionedabove it had been preceded in the 1930s by a one-million-word corpuscompiled by hand (De la Court 1937)26

Among the early computerized corpora with more than 100000 wordsthe following three seems to have been particularly significant (1) theEindhoven corpus of Dutch (2) the Nijmegen corpus of Modern Englishand (3) the Amsterdam Corpus of Old French (see Table 3)

The work with the Eindhoven corpus (second column in Table 3) startedin 1969 It was the result of the creation in 1968 of the inter-universityorganization Werkgroep Frequentie-Onderzoek van het Nederlands Themain partners in this collaboration were Technische Hogeschool teEindhoven (today the Eindhoven University of Technology) and Instituutvoor Dialectologie Volks- en Naamkunde van de Koninklijke NederlandseAkademie van Wetenschappen (a language institute within the Royal

Netherlands Academy of Arts and Sciences) Together they compiled acorpus of 768000 words Of these about 120000 words of spoken Dutchfrom the period 19601973 were provided by the Amsterdam partner whilethe Eindhoven team compiled the rest from 5 dailies 6 political weeklies9 popular weeklies 52 fiction books and 5 popular science books from theperiod 19691971 The project was supported by the NederlandseOrganisatie voor Zuiver-Wetenschappelijk Onderzoek (NZO) and resultedin the dictionary Woordfrequenties in geschreven en gesproken Nederlands(Uit den Boogaart 1975) At the time of the publication of the results theproject had a steering group with representatives of several Dutch academicinstitutions University of Amsterdam the Free University of Amsterdamthe Mathematical Centre of Amsterdam the University of Utrecht the

Table 3 Three Early Corpora in the Netherlands

Corpus Eindhoven Corpus Nijmegen Corpus Amsterdam

Corpus

Start year 1969 1976 Late 1970s

Principal

Investigator

PC Uit den Boogaart (b 1942) Jan Aarts (b 1933) Anthonij Dees

(b 1928)

Age of PI at

the start

27 43 sim40

Language Dutch English Old French

Institution Eindhoven University of

Technology

University of

Nijmegen

Vrije Universiteit

Amsterdam

Type of

material

Spoken and written material

19601973

Modern English Thirteenth-century

French charters

Corpus size

(million

words)

0768 0130 Small

Funding Nederlandse Organisatie voor

Zuiver-Wetenschappenlijk

Onderzoek (NZO)

NWO grant

university funding

University funding

Output Dictionary Syntactic parser Atlas of old

French dialects

References Uit den Boogaart (1975) Keulen (1986)

Oostdijk (1991)

Dees (1980)

Eindhoven University of Technology and the Royal Netherlands Academyof Arts and Sciences (KNAW) It thus appears to have been relatively wellintegrated in the wider academic community27

While the Eindhoven corpus was oriented toward Dutch the Nijmegencorpus (third column in Table 3) focused on English It was started as apilot project in 1976 as a collaborative between most English departmentsin the Netherlands (Oostdijk 1991 p 14) without any additional resourcesThe research group leader Jan Aarts (b 1933) was inspired to constructcomputerized corpora by visits in 19671968 to Randolph Quirkrsquos groupat University College London and in 1970 to Ohio University (Aarts 1994p 2) He was interested in the syntactic analysis of natural language andparticularly the rules that govern speech He came to collaborate with asecond group in Amsterdam interested in lexical and morphological ques-tions Together they compiled a 130000-word corpus consisting of 65 sam-ples of 2000 words which was manually annotated Aarts eventuallybecame Professor of Contemporary English in 1980 and could expand hisCCL activities In a second phase (from 1983) the Nijmegen group com-piled a one-million-word corpus and developed syntactic tools to automatethe annotation (TOSCA=Tools for Syntactic Corpus Analysis see AartsDitters amp Koster 1987 Oostdijk 1991) The resulting parser was also triedout in the International Corpus of English project but was eventually con-sidered too cumbersome for this large-scale corpus The research was sup-ported by the university faculty which provided funding for assistants andlater by a grant from the national funding agency NWO28 The corpus hasbeen used by other scholars on demand but is not due to copyright pro-blems open access The research came to an end in the 1990s when JanAarts retired and the group was dissolved The group leader in Amsterdamwho collaborated with the Nijmegen people left the Netherlands (see alsoAarts amp Meijs 1984)

The Amsterdam corpus (fourth column in Table 3) of French Charterswas developed at the end of the 1970s It was the result of an interest ofAnthonij Dees (b 1928) in dialectical patterns and a wish to solve a specificlinguistic problem The charters were first collected and compiled into acorpus with the help of two group members The result was a charter atlas(Dees 1980) In the early1980s the same group started to build another cor-pus the Amsterdam corpus of Old French literary texts (annotated andpart-of-speech tagged) This corpus was finished in 1987 (Dees Huber ampvan Rennen-Stein 1987) The project was supported by the universityboard which provided computer equipment and funding for technicalassistants The reason was that the university wanted to promote computer

use in the humanities and was looking for suitable projects However withdiminishing student numbers in French the French department was laterclosed and the charter corpus was lost forever With the help of a Dutchresearcher the Amsterdam corpus of Old French literary texts was takenover by international community members in Old French A group ofscholars from Stuttgart Ottawa Nancy and Zurich lemmatized andXML-formatted the corpus and created the Le Nouveau CorpusdrsquoAmsterdam (Kunstmann amp Stein 2007)29

Later corpus-building initiatives in the Netherlands occurred in the1980s as Dutch linguists created numerous small corpora for their ownpurposes Most of these corpora have been lost Often they were not cre-ated as databases for other linguists (eg copyrights were not secured)Furthermore the Instituut voor Nederlandse Lexicologie (the Institute forDutch Lexicology INL) started in the 1980s to create lexical databases(preforms of corpora) and now administers Dutch language corpora30

Even later corpus initiatives include video and audio files which were facili-tated by the advancement of both hardware and software computertechnology Furthermore a corpus of spoken Dutch was initiated in 1999and completed in 2003 Since then several large-scale corpus projects havestarted that included written Dutch These projects included more andmore computational linguists31 This occurred at a time when the huma-nities faculties at Dutch universities faced financial problems whichimplied that almost all Dutch historical linguistics chairs were cut and onlymodern Dutch linguistics chairs were kept Usually CCL was not a specificreason for giving a researcher a tenured position or a full professorshipOnly in one case was CCL considered sufficiently important by the facultyto create a chair in that area At the same time there have been significantpolitical decisions including governmental agreements between Dutch andFlemish governmental authorities on compiling large Dutch language cor-pora such as the now nine-million-word corpus of spoken Dutch (theDutch Spoken Corpus)32 Apart from that large European and also Dutchprojects have been set up by the government and funding agencies to buildinfrastructures for the humanities riding on the hype of e-humanitiesInterviewees complained that they were often left out of NWO infrastruc-ture funding they lobbied for a while and in recent years they have beenmore successful However none of these decisions is specifically made forcorpus linguistics

We can conclude that Dutch linguists took part in the internationaldevelopment of corpus-building from as early as the 1960s but they did soto a relatively limited extent The pioneers were men below 50 with

protected space obtained through tenured university posts mostly fundedby their university and with research council grants33 It is particularlyworth noting that the first large corpus was created in a university oftechnology a circumstance that again points to the cross-disciplinary char-acter of CCL Later on researchers in Dutch came to play a more promi-nent role through the creation of a large spoken-language corpus The1980s and early 1990s were filled with attempts by individual researchers tobuild small corpora Only in the second half of the 1990s did large-scalecorporus projects start with direct state support partly for politicallyinspired purposes

Sweden

In Sweden the three first CCL corpora were built up during a very shorttime in the 1960s (Table 4) They concerned present-day Swedish Germanand French respectively and were created at the universities of GothenburgLund and Stockholm The material for the first two consisted of news-paper texts while the third used bestselling novels For all three corporathe texts were taken from the 1960s with corpora sizes of 1 3 and 05 millionwords respectively The first two were externally funded while the smallestone was financed by the university All three resulted in dictionaries

The first major corpus in Sweden the Modern Swedish corpus (secondcolumn of Table 4) was created by Sture Allen (b 1928)34 After a disserta-tion based on seventeenth-century letters (Allen 1965) where he even atthat time took advantage of computer technology Allen set up a researchgroup to study modern Swedish by means of computers He was financedby the Bank of Sweden Tercentenary Foundation which was created in1964 (now called the Swedish Foundation for Humanities and SocialSciences) as well as the Swedish Council for Research in Humanities andmanaged to get access to texts from morning newspapers in StockholmGothenburg and Malmo in machine-readable form In this way hecompiled a corpus of one million words The output of the project was anumber of dictionaries (Allen 19701980 Allen 1972 Berg 1978) Inaddition the research led to the foundation in 1975 of the Language Bank(Sprakbanken) in Gothenburg as a national center of computational lexico-graphy35 At that time Allen had had a personal chair financed by theResearch Council since 1972 In 1980 he was elected as one of the 18 mem-bers of the Swedish Academy As the Secretary of the Academy19861999 he was able to link his own work to the long-time project of

the Academy from its foundation in 1786 to publish an extensive dictionaryof the Swedish language36

In 1967 Inger Rosengren (b 1934) of Lund University followed Allenwho had been the external examiner of her thesis on adjectives in MiddleHigh German (Rosengren 1966) by building up a corpus of ModernGerman (third column of Table 4) Her project on corpus linguisticswas supported by the Swedish Research Council for Humanities andSocial Sciences and two private foundations (Nathhorska stiftelsen andLangmanska kulturfonden see Rosengren 1972 p VI) For the corpus-building she managed to get access to the six-channel tapes that were

Table 4 Three Early Corpora in Sweden

Corpus Modern Swedish Modern German Modern French

Start year 1965a 1967 1968

Principal Investigator

Sture Allen (b 1928) Inger Rosengren

(b 1934)

Gunnel Engwall

(b 1942)

Age of PI at the start 37 33 26

Language Swedish German French

Institution Gothenburg University Lund University Stockholm

University

Type of material Morning dailies of the

1960s from Stockholm

Gothenburg and Malmo

Texts from Die

Welt and

Suddeutsche

Zeitung

19661967

25 French

bestsellers

19621968

words)

10 30 05

Funding The Bank of Sweden

Tercentenary Foundation

and the Swedish Research

Council

Research council

and private

foundations

University funding

Output Dictionaries and a data

Dictionaries Thesis and

dictionary

References Allen (19701980) Rosengren (1972

Engwall (1974

aAs mentioned in the text Allen had used computer technology already in his dissertation

(Allen 1965)

used for the typesetting of the two German newspapers Die Welt andSuddeutsche Zeitung From these she excluded certain categories andsampled texts for the period November 1 1966 to October 30 1967 endingup with a corpus of close to three million running words (2476560 for DieWelt and 500334 for Suddeutsche Zeitung (Rosengren 1972 p XXIV)The data processing used adaptations of programs that had been developedby the Allen research group (Rosengren 1972 p V) The output of theproject was two frequency dictionaries published in the 1970s (Rosengren1972 1977) As early as a year before the publication of the first volume(ie 1971) Rosengren had become professor of German languages at LundUniversity After the project was finished she turned to more generallinguistics publishing particularly in the field of pragmatics (cf egRosengren 1981 1984 1987)

The third corpus project in Sweden Modern French (fourth columnof Table 4) started in 196837 The corpus builder was Gunnel Engwall(b 1942) She was assigned the task as a thesis topic by her professor whoalso financed the work through an assistantship38 The project entailedcertain complications associated with the specific signs used in French andthe limited knowledge about corpus studies in the department Howeverafter Gunnel Engwall had accidentally met a librarian the KWIC programalso used by the Bonner Zeitungskorpus (see above) could be used for thecorpus work In order to cover important French novels in the 1960s sheused the best-seller lists published by the two French literary magazines LesNouvelles litteraires and Le Figaro litteraire for the period 19621968Together these lists contained 400 different titles 25 of which were singledout by means of a number of criteria (Engwall 1994 pp 6064) Fromeach of the 25 novels 20000 words were selected thus making the corpushalf a million words After lemmatization of the words in 10 of the 25novels she defended her dissertation in 1974 (Engwall 1974) This pavedthe way for a post-doctoral position during which period she could finishher work with all 25 novels and publish the results for the whole corpus(Engwall 1984) The material has also been included in the French corpuslibrary INaLF mentioned above Later on Gunnel Engwall continued withcolleagues to build corpora based on media material lrsquoExpress and leMonde as well as television broadcasts (Engwall 1996) Also in herresearch on Strindberg as a French author (see eg Engwall 1998 2009)corpus-building has been pursued With the passage of time the corpuswork appears to have been appreciated She was appointed to a chair inRomance Languages in Stockholm 1989 and has had various assignmentsin the Swedish academic community39

Among later corpus initiatives Jens Allwood (b 1947) took the track ofthe Quirk group focusing on spoken language while Ake Viberg (b 1945)starting out as a generativist and eventually funded by a state agencymoved into corpus linguistics through an interest in second-language acqui-sition Lars Borin (b 1957) a student of Slavic languages eventuallythrough external funding became a computational linguist and is todaythe head of Sprakbanken (cf above) For English finally Merja Kyto (b1953) trained in Finland with experience of building the Helsinki corpusis nowadays the leading scholar in Sweden40 She is well connected tothe international scientific community as secretary in the InternationalComputer Archive of Modern and Medieval English (ICAME)

Concluding the Swedish case we can note that pioneers representeddifferent universities as well as different languages As in Germany and theNetherlands the first corpus was directed toward the national languageIt was followed within a short time of corpus-building for German andFrench Funding for these came from foundations the major Swedishresearch council and universities with sufficient protected space providedby university posts and fellowships The first two projects took advantageof ready access to machine-readable texts from newspapers a circumstancethat also made copyright issues less problematic In contrast with Germanyand the Netherlands no national institute was created Instead theGothenburg group eventually was given the task to run the national databank for Swedish41

Switzerland

While researchers in the other three countries developed computer corporain the 1960s little such work was carried out in Switzerland until about 30years later (Table 5) The three early Swiss corpora also differed by focusingon English rather than any of the four official national languages They wereconstructed at the University of Zurich and were funded through a mix ofexternal and internal support with the major outputs being data banks42

The CCL pioneer in Switzerland appears to be Udo Fries (b 1942) Inhis own words in a response to his former colleague Gunnel Tottie43

Probably I was the first [corpus linguist] in other departments they were rather slow I

remember I was in Nijmegen in 1985 and heard about corpora from Jan Aarts and

Nelleke Oostdijk at a reception in Flor Aartsrsquo house Then I had some long talks with

Matti Rissanen about the planned Helsinki Corpus I forget where that was but he

invited me to come to the ICAME conference in HanasaariHanaholmen in 1987 but it

took some years before I hit on early newspapers in the early 1990s

Fries an immigrant to Switzerland from Austria studied at theUniversity of Vienna where he defended his thesis (Fries 1968) dealingwith the language of medieval mystery plays44 In 1973 he was awardedhis Habilitation and in 1974 a professorship of English Language andLiterature at the University of Innsbruck In 1977 he was recruited to theUniversity of Zurich as ldquoOrdinarius fur Englische Sprachwissenschaftrdquo Hisinvolvement in CCL concerned the production of the Zurich EnglishNewspaper Corpus (ZEN second column in Table 5) It covers Englishnewspapers in the period 16611791 which means that it contains materialfrom the early issues of The London Gazette until the first publication ofThe Times All in all the corpus consists of 349 complete newspaper issuesand contains 16 million words For obvious reasons these had to be keyed

Table 5 Three Early Corpora in Switzerland

Corpus Zurich English Newspaper

Corpus

COMPENG Corpus Fiji English

Corpus

Start year 1993 1998 2008

Principal

Investigator (PI)

Udo Fries (b 1942) Gunnel Tottie

(b 1937)

Marianne Hundt

(b sim1965)

Age of PI at the

51 61 sim43

Language English English English

Institution University of Zurich University of Zurich University of

Zurich

Type of material 349 newspaper issues

16611791

Old newspaper

texts Continuation

of ZEN

Spoken and

written Fiji

English

words)

16 150 10

Funding Stiftung fur Wissenschaftliche

Forschung Universitat

Zurich Schweizerischer

Nationalfonds

Der Schweizerische

Nationalfonds

University

funding

Output ZEN-Online Data bank Data bank

References httpwwwhelsinkifivarieng

CoRDcorporaZENindex

Fischer Tottie and

Lehmann (2002)

Biewer et al

(2010)

in manually The corpus is now available on CD-ROM and also through aweb-based system for searching the corpus (ZEN-Online)45 Over the yearsFries has also published several papers in relation to the ZEN corpus (egFries 1994 1997 2001 2002 2006) and an edited volume (Fries Muller ampSchneider 1997) He has been an active member of ICAME network andat the University of Zurich he was entrusted with leadership responsibilitiesby his colleagues46

At about the same time as Udo Fries started his work on corpus linguis-tics the Swede Gunnel Tottie (b 1937) joined the English Department at theUniversity of Zurich and later became involved in the second Swiss corpusCOMPENG (third column in Table 5) She had defended her dissertation(Tottie 1971) at Stockholm University in Sweden and had come acrossCCL through a research position in the department of the above-mentionedJan Svartvik at Lund University and a Longman fellowship in 1978 withQuirkrsquos group in London47 As a result she used the London-Lund as well asthe Brown corpus for a multivariate analysis of negations (Tottie 1991)Between 1998 and 2000 Gunnel Tottie took part in the researchCOMPENG project (Computerunterstutzte englische Korpuslinguistik) whichwas funded by the Swiss Nationalfonds It aimed at finalizing the ZEN cor-pus and creating a large newspaper corpus of 150 million words The corpuswas used for various projects on historical newspaper language48

Both Fries and Tottie were thus recruited from abroad The same is truefor the builder of the Fiji English corpus (fourth column in Table 5)Marianne Hundt since 2008 Professor of English Linguistics at University ofZurich49 She did her PhD (Hundt 1996) and Habilitation (Hundt 2002) atthe Albert-Ludwigs-Universitat in Freiburg During her Habilitation workshe worked with Professor Christian Mair who was involved in the creationof several corpora ldquoamong them F-LOB and Frown updates of the classicLOB and Brown corpora [hellip] and the Jamaican component of theInternational Corpus of Englishrdquo50 In other words Marianne Hundt wastrained in an environment where CCL was in focus As a result over the yearsshe also published a number of CCL papers with Christian Mair and others(eg Biewer Hundt amp Zipp 2010 Hundt amp Mair 1999 Mair HundtLeech amp Smith 2002)51 Currently Marianne Hundt and her colleagues areworking on a corpus of Fiji English in a new Centre for Linguistics52

As far as later developments are concerned in the view of MiriamLocher professor of English at the University of Basel ldquoin fact all of us[work with corpora] in one way or anotherrdquo53 Among her colleagues inBasel she mentions Annelies Hacki-Buhofer (long experience with large-scale projects and databases for German) Heike Behrens (EnglishGerman

language acquisition using CHILDES corpus) Miriam Locher (qualitativework on English corpora) Angela Ferrari and Anna-Maria De Cesare(text linguistics with Italian corpora) and Lorenza Mondada (vast multi-modal French corpora) Likewise interviews by one of us (Ramuz) at theUniversity of Geneva have identified corpus linguists

Before concluding the Swiss case it should be noted that althoughSwitzerland was late to develop CCL there has long been a Swiss interestin lexicography The oldest of these projects is Worterbuch der schweizer-deutschen Sprache (Swiss-German language dictionary also known asIdiotikon) which presents the German language in Switzerland since thethirteenth century It started in the nineteenth century and is expected to befinished by 2022 Three other continuing projects within VocabulairesNationaux (National Glossaries) were started in the early or mid-twentiethcentury Glossaire des patois de la Suisse romande (Glossary of French-speaking Switzerlandrsquos Dialects) Vocabolario dei dialetti della Svizzera itali-ana (Vocabulary of Italian-speaking Switzerlandrsquos Dialects) and DicziunariRumantsch Grischun (Romansh Graubunden Dictionary)54

All in all we can note that in relation to the development in other coun-tries CCL appears to have been adopted relatively late in SwitzerlandAnother significant feature is that the pioneers were immigrants into theSwiss academic system where they had been appointed to chairs after aresearch career abroad As a result they were somewhat older than the pio-neers in other countries and benefited from the protected space afforded toprofessors in the Swiss university system as well as access to universitysupport and funds from the national research council Finally it is worthnoting that the pioneers did not deal with any of the four national

Table 6 The First Three Corpora in the Four Countries

Corpus Germany Netherlands Sweden Switzerland

First 1962 Modern German

(encyclopedia and

newspaper)

1969 Modern Dutch

(spoken and written

19601973)

1965 Modern

Swedish

(newspapers)

1993 Historical

English

(newspapers)

Second 1964 Modern German

(newspapers from FRG

and GDR)

1976 Modern English 1967 Modern

German

(newspapers)

1998 Historical

English

(newspapers)

Third 1968 Modern German

(spoken public sources)

Late 1970s (thirteenth-

century French

charters)

1968 Modern

French

(bestsellers)

2008 Fiji

English (spoken

and written)

languages but instead with English although more recent corpora do dealwith them

CONCLUSIONS

The development of computer technology offered language scholars newopportunities for empirical studies through CCL As a result a number ofCCL pioneers appeared in France Italy the United Kingdom and theUnited States We expected similar developments to have occurred fromthe 1960s and onwards in the four countries under study (Germany theNetherlands Sweden and Switzerland) Our analysis has shown this to bethe case with one exception Switzerland (Table 6) In this country CCLdeveloped only in the 1990s after the recruitment of foreign linguists whileCCL started in the other three countries in the 1960s Germany andSweden even had three early CCL projects in the 1960s while the secondand third CCL projects in the Netherlands had to wait until the 1970s

This suggests that while linguists in Germany and Sweden followed inter-national developments closely those in the Netherlands and Switzerlandwere slower to adapt For Germany there is no doubt that the initiative tocreate das Institut der Deutsche Sprache (IDS) was significant for both theearly and the later development with a strong focus on studies of GermanAs in the case of BEC discussed by Laudel et al elsewhere in this volumethe substantial resources and protected space provided by such a state orga-nization enabled German researchers to develop significant CCL innovationsover a considerable period of time Nothing similar was launched in Swedenwhere instead the combination of an academic entrepreneur internationalinfluences and funding from a variety of agencies and foundations seems tohave initiated developments and stimulated followers The level of protectedspace needed to develop computerized corpora was provided by a successionof fellowships and university positions not all at the level of full professorEventually an institution similar to IDS developed in Gothenburg but thiswas more a bottom-up project than a top-down initiative

The later developments in the Netherlands and Switzerland appearaccording to our interviews to be the result of stronger alternative researchcommunities than in Germany and Sweden In the case of the Netherlandsthe generativists had a firmer grip and in Switzerland in addition to theChomskyans there were also strong groups of historical linguistics As wehave seen above CCL pioneers in the other countries had a less historicalorientation most of the corpora were directed toward language use in

modern times particularly through studies of media material (newspapertexts and broadcasted materials)

There are also reasons to believe that the late adoption of CCL inSwitzerland was a result of the Swiss language situation with four officiallanguages in stark contrast to the three early countries with one dominantlanguage each The pioneers in the other countries thus started out with themother tongue while in Switzerland none of the three of the early corporastudied any of the official four Swiss languages but instead English

Although it is difficult to determine the differences between the two groupsof countries might also be related to the contrasting levels of support for over-coming the technical problems and resource requirements for the building ofcorpora It is important to remember that computer technology at the timewas a far cry from what is offered today even by personal computers

While the major force behind the development of CCL in the four coun-tries as well as in other countries has been the development of computertechnology its diffusion was significantly stimulated by the success of inter-national role models such as the Brown corpus of Kucera and Francis(1967) and the work of the Quirk group in London We can also note thatthe adoption of CCL has to a large extent been the result of initiatives byacademic entrepreneurs most of them men in their early or mid-careerswho saw the potential of the new technology Furthermore there is no evi-dence that they have suffered from their going into CCL Rather it appearsthat they have had successful academic careers

Finally it should be noted that further technological advances havebeen very significant for later developments They have ensured that largecorpora are widely accessible and small corpora can be relatively easilycreated in the twenty-first century Therefore the divide between corpuslinguists and Chomskyans is no longer so dramatic What once was aninnovation has become an everyday tool for linguists It has also becomesignificant for modern information technology for instance in smart-phones In this way CCL is another demonstration of the unexpected useof basic research What was once looked down upon has become widelyused both in academic work and in practice

1 Today the field is labeled simply as ldquocorpus linguisticsrdquo (cf eg Taylor2008) Nevertheless due to the historical perspective of the present paper we willuse the term ldquocomputer corpus linguisticsrdquo throughout

2 In addition to the mentioned volumes Douglas Biber and Randi Reppenhave a contract with Cambridge University Press for The Cambridge Handbook ofCorpus Linguistics (personal communication by Merja Kyto)

3 See eg httpwwwehumanitiesnl4 Following Whitley (2000) integration is considered to be high within a scienti-

fic field when task uncertainty is low and the interdependence between scholars inthe field is high See also Engwall (1995)

5 See for instance Enkvist Ferguson Hajicova and Ladefoged (1992) and CSS(1995a 1995b 1997) which summarize developments in Sweden and Switzerlandrespectively

6 For an introduction to Chomskyrsquos work see eg Collins (2008) It shouldbe noted that Chomsky (1957) was published by the Dutch publisher MoutonAccording to linguists it had been rejected by US publishers

7 Cf eg httpwwwguardiancoukhigher-education-networkblog2011aug12father-roberto-busa-academic-impact

8 This paragraph is based on an interview with Bernard Quemada by GunnelEngwall on May 9 2011

9 Incidentally Bernard Quemada got the idea to use punched cards for his lan-guage studies by observing a serviceman from the electricity company using suchcards for registering meter readings10 The work at INaLF provided the basis for Le Tresor de la Langue Francaise

Informatise (TLF) which is a dictionary of the French language available onlineCD and as books (Tresor de la langue francaise informatise 2004)11 httpdsav-wissids-mannheimdeKORPORAPFPF_DOKUHTM and

Pfeffer (1964)12 See further Svartvik and Quirk (1980) and Svartvik (1990) Svartvik also con-

tributed to the integration of the field by arranging a Nobel symposium in 1991(Svartvik 1992)13 httpkhnthituibnoicamemanualslobindexhtm and Johansson (2008)

The year before the LOB corpus was completed (1977) the International ComputerArchive of Modern English (ICAME) had been founded by five key researchersamong them Nelson Francis Geoffrey Leech Stig Johansson and Jan Svartvikwith the stated purpose of assembling all available English corpora (httpicameuibnohistoryfounding_document_1977pdf) However the main reason for thefoundation of ICAME was to solve copyright problems with the LOB corpus (inter-view by Lars Engwall with Geoffrey Leech May 9 2013)14 Of course Chomsky has also been criticized see eg Gross (1979) Nist

(1965) and Ohman (2007)15 Interview with Robert Martin by Gunnel Engwall on May 11 201116 Another such project is the Harper Collins Dictionary See further Sinclair

(1987)17 httpwwwnatcorpoxacuk18 A notable venture in this context was the British Alvey program at the

University of Cambridge University of Edinburgh and Lancaster University in19831987 in order to provide tools for natural language processing The programwas a response to the Japanese Fifth Generation Computer Systems Program(Oakley amp Owen 1989)

19 httpwwwaclweborgarchivemiscHistoryhtml20 httpwwwleibniz-gemeinschaftdeeninstitute-museeneinrichtungenids It

is today one of the 86 institutions belonging to the Leibniz-Gemeinschaft21 httpwww1ids-mannheimdeklprojektekorporaarchivbzkhtml and

Hellmann (1984 1985 1992)22 httpdsav-oeffids-mannheimdeDSAvKORPORAFRFR_DOKUHTM

and Steger (1967)23 httpwww1ids-mannheimdeklprojektekorporaarchivmkhtml and

Hellmann (1976 pp 248249)24 httpwww1ids-mannheimdeklprojektekorporaarchivlimhtml and

Hellmann (1976 p 250)25 httpwww1ids-mannheimdeklprojektekorporaarchivhtml See also

Schaeder (1978) and Kupietz Belica Keibel and Witt (2010)26 See further Uit den Boogaart (1975 p 11) where it is also pointed out that

Willy Martin in Louvain Belgium as early as 1967 published the results from a100000-word corpus of Flemish newspapers and novels (Martin 1967) See alsoBaayen (1996)27 This paragraph is based on Uit den Boogaart (1975)28 The support from NWO (the Dutch Research Council for Advanced

Research) lasted from March 1 1985 until March 1 1989 (Oostdijk 1991 p 47footnote 17)

29 See also Glessgen and Vachon (2010) and httpwwwuni-stuttgartdelingromsteincorpus

30 ldquoThe Institute for Dutch Lexicology (INL) is a research institute financiallysupported by the governments of the Netherlands and (Flemish) Belgium Itsmission is to document the vocabulary and grammar of present-day and historicalDutch by creating maintaining and improving the accessibility of languageresources like dictionaries corpora computational lexica thesaurirdquo (httptranscriptoriumeuconsortiuminstitute-for-dutch-lexicology) In 2014 INL willarrange the 24th Meeting of Computational Linguistics in the Netherlands (httpwwwinlnl)

31 Corpus-building in the Netherlands from the 1990s onwards moved into thedomain of computational linguistics The ldquorealrdquo linguists participated in activitieslike ldquolinguistically enriching corporardquo although these seem to be only minor activ-ities in their overall research portfolio32 httplandsletrunlcgndoc_Englishtopicsprojectpro_infohtmintro33 Flor Aarts wrote in the 1994 Festschrift to Jan Aarts ldquoThe initiative that Jan

took more than 20 years ago has borne fruit Names like TOSCA and LDB havebecome household words in English corpus linguistics The reputation thatNijmegen now enjoys is clearly attested by the role it has been assigned in theInternational Corpus of English projectrdquo (Aarts F 1994 p 3) For Anthonij Deessee his Festschrift (Reenen amp Reenen-Stein 1988)34 This section is based on an interview of Sture Allen by Lars Engwall on

November 17 2011 the references provided in the text and the presentation ofSture Allen on the webpage of the Swedish Academy (httpwwwsvenskaakade-mienseenthe_academymembers938e01b1-b318-4c23-ba05-954127697d2a)

35 httpspraakbankenguse

36 Allenrsquos membership in the Swedish Academy also made it easier for him toarrange a Nobel symposium on text processing in 1980 (Allen 1982)37 In addition to the referred sources this section is based on an interview with

Gunnel Engwall by Tina Hedmo on April 19 201138 Also in this case there seems to have been international inspiration since the

professor in question had contacts with his Italian colleague Antonio Zampolli(19372003) a student of Father Busa mentioned above and an eminent CCLresearcher39 Gunnel Engwall was Pro Vice Chancellor (19942003) and Acting Vice

Chancellor (20032004) of Stockholm University She has also been member ofvarious bodies for research financing and President of the Royal Swedish Academyof Letters History and Antiquities (20062013) From 1988 to 1996 she was aboard member of Association for Literary and Linguistic Computing (ALLC)40 The Helsinki corpus was compiled in the period 19841991 and contains

1572800 words from the period from 730 to 1710 It was led by Matti Rissanen(httpwwwhelsinkifivariengCoRDcorporaHelsinkiCorpus)41 For more extensive information on the Swedish case see Engwall and

Hedmo (2012b)42 For more extensive information on the Swiss case see Ramuz (2012)43 See also Fries (1988) Note in the quote the significance of the Dutch pioneer

Jan Aarts and his student Nelleke Oostdijk (cf above on the Nijmegen corpus)44 The information in the following is taken from httpwwwesuzhch

aboutusteamufrieshtml45 httpwwwesuzhchSubsitesProjectszencorpushtml and httpes-zen

unizhch46 Udo Fries was an ICAME board member for many years and was the orga-

nizer of the 14th annual meeting in Zurich in 1993 (Fries Tottie amp Schneider1994) In terms of responsibilities in the university he was first dean of the Facultyof Philosophy in 19961998 then one of the Vice-Presidents in 19982004 He wasactive in CCL during these periods as well47 She describes the working atmosphere at Quirkrsquos unit in the following way

ldquo[We] had coffee and tea in Quirkrsquos office every day He rang a bell He would turnon the kettle because there was water in his office [and] when it was boiling therewas a buzz into the research room [We] were like four or five people perhapspeople working on the survey and we would go in and have coffee and he wouldsometimes join and sometimes notrdquo48 The relationship to Fries and corpora is further manifested through the

co-edited festschrift to him (Fischer Tottie amp Lehman 2002)49 This section is based on Marianne Hundtrsquos response to Gunnel Tottie and

her personal page httpwwwesuzhchSubsitesPersonalmhundthtml50 httpwwwhelsinkifiglobe13plenarieshtml51 She has also been linked to the international community through a visiting

professorship at Portland State University (httpwwwllcmanchesteracukresearchprojectsarcher) (summer term 1998) and involvement in the ICAMEorganization See further httpwwwesuzhchSubsitesPersonalmhundtPublicationsHundthtml

52 httpwwwesuzhchSubsitesPersonalmhundtProjectsHundthtml

53 Personal communication to Gunnel Tottie54 httpwwwsagwchsagwdie-akademieunternehmennwbhtml and http

wwwidiotikonchindexphpoption=com_contentandview=articleandid=98andItemid=145

ACKNOWLEDGMENTS

This article is an outcome of the research program ldquoRe-Structuring HigherEducation and Scientific Innovationrdquo (RHESI) within the EuropeanScience Foundation initiative ldquoHigher Education and Social Changerdquo(Euro-HESC) The research has been supported by the DeutscheForschungsgemeinschaft (DFG Grant Schi 5537-1) the NederlandseOrganisatie voor Wetenschappelijk Onderzoek (NWO) the SwedishResearch Council (VR Grant 90671701) and the Swiss National ScienceFoundation (SNF Grant 125814) We are very grateful to Grit Laudel forsharing information on the Netherlands based on her interviews

REFERENCES

Aarts F (1994) A tribute to Jan Aarts In J M G Aarts S Johansson B Altenberg J

Svartvik A-B Stenstrom N Oostdijk amp P Haan (Eds) Corpus-based research into

language In honour of Jan Aarts (pp 13) Amsterdam Rodopi

Aarts J amp Meijs W (Eds) (1984) Corpus linguistics [1] recent developments in the use of

computer corpora in English language research Amsterdam Rodopi

Aarts J M G Ditters E amp Koster C H A (1987) TOSCA The Nijmegen research group

for corpus linguistics Nijmegen English Department University of Nijmegen

Allen S (1965) Grafematisk analys som grundval for textedering med sarskild hansyn till

Johan Ekeblads brev till brodern Claes Ekeblad 16391655 Almqvist amp Wiksell

Goteborg (Nordistica Gothoburgensia 1 Diss)

Allen S (19701980) Nusvensk frekvensordbok baserad pa tidningstext Frequency dictionary

of present-day Swedish based on newspaper material Data linguistica 1-4 Almqvist amp

Wiksell International Stockholm

Allen S (1972) Tiotusen i topp ordfrekvenser i tidningstext Data linguistica 6 Almqvist amp

Wiksell Stockholm

Allen S (1982) Text processing Text analysis and generation text typology and attribution

Proceedings of Nobel symposium 51 Almqvist amp Wiksell International Stockholm

Baayen R H (1996) The randomness assumption in word frequency statistics Research in

Humanities Computing 5 6996

Berg S (1978) Olika lika ord svenskt homograflexikon Data linguistica 12 Stockholm

Almqvist amp Wiksell International

Biber D Conrad S amp Reppen R (1998) Corpus linguistics Investigating language structure

and use Cambridge Cambridge University Press

Biewer C Hundt M amp Zipp L (2010) lsquoHowrsquo a Fiji corpus Challenges in the compilation

of an ESL ICE component ICAME Journal 34 523

Buchanan M A (1931) A graded Spanish word book American and Canadian Committees

on Modern Languages Series 3 Toronto University of Toronto Press

Busa R (1951) Sancti Thomae Aquinatis hymnorum ritualium varia specimina concordan-

tiarum Primo saggio di indici di parole automaticamente composti e stampati Archivum

philosophicum Aloisianum Series 2 Milan Fratelli Bocca

Cheydleur F D (1934) French idiom list based on a count of 1183000 running words

American and Canadian Committees on Modern Languages Series 16 New York NY

Macmillan

Chomsky N (1957) Syntactic structures The Hague Mouton

Chomsky N (1965) Aspects of the theory of syntax Cambridge MA MIT Press

Chomsky N (1972) Studies on semantics in generative grammar The Hague Mouton

Chomsky N (1982) Noam Chomsky on the generative enterprise A discussion with Riny

Huybregts and Henk van Riemsdijk Dordrecht Foris

Chomsky N (1995) The minimalist program Cambridge MA MIT Press

Chomsky N (2002) On nature and language Cambridge Cambridge University Press

Collins J (2008) Chomsky A guide for the perplexed London Continuum

CSS (1995a) Evaluation de la recherche en sciences humaines en Suisse Rapport de base pour

langues et litteratures romanes Politique de la recherche Berne FOP

CSS (1995b) Evaluation de la recherche en sciences humaines en Suisse Rapport de base pour

langue et litterature allemandes Politique de la recherche Berne FOP

CSS (1997) Evaluation de la recherche en sciences humaines en Suisse La recherche suisse en

sciences humaines Qualites contraintes et perspectives drsquoavenir Politique de la recherche

Berne FOP

De la Court J F H A (1937) De meest voorkomende woorden en woordcombinaties in het

Nederlandsch Batavia Volkslectuur

Dees A (Ed) (1980) Atlas des formes et des constructions des chartes francaises du 13e siecle

Beihefte zur Zeitschrift fur romanische Philologie Series 178 Tubingen Niemeyer

Dees A Huber O amp van Rennen-Stein K (Eds) (1987) Atlas des formes linguistiques des

textes litteraires de lrsquoancien francais Beihefte zur Zeitschrift fur romanische Philologie

Series 212 Tubingen Niemeyer

Eggers H (1964) Erstes Kolloquium uber Syntax naturlicher Sprachen und Datenverarbeitung

Saarbrucken 2930 April 1963 Wiesbaden F Steiner

Eggers H (Ed) (1969) Elektronische Syntaxanalyse der deutschen Gegenwartssprache ein

Bericht Tubingen Niemeyer

Engwall G (1974) Frequence et distribution du vocabulaire dans un choix de roman francais

Dissertation Stockholm Skriptor sprakforlag

Engwall G (1984) Vocabulaire du roman francais 19621968 dictionnaire des frequences

Data linguistica 17 Stockholm Almqvist amp Wiksell International

Engwall G (1994) Not chance but choice Criteria in corpus construction In S B T Atkins amp

A Zampolli (Eds) Computational approaches to the Lexicon (pp 4982) New York

NY Oxford University Press

Engwall G (1996) Corpus de francais etablis en Suede Revue Francaise de Linguistique

Appliquee 1(2) 8990

Engwall G (1998) Strindberg et ses contacts francais In S Mellet amp M Vuillaume (Eds)

Mots chiffres et dechiffres Melanges offerts a Etienne Brunet (pp 473501) Paris

Honore Champion Editeur

Engwall G (2009) Orthonet et les Vivisections drsquoAugust Strindberg In C Delcourt amp

M Hug (Eds) Melanges offerts a Charles Muller pour son centieme anniversaire 22

septembre 2009 (pp 169181) Paris CILF

Engwall L (1995) Management research A fragmented adhocracy Scandinavian Journal of

Management 11(3) 225235

Engwall L amp Hedmo T (2012a) The emergence and organizing of scientific fields The case

of corpus linguistics Paper for sub-theme 17 Organizing Science at the 28th EGOS

Colloquium Helsinki July 57 2012

Engwall L amp Hedmo T (2012b) Corpus linguistics in Sweden Paper for the RHESI work-

shop in Berlin April 1819 2012

Enkvist N E Ferguson C A Hajicova E amp Ladefoged P (1992) Linguistic research in

Sweden The Swedish Council for Research in the Humanities and Social Sciences

Stockholm

Filmore C J (1992) lsquoCorpus Linguisticsrsquo or lsquoComputer-aided Armchair Linguisticsrsquo In

J Svartvik (Ed) Directions in corpus linguistics Proceedings of Nobel symposium

Stockholm August 48 1991 Berlin Mouton de Gruyter (Vol 82 pp 3560)

Fischer A Tottie G amp Lehmann H M (Eds) (2002) Text types and corpora Studies in

honour of Udo Fries Tubingen Gunter Narr

Francis N W (1986) Dinner speech given at the 5th ICAME conference on computers in

English language research Windermere England May 21 1985 ICAME News Vol

10 May pp 57 Retrieved from httpicameuibnohistoryICAME_no_10_dinner_

speechpdf

Fries U (1968) Zur Syntax der Chester Plays Dissertationen der Universitat Wien 6

Notring Wien

Fries U (1988) The use of computers in English linguistics In M Markus (Ed) Historical

English On the occasion of Karl Brunnerrsquos 100th birthday Innsbrucker Beitrage zur

Kulturwissenschaft Anglistische Reihe (pp 4562) Innsbruck Institut fur Anglistik

Universitat Innsbruck

Fries U (1994) ZEN minus Zurich English newspaper corpus In M Kyto M Rissanen amp S

Wright (Eds) Corpora across the centuries Proceedings of the first international collo-

quium on English diachronic corpora St Catherinersquos College Cambridge March 2527

1993 Language and Computers Studies in Practical Linguistics Series 11 (pp 17 18)

Amsterdam Rodopi

Fries U (1997) The vocabulary of ZEN Implications for the compilation of a corpus

In R Hickey M Kyto I Lancashire amp M Rissanen (Eds) Tracing the trail of

time Proceedings from the second diachronic corpora workshop Language and

Computers Studies in Practical Linguistics Series 18 (pp 153166) Amsterdam

Rodopi

Fries U (2001) Text classes in early English newspapers European Journal of English Studies

5 167180

Fries U (2002) StudyingEnglish linguisticsin Switzerland In W Zach (Ed) English stu-

dies Pastpresentfuture (pp 6273) Hamburg Verlag Dr Kovac

Fries U (2006) Death notices The birth of a genre In R Facchinetti amp M Rissanen (Eds)

Corpus-based studies of diachronic English (pp 157170) Bern Peter Lang

Fries U Muller V amp Schneider P (Eds) (1997) From AEliglfric to the New York Times

Studies in English corpus linguistics Language and Computers Studies in Practical

Linguistics Series 19 Amsterdam Rodopi

Fries U Tottie G amp Schneider P (Eds) (1994) Creating and using English language cor-

pora Papers from the fourteenth international conference on English language research

on computational corpora Zurich 1993 Language and Computers Studies in Practical

Glessgen M-D amp Vachon C (2010) Repertoire bibliographique du nouveau corpus

drsquoAmsterdam etabli par Anthonij Dees et Piet Van Reenen Amsterdam 1987 revu et

elargi par M-DG et CV Stuttgart Institut fur LinguistikRomanistik

Gross M (1979) On the failure of generative grammar Language 55(4) 859885

Halliday M A K (2004) Lexicology and corpus linguistics An introduction London

Continuum

Hellmann M W (1976) Moglichkeiten und Probleme bei vergleichenden Wortschatzunter

suchungen zum offentlichen Sprachgebrauch in der Bundesrepublik und der DDR In

Institut fur Deutsche Sprache (Ed) Probleme der Lexikologie und Lexikographie

Sprache der Gegenwart Band Series 39 (pp 242274) Dusseldorf Schwann

Hellmann M W (Ed) (1984) Ost-West-Wortschatzvergleiche maschinell gestutzte

Untersuchungen zum Vokabular von Zeitungstexten aus der BRD und der DDR

Tubingen Gunter Narr Verlag

Hellmann M W (1985) Das Bonner Zeitungskorpus Teil 1 Informationen fur den Benutzer

Mitteilungen des IDS Band Series 11 (pp 94157) Mannheim IDS

Hellmann M W (1992) Worter und Wortgebrauch in Ost und West Ein rechnergestutztes

Korpus-Worterbuch zu Zeitungstexten aus den beiden deutschen Staaten die Welt und

Neues Deutschland 3 Bande Tubingen Gunter Narr Verlag

Henmon V A C (1924) A French word book based on a count of 400 000 running

words Bureau of Educational Research Bulletins Series 3 University of Wisconsin

Madison WI

httpdsav-oeffids-mannheimdeDSAvKORPORAFRFR_DOKUHTM

httpdsav-wissids-mannheimdeKORPORAPFPF_DOKUHTM

httpes-zenunizhch

httpicameuibnohistoryfounding_document_1977pdf

httpkhnthituibnoicamemanualslobindexhtm

httplandsletrunlcgndoc_Englishtopicsprojectpro_infohtmintro

httpspraakbankenguse

httptranscriptoriumeuconsortiuminstitute-for-dutch-lexicology

httpwwwaclweborgarchivemiscHistoryhtml

httpwwwehumanitiesnl

httpwwwesuzhchaboutusteamufrieshtml

httpwwwesuzhchSubsitesPersonalmhundthtml

httpwwwesuzhchSubsitesPersonalmhundtProjectsHundthtml

httpwwwesuzhchSubsitesPersonalmhundtPublicationsHundthtml

httpwwwesuzhchSubsitesProjectszencorpushtml

httpwwwguardiancoukhigher-education-networkblog2011aug12father-roberto-busa-

academic-impact

httpwwwhelsinkifiglobe13plenarieshtml

httpwwwhelsinkifivariengCoRDcorporaHelsinkiCorpus

httpwwwhelsinkifivariengCoRDcorporaZENindexhtml

httpwwwidiotikonchindexphpoption= com_contentandview=articleandid=98andItemid=145

httpwwwinlnl

httpwwwleibniz-gemeinschaftdeeninstitute-museeneinrichtungenids

httpwwwnatcorpoxacuk

httpwwwsagwch

httpwwwsagwchsagwdie-akademieunternehmennwbhtml

httpwwwsvenskaakademienseenthe_academymembers938e01b1-b318-4c23-ba05-954127

697d2a

httpwwwuni-stuttgartdelingromsteincorpus

httpwww1ids-mannheimdeklprojektekorporaarchivhtml

httpwww1ids-mannheimdeklprojektekorporaarchivbzkhtml

httpwww1ids-mannheimdeklprojektekorporaarchivmkhtml

Hundt M (1996) New Zealand English and its relation to British and American English

Dissertation Freiburg Albert-Ludwigs-Universitat

Hundt M (2002) English mediopassive constructions A cognitive corpus-based study of their

origin spread and current status habilitation Freiburg Albert-Ludwigs-Universitat

Hundt M amp Mair C (1999) lsquoAgilersquo and lsquouptightrsquo genres The corpus-based approach to lan-

guage change in progress International Journal of Corpus Linguistics 42(2) 221242

Institut fur Deutsche Sprache (Ed) (1974) Gesprochene Sprache Dusseldorf Schwann

Ijiri Y amp Simon H A (1977) Skew distributions and the sizes of business firms Amsterdam

North-Holland

Johansson S (2008) Some aspects of the development of corpus linguistics in the 1970s and

1980s In A Ludeling amp M Kyto (Eds) Corpus linguistics An international handbook

(pp 3353) Berlin Mouton de Gruyter

Juilland A Brodin D amp Davidovitch C (1970) Frequency dictionary of French words the

romance languages and their structures First series F1 The Hague Mouton

Juilland A amp Chang-Rodriguez E (1964) Frequency dictionary of Spanish words the

romance languages and their structures First series S1 The Hague Mouton

Juilland A Edwards P M H amp Juilland I (1965) Frequency dictionary of Rumanian

words the romance languages and their structures First series R1 The Hague Mouton

Juilland A amp Traversa V P (1973) Frequency dictionary of Italian words the romance lan-

guages and their structures First series I1 The Hague Mouton

Kaeding F W (18971898) Haufigkeitsworterbuch der deutschen Sprache 1-2 Steglitz bei

Berlin Selbstverlag des Herausgebers

Kennedy G (1998) An introduction to corpus linguistics London Longman

Keulen F (1986) The Dutch computer corpus pilot project Some experiences with a semi-

automatic analysis of contemporary English In J Aarts amp W Meijs (Eds) Corpus

linguistics volume 2 New studies in the analysis and exploitation of computer corpora

(pp 127161) Amsterdam Rodopi

Kucera H amp Francis N (1967) Computational analysis of present-day American English

Providence RI Brown University Press

Kunstmann P amp Stein A (2007) Le nouveau corpus drsquoAmsterdam In P Kunstmann amp

A Stein (Eds) Le Nouveau Corpus drsquoAmsterdam Actes de lrsquoatelier de Lauterbad

2326 fevrier 2006 (pp 927) Stuttgart Steiner

Kupietz M Belica C Keibel H amp Witt A (2010) The German reference corpus

DeReKo A primordial sample for linguistic research In N Calzolari et al (Eds)

Proceedings of the seventh conference on international language resources and evaluation

(LREC 2010) ELRA Malta 18481854 Retrieved from httpwwwlrec-conforg

proceedingslrec2010pdf414_Paperpdf)

Ludeling A amp Kyto M (Eds) (20082009) Corpus linguistics An international handbook

Mair C Hundt M Leech G amp Smith N (2002) Short-term diachronic shifts in part-of-

speech frequencies A comparison of the tagged LOB and F-LOB corpora

International Journal of Corpus Linguistics 7(2) 245264

Martin W (1967) De inhoud van krant en roman Een frequentieonderzoek Antwerp

Uitgeverij Plantyn

McEnery T amp Hardie A (2011) Corpus linguistics Method theory and practice Cambridge

Meyer C F (2002) English corpus linguistics An introduction Cambridge Cambridge

University Press

Morgan B Q (1933) German frequency word book based on Kaedingrsquos Haufigkeitsworterbuch

der deutschen Sprache Arranged and edited by B Q Morgan American and Canadian

Committees on Modern Languages Series 9 New York NY Macmillan

Morris W (Ed) (1969) The American Heritage dictionary of the English language New

York NY American Heritage

Nist J (1965) A critique of generative grammar Linguistics 3(19) 102110

Oakley B amp Owen K (1989) Alvey Britainrsquos strategic computing initiative Cambridge MA

MIT Press

Ohman S (2007) The essence of language A philosophical problem Why Noam Chomsky was

never a linguist Nora Nya Doxa

OrsquoKeeffe A amp McCarthy M (Eds) (2010) The Routledge handbook of corpus linguistics

Abingdon Routledge

Oostdijk N (1991) Corpus linguistics and the automatic analysis of English Amsterdam Rodopi

Pfeffer J A (1964) Grunddeutsch Basic (Spoken) German word list grundstufe Englewood

Cliffs NJ Prentice-Hall

Quemada B (Ed) (19591993) Materiaux pour lrsquohistoire du vocabulaire francais datations et

documents lexicographiques Part 130 Besancon Centre drsquoetude du vocabulaire

francais

Quemada B (1968) Les dictionnaires du francais moderne 15391863 etude sur leur histoire

leurs types et leurs methodes Dissertation Didier Paris

Quirk R (1957) Relative clauses in educated spoken English English Studies 38(1)

Quirk R Greenbaum S Leech G amp Svartvik J (1985) A comprehensive grammar of the

English language London Longman

Quirk R amp Svartvik J (1972) A grammar of contemporary English London Longman

Quirk R amp Svartvik J (1978) A corpus of modern English Lund University of Lund

Ramuz R (2012) Corpus linguistics in Switzerland Working Paper 012012 Paper for the

RHESI Workshop in Berlin April 1819 2012

Rosengren I (1966) Semantische Strukturen eine quantitative Distributionsanalyse einiger

mittelhochdeutscher Adjektive Dissertation Gleerup Lund

Rosengren I (1972) Ein Frequenzworterbuch der deutschen Zeitungssprache Die Welt

Suddeutsche Zeitung 1 Lunder germanistische Forschungen 41 Lund LiberLaromedel

Gleerup

Rosengren I (Ed) (1981) Sprache und Pragmatik Lunder Symposium 1980 Lunder germa-

nistische Forschungen 50 Lund LiberLaromedelGleerup

nistische Forschungen 53 Stockholm Almqvist amp Wiksell International

de Saussure F (1916) Cours de linguistique generale Lausanne Payot

Schaeder B (1978) Eine Dokumentation maschinenlesbarer Textcorpora der deutschen

Gegenwartssprache In D Mohn H Weinrich W Mentrup amp H Henne (Eds)

Interdisziplinares deutsches Worterbuch in der Diskussion (Sprache der Gegenwart

Schriften des Instituts fur deutsche Sprache 45) (pp 233254) Dusseldorf Schwann

Simon H A (1955) On a class of skew distributions Biometrika 52(34) 425440

Sinclair J M (1987) Looking up An account of the COBUILD project in lexical computing

and the development of the collins COBUILD English language dictionary London

HarperCollins

Steger H (1967) Gesprochene Sprache Zu ihrer Typik und Terminnologie In Institut fur

Deutsche Sprache (Ed) Satz und Wort im heutigen Deutsch (Sprache der Gegenwart 1)

(pp 259291) Dusseldorf Schwann

Svartvik J (1966) On voice in the English verb Dissertation The Hague Mouton

Svartvik J (Ed) (1990) The London-Lund corpus of spoken English Description and research

Lund Studies in English Series 82 Lund Lund University Press

Svartvik J (Ed) (1992) Directions in corpus linguistics Proceedings of Nobel symposium 82

Stockholm August 48 1991 Berlin Mouton de Gruyter

Svartvik J (2007) Corpus linguistics 25+ years on In R Facchinetti (Ed) Corpus linguistics

25 years on (pp 1125) Amsterdam Rodopi

Svartvik J amp Quirk R (Eds) (1980) A corpus of English conversation Lund Studies in

English Series 56 Lund Liber LaromedelGleerup

Svartvik J amp Sager O (1971) Modern engelsk grammatik (A modern English grammar)

Stockholm Laromedelsforlaget

Taylor C (2008) What is corpus linguistics What the data says ICAME Journal 32

179200

Teubert W amp Cermakova A (2007) Corpus linguistics A short introduction London

Continuum

Tottie G (1971) Have to A study and usage and acceptability in present-day British English

Dissertation Stockholm Stockholm University

Tottie G (1991) Negation in English speech and writing A study in variation London

Academic Press

Tresor de la langue francaise informatise (2004) Analyse et traitement informatique de la lan-

gue francaise Centre national de la recherche scientifique Universite de Nancy II Paris

CNRS editions

Uit den Boogaart P C (Ed) (1975) Woordfrequenties in geschreven en gesproken Nederlands

Utrecht Oosthoek Scheltema and Holkema

van Berckel J A Th M Corstius H B Mokken R J amp Van Wijngaarden A (1965)

Formal properties of newspaper Dutch Amsterdam Mathematisch Centrum

Vander Beke G E (1929) French word book American and Canadian Committees on

Modern Languages Series 15 New York NY Macmillan

Van Reenen P T amp van Reenen-Stein K (1988) Distributions spatiales et temporelles con-

stellations des manuscrits etudes de variation linguistique offertes a Anthonij Dees a

lrsquooccasion de son 60eme anniversaire Amsterdam John Benjamins

Yule G U (1944) The statistical study of literary vocabulary London Cambridge University

Zampolli C Cignoni L amp Peters C (Eds) (1981) Computational lexicology and lexico-

graphy Special issue dedicated to Bernard Quemada I-II Linguistica Computazionale VI

and VII Pisa Giardini

Zipf G K (1932) Selected studies of the principle of relative frequency in language

Cambridge MA Harvard University Press

HOW DO INSTITUTIONAL

CHANGES AFFECT SCIENTIFIC

INNOVATIONS THE EFFECTS

OF SHIFTS IN AUTHORITY

RELATIONSHIPS PROTECTED

SPACE AND FLEXIBILITY

Richard Whitley

ABSTRACT

Recent changes in the funding and governance of academic research inmany OECD countries have altered established authority relationshipsgoverning research priorities and judgements These shifts in the influ-ence of a variety of groups and organisations over scientific choices andcareers can be expected to affect the development of different kinds ofintellectual innovations by changing the level of protected space they pro-vide researchers and the flexibility of dominant intellectual standardsgoverning the allocation of resources and evaluation of research out-comes Variations in these features of public science systems influencescientistsrsquo willingness to pursue unusual and risky projects over many

ISSN 0733-558Xdoi101108S0733-558X20140000042012

years and help to explain cross-national differences in the rate and modeof development of four innovations in the physical biological and humansciences

Keywords Institutional changes authority scientific innovationsprotected space flexibility

INSTITUTIONAL CHANGE AUTHORITY

RELATIONSHIPS AND LEVELS OF PROTECTED

SPACE AND FLEXIBILITY IN THE PUBLIC SCIENCES

The conditions under which scientific research undertaken for publicationis conducted and assessed have changed considerably in many OECDcountries over recent decades The general pattern of expanding highereducation and public support for academic research that was quite wide-spread in the early decades after the end of the Second World War has lar-gely come to an end if not being reversed in terms of support per studentand many governments have become much more active in seeking to ensurethat publicly funded research contributes to social and economic goalsoften through the establishment of new agencies and procedures for steer-ing priorities and evaluating outcomes (Whitley amp Glaser 2007 WhitleyGlaser amp Engwall 2010 Ziman 2000)

These shifts in the nature and organisation of state support for scienti-fic research conducted primarily for publication in peer reviewed journalsand books can be expected to have a variety of consequences for scien-tistsrsquo choices of what to study and how to do so and thus for the diversityand kinds of intellectual innovations developed For instance as research-ers have to compete more intensively for scarce resources from state agen-cies on the basis of their project proposals being likely to contributesignificantly to mainstream disciplinary goals within two to three yearsthey are less likely to work on unconventional problems with highlyuncertain outcomes involving long-term commitments Such pressures tofollow current priorities and focus on problems that can be expected tolead to publishable results within project funding periods are likely to bereinforced by the institutionalisation of strong research evaluation systemsand elaboration of public rankings that judge universitiesrsquo performance interms of their staffsrsquo scientific achievements (Weingart amp Maasen 2007Whitley 2007)

368 RICHARD WHITLEY

In contrast where the state takes an active role in steering research prio-rities towards the resolution of societal problems and contributing to pub-lic policy goals more generally we might expect researchers to undertakeprojects that involved cross-disciplinary approaches and addressed a widerrange of concerns than those central to any one established field This ismore probable if governments commit large resources to such researchgoals over sufficiently long periods for scientists to believe that they couldbuild careers around them Such broadening of research topics andapproaches beyond current disciplinary priorities might also be supportedby the growing legitimacy of commercial interests in selecting problemsand deciding how to tackle them

These varied and sometimes opposite consequences of recent restruc-turings of funding and governance arrangements for public science systems(PSS) ie those where research is primarily undertaken for publicationand organised around the competitive pursuit of reputations for contribu-tions to collective intellectual goals (Whitley 2000) highlight two moregeneral points about how institutional reforms are likely to influenceactorsrsquo behaviour in the sciences and other spheres of activity First suchreforms are rarely so consistent in their nature and implementation thattheir incentives and constraints generate a single set of outcomes especiallywhen they are introduced over a considerable period of time Second theirimpact can vary considerably between societal contexts as exemplified bythe quite different results of implementing financial market deregulation inthe United Kingdom and Japan in the 1980s and 1990s (Kushida ampShimizu 2013 Laurence 2001) The overtly rather similar reforms ofhigher education systems undertaken by many European states often sum-marised as lsquoNew Public Managementrsquo (De Boer Enders amp Leistye 2007Schimank 2005) have had contrasting effects on universities and the orga-nisation of national PSS in different countries as a reflection of major his-torical differences in the role of the state and modes of academicgovernance (Paradeise Reale Bleiklie amp Ferlie 2009 Regini 2011a2011b Whitley 2012) as well as affecting different kinds of scientific fieldsin different ways

In examining the impact of recent institutional reforms on scientistsrsquoresearch strategies and the kinds of contributions they make to intellectualdevelopment in different countries it is important then to identify boththe key processes through which such changes are likely to affect the devel-opment of different kinds of research projects on the one hand and theconditions under which these processes operate in particular sciences in dif-ferently organised PSS on the other hand These conditions include the

369How Do Institutional Changes Affect Scientific Innovations

overall organisation and control of resource allocation processes orderingof research priorities and structuring of scientistsrsquo careers in national PSSas well as the nature of prevalent research practices and epistemic culturesin different fields (Glaser Lange Laudel amp Schimank 2010) They affectthe kinds of commitments that different scientists are likely to make inworking on particular kinds of problems and so how they respond to shiftsin the funding and governance of academic research and their developmentof different sorts of intellectual innovations Innovations here refer to sub-stantial changes of research practices typically involving the developmentof new skills and techniques working on new models and materials andorintegration and elaboration of new kinds of knowledge

A particularly important process through which institutional reformsare likely to affect research strategies involves their shifting the relativeauthority of different groups and organisations over intellectual prioritiespreferred ways of dealing with problems and the making of epistemic jud-gements (Glaser 2010 Whitley 2010) Increasing the authority of univer-sity managers over academic staff in situations of intense competition forresources and scientific reputations for instance may well encourageresearchers to focus on relatively incremental innovations that promise togenerate publishable results within one or two years rather than undertakemore complex and difficult projects with highly uncertain outcomes

The impact of such changes in authority relationships and more generalinstitutional reforms on the direction of scientific research is mediated bytwo critical features of PSS the amount of protected space they affordresearchers and the flexibility of the standards that govern the allocation ofresources and reputations for contributions to collective intellectual goalsProtected space (PS) refers to the period of time in which scientists havediscretion over the use of needed resources including their own efforts topursue particular problems and approaches before having to producepublishable and collectively valued results It incorporates authority overthe choice and formulation of which topics to study how to do so and howto obtain and manage resources It can be afforded either through theorganisational delegation of control of time and other resources to indivi-dual researchers andor through them gaining multi-year project financeexternally

Protected space is important for scientific innovation because it enablesresearchers to tackle difficult problems that might require considerabletime and other resources to resolve Where it is limited in terms of theirspans of discretion over their use of time and other resources scientists areless likely to invest in the development of new skills and techniques or to

370 RICHARD WHITLEY

work on complex problems that could result in apparent failure Ceterisparibus decreasing levels of protected space can be expected to inhibitresearchersrsquo development of innovations that involve long-term commit-ments to new skill and knowledge acquisition and the articulation of newtheoretical frameworks whose intellectual payoffs are highly uncertain inmuch the same way as the lack of protected space can inhibit the develop-ment of path-breaking commercial innovations (Smith amp Raven 2012)

As well as the level or lsquoamountrsquo of protected space afforded to research-ers varying between PSS and affecting investment in long term and uncer-tain projects its distribution between qualified researchers in a particularorganisation or scientific community also differs across academic systemsFor example in highly centralised patronage-based PSS where elite aca-demics may have high degrees of protected space this clearly does notapply to the bulk of the scientific workforce and so the overall range ofactors with significant levels of protected space in such a system is lowEqually where senior professors head institutes that are allocated publicfunds each year as a block grant to pay for assistants and other resourcesthey may enjoy considerable levels of protected space but junior colleaguesdo not since they are usually employed to assist her or him rather than topursue their own projects In such cases the combination of increasingdependence on external project finance with the broadening of access tosuch funding could reduce senior professorsrsquo PS while enhancing thatof their staff as Louvel (2010) suggests has happened in some Frenchlaboratories

Flexibility here refers to the openness of the scientific communityemployers funding agencies and other authoritative groups and organisa-tions to legitimising and supporting novel and unusual ways of framingproblems developing unorthodox ways of dealing with them and interpret-ing evidence It incorporates the ease with which resources can be obtainedfor unusual projects and approaches as well as toleration of a variety ofintellectual approaches and ways of conducting research For scientists tobe able and willing to work on unusual problems with skills and techniquesthat cross established disciplinary boundaries the standards and proce-dures governing the allocation of resources and reputations need to havesome flexibility and tolerance of deviant and unorthodox approaches Thisseems more likely in PSS where research funding is relatively readily avail-able from a variety of agencies and academic careers are not tightlycontrolled by national disciplinary elites

Ceteris paribus the degree to which authority over resource allocationcareers and reputational standards is concentrated in a few organisations

and groups is likely to affect the level of flexibility in a PSS quite stronglyand so is a key factor in distinguishing those that could support the devel-opment of radical innovations from those where this would be moredifficult Where governance changes enhance the concentration of controlover epistemic norms exercised by disciplinary elites flexibility and open-ness to unorthodox problems and approaches can be expected to declineas perhaps in the case of economics in the United Kingdom (Lee 2007Lee Pham amp Gu 2013)

Although the degree of protected space and flexibility may be positivelycorrelated in the sense that PSS with low levels of protected space are unli-kely to exhibit much flexibility in allocating resources and reputationsthere are a considerable variety of combinations of differing levels of thesefeatures of PSS that can be expected to be associated with variations in thekinds of research projects undertaken by scientists in different countriesThese vary in terms of two characteristics First the diversity of topicsinvestigated and approaches adopted especially their divergence from cur-rent disciplinary priorities and frameworks Second the expected level ofinvestment of time and other resources needed to tackle particular researchproblems or commitment costs As governance and authority relationshipschange in PSS so too will the levels of protected space and flexibility theyafford scientists and so their commitment to pursue scientific projects thatdiffer on these two dimensions

In this paper I suggest how we can analyse the impact of changingauthority relationships resulting from institutional reforms on scientistsrsquoresearch strategies and hence the development of new research areas andinnovations by examining their consequences for levels of protected spaceand flexibility of intellectual standards in differently organised nationalPSS Where these change substantially they are likely to affect the varietyof problems tackled and approaches adopted as well as researchersrsquo willing-ness to undertake long-term projects with highly uncertain outcomesBecause the ways in which scientific research undertaken for publication isorganised supported and controlled continue to vary significantly betweennational academic systems and recent funding and governance changeshave been instigated and implemented predominantly by national govern-ments it is appropriate to focus on the national level of analysis hereHowever the opportunities to pursue academic careers internationallyand obtain support from foreign sources have reduced the impact of chan-ging national institutions and authority structures in some sciences as weshall see

372 RICHARD WHITLEY

Initially I summarise the major shifts in authority relationships thathave taken place in many OECD countries before suggesting how they arelikely to affect the degree of protected space and flexibility provided by dif-ferent PSS and consequent patterns of intellectual change These relation-ships are then illustrated by describing how four scientific innovations thatvaried in the level of protected space and flexibility needed to support theirdevelopment were realised in four European countries Germany theNetherlands Sweden and Switzerland These differed considerably in theextent to which funding and governance changes have altered the authorityof different groups and organisations over scientific goals and evaluationsin recent decades as well as in the processes through which scientists wereable to gain sufficient protected space and resources to pursue these inno-vations Some implications of these examples for understanding how recentand current institutional changes are likely to affect the development ofinnovations are summarised in the concluding section

IMPACT OF CHANGING AUTHORITY

RELATIONSHIPS ON LEVELS OF PROTECTED SPACE

AND FLEXIBILITY OF INTELLECTUAL STANDARDS

IN PUBLIC SCIENCE SYSTEMS

Four Major Changes in Authority Relationships

While the recent changes in funding and governance of academic research inmany OECD countries have had a variety of consequences for the authorityrelationships governing research priorities scientific reputations andcareers most can be summarised under four broad headings (Whitley et al2010) First there has been an increase in the degree of researcher depen-dence upon peer reviewed project proposals and the scientific elites judgingthese Second more national and regional states have attempted to steer theresearch priorities pursued by scientists with public funding towards publicpolicy objectives Third many governments have reformed the governanceof universities and state research institutes to increase the authority andresponsibilities of central managerial roles and compelled them to competemore intensively for resources and reputations as distinct organisationsFinally commercial interests broadly understood have gained increasinglegitimacy and authority over research goals and programmes While not all

of these changes have taken place to the same extent in each of the OECDmember countries and their impact continues to vary significantly betweenPSS they do encapsulate most of the reforms undertaken by many statessince the 1970s as they tried to establish pseudo-markets for prestige andresources at the same time as steering the direction of scientific researchtowards public policy objectives particularly economic growth

Considering first how authority over the allocation of resources forresearch and the evaluation of performance has altered there has been awidespread shift towards making academic research more dependent on theextra-mural funding of two to three year projects that is allocated on thebasis of highly competitive peer review This competition for projectfinance is reinforced in some countries by universities and other PROsusing the amount of money thus raised as an indicator of quality and highperformance and so affecting their reward and promotion policies As aresult researchers are becoming more dependent on the national and inter-national scientific elites who advise funding councils and help to set theirintellectual priorities Their growing authority over resource allocation andreputations is further enhanced by their involvement in the post facto eva-luations of the research performance of groups and institutes that manystates have established in recent years especially where these are organisedon a disciplinary basis (Martin amp Whitley 2010 Morris 2010 Whitley ampGlaser 2007)

As well as national and international scientific elites becoming moredirectly influential in the setting of research priorities the selection of pub-licly funded projects is also being increasingly overtly steered by state min-istries and agencies towards contributing to public policy goals that maywell conflict with disciplinary priorities At the same time then as peer-reviewed project financing of research by state agencies is increasing thedirect influence of scientific elites it also enables ministries and agencies toaffect at least in principle the kinds of work that is supported and thustheir authority over research goals and priorities Additionally many gov-ernments are also pressing PROs to support their strategies as they estab-lish targets and key performance indicators for evaluating PRO efficiencyand effectiveness that include such contributions through mission-basedcontracts and similar steering mechanisms (Aljets amp Lettkemann 2011Regini 2011b Wright amp Oslashrberg 2009)

The third change in authority relationships in PSS concerns those withinresearch organisations particularly universities In their attempts to reducecosts and increase the effectiveness of PROs in being both scientifically

374 RICHARD WHITLEY

lsquoexcellentrsquo (Weingart amp Maasen 2007) and contributing to public policygoals many states have sought to increase the authority of deans presi-dents and rectors over resource allocation the direction of researchprogrammes and personnel decisions usually at the expense of localdepartmental scientific elites leading research institutes (de Boer et al2007 Meier amp Schimank 2010 Muller-Camen amp Salzgeber 2005Whitley amp Glaser this volume) As universities in more PSS become directemployers of researchers and are granted some autonomy in making per-sonnel decisions on an organisation-wide basis senior managers becomemore able to exert some authority over these choices as representatives ofthe collective interests of the university

By restructuring authority relationships within universities and encoura-ging them to compete as responsible organisations for public funding andscientific renown governments have tried to create more lsquocompletersquo organi-sations (Brunsson amp Sahlin-Andersson 2000) that can be held to accountfor their contributions to public policy goals and their efficiency in meetingagreed objectives The more that universities become the dominant organi-sational unit being held accountable for academic research and teachingperformance those representing and claiming to manage them or beingexternally viewed as the responsible agents become more authoritativeand legitimate actors for strategic decision making including (re)allocatingresources to achieve state targets and agreed lsquoprofilesrsquo as in some Germanregions (Meier amp Schimank 2010) This change is especially likely to be sig-nificant in institute-based universities where authority over the organisationand direction of research is combined with resource allocation and adminis-trative responsibilities in the hands of institute heads as in many continen-tal European PSS until recently

Finally the increasing pressure by many governments to increase thedirect economic benefits generated by academic research through suchmeans as contract research consulting and company spin offs has been rein-forced in some PSS by significant real-term reductions in state funding ofuniversities that have led some to become much more proactive in seekingsupport from private agencies and organisations (Regini 2011a) Whilesome of the growth of technology transfer offices may have been more aresponse to public incentives for them than derived from business demand(Krucken 2003 Krucken amp Meier 2006) and much research commerciali-sation through patents and new firm formation has proved of limitedfinancial value (Mowery Nelson Sampat amp Ziedonis 2004 Regini 2011a)such pressures have increased the authority of those representing commercial

interests in the setting of research priorities and guiding academic careers byencouraging proactive strategies for exploiting research results

Their Impact on Levels of Protected Space and Flexibility

These four major shifts in authority relationships can be expected to affectthe level of protected space and flexibility in national PSS in different waysI have summarised their likely impact on these features in general terms inTable 1 but the extent to which these effects actually occur depends greatlyon the nature of the PSS being considered especially how academic careersare organised and controlled and how funding and governance changesare implemented Particularly important factors here are the degree of con-centration of control over resources and of the stratification of PROs andfunding agencies As can easily be seen the broad implications of manychanges is that they can be expected to reduce protected space at least forsenior scientists while potentially increasing flexibility by reducing the con-trol of established disciplinary elites over careers

Considering first the impact of funding and governance changes thatincrease the authority of scientific elites by requiring many researchers toapply for peer-reviewed project grants every two to three years this seemslikely to reduce the degree of protected space available to researchers Sucha reduction will be especially significant in fields where problems cannotreadily be broken down into distinct project modules that can be studiedsequentially and require long periods of uninterrupted research time(Glaser et al 2010)

Increasing dependence on external project-based funding could thoughbroaden the range of researchers able to benefit from some protected space ifthey are enabled to apply for separate support for their own projectsDepending on the implementation of other governance changes particularly adecline in the patronage powers of local scientific elites such competitive allo-cation of research support on a project basis might increase the number andvariety of scientists becoming principal investigators on their own Howeversuch an outcome remains of course highly context dependent and cannot beassumed to follow directly from increases in the proportion of project-basedfunding in all cases especially where traditional academic oligarchies remainmajor controllers of academic careers as perhaps in Italy (Capano 2009)

Similar contextual factors affect the impact of increasing project fundingdependence on flexibility especially the extent to which such support isdominated by established disciplinary elites Where funding can be

376 RICHARD WHITLEY

obtained from a diversity of research agencies and foundations that arewilling to support academic research for a variety of purposes and rely ona wide range of advisors to assess proposals we would expect suchdependence to increase the degree of flexibility relative to that found intraditionally oligarchic academic systems According to Stokes (1997) thispluralism of funding channels and goals is a key feature of the US researchsupport system that has helped to maintain a high level of flexibility inresource allocation and diversity of research themes

Where however such project funding for publishable research is con-centrated in a single or very small number of national agencies is severely

Table 1 Expected Consequences of Changes in Authority Relationshipsfor Protected Space and Flexibility

Authority Changes Protected Space Flexibility of Standards Governing

Degree Resources Reputations

Increasing

dependence on peer-

reviewed project

proposals and

national and

international

scientific elites

Reduced for senior

academics especially in

institute-based PSS

increased for junior

academics able to attract

long-term funding directly

Increased where

there are diverse

agencies reduced

where the

concentration of

control is high and

the amounts

available decline

Increased where

there are diverse

peer review groups

reduced otherwise

Increasing state

steering of research

priorities

Reduced especially for

senior academics in

institute-based universities

Increased if

substantial resources

are allocated to

diverse goals and

career paths

Increased if

substantial

resources are

allocated to diverse

goals and career

Increasing central

managerial authority

within PROs in a

more competitive

environment

Reduced for senior

institute-based

universities where

resource competition is

intense and project-based

Reduced where

resources are highly

constrained and

control over their

allocation is

concentrated

Increased in

pluralistic

environments

reduced in highly

stratified and

concentrated ones

Increasing authority

of commercial

interests

Reduced except where

research

commercialisation enables

scientists to control

significant resources

Increased where it

diversifies sources of

substantial support

Increased where it

broadens career

limited in quantity relative to the number of qualified applicants and isawarded on the basis of advice from a small number of elite scientistsstandards governing resource allocation and reputations are likely to berather inflexible Scientists in such situations can be expected to be quiteconstrained to work on mainstream topics in ways that should result incontributions to disciplinary goals that do not challenge orthodox viewsand many have to reduce their ambitions as seems to have happened insome sciences in Australia (Glaser amp Laudel 2007)

Considering next the greater authority of state ministries and other pub-lic policy agencies over research priorities as they seek to steer academicresearch towards socio-economic objectives this might also be expected toreduce the degree of protected space afforded to researchers as they becomepressured to generate relevant results from each project However such areduction depends on the specificity of these objectives and the extent towhich they are expected to be achieved in short project cycles If they tendto be very broadly set such as contributing to the lsquowar on cancerrsquo andprovide a basis for funding a considerable number of projects over manyyears state steering may not have such restrictive consequences By provid-ing finance for wide-ranging purposes they may actually widen the distri-bution of protected space by enabling junior researchers to access supportthat would not otherwise be available to them

Similar considerations affect the impact of increased state steering onflexibility In general such broadening of the goals to be achieved by pub-licly supported scientific research should increase the variety of researchproblems investigated relative to those focused on tightly defined disciplin-ary objectives especially where public policy interests encourage cross-disciplinary collaboration and provide significant amounts of continuingfunding for novel approaches that contribute to nonmainstream topicsThe more such steering facilitates the establishment of new academic careerpaths through for example the development of new research institutesand long-term finding programmes for policy focused purposes the moreit will increase the flexibility of both resource allocation and reputationalstandards as in the case of ministry-supported applied social sciencesin Germany (Glaser et al this volume) If however it remains largelyimplemented through established procedures and agencies especially thoserelying on scientific elite judgements of quality and does not lead tosubstantial increases in the amount of funding provided it is unlikelyto encourage greater reputational flexibility even if some resources are allo-cated to a more diverse set of problems than would otherwise have beenthe case

378 RICHARD WHITLEY

Turning to consider third the impact of changing authority relation-ships within PROs particularly universities the increasing powers andresponsibilities of senior managers are likely to limit the amount of pro-tected space available to senior academics as they attempt to organise thelattersrsquo activities towards organisational objectives In the increasingly com-petitive environment experienced by many universities towards the end ofthe 20th century these typically focus on increasing their scientific andsocial prestige while also increasing the amount of money raised from avariety of sources Managers can be expected to encourage academics toresearch topics that will pay off in both intellectual and financial terms inthe relatively near future rather than support them in pursuing difficultproblems whose resolution cannot be assumed to be realised in the short tomedium term

As they gain more authority over resource allocation and personneldecisions and come under external pressures to demonstrate their ownperformance university presidents are likely to reinforce funding counciland other groupsrsquo demands for publishable results and fairly predictablecontributions to current priorities Particularly in PSS where senior profes-sors have previously been able to control research resources for their ownprogrammes as in much of continental Europe and Japan this shift inauthority could restrict their protected space quite sharply

Equally the growing use by senior managers of formal indicators ofresearch performance and elaboration of control procedures to demon-strate their efficiency is unlikely to increase the level of protected spaceafforded to junior academics Being unable to make well-informed judge-ments of who should be supported to research complex and difficult pro-blems in fields they usually know nothing about (Glaser et al 2010Whitley amp Glaser this volume) managers are likely to respond to externalcompetitive pressures by encouraging all academic staff to pursue relativelyshort-term objectives that can be expected to generate concrete achieve-ments within each budgetary cycle

Similar factors can be expected to influence the impact of increasingcentral managerial authority on flexibility Granting university seniormanagers more autonomy and influence over personnel decisions and theallocation of resources should enable them to respond more rapidly by real-locating posts and supporting materials to new fields and problems thanwas feasible in traditional oligarchic PSS albeit to a more limited extentthan some might have expected (Meier amp Schimank 2010) However theirignorance of the details of most scientific fields means that they find it diffi-cult to justify major shifts of resources without significant support from

established scientific elites Together with their competitive pursuit of scien-tific prestige and hence reliance on the priorities and judgements of theseelites in making strategic choices this means that they are unlikely to sup-port research projects that challenge current orthodoxies and reputationalstandards The more authority top PRO managers are granted and themore pressured they become to meet societal demands in a highly competi-tive and constricted resource environment the more they are likely to func-tion as transmission if not indeed amplification mechanisms translatingexternal priorities into internal goals and targets (Glaser amp Laudel 2007)

An instance of how university top managers can use external pressuresand access to resources to reorganise both research and teaching substan-tially is described by Duncan Wilson (2008 Wilson amp Lancelot 2008) inhis history of the biological sciences at Manchester University In this casea new Vice-Chancellor with a background in medical microbiology wasable to build on both internal and external dissatisfaction with the achieve-ments of existing departments to push through a major restructuring thatresulted ultimately in the creation of a unified Faculty of Life Sciences inwhich molecular biology and biochemistry dominated with the active sup-port of charities and the Medical Research Council (MRC) A crucial rolein these reforms was played by a group of young professors nicknamed thegang of four who had begun agitating for radical changes before the fund-ing cuts of 1981 and the reorganisation of university governance later inthe 1980s This group was able to use these shifts in the external environ-ment as well as gaining the support of the new Vice-Chancellor who notcoincidentally had close connections with the University GrantsCommittee and the MRC to establish their vision of biological and bio-medical research at Manchester

This case highlights an important general point about the impact of gov-ernance and authority changes on research priorities and approaches itdepends greatly on the interests and resources of different groups of scien-tists and other collective actors in each PSS As with any kind of institu-tional change social agents are not simply passive recipients of new rolesand norms but are variously capable of directly influencing their develop-ment and implementation as well as using them to advance their owninterest in ways that may or may not coincide with the intentions of theirinitiators This is of course especially so in the public sciences where knowl-edge about goals resources and personal competences is so asymmetricallydistributed

In many respects this increase in managerial authority is mirrored by thedeclining authority of local scientific elites particularly in institute-based

380 RICHARD WHITLEY

universities where senior professors collectively controlled resources andpersonnel decisions as well as intellectual priorities and standards Theirprotected space becomes reduced as their control of junior staff declinesand resources have to be obtained through competitive bidding to externalagencies However as their former assistants become less dependent ontheir patronage for access to resources and permanent positions and areable to obtain project support directly from funding agencies the distribu-tion of protected space across the PSS as a whole becomes wider albeitmore limited in degree than previously

Similarly in PSS where senior professorial oligarchies also exercisedstrong central authority at the national level over resource allocation andcareers their weakened position as a result of governance reforms shouldincrease the overall flexibility of decision making standards However for-mal changes in regulations and funding arrangements do not always resultin significant shifts in authority as the recent history of Italian governancereforms indicates (Ballarino amp Perotti 2011 Capano 2009 Moscati2001) Here quite substantial changes in formal structures do not seem tohave greatly reduced the patronage powers of national disciplinary elitesover academic careers and so intellectual and resource flexibility have yetto increase significantly

Considering finally the likely impact of pressures to increase revenuesfrom the commercialisation of research and providing services to the busi-ness sector these can be expected to reduce the overall level of protectedspace as companies seek relatively short-term benefits from such invest-ments and are unlikely to support research where the pay offs are highlyuncertain It is also worth pointing out that increasing universitybusinesscollaboration and mutual support beyond consultancy and short-term tech-nical problem solving depends greatly on the structure of national econo-mies especially the nature of major industries and size distribution offirms as Regini et alrsquos (2011b) national case studies of European universi-ties illustrate

There may be some cases where substantial contributions to scientificgoals can lead to valuable patents and the formation of successful firms toexploit them as in some parts of biotechnology and biomedical researchbut these are far fewer than politicians and university presidents haveclaimed and the growth of university patenting following the passage ofthe Bayh-Dole Act in 1980 has proved less profitable for most organisa-tions than hoped (Geuna amp Nesta 2006 Mowery et al 2004) In anycase while research universities competing for scientific prestige mayinvest in service functions to facilitate the commercialisation of scientific

knowledge their ability to steer research priorities towards such outcomesis limited by managersrsquo ignorance of scientific fields and of the suitabilityof particular groups for work on specific topics Equally scientists seekingcommercial pay offs from their work notoriously underestimate theresources and time needed to reap financial benefits which are also usuallydifficult to estimate reliably

While the growing influence of commercial interests over research prio-rities and university policies may help to increase the flexibility of resourceallocation and reputational standards in PSS dominated by disciplinaryoligarchies this depends greatly on how such influence is exercised andthe nature of the context involved If such interests are incorporated intothe decision making processes of highly centralised funding agenciesdominated by current scientific elites it seems unlikely that they will sig-nificantly increase the diversity of research programmes supported orencourage the establishment of new career paths in new fields On theother hand if commercial interests lead to substantial support being pro-vided for research over some time in a more pluralistic funding environ-ment with a variety of public and private foundations they could help toincrease flexibility as perhaps happened with the development ofcomputer-based corpus linguistics research where publishers and compa-nies providing translation services became interested in its commercialapplications

Likely Consequences for Intellectual Diversity and Scientistsrsquo Commitments

These expected changes in the levels of protected space and flexibility pro-vided by different PSS as a result of shifts in prevalent patterns of authorityare likely to influence both the diversity of problems and approaches devel-oped by researchers especially their deviation from current disciplinaryorthodoxies and priorities and scientistsrsquo willingness to commit substantialamounts of time and other resources to working on projects whose out-comes are highly unpredictable and may not produce significant publica-tions in the relatively short term Where the degree of protected space islimited in terms of researchersrsquo spans of discretion over their use of timeand other resources scientists are less likely to invest in the development ofnew skills and techniques or to work on complex problems that could resultin apparent failure Ceteris paribus relatively low levels of protected spacecan be expected to inhibit researchersrsquo development of innovations thatinvolve long-term commitments to new skill and knowledge acquisition

382 RICHARD WHITLEY

and the articulation of new theoretical frameworks whose intellectual pay-offs are highly uncertain

Flexibility is obviously key to scientistsrsquo willingness and ability to workon the development of deviant and unfashionable innovations Whereresources and reputations are tightly controlled by disciplinary elites some-times tacitly operating in coalition with policy makers and bureaucrats it isvery difficult if not indeed effectively impossible for most scientists todevelop a research career by studying problems outside the mainstreamwith unorthodox approaches Even for those working within establishedintellectual boundaries it can be difficult to obtain support for projectsthat seem infeasible to elite advisors and require substantial investment inapparatus and skills

Four particular combinations of limited and considerable levels of pro-tected space and flexibility can readily be distinguished in which differentkinds of research projects are likely to predominate varying in terms oftheir intellectual diversity and unorthodoxy on the one hand and theirinvestment costs and uncertainty about achieving publishable results onthe other hand As governance and authority relationships change in PSSso too will the levels of protected space and flexibility they afford scientistsand so their commitment to pursue these different kinds of scientificprojects

The combination of limited levels of protected space with limited flex-ibility for instance in PSS with highly concentrated control over resourceallocation and sharply reduced public funding of academic research islikely to encourage predominantly incremental innovations within estab-lished intellectual trajectories or what could be termed puzzle solvingbecause of the need to make acceptable contributions to current intellec-tual goals with limited resources especially research time Most scientistsin such PSS do not have a great deal of time and other resources inwhich to learn or acquire new skills and knowledge before generatingpublishable results and so are unlikely to tackle unfashionable problemsthat involve considerable switching costs especially where these requireexpensive apparatus and materials and research outcomes are highlyuncertain

Where limited protected space is combined with greater flexibility of stan-dards governing resource allocation and reputation perhaps because statesprovide substantial resources for research that contributes more to publicpolicy goals than to those of individual established disciplines we wouldexpect a greater diversity of research topics and approaches to developincluding those involving skills and knowledge from different fields

However if protected space remains quite limited for instance by the needto produce publishable results that are recognised as significant contribu-tions every two to three years most scientistsrsquo willingness to investigatemajor complex and difficult problems that are likely to take much time andeffort to resolve will be restricted Innovations in these kinds of PSS thenmay deal with more varied problems and adopt a range of differentapproaches to them than in the previous case but are likely to remainachievable within short project cycles and currently available skills that donot involve high commitment costs

In PSS that combine higher degrees of protected space for seniorresearchers with limited flexibility as in the traditional continentalEuropean chair system without substantial extra-mural project fundinginstitute heads are more able to pursue long-term uncertain and difficultproblems that may not result in significant results for quite some timeHowever these are unlikely to involve major challenges to established dis-ciplinary goals or boundaries or to contribute more to public policy goalsthan to current intellectual priorities unless substantial new resources areallocated to them and possibly new organisational frameworks establishedInnovations that involve high development costs are here more feasiblethan in the two previous situations especially where they require long per-iod of uninterrupted research time andor depend highly on personal inter-pretations of problems and evidence but will typically remain withindominant intellectual paradigms

Finally increasing flexibility in PSS that provide considerable levels ofprotected space for senior staff and perhaps for more junior ones as wellafter a period of probation can be expected to encourage greaterdiversity of research goals and approaches while continuing to supportlong-term investments in tackling difficult problems that may not generatepublishable outcomes in the short to medium term Cross-disciplinary pro-blem solving should become easier in this situation as may organising mul-tidisciplinary teams to contribute to public policy objectives Innovationsthat challenge current orthodoxies and involve unusual approaches tounderstanding complex processes are more likely to be developed in thesesituations which typically occur in relatively beneficent funding environ-ments with a variety of opportunities for scientific careers

It is important to recognise the crucial role of established research prac-tices in mediating these connections between protected space flexibilityand innovation patterns in different sciences The general impact of gov-ernance and authority changes in different PSS on levels of protected

384 RICHARD WHITLEY

space and flexibility and on scientistsrsquo investments in developing particularkinds of innovations varies between scientific fields that have differentways of conducting research Particularly critical here is the decomposabil-ity of problems since it can alleviate some of the pressures arising fromreductions in the degree of protected space by breaking complex problemsdown into sub-projects that can be carried out in two to three years witha reasonable expectation that they will generate sufficiently interestingresults to support further applications for funding While such modularisa-tion may involve frequent bids for support on a more or less continualbasis it can enable scientists to pursue longer term research programmesin a more restrictive environment and produce publications for short-termevaluations

It can also help them deal with the consequences of reductions in flex-ibility by facilitating the division of research activities between those work-ing within the mainstream of established priorities and those challengingthem Scientists wishing to pursue unfashionable and difficult researchtrails that may not payoff in terms of major publications for some timemay be able to combine such projects with shorter term ones that can drawon existing skills and knowledge to generate contributions to establishedpriorities Such a combination of relatively incremental innovation withinmainstream disciplinary frameworks with more deviant interests and skillstackling relatively unusual problems seems to be quite a common meansfor scientists to develop unorthodox approaches even in fields dominatedby strong hierarchies

In general such decomposability and modularisation of problemsdepends on concepts descriptions and techniques being quite highlycodified and standardised so that results from different studies can be sys-tematically evaluated and integrated across groups and locations This inturn means that the formulation of problems use of knowledge and techni-ques and assessment of evidence cannot rely greatly on personal interpreta-tions Where such understandings do depend considerably on the personaldescription and use of empirical materials with more tacit skills developedover a considerable period as is often the case in historiographicalresearch (Glaser et al 2010) it is obviously much more difficult to breakproblems down into discrete packages to be undertaken by inexperiencedresearchers For such fields then reducing the level of protected spaceafforded to scientists is likely to have more significant and long-lastingeffects than in those where problems are more easily and meaningfullydecomposable

LEVELS OF PROTECTED SPACE AND FLEXIBILITY

SUPPORTING THE DEVELOPMENT OF FOUR

Characteristics of Four Innovations in Different Sciences

These connections between institutional reforms authority shifts levels ofprotected space and flexibility and the development of scientific innovationscan be illustrated by considering how four quite different intellectual devel-opments took place in four European countries that varied in the extent oftheir funding and governance changes drawing on the much more detailedaccounts provided in the articles by Engwall et al Glaser et al and Laudelet al in this volume These developments represent different kinds of inno-vation in the physical biological and human sciences that varied in (a) theextent to which they competed with and challenged current orthodoxies(b) their intellectual and material costs (c) their perceived scientific poten-tial and (d) extent of extra-scientific relevance as summarised in Table 2

The first concerns the realisation of Bose-Einstein condensates (BEC) atultra low temperatures The second focuses on the influence of developmen-tal mechanisms in different organisms on their pattern of evolution in theemerging field of evolutionary developmental biology (ED) The third dealswith the comparative analysis of the effectiveness of national school sys-tems as revealed by large-scale assessments of student performance (ILSA)while the fourth concerns the construction and analysis of digitalised writ-ten and spoken linguistic corpora with computers (computerised corpus lin-guistics or CCL) In this section I suggest how their key characteristicsimply different degrees of protected space and flexibility of standards gov-erning the allocation of resources and reputations were needed for theirdevelopment The following section describes how researchers in four coun-tries were able to gain sufficient degrees of PS and support to contributeto these developments

As Laudel et al explain in their article titled ldquoCold Atoms HotResearch High Risks High Rewards in Five Different Authority Structuresrdquothe existence of BEC was theoretically predicted in 1924 by Bose andEinstein When matter is cooled very close to absolute zero (100 nanokelvin)particles become practically immobile They lose their individual identitiesand coalesce into a single blob This state of matter is called a Bose-Einsteincondensate Although widely accepted as a theoretical phenomenon theexperimental realisation of BEC was regarded by many physicists as being

386 RICHARD WHITLEY

Table 2 Summary Characteristics of Four Scientific Innovations

Characteristics Experimental

Bose-Einstein

Condensates

Experimental

EvolutionaryDevelopmental

Biology

Assessments of Student

Performance

Computerised Corpus

Linguistics

Competition with

established research

priorities and

approaches

Low Considerable Considerable with humanist

education studies less with

psychology and economics

of education

High with transformational

grammarians lower with

natural language students

Resources needed High Medium to high Low for psychologists and some

sociologists high for humanist

education researchers

High for corpus builders

low to medium for most

corpus users

Research potential High High Limited for established

disciplines higher for policy-

related fields

High for empirical language

studies

External relevance and

support

Low Low High Considerable for publishers

and translators

HowDoInstitu

tionalChanges

Affect

tificInnovatio

very difficult if not impossible to achieve with current technologies in the1990s Attempting such realisation was not then a deviant or radical researchproject that competed with and strongly challenged the dominant theories inphysics but did constitute a significant experimental problem that manyestablished scientists considered very difficult to resolve if not impossible ontheoretical grounds The results of such projects were quite uncertain andcould reasonably be expected to fail for experimental reasons as well as beingthought unlikely by some on theoretical grounds However the standards fordeciding what constituted success or failure were widely accepted and codifiedso that the first realisations were quickly acknowledged by most physicistsand by the award of a Nobel Prize despite earlier scepticism

These successes did though require considerable time and effort as wellas substantial investment in developing complex and expensive apparatusEven after the first achievements in 1995 it often took two to three years toconstruct workable apparatus and frequently even longer to generate pub-lishable results for different materials with many researchers having tolearn quite new skills and techniques As many of those interviewed byLaudel and her colleagues confirmed these were expensive and long-termprojects even by the standards of experimental physics that required aconsiderable amount of time and technical resources

While some physicists thought that such states of matter once achievedwere of limited further intellectual interest and not worth expendingmajor resources on their exploration most seem to have regarded them asoffering rich opportunities for future projects and considerable intellectualpotential As yet though they have elicited little commercial interest and sup-port for their research has almost entirely been provided by public researchorganisations dedicated to supporting intellectually driven research goals

In contrast much work in evolutionary developmental biology isregarded by many population geneticists and other researchers studyingpatterns of organism evolution as deviating significantly from mainstreambiology and challenging much accepted orthodoxy by researching howparticular features of organism development can affect their evolutionAdditionally experimental ED research often involves the coordination ofa wide range of people with different backgrounds and skills and invest-ment in expensive infrastructure to maintain and study particular creaturessuch as snakes and fish over many generations which takes a considerableamount of time As Laudel and colleagues point out in their article titledldquoHighly Adaptable but Not Invulnerable Necessary and FacilitatingConditions for Research in Evolutionary Developmental Biologyrdquo thismeans that undertaking such work can be a relatively long-term and costly

388 RICHARD WHITLEY

activity that could take several years to come to fruition resulting in manyscientists preferring to adopt less costly or lsquoevo-devo lightrsquo strategies basedon publicly available information andor theoretical modelling Howeverresearchers regard this as an area of research with high intellectual poten-tial that could lead to a significant reshaping of the biological sciencesLike BEC though it has little commercial relevance and is unlikely toattract support for non-scientific purposes

Just as judgements about the scientific status of ED research canvary considerably between scientists from different backgrounds so tooevaluations of research on the comparative performance of national school -systems using standardised surveys differ greatly between education resear-chers economists psychologists and sociologists Where researchers havebeen used to dealing with large-scale data sets as in many of the empiricalsocial sciences they were unlikely to regard PISA and similar results particu-larly distinctive or novel and challenging established approaches It was insciences dominated by traditional humanist understandings of educationalresearch such as those pursued in many German universities and the philo-sophy of education in Anglophone academia that these surveys representedan opportunity for a radical shift towards empirical social scientificapproaches to educational research and also for more policy focused studies

Much of the impetus for comparative analyses of school performanceusing these large-scale surveys came from bureaucratic and political elitesrather than established social scientists and many academics remainedrather disdainful of their purposes and the techniques employed Theintellectual potential of such surveys and research based on them was com-monly regarded as being rather limited given their highly standardised anddecontextualised nature except for posing many puzzles for morepolicy-oriented researchers Insofar as they did lead to new chairs beingestablished and new research streams becoming institutionalised in somenational PSS they helped to construct a new policy-focused area of thesocial sciences crossing variously differentiated fields of educational studiesand the economics psychology and sociology of education They representthen an example of how state-directed investments in particular areas ofpolitical concern can stimulate the development of novel intellectual strate-gies although it remains to be seen how substantial and long-lived suchprojects and programmes remain in many national PSS

Considering finally computerised corpus linguistics this too varies inthe extent of its innovative character and implications for establishedresearch practices according to the backgrounds and goals of researchersConstructing and analysing text corpora in different languages have been

accepted activities for many students of natural languages since the 19thcentury if not earlier While computerised corpus construction has enabledmuch larger ones to be created and more systematic analyses to be underta-ken many linguists consider this to constitute an incremental change intechnique and size of database or a considerable increase in lsquobad datarsquo(Durrell 2012) rather than a radical intellectual innovation Othershowever regard such corpora as forming the basis for a new empiricallybased science of linguistics (Hardie amp McEnery 2010 Mukherjee 2010Teubert 2010) while those studying the deep structures of generativegrammars have been prone to dismiss most such corpora studies as superfi-cial analyses of surface phenomena (Fries 2010)

The degree to which computerisation per se represents a substantiveintellectual challenge to current approaches and involves significant shiftsin research practices depends then greatly on onersquos conceptualisation oflinguistics research as a scientific endeavour its priorities and key skillsWhere leading departments institutes and graduate schools were domi-nated by followers of Chomsky as seems to have been the case in much ofthe United States constructing and analysing corpora whether compu-terised or not could be widely seen as being unscientific and not likely tolead to any significant intellectual contribution (Fries 2010 Hardie ampMcEnery 2010) On the other hand in countries where language researchinstitutes were integrated into organisational units tied to the training ofsecondary school teachers of natural languages and were more open to avariety of approaches as in much of Europe such stigmatisation andexclusion of corpus-based analyses were less likely

The resources needed for constructing such computerised corpora wereand remain considerable although the widespread availability of personalcomputers has reduced these costs for small corpora With large-scale cor-pora still being built for many written languages and dialects as well asspoken ones obtaining the necessary substantial resources continues to bea significant challenge for many linguists Corpus construction and analysishave though been seen as quite relevant for some commercial interestsparticularly dictionary and grammar publishers More recently the grow-ing availability of many software packages and relatively easy access tomost previously constructed corpora in a number of languages since thediffusion of personal computers mean that many corpus users are able toconduct their research without needing unusually large amounts of projectfunding once they have acquired the basic analytical skills and techniques

For those committed to the empirical analysis of language performancecomputerised corpora do of course offer considerable research potential

390 RICHARD WHITLEY

This is both for contributing to current intellectual goals and problems andfor developing new ones that were infeasible with previous technologies Itis especially the belief that these new techniques will lead to the pursuitof new questions and intellectual concerns that has encouraged someresearchers to hail them as constituting the basis for building a new disci-pline of linguistics in a comparable manner to the impact of other researchtechnologies on other sciences (Mukherjee 2010 compare Joerges ampShinn 2001)

Levels of Protected Space and Flexibility Supporting theDevelopment of these Innovations

These summaries of four intellectual developments in different scienceshighlight the varied amounts of different kinds of investments neededfor achieving particular research goals as well as the differences in thesebetween scientists from different backgrounds which obviously affect whobecomes committed to them and how they work on them The levels ofprotected space and flexibility of standards that were necessary for theirdevelopment by different researchers are summarised in Table 3 and willnow be discussed in a little more detail

In the case of realising BECs scientists clearly required quite high levelsof protected space since it took most of them several years to construct thenecessary experimental apparatus and understand how it worked to pro-duce them in different materials During this time they were unable to gen-erate publishable experimental results and so ran considerable risks ofbeing regarded as unproductive In some cases journals did publish detailsof their experiments including failed ones but interviewees were acutelyaware of the reputational dangers involved in pursuing such long-termresearch projects that many physicists thought would continue to result inan inability to produce these condensates

Given the resources needed to undertake these experiments it was alsonecessary for national and international funding agencies to provide majorsupport for them even when some elite physicists considered them likely tofail Without such flexibility in allocating resources they simply could notbe conducted which meant that scientists in PSS where the major agenciescould or would not provide substantial funds were unable to work onBEC A further aspect of such flexibility concerned its extension to sup-porting the development of a variety of different approaches rather thanconcentrating resources on a single one Apparently this pattern of parallel

Table 3 Levels of Protected Space and Flexibility Needed for the Development of Four ScientificInnovations

Conditions Supporting

Development

Experimental Bose-

Einstein Condensates

Experimental EvolutionaryDevelopmental Biology

Performance

Computerised Corpus

Linguistics

Protected space High (although some

modularity in

publications)

Medium to high High for conducting the surveys

and for their use by humanist

educational researchers lower for

economists psychologists and

many sociologists using them

High for constructing

corpora and users having

to learn new skills lower

for users with access to

computer expertise

Flexibility of standards

governing the allocation

of resources

Considerable as long

as elites are sceptical of

success

Medium to high Considerable in many

educational research fields in

some countries lower in most

other social sciences

High for resources for

building corpora lower

for most corpus users

governing reputations

and the worth of

contributions

Limited given its

centrality to AMO

physics

educational research fields

depending on national epistemic

conditions lower in most other

social sciences

Low for natural language

departments high for

general linguistics

investments was an important factor in the early success of some USgroups In contrast flexibility in the standards applied to evaluations ofresearch results and hence reputations was less crucial since the phenom-enon of BEC was widely accepted in atomic physics although its experi-mental accomplishment was not thought feasible by many physicists

Experimental research in evolutionary developmental biology alsorequired considerable levels of protected space given the complexitiesinvolved in researching development in different organisms and their evolu-tion over generations As Laudel et al make clear in their article titledldquoHighly Adaptable but Not Invulnerable Necessary and FacilitatingConditions for Research in Evolutionary Developmental Biologyrdquo colla-boration between scientists with diverse knowledge and skills over sometime was crucial for much of this kind of research which took longer toproduce publishable results than more conventional studies of evolutionaryphenomena Some researchers however were able to continue to work onprevious projects that enabled them to continue to publish papers whileestablishing facilities and cooperative networks for ED research as well asbreaking long-term projects down into discrete packages that could eachgenerate some publications reflecting the codified and modular nature ofmuch biological research

As in the case of BEC research obtaining the necessary resources forexperimental ED work has been a major concern for scientists particularlymaintaining support for needed facilities over a number of years Onlywhere public research councils charities and foundations have been willingto provide substantial funds for such unconventional research has ED beenfeasible Thus both a relatively munificent financial environment and awillingness by some agencies and their advisors to be flexible in allocatingresources to diverse projects and approaches as well as in allowing consid-erable discretion to scientists in how they used such funding have beennecessary for the development of this kind of research Equally the possibi-lity of publishing research results and gaining reputations from scientificaudiences for significant contributions has been crucial for ED work to beattractive for researchers and form the basis for successful scientific careersWhere traditional evolutionary views and methodological standards over-whelmingly dominated leading journals conferences and academic posi-tions ED research was difficult to undertake as a core component ofscientistsrsquo careers although it has become more feasible in the 2010s incountries where resources are more widely available

In contrast once the results of large-scale assessments of school studentperformance were made publicly available for researchers social scientists

used to analysing large data sets did not require long periods of time togenerate publishable outputs On the other hand researchers who did nothave such knowledge and skills needed more protected space to acquiresuch competences and in some cases to alter their intellectual goals andapproaches quite radically Similarly most scientists experienced in com-parative performance analysis from different social sciences were able togain the needed resources once the ILSA data sets were released for aca-demic research but where traditional educational research goals andapproaches remained dominant this was more difficult In such academicsystems without major changes in the kinds of work regarded as central toeducational studies and associated academic careers such as those takingplace in Germany during the institutionalisation of empirical educationalresearch in many universities the widespread development of ILSA-basedstudies was inhibited While then only limited flexibility in the standardsgoverning resource allocation and reputations was necessary for mostsocial scientists interested in using ILSA data sets much greater levels offlexibility would be required for educational researchers where this fieldwas dominated by more humanistic goals and research styles

Similar differences in the required levels of protected space and flexibilityfor scientists from different backgrounds and orientations to becomeinvolved in developing computer-based language corpora occurred in lin-guistics For many students of natural languages who were used to workingon corpora with standardised analytical techniques the availability of com-puterised corpora greatly facilitated their research for relatively little invest-ment in acquiring new skills This was even less for scholars able to usePhD students or other colleagues for the more technical aspects Thus thedegree of protected space needed for such scientists to use electroniccorpora was rather limited

In contrast those involved in building these large corpora did require highlevels of protected space especially in the early days as this was a lengthyprocess often needing the development of new skills without much technicalassistance As Fries (2010) has remarked most of the initial English languagecorpora were constructed by established scholars who could afford to spendmany years doing so without worrying too much about their lack of publica-tions during that period Today it would be difficult for younger researchersto consider seriously spending several years building a new corpus givenintensified pressures to publish papers in many PSS

The substantial resources required to construct linguistic corpora oversome years obviously mean that funding bodies had to be able and willingto support such work on a significant scale for scientists to undertake such

394 RICHARD WHITLEY

tasks In the 1960s and 1970s this required major investments in computerfacilities that were quite limited in many national PSS especially in univer-sity humanities departments Equally where analytical techniques had tobe specifically developed for linguistic purposes foundations and researchcouncils supporting language studies needed to be quite flexible in the kindsof work they were willing to fund if corpus users were to take full advan-tage of the newly built corpora

Flexibility in reputational standards was less important for many naturallanguage students since such corpus building and analysis had been legiti-mate activities for many years This was not the case for those in manydepartments of linguistics andor researchers pursuing more formal ana-lyses of deep linguistic structures along Chomskyan lines Such scholarsfound it more difficult to publish articles about and gain positive reputa-tions for work based on electronic corpora This was especially so in theUnited States since generative linguistics became dominant in many gradu-ate schools Although toleration of diverse conceptions of linguisticsresearch and their use of computerised corpora seems rather greater inmost European countries disputes about the epistemological status of suchwork remain widespread and strongly argued (Barlow 2011 Hardie ampMcEnery 2010 Mukherjee 2010)

CONDITIONS PROVIDING PROTECTED SPACE AND

FLEXIBILITY FOR DEVELOPING INNOVATIONS IN

The varied levels of protected space and flexibility needed to supportscientists working on these four innovations have been provided indifferent ways and to varying degrees by particular conditions infour European PSS that varied considerably in the extent to which theyhad undergone substantial funding and governance reforms Simplifyinggreatly these were most noticeable in the Netherlands particularly inthe 2000s less significant in Sweden and Switzerland and relatively lowin much of Germany although with some variation between federalstates The main conditions providing protected space and flexibility ofstandards governing resource and reputation allocation are listed inTable 4 together with those countries in which they were at least partlysignificant for scientists being able to develop each of the four innova-tions discussed here

Table 4 Conditions in Four National Public Science Systems Providing Protected Space and Flexibility forthe Development of Four Scientific Innovations

Supporting Conditions BEC High and Medium

Cost ED

ILSA Computerised Corpus

Linguistics

Providing protected space

Tenured posts CH DE NL CH DE NL

(up to 2000s) SE

CH DE NL SE CH DE NL SE

International careers and learning

opportunities

CH DE NL CH DE NL

(up to 2000s) SE

CH NL SE

Providing resource flexibility

Local research capacity (eg technicians

experimental facilities)

CH DE NL

(declining)

CH DE NL

(up to 2000s) SE

DE CH NL DE SE

Recurrent funding locally CH DE NL SE

(up to 2009)

CH DE NL

(up to 2000s) SE

DE CH SE

Diversity of funding agencies and their

advisors

DE SE DE SE CH DE NL SE DE NL SE

Diversity of funding programmes CH DE CH CH

Political support and funding CH DE NL SE CH NL

Providing reputational flexibility

Career and employment opportunities in

different organisational units

DE CH DE DE CH DE SE

Diverse audiences and publications CH DE SE CH DE SE NL CH DE SE

The most common means of gaining some protected space has ofcourse been appointment to a tenured post in a university or state researchinstitute and this was important for most contributors to all four innova-tions in the countries studied here Moreover where such posts werenot widely institutionalised in a national academic system andor becamereduced through restructuring as in the Netherlands for ED the continueddevelopment of these innovations became much more difficult

In some fields especially the laboratory sciences where post-doctoralposts have become widely established in a number of OECD countriesresearchers could obtain fellowships dedicated to particular problems inthe context of the innovations This was particularly important for theBEC problem in Germany and Switzerland but also helped work on highcost ED and CCL in Sweden Significant features of many of these fellow-ships as well as shorter-term appointments in different laboratories anddepartments were their international availability and the opportunitiesthey offered to learn new knowledge and skills In three of the four casesthe opportunities to work with internationally renowned researchers werevery important for many scientists such that it seems unlikely they wouldhave been able to make significant contributions without them In someinstances the availability of positions and research facilities in foreigncountries was also important for scientists to continue to work on problemsand approaches that required long-term commitments particularly in evo-lutionary developmental biology Switzerland was especially hospitable inthis regard (Benninghoff et al this volume) At least for some researchersthen international careers and learning opportunities have helped them togain sufficient protected space to make major investments in dealing withmajor intellectual problems

Considering next the conditions that enabled scientists to obtain thenecessary and substantial resources for developing these innovations ineach country these can be summarised under three broad headings (a)local flexibility and capacity (b) national and international diversity offunding agencies and programmes and (c) extra-scientific sources of sup-port In all four scientific cases the availability of local technical supportsuch as workshops computers experimental apparatus and technicianswas important and often crucial for scientists to work effectively on theseinnovations Often such support was provided by departments and insti-tutes although not always by the home department of the researcher con-cerned but in a few instances the resources were allocated by the largerorganisation The key factor was organisational flexibility that enabledprofessors institute heads deans and other senior staff to support this kind

of research with locally available human and material resources withoutneeding to gain external project funding or approval from extra-muralorganisations This requires both the existence of such resources and discre-tionary authority over their use at the local level where the particular kindof research is being conducted because skills and facilities are often specificto certain types of scientific work

Resource flexibility is also encouraged where researchers are able toobtain project funding from different agencies whether national or inter-national public bodies charities or private foundations that pursuevaried objectives While such diversity may not always enable unorthodoxor interdisciplinary research to be funded it is likely to involve the use of awider range of advisors and referees than where all public research supportcomes from a single national agency and there are very few alternativesources of support for academic research Even where there is only onesuch national research council as in Switzerland it may support a varietyof funding programmes with diverse purposes and reviewers that can alsoprovide some resource flexibility especially where the public funding ofacademic research is relatively generous An additional factor supportingBEC research in Switzerland but not ED projects was the common prac-tice of the SNF to roll over grants for continuing work in the physicalsciences without scientists having to re-apply and compete for supportanew every few years The key factor here is diversity of agencies pro-grammes purposes and advisors A further important feature of fundingsystems that provided some resource flexibility for these innovations wasthe willingness of many agencies to delegate substantial discretion to recipi-ents of project grants over how they used the money This enabled scientiststo deal with unforeseen results and to combine resources from differentgrants when necessary

Extra-scientific interests can also provide some resource flexibility forparticular kinds of research Especially important has been the role of pri-vate companies and other commercial interests in supporting the develop-ment of computer-based linguistic corpora at least in the United Kingdomand some other countries Publishers of dictionaries and materials for lan-guage teaching have been active in providing funding for building corporasometimes on a very long term basis and even to the extent of threateningthe firmrsquos existence Likewise political and bureaucratic interests haveplayed a significant role in the development of comparative studies ofschool performance in some countries Even if the extent and standing ofsuch research has been limited among established academic elites thisevidence of the significance of both political and commercial interests in

398 RICHARD WHITLEY

supporting specific intellectual innovations in the human sciences suggeststhat biomedical and technological fields are by no means the only scienceswhere such support can be important influences on research strategies inacademia albeit usually in combination with public agencies dedicated tolargely scientific goals

Flexibility of the standards governing the evaluation of scientific contri-butions and reputations is also enhanced by institutional and organisa-tional diversity particularly when this supports multiple career andemployment opportunities The Max Planck Institutes and other publicresearch institutes in Germany for example enable senior staff to pursuelong-term projects that may be unfashionable among funding agenciesandor difficult to support on a two to three year project basis as in thecases of BEC and ED discussed by Glaser et al in their article (thisvolume) Similarly the existence of different academic departments withinuniversities whose research and teaching domains overlap can create moreopportunities or organisational slack for scientists to undertake multidis-ciplinary research with unorthodox approaches than where a single organi-sational unit controls all posts and careers in particular fields Just as theexpanding business schools in many research universities provided alterna-tive career paths for social scientists in the 1970s and 1980s in many OECDcountries so too have departments and institutes of palaeontology plantbiology and related fields helped to provide opportunities for ED research-ers and a variety of natural language departments supported computer-based corpus linguistics research in many European PSS

The growth of publication outlets and scientific audiences for researchconcerned with novel goals and approaches has also been important for thedevelopment of many innovations In the case of ED for instance scien-tists have been able to publish many of their results in a variety of journalseven if some of these were not the most prestigious or had the most impactThe expansion of higher education and public support for scientificresearch in many OECD countries since 1945 has facilitated this growth innew conferences journals and other media for disseminating the results ofwork on new problems andor developing new approaches Such expansionhas helped scientists pursuing unorthodox topics with novel research prac-tices to establish themselves as distinct reputational communities as theInternational Computer Archive of Modern and Medieval English confer-ences and journal has done since 1977 in CCL (Svartvik 2007)

Not all these communities have though become fully entrenched in uni-versity departments faculties and institutes especially in PSS dominatedby the institute model of academic organisation but the relative ease with

which new journals and conference themes have been set up in many scien-tific fields has helped to create new audiences for intellectual innovationsand supported those wishing to pursue research goals that were not centralto established disciplines This was especially the case for innovations inareas where policy- and business-related interests were involved such asschool effectiveness research and CCL and more widely in business andmanagement studies (Engwall Kipping amp Usdiken 2010)

This analysis of the varied ways in which particular features of nationalPSS have enabled or inhibited the development of different kinds of intel-lectual change through their impact on levels of protected space and flex-ibility suggests the following five conclusions could be drawn about howinstitutional reforms are likely to affect the pursuit of different researchgoals and approaches First reductions in the discretion afforded to aca-demic staff over which kinds of research to pursue through for instancechanging the conditions governing tenured appointments and increasingthe frequency of performance evaluations tied to short-term outputs andproject funding success will decrease the levels of researchersrsquo protectedspace and probably inhibit them from undertaking long-term projects withhighly uncertain outcomes

It seems though that the period of protected space needed for research-ers to develop new skills and knowledge can be reduced to a limited extentby scientists being able to work with research leaders in different locationsoften abroad Relatedly in at least some cases innovating scientists havebeen able to pursue long-term projects supported by a succession ofresearch fellowships frequently in different organisational settingsInternational mobility and the widespread availability of such fellowshipscan then be helpful in enabling some researchers to undertake projectsrequiring extensive periods of learning especially in sciences where pro-blems can be decomposed into discrete modules They do not howeverguarantee success in dealing with complex problems with highly uncertainexperimental outcomes

Second the availability of local research capability is often crucial forscientists to work on complex problems requiring substantial technicalinfrastructure If animal houses university workshops and other forms oforganisationally controlled infrastructure decline andor are closed downthe flexibility they provide will be lost and this may well discourageresearchers from tackling problems that require access to such generalpurpose facilities especially in countries where national research funding isdifficult to obtain and rarely supports infrastructure provision Similarlyreplacing university block grant funding with project-specific supportreduces flexibility at the local level and is likely to inhibit the pursuit of

400 RICHARD WHITLEY

unfashionable research goals that involve cross-disciplinary coordinationespecially where project proposals are judged by discipline-based panelsEqually where states make major changes in their funding prioritiestowards supporting research that promises to contribute more directly tosocio-economic goals they can have quite a significant impact on the direc-tion of academic research as in the case of ED in the Netherlands in the2000s

Such consequences of substantial funding and governance reforms ofPSS are particularly probable where state support for academic research isbecoming tightly constrained and is highly concentrated in a few agenciesorganised around established disciplines While universities may be equallygoverned by current scientific elitesrsquo perceptions of worthwhile projectstheir multiplicity and varied histories seem more likely to encourage intel-lectual diversity and the pursuit of different innovations than when projectfinance is allocated by a single central agency reliant on the assessments ofdiscipline-based advisors In general then third the more reliant scientistsbecome on project-based funding from a single agency with a limited rangeof programmes and the less able that agency is to support researchers interms of available resources per qualified applicant the more difficult itwill be for scientists to pursue complex research goals requiring long-termsupport with expensive apparatus and infrastructure

To some extent fourth such concentrated control of resources may beoffset by support from policy makers and commercial interest as illustratedby the cases of ILSA and CCL However it is important to note that thedevelopment of intellectual innovations with resources provided by thesekinds of extra-scientific interests depends on the level of support being sub-stantial and relatively long term as in the creation of established chairs andresearch institutes or helping to fund the construction of large corpora andoften is combined with funding from more conventional intra-scientificorganisations as in the case of some British corpora building

Finally the reduction of higher education budgets in many OECDcountries and retrenchment of public support for academic research in thepast two or so decades seem likely to make it more difficult to establishnew journals and conferences around unfashionable intellectual innova-tions especially where these conflict with institutionalised priorities andapproaches in prestigious sciences In particular it may well become moredifficult to build academic careers around deviant goals and frameworks asuniversities and other PROs compete for scarce resources and reputationsfor excellence as understood by current national and international elitesparticularly where these elites also control the assessment of research con-tributions in lsquostrongrsquo research evaluation systems (Whitley 2007)

CONCLUDING REMARKS

Considering more general concerns about how major funding and institu-tional reforms are affecting intellectual developments in different sciencesthis discussion highlights five points First it is critical to identify theirimpact on the authority relationships governing the selection and formula-tion of research problems the skills and knowledge used and how collectivejudgements of scientific merit are made Such changes affect the degree anddistribution of protected space afforded to researchers and the flexibility ofstandards governing the allocation of resources and reputations which inturn affect scientistsrsquo decisions about how to invest their time skills andknowledge in dealing with particular kinds of problems for reputationalpurposes

Second many of the changes that have taken place particularly inOECD countries are likely to have divergent and sometimes contradictoryconsequences for scientistsrsquo behaviour at least in terms of how they havebeen justified For example intensifying post hoc evaluations of depart-ments and universitiesrsquo performance in terms of their contributions toscientific knowledge through discipline-based peer review reinforces theauthority of disciplinary elites over priorities and careers that is also beingstrengthened by researchersrsquo increasing dependence on peer-reviewedproject-based finance Such reinforcement makes it more risky for scientiststo undertake cross-disciplinary research into problems that are beingemphasised by state agencies but are regarded as uninteresting by suchelites who dominate peer group assessments Equally enhancing the powersand authority of university administrators in situations of tight resourceconstraints and intense competition for project support can increasepressures on scientists to pursue mainstream research and concentrate ongenerating publishable results in the short to medium term rather thanencouraging universities to pursue distinct profiles and strategic diversity

Third the impact of changes in authority relations on protected spaceflexibility and innovation patterns varies considerably between differentlyorganised PSS The stratification of higher education institutions con-centration of control of key resources and traditional stateuniversity rela-tionships all affect how these changes develop and are implementedespecially the roles played by scientific elites and university managersOutcomes are also influenced by broader institutional and politicaleconomicfeatures of each society such as the historical structure and role of the statein managing economic and social change the nature of dominant economic

402 RICHARD WHITLEY

sectors the organisation of elite labour markets and the size distribution ofmajor employers (Regini 2011a Whitley 2012) These contextual featuresinfluence how states introduce governance changes how universities andother major actors respond to them and how authority relations alter withinscientific communities

Fourth there are important variations in the dominant research prac-tices of different kinds of science that affect the impact of governancechanges on innovation patterns Decomposability and modularisation ofresearch problems can mitigate the consequences of reduced protectedspace for the selection of topics that involve considerable switching costswhile the need for substantial uninterrupted research time and dependenceon personal interpretations of problem formulations and evidence are likelyto increase them Within limits decomposability can also help scientistsovercome some of the effects of reduced flexibility on intellectual diversityand challenge by facilitating their diversification of problem choices whilehigh levels of capital intensity tend to reinforce them

Finally some groups of scientists can under particular circumstancesuse external pressures to achieve quite radical change at the local level andeven attempt to reorganise and redirect national scientific communitiesWhile many of the changes discussed in the literature on NPM and highereducation stem from political and administrative imperatives and resourceconstraints at the national level some have developed in tandem withmoves by scientific elites to preserve core support for their ideas of the bestscience by encouraging competition for resources and peer reviewbasedevaluations of performance as in the United Kingdom (Martin amp Whitley2010) By establishing strong evaluation systems and procedures for allo-cating resources that build on prevalent research practices in particularkinds of science such as biological and biomedical ones these elites areprivileging their kinds of scientific research at the expense of those basedupon different practices

ACKNOWLEDGEMENTS

Earlier versions of this paper were presented to conferences and workshopsin Helsinki Berlin and Stockholm I am most grateful for comments andsuggestions received on those occasions as well of course to my colleaguesin the RHESI project

REFERENCES

Aljets E amp Lettkemann E (2011 July) New demands for strategic action Research in

Germany Paper presented to Subtheme 45 on Reconstructing Universities as

Organizations Increasing Authority with Limited Strategic Capabilities at the 27th

EGOS Colloquium Gothenburg

Ballarino G amp Perotti L (2011) Italy Gradual changes and an uncertain autonomy In

M Regini (Ed) European universities and the challenge of the market (pp 168182)

Barlow M (2011) Corpus linguistics and theoretical linguistics International Journal of

Corpus Linguistics 16 344

Brunsson N amp Sahlin-Andersson K (2000) Constructing organizations The example of

Capano G (2009) Looking for serendipity The problematical reform of government within

Italyrsquos universities Higher Education 55 481504

De Boer H Enders J amp Leistye L (2007) Public sector reform in Dutch higher education

Durrell M (2012 December) lsquoRepresentativenessrsquo lsquobad datarsquo and legitimate expectations

What can an electronic historical corpus tell us that we didnrsquot actually know already (and

how) Paper presented to a conference on historical corpora held at the Johann Goethe

University Frankfurt

Engwall L Kipping M amp Usdiken B (2010) Public science systems higher education and

the trajectory of academic disciplines Business studies in the United States and Europe

Fries P H (2010) Charles C Fries linguistics and corpus linguistics ICAME Journal 34

Geuna A amp Nesta L (2006) University patenting and its effects on academic research The

emerging European evidence Research Policy 35 790807

Glaser J Lange S Laudel G amp Schimank U (2010) The limits of universality How

field-specific epistemic conditions affect authority relations and their consequences In

lae on Australian University research In R Whitley amp J Glaser (Eds) The changing

governance of the sciences The advent of research evaluation systems (Vol 26

pp 127152) Dordrecht Springer (Sociology of the Sciences Yearbook)

Hardie A amp McEnery T (2010) On two traditions in corpus linguistics and what they have

in common International Journal of Corpus Linguistics 15 384394

Joerges B amp Shinn T (Eds) (2001) Instrumentation between science state and industry

(Vol 22) Dordrecht Kluwer (Sociology of the Sciences Yearbook)

404 RICHARD WHITLEY

sity structures Higher Education 46 315339

G S Drori J W Meyer amp H Hwang (Eds) Globalization and organization World

society and organizational change (pp 241257) Oxford Oxford University Press

Lee F (2007) The research assessment exercise the state and the dominance of mainstream

economics in British universities Cambridge Journal of Economics 31 309325

Lee F Pham X amp Gu G (2013) The UK research assessment exercise and the narrowing

of UK economics Cambridge Journal of Economics 37 693717

Morris N (2010) Authority relations as condition for and outcome of shifts in governance

The limited impact of the UK research assessment exercise on the biosciences In

Moscati R (2001) Italian professors in transition Higher Education 41 103129

Mowery D C Nelson R R Sampat B N amp Ziedonis A A (2004) Ivory tower and indus-

trial innovation University-industry technology transfer before and after the Bayh-Dole

Act Stanford CA Stanford University Press

Mukherjee J (2010) Corpus linguistics versus corpus dogmatism Pace Wolfgang Teubert

International Journal of Corpus Linguistics 15 370378

Muller-Camen M amp Salzgeber S (2005) Changes in academic work and the chair regime

The case of German business administration academics Organization Studies 26

271290

Regini M (2011a) The challenge of the market In M Regini (Ed) European universities and

the challenge of the market (pp 8090) Cheltenham Edward Elgar

Regini M (Ed) (2011b) European universities and the challenge of the market Cheltenham

Edward Elgar

Smith A amp Raven R (2012) What is protective space Reconsidering niches in transitions

to sustainability Research Policy 41 10251036

Stokes D E (1997) Pasteurrsquos quadrant Basic science and technological innovation

Washington DC Brookings Institution Press

Teubert W (2010) Our brave new world International Journal of Linguistics 15 354358

Weingart P amp Maasen S (2007) Elite through rankings the emergence of the enterprising

The advent of research evaluation systems (Vol 26 pp 75100) Dordrecht Springer

(Sociology of the Sciences Yearbook)

ing research evaluation systems for knowledge production in different countries and

scientific fields In R Whitley amp J Glaser (Eds) The changing governance of the

sciences The advent of research evaluation systems (Vol 26 pp 327) Dordrecht

Springer (Sociology of the Sciences Yearbook)

research evaluation systems (Vol 26) Dordrecht Springer (Sociology of the Sciences

Yearbook)

Wilson D (2008) Reconfiguring biological sciences in the late twentieth century A study of the

University of Manchester Manchester Faculty of Life Sciences University of

Manchester

Wilson D amp Lancelot G (2008) Making way for molecular biology Institutionalising and

managing reform of biological science in a UK university during the 1980s and 1990s

Studies in the History and Philosophy of the Biological and Biomedical Sciences 39

Wright S amp Oslashrberg J (2009) Prometheus (on the) rebound Freedom and the Danish steer-

ing system In J Huisman (Ed) International perspectives on the governance of higher

education (pp 6987) London Routledge

406 RICHARD WHITLEY

FRONT COVER

ORGANIZATIONAL TRANSFORMATION AND SCIENTIFIC CHANGE THE IMPACT OF INSTITUTIONAL RESTRUCTURING ON UNIVERSITIES AND INTELLECTUAL INNOVATION

COPYRIGHT PAGE

CONTENTS

ADVISORY BOARD

EDITORS INTRODUCTION

INSTITUTIONAL CHANGE AND THE TRANSFORMATION OF UNIVERSITIES AS STRATEGIC ACTORS

THE IMPACT OF INSTITUTIONAL REFORMS ON THE NATURE OF UNIVERSITIES AS ORGANISATIONS

INTRODUCTION


VARIETIES OF PRE-REFORM UNIVERSITIES AND CHARACTERISTICS OF FIRMS

THE IMPACT OF INSTITUTIONAL REFORMS ON UNIVERSITY AUTONOMY

THE IMPACT OF INSTITUTIONAL REFORMS ON UNIVERSITY ORGANISATIONAL CAPABILITIES

CONCLUDING REMARKS

REFERENCES

EMPOWERMENT OF FRENCH UNIVERSITIES BY FUNDING AND EVALUATION AGENCIES

INTRODUCTION

ORGANIZATIONAL CHANGE AND TRANSFORMATION OF UNIVERSITY GOVERNANCE

FRENCH HIGHER EDUCATION REFORMS THE LOPRI (2006) AND THE LRU (2007)

EXTERNAL PEER REVIEW ENHANCES CENTRALIZATION AND COORDINATION WITHIN UNIVERSITIES

EXTERNAL PEER REVIEW ALLOWS LESS EGALITARIAN ALLOCATION OF RESOURCES AND RESTRUCTURATIONS

A RATHER LARGE APPROPRIATION OF THE NORMS AND EXPECTATIONS DEVELOPED BY THE EVALUATION AGENCY AND THE RESEARCH COUNCIL

CONCLUSION

NOTES

REFERENCES

FUNDING INDIVIDUALS ndash CHANGING ORGANISATIONS THE IMPACT OF THE ERC ON UNIVERSITIES

INTRODUCTION

CONCEPTUAL MODEL

METHODOLOGY


DISCUSSION

CONCLUSION THE IRONY OF ORGANISATIONAL ACTORHOOD AND ERC SUCCESS

NOTES

ACKNOWLEDGEMENTS

REFERENCES

WHERE HAVE ALL THE SCIENTISTS GONE BUILDING RESEARCH PROFILES AT DUTCH UNIVERSITIES AND ITS CONSEQUENCES FOR RESEARCH

INTRODUCTION


THE EVOLUTION OF DUTCH STATE EXPECTATIONS CONCERNING UNIVERSITY RESEARCH PROFILES


THE IMPACT OF PROFILE BUILDING ON RESEARCH FIELDS

CONCLUSIONS CAUSES AND CONSEQUENCES OF PROFILE BUILDING

NOTE

ACKNOWLEDGEMENTS

REFERENCES

MERGER MANIA IN SCIENCE ORGANIZATIONAL RESTRUCTURING AND PATTERNS OF COOPERATION IN AN ACADEMIC RESEARCH CENTRE

INTRODUCTION

BEYOND MYTH AND CEREMONY HOW FORMAL STRUCTURES MAY MATTER IN ACADEMIC SETTINGS

DATA AND METHODS

BACKGROUND


EXTERNAL AND INTRA-ORGANIZATIONAL IMPLICATIONS OF THE MERGER


NOTES

ACKNOWLEDGEMENTS

REFERENCES

CONTRADICTORY CONSEQUENCES OF INSTITUTIONAL CHANGES ON INTELLECTUAL INNOVATION IN THE PUBLIC SCIENCES

INSTITUTIONAL CONDITIONS AND CHANGING RESEARCH PRACTICES IN SWITZERLAND

INTRODUCTION



CHANGING RESEARCH PRACTICES IN FOUR SCIENTIFIC DOMAINS IN SWITZERLAND

THE INFLUENCE OF INSTITUTIONAL CONDITIONS ON CHANGING RESEARCH PRACTICES IN THE SWISS RESEARCH SYSTEM

CONCLUSIONS

NOTES

REFERENCES

COLD ATOMS ndash HOT RESEARCH HIGH RISKS HIGH REWARDS IN FIVE DIFFERENT AUTHORITY STRUCTURES

INTRODUCTION


METHODS AND DATA

EPISTEMIC PROPERTIES OF EXPERIMENTAL BEC RESEARCH AND ITS REQUIRED PROTECTED SPACE

INTERNATIONAL AND NATIONAL DYNAMICS OF BEC RESEARCH

HOW VARIOUS SETS OF AUTHORITIES ALLOW HINDER PREVENT BEC RESEARCH

MACRO-MICRO LINKS IN THE DEVELOPMENT OF SCIENTIFIC INNOVATIONS

CONCLUSIONS

NOTES

ACKNOWLEDGEMENT

REFERENCES

HIGHLY ADAPTABLE BUT NOT INVULNERABLE NECESSARY AND FACILITATING CONDITIONS FOR RESEARCH IN EVOLUTIONARY DEVELOPMENTAL BIOLOGY

INTRODUCTION


METHODS AND DATA

lsquoTHE SNAKE TOOK US ABOUT THREE YEARSrsquo ndash TRANSITIONS TO EVO-DEVO

CONCLUSIONS

THE IMPACT OF AUTHORITY RELATIONS ON PATHS TO EVO-DEVO

NOTES

ACKNOWLEDGEMENTS

REFERENCES

PATH DEPENDENCE AND POLICY STEERING IN THE SOCIAL SCIENCES THE VARIED IMPACT OF INTERNATIONAL LARGE SCALE STUDENT ASSESSMENT ON THE EDUCATIONAL SCIENCES IN FOUR EUROPEAN COUNTRIES

INTRODUCTION

COMPARING THE IMPACT OF INTERNATIONAL LARGE SCALE STUDENT ASSESSMENTS


EXPLAINING COMMONALITIES AND DIFFERENCES IN NATIONAL RESPONSES TO ILSA STUDIES

CONCLUSIONS TENSIONS BETWEEN SERVICE DELIVERY AND ENDOGENOUS RESEARCH DYNAMICS

NOTE

REFERENCES

WHERE TO GO FOR A CHANGE THE IMPACT OF AUTHORITY STRUCTURES IN UNIVERSITIES AND PUBLIC RESEARCH INSTITUTES ON CHANGES OF RESEARCH PRACTICES


APPROACH

BEGINNING NEW LINES OF RESEARCH IN DIFFERENT ORGANISATIONAL SETTINGS

PATTERNS OF AUTHORITY SHARING AND OPPORTUNITIES TO CHANGE RESEARCH PRACTICES IN TWO TYPES OF RESEARCH ORGANISATIONS

CONCLUSIONS

NOTES

REFERENCES

COMPUTER CORPUS LINGUISTICS AN INNOVATION IN THE HUMANITIES

INTRODUCTION


THE EMERGENCE OF CCL IN FOUR EUROPEAN COUNTRIES

CONCLUSIONS

NOTES

ACKNOWLEDGMENTS

REFERENCES

HOW DO INSTITUTIONAL CHANGES AFFECT SCIENTIFIC INNOVATIONS THE EFFECTS OF SHIFTS IN AUTHORITY RELATIONSHIPS PROTECTED SPACE AND FLEXIBILITY

INSTITUTIONAL CHANGE AUTHORITY RELATIONSHIPS AND LEVELS OF PROTECTED SPACE AND FLEXIBILITY IN THE PUBLIC SCIENCES

IMPACT OF CHANGING AUTHORITY RELATIONSHIPS ON LEVELS OF PROTECTED SPACE AND FLEXIBILITY OF INTELLECTUAL STANDARDS IN PUBLIC SCIENCE SYSTEMS

LEVELS OF PROTECTED SPACE AND FLEXIBILITY SUPPORTING THE DEVELOPMENT OF FOUR SCIENTIFIC INNOVATIONS

CONDITIONS PROVIDING PROTECTED SPACE AND FLEXIBILITY FOR DEVELOPING INNOVATIONS IN FOUR EUROPEAN COUNTRIES

CONCLUDING REMARKS

ACKNOWLEDGEMENTS

REFERENCES