227555475-1848558449

TRANSPORT SURVEY METHODS:

KEEPING UP WITH A CHANGING WORLD

Related Books

HENSHER Trafc Safety and Human Behavior

HENSHER & BUTTON Handbooks in Transport 6 Volume set

STOPHER & STECHER Travel Survey MethodsHIMANEN, LEE-GOSSELIN& PERRELS

Building Blocks for Sustainable Transport

ELVIK & VAA The Handbook of Road Safety MeasuresBEN-AKIVA, MEERSMAN &VAN DE VOORDE

Recent Developments in TransportModeling: Lessons for the Freight Sector

TRANSPORT SURVEY METHODS:KEEPING UP WITH A CHANGINGWORLD

EDITED BY

PATRICK BONNELENTPE-LET, Lyon, France

MARTIN LEE-GOSSELINUniversite Laval, Quebec, Canada

JOHANNA ZMUDNuStats, USA

JEAN-LOUP MADREINRETS, Paris, France

United Kingdom North America JapanIndia Malaysia China

Emerald Group Publishing Limited

Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2009

Copyright r 2009 Emerald Group Publishing Limited

Reprints and permission service

Contact: [email protected]

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. No responsibility is accepted for the accuracy of

information contained in the text, illustrations or advertisements. The opinions expressed

in these chapters are not necessarily those of the Editor or the publisher.

British Library Cataloguing in Publication Data

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

ISBN: 978-1-84855-844-1

Contents

List of Contributors ix

Preface xiii

Acknowledgements xv

PART I: INTRODUCTION

1. Keeping Up with a Changing World: Challenges in the Design ofTransport Survey Methods 3Patrick Bonnel, Martin Lee-Gosselin, Jean-Loup Madre andJohanna Zmud

2. The Travel Survey Toolkit: Where to From Here? 15Peter R. Stopher

PART II: SUSTAINABILITY AND TRAVELLER ADAPTATION

3. What is so Special about Surveys Designed to Investigate theEnvironmental Sustainability of Travel Behaviour? 49Peter Bonsall

4. Evolving Behaviour in the Context of Interest in EnvironmentalSustainability: Synthesis of a Workshop 71Mark Bradley

5. Myths, (Mis)Perceptions and Reality in Measuring VoluntaryBehavioural Changes 81Werner Brog and Ian Ker

6. Surveys for Behavioural Experiments: Synthesis of a Workshop 113Peter Jones, Regine Gerike and Giorgia Servente

PART III: GLOBAL SOCIAL ISSUES

7. Surveying Hard-to-Reach Groups 127Benot Riandey and Martine Quaglia

8. The Challenges of Surveying Hard to Reach Groups:Synthesis of a Workshop 145Roger Behrens, Mark Freedman and Nancy McGuckin

9. Emerging Methods and Technologies for Tracking PhysicalActivity in the Built Environment 153Sean T. Doherty

10. Physical Activity in the Built Environment: Synthesis of a Workshop 191Kelly J. Clifton

11. Acquiring Data on Travel Behaviour During Emergencies andExceptional Events 197Earl J. Baker

12. Capturing Travel Behavior during Exceptional Events:Synthesis of a Workshop 213Chester Wilmot and Matthieu de Lapparent

13. Tourist Flows and Inows: On Measuring Instruments andthe Geomathematics of Flows 219Christophe Terrier

14. Surveys of Tourists and Transients: Synthesis of a Workshop 243Carlos H. Arce and Alan Pisarski

PART IV: FREIGHT AND TRANSIT PLANNING

15. How to Improve the Capture of Urban Goods Movement Data? 251Daniele Patier and Jean-Louis Routhier

16. Surveys on Urban Freight Transport: Synthesis of a Workshop 289Matthew Roorda, Arnim Meyburg and Michael Browne

17. The Collection of Long-Distance Road Freight Data in Europe 295Alan McKinnon and Jacques Leonardi

18. The Acquisition of Long-Distance Freight Data:Synthesis of a Workshop 311Kara Kockelman, Michael Browne and Jacques Leonardi

vi Contents

19. Identifying and Reconciling the Data Needs of PublicTransit Planning, Marketing and Performance Measurement 321Gerd Sammer

20. Data for Public Transit Planning, Marketing and ModelDevelopment: Synthesis of a Workshop 349Orlando Strambi, Martin Trepanier and Linda Cherrington

PART V: TECHNOLOGY APPLICATIONS

21. Collecting and Processing Data from Mobile Technologies 361Peter R. Stopher

22. Mobile Technologies: Synthesis of a Workshop 393Jean Wolf

23. What Is Different About Non-Response inGPS-Aided Surveys? 403Stacey Bricka

24. Nonresponse Challenges in GPS-Based Surveys:Synthesis of a Workshop 429Heather Contrino

25. Electronic Instrument Design and User Interfaces forActivity-Based Modeling 437H. J. P. Timmermans and E. Hato

26. Electronic Instrument Design and User Interfaces:Synthesis of a Workshop 463Elaine Murakami, Sharon OConnor and Jane Gould

27. Visualize This: Opportunities and Challenges for theTravel Survey Community 473Catherine T. Lawson

28. Data Visualization Techniques: Synthesis of a Workshop 497Michael Manore and Stephan Krygsman

PART VI: EMERGING/PERSISTENT SURVEY ISSUESAND DATA HARMONISATION

29. Large-Scale Ongoing Mobility Surveys: The State of Practice 505Elizabeth S. Ampt, Juan de Dios Ortuzar and Anthony J. Richardson

Contents vii

30. Moving from Cross-Sectional to Continuous Surveying:Synthesis of a Workshop 533Dirk Zumkeller and Peter Ottmann

31. Moving Towards Continuous Collection of Large-Scale MobilitySurveys: Are There Compelling Reasons? A Discussant Response 541Tim Raimond

32. Vehicle-Based Surveys: Towards More Accurate and ReliableData Collection Methods 549Dominika Kalinowska and Jean-Loup Madre

33. Vehicle-Based Surveys: Synthesis of a Workshop 577Klaas van Zyl, Dominika Kalinowska, Jean-Loup Madre andBob Leore

34. Survey Mode Integration and Data Fusion: Methods and Challenges 587Caroline Bayart, Patrick Bonnel and Catherine Morency

35. Best Practices in Data Fusion: Synthesis of a Workshop 613John W. Polak and Eric Cornelis

36. Lessons from an Overview of National Transport Surveys, fromWorking Group 3 of COST 355: Changing Behavior Toward a MoreSustainable Transport System 621Jimmy Armoogum, Kay W. Axhausen and Jean-Loup Madre

Subject Index 635

viii Contents

List of Contributors

Elizabeth S. Ampt Sinclair Knight Merz, Adelaide, Australia

Carlos H. Arce Mygistics-NuStats, Austin, TX, USA

Jimmy Armoogum French National Institute of Research on Transportsand Safety (INRETS-DEST), Noisy, France

Kay W. Axhausen Swiss Federal Institute of Technology Zurich Institut for Transport Planning and Systems(ETHZ-IVT), Zurich, Switzerland

Earl J. Baker Florida State University, Tallahassee, FL, USA

Caroline Bayart LET-ENTPE-CNRS-Universite Lyon 2, Lyon, France

Roger Behrens University of Cape Town, Cape Town, South Africa

Patrick Bonnel LET-ENTPE-CNRS-Universite Lyon 2, Lyon, France

Peter Bonsall Institute for Transport Studies, University of Leeds,Leeds, UK

Mark Bradley Bradley Research and Consulting, Santa Barbara,CA, USA

Stacey Bricka NuStats, Austin, TX, USA

Werner Brog Socialdata, Munich, Germany

Michael Browne University of Westminster, London, UK

Linda Cherrington Texas Transportation Institute, The Texas A&MUniversity System, Houston, TX, USA

Kelly J. Clifton University of Maryland, College Park, MD, USA

Heather Contrino Federal Highway Administration, USA

Eric Cornelis FUNDP, University of Namur, Belgium

Matthieu de Lapparant INRETS, Noisy-le-Grand Cedex, Paris, France

Sean T. Doherty Department of Geography & Environmental Studies,Wilfrid Laurier University, Waterloo, Ontario,Canada

Mark Freedman Westat, Rockville, MD, USA

Regine Gerike University of Technology, Munich, Germany

Jane Gould University of California, Los Angeles, CA, USA

E. Hato Behavior in Networks Studies Unit, University ofTokyo, Tokyo, Japan

Peter Jones University College London, London, UK

Dominika Kalinowska German Institute for Economic Research, DIW Berlin,Germany

Ian Ker CATALYST, WA, Australia

Kara Kockelman University of Texas at Austin, TX, USA

Stephan Krygsman University of Stellenbosch, South Africa

Catherine T. Lawson University at Albany, State University of New York,Albany, NY, USA

Martin Lee-Gosselin Universite Laval, ESAD-CRAD, Quebec, Canada

Jacques Leonardi University of Westminster, London, UK

Bob Leore Department of Transport, Ottawa, Canada

Jean-Loup Madre INRETS, Paris, France

Michael Manore Vispective Management Consulting, USA

Nancy McGuckin Travel Behavior Associates, South Pasadena,CA, USA

Alan McKinnon Heriot-Watt University, Edinburgh, UK

Arnim Meyburg Cornell University, Ithaca, NY, USA

Catherine Morency Ecole Polytechnique de Montreal,MADITUC-CIRRELT, Canada

Elaine Murakami Federal Highway Administration, Seattle, WA, USA

Sharon OConnor Resource Systems Group, White River Junction,VT, USA

Juan de Dios Ortuzar Department of Transport Engineering and Logistics,Ponticia Universidad Catolica de Chile, Santiago,Chile

x List of Contributors

Peter Ottmann Institute for Transport Studies, Karlsruhe, Germany

Danie`le Patier LET, Universite de Lyon, France

Alan Pisarski Consultant, Falls Church, VA, USA

John W. Polak Centre for Transport Studies, Imperial CollegeLondon, UK

Martine Quaglia INED-SFS, Paris, France

Tim Raimond Transport Data Centre, NSW Transport andInfrastructure, Australia

Benot Riandey INED-SFS, Paris, France

Anthony J. Richardson The Urban Transport Institute (TUTI), Victoria,Australia

Matthew Roorda University of Toronto, Toronto, Canada

Jean-Louis Routhier LET, Universite de Lyon, France

Gerd Sammer Institute for Transport Studies, University of NaturalResources and Applied Life Sciences, Vienna, Austria

Giorgia Servente DITIC, Politecnico di Torino, Torino, Italy

Peter R. Stopher The Institute of Transport and Logistics Studies,The University of Sydney, Australia

Orlando Strambi Escola Politecnica da Universidade de Sao Paulo,Brazil

Christophe Terrier INSEE, France

H. J. P. Timmermans Urban Planning Group, Eindhoven University ofTechnology, The Netherlands

Martin Trepanier Ecole Polytechnique, University of Montreal, Canada

Klaas van Zyl SSI Engineers and Environmental Consultants,South Africa

Chester Wilmot Louisiana State University, Baton Rouge, LA, USA

Jean Wolf GeoStats, Atlanta, GA, USA

Johanna Zmud NuStats, USA

Dirk Zumkeller Institute for Transport Studies, Karlsruhe, Germany

List of Contributors xi

Preface

About every three years the international transport survey community gets togetherto discuss innovation and quality in transport survey methods and document thosediscussions in a publication. As this has been happening since 1979, one has to ask:Do we really need another book on survey methods in transport? The answer is Yes,Ify with the If involving whether there have been recent changes that make thecapture of transportation data through surveys materially different than they wereyears ago when other books on the topic were published. One obvious example of arecent change is the need for new data, models, and other analytical tools to supportgreenhouse gas (GHG) reduction and energy efciency policies in nations around theglobe. Another example is the proliferation of new and affordable informationtechnologies that survey designers can employ to collect and process data, helpingthem confront increasing barriers to participation in surveys, barriers that in somecases arise from the same technologies, such as developments in telephony. This bookfocuses on such changes, and on the opportunities and challenges they represent,both for improved survey methods and for the comparability of the data that theyprovide to different agencies and countries.

As statistical surveys attempt to address GHG issues, and other importanttransportation policy and planning challenges, they exhibit evidence of success, yet atthe same time they frequently come under threat. The evidence of success is thatsurveys are ubiquitous in the transport world. Almost all countries in the world usethem to measure passenger travel, freight movements, or public transit ridership.Leaders in government use survey results to guide policy, and the call from theiradvisors for more data to address greater challenges is increasing. However, fundingfor data collection is too often an easy target in difcult economic times, such as thepresent, and surveys must be credible, transparent, and of assured quality. Thepapers in this book are thus relevant to government, transport industry practitioners,academic scientists, and commercial researchers.

The book provides a review of the current state of transport survey methods forcapturing data in several key areas: freight, personal travel, tourism, evacuations andrelated travel, and the environmental footprint of transport, among others. Itcaptures the essence of discussions at the 8th International Conference on SurveyMethods in Transport that took place in Annecy, France, in 2008. Conferenceparticipants from over 25 countries included leading survey researchers and transport

professionals representing industry and government policy makers, as well asacademic scholars and researchers.

The Annecy Conference succeeded in its main objectives: sharing up-to-dateinformation and experiences on transport survey methods; fostering discussion ofmutual problems and issues that affect survey design, data processing, and reporting;proposing and suggesting new initiatives and future approaches for the measurementof critical transportation system indicators; and feeding the results of thesediscussions into a permanent record in the form of this peer-reviewed book. Thebook is not a proceedings volume, but a peer-reviewed selection of about one-third ofthe papers that were presented, as well as a synthesis of 16 workshops.

An editorial committee guided the work that led to this book. It consisted of thefour co-editors: Patrick Bonnel, Martin Lee-Gosselin, Jean-Loup Madre, andJohanna Zmud, who also served as co-chairs of the Conference. These four, togetherwith the help of Jimmy Armoogum, divided up the editorial oversight. They built onthe considerable efforts of the many people and organizations, recognized in theAcknowledgements that follow. We are indebted to all those who donated their timeand energy to review, critique, and add to our body of knowledge about transportsurvey methods, in order to continuously improve the quality of transport surveysand enhance the value and utility of the data that such surveys provide for transportpolicy and decision-making.

Finally we, the co-chairs of the ISCTSC, thank all the Annecy authors for theirdiligence and hard work. We are condent that their continued diligence will lead tonew insights for, and new approaches in, transport survey methods.

Martin Lee-GosselinJohanna Zmud

ISCTSC Co-Chairs

xiv Preface

Acknowledgements

The series of international transport survey conferences and the published record isunder the responsibility of the ISCTSC. For the period leading up to the AnnecyConference through to the completion of this book, the members were:

International Steering Committee for Travel Survey Conferences (ISCTSC)

Tom Adler Resource Systems Group, Inc. U.S.A.

Carlos Arce NuStats, Inc. U.S.A.

Chandra Bhat2 University of Texas U.S.A.

Werner Brog Socialdata, GmbH Germany

Patrick Bonnel ENTPE France

Kelly Clifton2 University of Maryland U.S.A.

Peter Jones University College, London United Kingdom

Ryuichi Kitamura1 Kyoto University Japan

Stephan Krygsman2 Stewart Scott, Inc. South Africa

Martin Lee-Gosselin (Co-Chair) Universite Laval Canada

Jean-Loup Madre INRETS France

Arnim Meyburg Cornell University U.S.A.

Catherine Morency2 Polytechnique de Montreal Canada

Elaine Murakami1 FHWA U.S.A.

Juan de Dios Ortuzar Ponticia Universidad Catolica Chile

Alan Pisarski Consultant U.S.A.

Tony Richardson The Urban Transport Institute Australia

Gerd Sammer Universitat fur Bodenkultuur Austria

Cheryl Stecher1 IBM U.S.A.

Orlando Strambi Universidade de Sao Paulo Brazil

Peter Stopher University of Sydney Australia

Harry Timmermans Technical University of

Eindhoven

Netherlands

Klaas van Zyl1 Stewart Scott, Inc. South Africa

Chester Wilmot Louisiana State University U.S.A.

Johanna Zmud (Co-Chair) NuStats, Inc. U.S.A.

Dirk Zumkeller Universitat Karlsruhe Germany

1Member until October 2008.2Member from October 2008.

It was with great sadness that we learnt, in February 2009, of the death of RyuichiKitamura. Ryuichi was a source of endless energy and inspiration among theinternational community of those concerned with data on travel, communication andactivities. He saw well beyond transport networks, caring deeply about the data neededto help communities become more responsive to human values. He was one of a kind andwe miss him very much.

The success of the Annecy Conference was owed to nearly two years of painstakingpreparations by the Local Organising Committee (LOC) in France, co-chaired byPatrick Bonnel and Jean-Loup Madre, both of whom have also been longstandingmembers of the International Steering Committee (ISCTSC), and who took a veryactive role in the scientic programme as well. The LOC comprised:

Patrick Bonnel, Laboratoire dEconomie des Transports (LET-ENTPE) (Co-Chair)

Jean-Loup Madre, Institut National de la Recherche sur les Transports et leurSecurite (INRETS) (Co-Chair)

Jimmy Armoogum, Institut National de la Recherche sur les Transports et leurSecurite (INRETS)

Gerard Brun, Ministe`re franc-ais de lEcologie, de lEnergie du developpementdurable et de la Mer, direction de la recherche et de lanimation scientique ettechnique (DRAST)

Marc Christine, Institut National de la Statistique et des Etudes Economiques(INSEE)

Mary Crass, International Transport Forum (ECMT/OECD)Marie-Odile Gascon, Centre dEtudes sur les Reseaux, les Transports, lUrba-

nisme et les constructions (CERTU)Emmanuel Raoul, Ministe`re franc-ais de lEcologie, de lEnergie du developpe-

ment durable et de la Mer

In addition, our sincere thanks to Christophe Rizet of INRETS for all thearrangements needed to make possible holding the nal meeting of COSTProgramme 355 in collaboration with the ISCTSC meeting.

A meeting of the size of the Annecy Conference also depended on the support ofmany staff. We are particularly indebted to Florence Toilier of ENTPE for herpatient resolution of many logistical questions and the production of the conferenceCDROM, Peter Endemann of Planungsverband Ballungsraum Frankfurt/Rhein-Main and the ETC for so effectively running the Conference secretariat at Annecy (inve languages), to Pierre-Olivier Flavigny and Philippe Marchal of INRETS fororganising the most hospitable Internet facilities that any of us have seen heretoforeat a conference, and to the staff of Les Balcons du Lac dAnnecy.

The ISCTSC would also like to recognise that the Annecy Conference would nothave been possible without the generosity of our sponsors, who provided nancial,material and/or staff support. In addition, the sponsors made it possible to continueISCTSCs commitment to invest in our future by providing scholarships to

xvi Acknowledgements

participants from lower-income countries and some support to new researchers andstudents. They were:

La Direction de la recherche et de lanimation scientique et technique (DRAST),Ministe`re franc-ais de lEcologie, de lEnergie du developpement durable et de laMer

Ecole Nationale des Travaux Publics de lEtat (ENTPE), France Institut National de la Recherche sur les Transports et leur Securite (INRETS),France

Laboratoire dEconomie des Transports (LET), France NuStats, United States PTV AG, Germany Service de lObservation et des Statistiques (SOeS), Ministe`re franc-ais delEcologie, de lEnergie du developpement durable et de la Mer

The Urban Transport Institute (TUTI), Australia

In addition to these sponsors, we are most grateful for the many individuals in theworld community of transport survey researchers and practitioners who donatedtheir time and expertise as authors, reviewers, chairs, discussants and rapporteurs.More than 100 extended abstracts and papers were reviewed, and in the case ofpapers selected for this book, by three reviewers.

The preparation of the book was greatly aided by the counsels of Zoe Sanders andClaire Ferres of Emerald Group Publishing Limited, and the work of ISCTSCscopy-editor Christopher Parker. We thank them for their attention to detail.

Acknowledgements xvii

PART I

INTRODUCTION

Chapter 1

Keeping Up with a Changing World: Challengesin the Design of Transport Survey Methods

Patrick Bonnel, Martin Lee-Gosselin, Jean-Loup Madre andJohanna Zmud

Abstract

At the 2008 International Conference on Transport Survey Methods inAnnecy, France, transport survey methodologists and practitioners sharedtheir experience with keeping abreast of the data needs of a rapidly changingworld. Over the past decade, this has translated into the need for: an expandedtravel survey toolkit; methodological innovation for surveys of freight andpublic transport operations; a growing use of data collection and processingtechnologies; a need to align surveys with other data streams; and an increasedinterest in the comparability of international datasets on personal traveland commodity movements in an era of globalisation. We discuss how theseguided the choice and scope of the ve themes around which both the AnnecyConference and this book were organised.

The International Steering Committee for Travel Survey Conferences (ISCTSC)organises periodic international conferences on the survey methods that supportplanning, policy development, modelling and evaluation through the observationof person, vehicle and commodity movements at the urban, rural, regional, intercityand international scales. The evolution of the underlying issues, and the methodo-logical response, can be seen in the series of publications that drew on previousconferences, most recently the 1997 Grainau Conference (Stopher & Jones, 2000),

Transport Survey Methods

Copyright r 2009 by Emerald Group Publishing Limited

All rights of reproduction in any form reserved

ISBN: 978-1-84855-844-1

the 2001 Kruger Conference (Stopher & Jones, 2003) and the 2004 Costa RicaConference (Stopher & Stecher, 2006). That evolution includes:

a gradual expansion of the travel survey toolkit beyond the needs of core businessurban area household travel surveys and national travel surveys;

a growing recognition of the need for new approaches to collecting data on freightmovements and public transport operations;

an increasing expectation that data from transport user surveys should be alignedwith other data streams from administrative and commercial sources;

a growing application of digital technologies to aid data collection and processing; an increased attention to international (and within nation) comparisons of dataon personal travel and commodity movements, and to international ows in thecontext of a progressive globalisation of national economies.

With this in mind, the Transport Survey Methods Conference held in Annecy,France, in May 2008 was designed to continue the emphasis of previous meetings ontransport survey quality and on the standards for assessing and maintaining quality(e.g. Stopher, Wilmot, Stecher, & Alsnih 2006), and also to look ahead to transportsurvey data harmonisation and comparability within and across countries. It was aconcerted response to the evolving need to track and compare key policy measuresand statistics, and their implications for sustaining mobility, in todays global,interconnected world. For example, how can we track and compare long distancemobility in Belgium versus the United States when the two nations employ differentdenitions of a long distance trip? (Bonnel, Madre, & Armoogum, 2005) What doesit mean that the mobility rate is around three trips per day in a Netherlandsmetropolitan area and around four trips per day in the Grenoble metropolitan area?Can we attribute this difference to policy measures in respective areas, true travelbehaviour differences, different survey methodologies or different spatial boundarydenitions? (Bonnel, 2003).

As a community, transport survey researchers, practitioners and planners needadvance knowledge of the components of survey and data collection design that areupstream from reliable and accurate intra-national and international comparisons.The topics of the Annecy Conference were intended to facilitate such discussions, andin a few cases to initiate them. It was also hoped that progress would be madetowards the development of a framework for harmonising passenger transportsurvey data and statistics along the lines of what has been done for road freight dataat European level (Pasi, 2008). At the same time, it was recognised that some classesof transport surveys are not yet ready for such a framework, and may require somefairly fundamental methodological research in the shorter term. Other classes forexample those that explore hypothetical travel behaviour under a range of possiblefuture environmental pressures are not intended to generate national statistics, butmerit our attention for other sound reasons.

The Annecy Conference thus sought a balance between the themes of dataharmonisation and the data quality. With its workshop format, the conferencecontinued the ISCTSC tradition to create an opportunity for networking, collaboration

4 Patrick Bonnel et al.

and sharing of knowledge. At the same time, this particular conference was unique inthat it was held in conjunction with the nal meeting of the European COST Action355: Changing Behaviour Towards a More Sustainable Transportation System, and areview of the activities and recommendations of this COST action pertaining tonational surveys is presented in this volume. In Annecy, experts presented papers thatprovided insight into the many-faceted conceptual dimensions of 16 workshop topics,as well as the practical issues of the impact of data needs, socio-political issues andfuture policy issues affecting transport surveys. These papers and syntheses of theassociated workshops comprise the majority of the chapters of this book.

Conference participants included leading survey researchers and transport profes-sionals from over 25 countries in the private and public sectors, as well as academicscholars and researchers. Three keynote addresses were presented. Simo Pasi provideda snapshot of Eurostatss priorities and activities related to data harmonisation in thefreight area. William Dutton, Oxford Internet Institute, provided insight into newresearch in the areas of communication and information technologies and itsimplications for data collection in the transport domain. Peter Stopher, Universityof Sydney, identied his proposal for an innovative, multi-approach transport surveytoolkit, and his keynote presentation is included as Chapter 2 of this volume.

The book is organised into ve themes that represent the key challenges andopportunity spaces in transport surveys:

I. Sustainability and Traveller Adaptation, Chapters 36, providing currentthinking on survey protocols related to travel behaviour change;

II. Global Social Issues, Chapters 714, describing survey practice on current socialchallenges tourist ows, evacuations, physical activity and excluded populations;

III. Freight and Transit Planning, Chapters 1520, addressing the perennialtransportation system challenges of capturing data on freight movements andpublic transit ridership;

IV. Technology Applications, Chapters 2128, identifying the new challenges andnew opportunities for applying technology solutions to improve survey dataquality;

V. Emerging/Persistent Survey Issues and Data Harmonisation, Chapters 2936,describing best practices for overcoming persistent problems with transport surveydata.

The following sections provide more information about each of these ve thematicareas by highlighting the pertinent transport survey methods questions, issues,challenges and recommendations that were presented in the resource papers for eachworkshop, or which came out of the workshop discussions themselves.

1.1. Sustainability and Traveller Adaptation

Sustainability is on the political agenda of every nation. It has spawned numerouspolicy measures in order to inuence travel behaviour towards more environmentally

Challenges in the Design of Transport Survey Methods 5

friendly modes of transport. One such policy, voluntary behavioural changeprogrammes, has generated great interest. These programmes include a diverse rangeof initiatives that vary in both who is targeted and the nature of the intervention.While transport survey researchers have long been interested in travellers responsesto external factors such as the price of fuel or provision of new infrastructure, theyare on less familiar ground when their traditional dependent variables the amountof travel and its attributes are the direct target of an intervention.

Various methods are used to monitor and evaluate these programmes, but all facesome common difculties in applying best practice. For instance, measurementunits impact the analysis that can be performed on data. Is it more important tomeasure the change in behaviour or to measure the impact of the change (e.g. interms of emissions)? Do we need to analyse the factors that inuence change in orderto monitor the process of change and improve programme efciency? What is thebest timing for measurement: shortly after programme introduction in order tomeasure the direct relation between programme and behavioural change or after alonger time in order to observe snowball effect and other changes that take longertime to observe (residence location, for example) but that can have strong effects?How do we measure behavioural change in order to minimise potential bias?

Sustainable initiatives are often part of a more global programme and areintroduced in an evolving environment. Thus, their evaluation faces the usualchallenges of impact assessments: trying to disentangle outcomes, and determiningwhat is related to the targeted initiatives versus what is related to other changes (e.g.changes in life cycle). The need to adequately and consistently dene experimentaland control groups to circumvent potential difculties and alleviate biases wasdiscussed. Difculty in doing so, however, is increased by the fact that sustainabilityinitiatives might be developed on limited scale and with limited budget. It is thereforenot always possible to achieve statistically signicant results. Meta studies could bedeveloped to overcome this problem but they may only accentuate the biases of eachdiscrete study and may not always be well documented. It is therefore recommendedto triangulate, using various survey methods or gathering data from multiple sourcesin order to apply triangulation, to deal with or to complement conicting results,when possible. Transport survey researchers would also benet from experiencesfrom other disciplines like psychology, health/epidemiology and sociology to gleanbest practices for effective survey science experiments and programme evaluations.

As noted above, behavioural change does, of course, occur in response to manysituations other than programmes that actively seek it. As well as costs and newinfrastructure, these can include rising congestion, energy shortages, new vehicle,fuel or communications technologies, and a wide range of policy shifts to addresspublic health, safety or environmental problems. The Annecy discussions focussedin particular on surveys of changing behaviour in the context of an interest inenvironmental sustainability policy. Importantly, they included stated responsesurveys to explore potential behaviour under hypothetical situations, and thechallenge of obtaining authentic responses about past and future behaviour alike.Social desirability bias is a term used to describe the tendency of respondents toreply in a manner that will be viewed favourably by others, often resulting in


respondents over-reporting of good behaviour or under-reporting of badbehaviour. The potential for social desirability effects in the case of environmentallyfriendly behaviours, which are generally considered as desirable, is great. Once again,the triangulation of more than one method may be needed to interpret adaptivebehaviour.

1.2. Global Social Issues

No discipline can defer addressing contemporary social, political, economic andenvironmental issues within the course of its scientic discourse. Transport surveyresearch is no different. Surveys are executed in a socio-politico-economic context.Who actually participated in a survey is often more important than what the surveyfound. Engendering the participation of some population segments is more difcultthan others (i.e. illiterate, immigrant, low or high income persons). Often such non-participation is systematic rather than episodic. Hence, a generalised term hard-to-reach population has entered into the transport survey literature. Consequences ofthis lack of coverage are discussed in this section of the book, especially withregard to the social dimension of sustainable development. Strategies to overcomepotential biases are offered and include multi-mode or multi-method surveys, as bothunder-representation and its impact on data are different depending on populationsegments concerned.

Public health and tness are growing 21st century social concerns in bothdeveloped and developing countries. Within the transport survey world, physicalactivity in the built environment is receiving a lot of attention in relation to healthconcerns, especially obesity. Topics addressed include the potential of travel surveysto inform this social issue in general, as well as methodological issues that affectthe utility of surveys to answer questions specic to transport policy and planning.This is an issue that generates opportunities and challenges for interdisciplinarycollaboration. Particular attention must be given to the effective capture of short andwalking or cycling trips, which are sometimes neglected in travel surveys. We alsoneed to better understand the motivation underlying physical activities within travel,and relate this to the transport system and the built environment.

How to more effectively conduct surveys focussed on transport during exceptionalevents continues to be of great interest. A major dimension of this is emergencymanagement and evacuations. But this is not just about getting people out of harmsway. There is great variability in the responsiveness of transportation systems, thehazard events, and the key players or detectors of those events. Unfortunately, thereare not many, if any, standards to systematically address these factors in a surveycontext. As a starting point, this section of the book addresses this situation byidentifying the variety of exceptional events and proposing a typology related tospecic dimensions of an event. Data needs cover both travel behaviour data and itsdeterminants, and also the context of the event. Preference surveys, especially statedchoice surveys, are considered important.


Economic sustainability has often been associated with tourism infrastructure,and so effective methods to survey tourists and transients are quite important.Content related to this topic attempts to clarify the use of survey terminologyconcerning survey accuracy and reliability in tourism research. Denition of touristsand transients is still problematic and depend often on the spatial scale used. Theneed for standards to assess accuracy is introduced. A unique survey solutionappears unrealistic. Therefore data need to be collected from various methods andfrom various sources other than transport-related sources. Research is thereforenecessary to better integrate these multiple sources of data.

The workshops in this global social issues area appear quite different but somecommon ground can be emphasised. Interdisciplinary surveys appear promising, andthe need to cross traditional boundaries between academic disciplines or schoolsof thought surfaced in a number of discussions. The need was felt to investigate newparadigms, new approaches and new conceptual frameworks to better understandthe determinants of behaviour or the determinants of survey response. The study ofglobal social issues has also emphasised the need for gathering data from varioussources (within and outside of the transportation domain), which in turn increasesthe need to develop methods to integrate and enrich (or merge) these data. In allcases, the potential of technology (i.e. GPS tracking, sensors, RFID) to improve dataquality is raised, even if more research is needed to better identify the benets of andlimitations in the information collected using these technologies. Moreover, there arepotential roles for several of the survey methods discussed under this theme tocontribute to localised alerts in real time, such as to health authorities or incidentresponse agencies. It follows that they would also contribute to collective andcumulative assessments of hazardous conditions.

1.3. Freight and Transit Planning

The interaction between freight ows and transportation infrastructure continues toincrease in both importance and complexity. Signicant changes in the economy haveimpacts on the level, type and geographic distribution of freight ows. The variety ofuses for freight transportation data is broad and increasing. Estimating the likelyimpacts of proposed regulations on air quality, greenhouse gas emissions and climatedepends greatly on the availability of improved freight ow data. Yet, our traditionalsources of freight transportation data struggle to meet new requirements. There area number of challenges in freight surveys, not the least of which is denitional exactly what is freight? What do end users want to measure, model, forecast, etc.,with freight survey data? Are we interested in surveying what is being moved, how itis moved, or both? What about why it is being moved? What information domodellers need to develop the tools for measuring and predicting freight? These arejust some of the questions that need to be answered to allow survey designers todevelop and implement surveys that provide not only meaningful information, butalso the data for policy makers and planners to answer their questions.


Advancing change and innovation in these types of surveys requires addressingdenitional and measurement issues, institutional constraints, private-sectorconcerns, nancial constraints and multi-agency coordination across modes. Inter-modality issues are of particular importance: for example, more and more often, apackaged good is not moved by a single vehicle. Thus, the alternative is betweenshipment surveys (e.g. CFS in the United States, ECHO in France) versus vehicle-based surveys (as mentioned in the keynote presentation by Pasi on truck surveysharmonised by EUROSTAT). There are a number of links to other discussions.In general, it would be fruitful to compare the way freight and travel surveymethodologists approach the above fundamental questions. Also, there are newdevelopments in vehicle-based surveys, discussed under the Emerging/PersistentIssues theme, that are relevant here, even though the sampling frame here is mostlyheavy-duty vehicles rather than vehicles in general.

Public transit, and transit-supportive land use, is becoming increasingly importantto climate change policy as national, regional and local governments developand implement initiatives to reduce energy use and greenhouse gas emissions. Dataneeds for public transport planning cover a wide range of domains (e.g. ridership,performance, service quality, customer satisfaction, non-users expectations). Avariety of methods are therefore needed to investigate these domains. Unfortunately,scientic literature about data collection and analysis methods specically aimed atpublic transit is not plentiful. Key questions remain, such as where has innovative orbest practice been employed? What are the best methods, innovations, hardware andsoftware tools? What administrative data systems have been developed to full thespecic requirements of a single study? There has not been a systematic or empiricalassessment of methodological improvements in areas such as these. Improvementsare needed in areas that include, but are not limited to: questionnaire wording andgraphic design to improve comprehension and response quality; selection, trainingand continuous monitoring of surveyors to assure high performance standards; real-time geocoding and verication to improve accuracy and precision; and a variety ofalternative methods for data capture. Of particular interest are new methods forweighting and expanding survey data that account for biases such as short versuslong surveyed trips, multi-transfer journeys and multi-modal journeys. The industryis hopeful that various new technologies (i.e. ticketing systems, automatic vehiclepositioning, automatic passenger counting, GPS tracking) appear very promising toautomatically generate data. But this benet is currently offset by the challenge facedin processing and organising the huge quantities of data that result.

1.4. Technology Applications

As noted above, technological aids to surveys is a growth area for the transportsurvey community. There were four workshops on technology topics at the AnnecyConference (see Chapters 2128 of this book), and technology was actively discussed


in several more. This compares to one small session at the 1997 Conference, oneworkshop at the 2001 Conference and two workshops at the 2004 Conference.

In the case of travel surveys, researchers are facing increasing difculties,including response rate decreases, sampling frame limitations and privacy concernsat a time when data requirements are getting more complex. New technologies areoften seen as a solution to solve or at least mitigate these difculties. There is noargument that the potential benets are high. But challenges exist. For example,good user interfaces to help reduce respondent burden, and therefore to increaserespondent participation and data reliability, are still not the norm. Research isneeded to fully document the strengths and weaknesses of specic technologyapplications. GPS/GSM helps to produce more accurate data (i.e. fewer tripomissions, more accurate locations and timing, collection of individualised scheduleand planning elements) for longer travel periods, but at what cost? Web-basedsurveys in combination with other modes facilitate response rate increases amongcertain populations at lower costs, but with what biases? Visualisation techniques areespecially promising for generating a better understanding of travel behaviour butrequire more examination of uses and consequences. Automated data productionsystems (i.e. fare collection systems, passenger counting systems, mobile phonelocation systems, credit card transactions, plate number recognition systems) collecthuge quantities of data that may have great potential for transport researchcommunity. The potential of each technology needs further research to betteridentify the expected benets, especially in terms of data quality and level ofprecision, with implications for the return on investment.

But use of new technologies comes with new problems and new challenges,especially the techniques for processing and enriching the data. New technologies areoften seen as a way to reduce non-response in surveys. Benets are obvious regardingthe more complete reporting of trips, or of some of their characteristics. But someindividuals are also reluctant to use the technologies, or may not yet have accessto them, generating other non-response biases. We need to explore the potential ofeach technology and each mode to recruit and to motivate respondents, in order todevelop strategies to overcome non-response. In this respect, new technologies arelike any other methodological innovation with risks and benets to be empiricallyassessed.

Electronic instrument design offers the promise to simplify and customisequestionnaires, and therefore to reduce respondent burden and increase participa-tion. But research is still needed to develop interfaces that are congruent with themental models of how respondents think about the questions embedded in surveys.

New technologies generate huge quantities of data that generally need an a prioriprocessing plan in order to be useful for data analysis and planning. This is typicallythe case for GPS/GSM or cell phone data. Furthermore, such primary data aregenerally incomplete as survey responses (e.g. only space, time and their derivativesfor GPS data). We therefore need to further develop tools to enrich these data,generating mode and purpose from GPS data for example. Research is stillnecessary to develop more efcient data processing, and where possible, proposesome standardisation principles and software tools. A promising avenue seems to be


the use of learning algorithms to build essentially passive interpretation of datastreams following a period of active involvement of respondents (or at least a sub-sample) in the interpretation or their own data. While improvements have beenachieved in the technical performance of survey technology, especially the powermanagement and storage capacities of mobile technologies, or the precision oflocation and timing, more testing and experimentation is needed. Privacy concernsand ethical rules are expected to assume increasing importance as applications moveto larger and larger samples. When considered within the larger context of the issuesraised at the Annecy Conference and in this publication, it is clear that technologyapplication will not be a silver bullet. None of the available technologies appears ableby itself to cover adequately the growing list of transport data needs. Future work ondeveloping innovative combinations of technologies, survey modes and data fusionmethods needs to be conducted, documented and disseminated.

1.5. Emerging/Persistent Survey Issues and Data Harmonisation

Transportation systems are large, complex, multi-component, multi-player collec-tions of interacting elements and subsystems. These systems play critical roles insocieties and economies, providing accessibility (and thus value) to places andmobility to people and goods. Decisions about development and operation of thetransportation system are of central importance to leadership at all levels, in both thegovernment and the private sectors. Data and the information produced from dataare key assets of transportation systems because of the roles they play in support ofdecision-making: problem identication, design of options and priority setting. Thesedata include not only the measures of travel behaviour information but also itsdeterminants. Yet one of the key sources of these data, transport surveys, tends to beperformed at longer intervals due to nancial and other institutional constraints.It becomes harder for decision makers, policy makers and planners to have at theirdisposal up-to-date data and information about the dynamics of travel behaviourthat are the necessary underpinnings of good transport policy and planning.

Continuous surveys appear to be a promising solution to address the concomitantproblems of both producing up-to-date data and the resulting trend data series.Continuous surveys are a category of survey design that encompass longitudinaland panel methods. With the interest in sustainability and behaviour changeprogrammes, continuous surveys present many advantages, not the least of which isthe enabling of a more accurate calculation of time-series for auto-correlatedvariables (e.g. mobility or car ownership). The growing interest in these types ofsurveys is illustrated by specic experiences gained from various case studyapplications around the world. Among the persistent challenges identied, twoappear crucial needing far more research. First, fatigue of staff and interviewershas lead in some countries to declining quality, notably as measured throughresponse rates: permanent monitoring and incentive programmes are necessary tosustain motivation and data quality. Second, and perhaps more important, methods


need to be developed to account for the dynamics present in data collected overseveral years, and resulting from exceptional or unpredictable external factors.

As discussed above, the role of surveys that use vehicles as the sampling frame is notconned to freight and transit. In Canada, for example, vehicle-based surveys of light-duty vehicles have provided valuable national data on vehicle and passenger kilometresof travel in the absence of a national travel survey. At Annecy, there was a wide-ranging discussion of the state of the art of vehicle-based surveys in France, Germany,Finland, the UK, Bangladesh, the USA and Canada. The promise of vehicle-basedsurveys was assessed, in particular, for the advantages of the parallel collection of dataon duty cycles, fuel use, emissions and safety systems, using increasingly sophisticatedtechnological aids that work well on vehicle platforms. The importance of suchmicrodata in a future that may include the more widespread use of automatic systemsfor charging individual users did not escape the discussions of this workshop.

Most of the workshops have identied the need to combine methodologiesbecause single methods or surveys generally failed to sufciently cover the problemthat motivated the surveys. We also have more and more data that are automaticallyproduced, especially from new technology systems, and which need to be processedand enriched with other sources. Combining survey modes or methods within surveyappears to be a good strategy, both to reduce the non-response problem and toimprove the representativeness of the population, especially for hard-to-reachgroups. This problem is not new, and several methods are available in a number ofsurvey elds besides transport. In addition, progress is being made on the alignmentor fusion of data from surveys of users and other sources, notably those from trafcand transit monitoring, administrative data and even nancial transactions.Nevertheless, there is a need to better assess the potential of each method on realcase studies. There is a particular need to identify methodology to validate fuseddata, and to assess the implications for data analysis and modelling, especiallyregarding level of condence.

The holding of the nal COST 355 meeting in conjunction with the AnnecyConference gave participants from around the world the opportunity to appreciatethe breadth of European cooperation on the role of behaviour in a more sustainabletransport system. While the rst two of the COST 355 themes, Freight transport andenergy consumption and Automobile: panel data analysis were of great interestto the participants, the third Overview of national transport surveys was of directrelevance to the conference theme, the more so because of its focus on thecomparability of survey data. Our nal chapter provides a tting conclusion to thisbook as it summarises the lessons from more than three years of analysis of theEuropean experience.

1.6. Conclusion

A number of ideas and proposals emerged during the conference, and participantsreached some conclusions about how to improve the quality of transport survey data.


The emphasis in most of the workshop discussions was on developing practical,achievable and affordable strategies for collection of essential travel information thatwould be less contingent upon shifting political and funding priorities. While suchpoints are addressed in more detail in the chapters, some of the priorities that cutacross the different workshop topics are presented below:

continuous longitudinal surveys to improve the frequency, timeliness andresponsiveness of personal or freight data to policy and planning issues;

creation of national mobility panels to obtain better measures of changes in traveldemand relative to economic, social and environmental factors;

consideration of digital methods of data capture and organisation to enhancequality and cost-effectiveness of data capture;

automated vehicle movement data collection (through ITS, probe vehicles, cellphone probes or other telematic techniques);

supplementation or replacement of primary micro-level data collection withsimulation and modelling approaches to improve the accessibility of informationfor decision making;

greater application of computer-assisted techniques for visualising data tofacilitate communication and decision making.

It was not difcult to draw an overarching conclusion from the AnnecyConference that international collaboration is fruitful, but that a strategy for thatcollaboration would be even better. Harmonisation of data across our countrieswould be of great value, but so would be a fuller appreciation of the variety ofexperimentation. The ISCTSC Conference Series is committed to continue thesharing of our experiences: of serving better those who pay for surveys; of workingwith the tools and technologies that can help surmount the new challenges to our art;and of helping identify the opportunities to make the transport system more sociallyand environmentally responsible.

References

Bonnel, P. (2003). Postal, telephone, and face-to-face surveys: How comparable are they? In:

P. R. Stopher & P. M. Jones (Eds), Transport survey quality and innovation (pp. 215237).

Oxford: Pergamon Press.

Bonnel, P., Madre, J. -L., & Armoogum, J. (2005). National transport surveys: What can we

learn from international comparisons? European Transport Conference, Strasbourg.

Pasi, S. (2008). Harmonisation and comparability of European Union road freight statistics,

Eurostat G-5-Transport Statistics, Luxembourg, available at: http://www.isctsc.let.fr/

2008Conf/keynote.html (consulted May 2009).

Stopher, P. R., & Jones, P. M. (Eds). (2000). Transport surveys: Raising the standard.

Washington, DC: Transportation Research Circular E-C008, Transportation Research

Board.


Stopher, P. R., & Jones, P. M. (Eds). (2003). Transport survey quality and innovation (p. 646).

Oxford: Pergamon Press.

Stopher, P. R., & Stecher, CC. (Eds). (2006). Transport survey methods: Quality and future

directions (p. 685). Oxford: Elsevier.

Stopher, P. R., Wilmot, C. G., Stecher, C. C., & Alsnih, R. (2006). Household travel surveys:

Proposed standards and guidelines. In: P. R. Stopher & C. C. Stecher (Eds), Transport

survey methods: Quality and future directions (pp. 1974). Oxford: Elsevier.


Chapter 2

The Travel Survey Toolkit: Where toFrom Here?$

Peter R. Stopher

Abstract

The purpose of this chapter is to provoke thinking about the directions inwhich the travel survey toolkit should move in the near future based on theauthors personal experience and as an outcome of the Travel Survey Methodsconference. The chapter begins with a brief historical review that attempts toshow some of the major elements of change that have occurred in travel surveymethods over the past 4050 years. A more detailed review is provided aboutdevelopments over the past 1015 years. The chapter then explores a number ofemerging challenges, including telephone contact of potential respondents,computer-assisted surveys, Internet surveys, mixed-mode surveys, the impactsof language and literacy and the potentials of mobile technologies. Based onthis, the chapter then considers future directions that should be pursued. Thechapter suggests that it has been changes in survey methodology that have,in the past, sometimes enabled and at other times led to changes in modellingparadigms, and that this may be an appropriate time for travel surveymethodology again to enable changes in modelling paradigms. A speculativespecication of a new household travel survey that makes use of a number ofthese developments is then offered. The chapter ends with some concludingremarks that issue a challenge to the travel survey community to think outsidethe box and foster change and improvement in the accuracy and representa-tiveness of travel surveys.

$Keynote Paper prepared for the 8th International Conference on Transport Survey Methods, Annecy,

France, May 2008.

Transport Survey Methods

Copyright r 2009 by Emerald Group Publishing Limited

All rights of reproduction in any form reserved

ISBN: 978-1-84855-844-1

2.1. Introduction

To ask where we are going from here requires, rst, that we know where here is. Tothat end, at the outset of this paper, it seemed worthwhile to reect on where we havecome from in the area of travel survey methods, before setting out to try to see wherewe are going in the early 21st century. Following that, this paper discusses some ofthe emerging challenges and also opportunities that appear to be available to thoseintent upon undertaking travel surveys in the early 21st century. For the most part,this paper focuses on travel surveys of human populations and the main focus withintravel surveys of human populations is on the household travel survey. However,some mention is made in passing of other travel survey needs, especially in relation toon-board transit surveys and roadside or intercept surveys.

The purpose of this paper is not to be exhaustive in the treatment of the travelsurvey toolkit, but rather to open up some ideas for consideration during the courseof the 8th International Conference on Travel Survey Methods in Annecy, so thatdelegates to the conference may be inspired to think outside the box in lookingat where we are going and where we need to go in the future. There is little doubt thatthere are many threats to conducting travel surveys at this time, and that some ofthese threats are likely to become more serious in the future. There are also manynew opportunities to do things better than they have been done in the past, althoughthere are potential pitfalls that could lead towards negative impacts on travel surveyquality. It is hoped that this paper will stimulate thinking and also help to ensure thatthe paths followed will bring about improvements in data quality.

2.2. A Historical Review

Although it was before probably any of those currently active in travel surveymethods were actually working in the eld, the beginnings of travel survey methodswere rmly founded in the early stages of urban transport planning, especially as itwas practiced in the United States in the 1950s and 1960s. At that time, urbantransport planning was primarily a concern of the largest urban areas in the country,and also of major cities in a few other countries around the world, where USconsultants were largely employed to bring this new eld of transport planning tothose cities, for example, the London Trafc Study of the early 1960s (LondonCounty Council, 1964). At the core of the emerging urban transport planning eldwas the home interview survey, with its associated cordon and screenline surveys,often involving roadside interview surveys. The principal method used to obtain dataabout the travel movements of people was, therefore, an interview.

Roadside interviews were conducted by having the help of local police to agdown a systematic sample of cars, which were required to drive into a marked-offbay, where a short interview, usually lasting no more than about 2min, took place.The interview determined vehicle occupancy, origin location, destination locationand the purpose of the trip. The frequency with which the trip was made was also

16 Peter R. Stopher

often asked. The roadside interview also involved counts being made of the vehiclespassing the interview location, so that the total volume on the roadway could bedetermined. Samples were usually of the order of one vehicle in ten. The roadsideinterview is still alive and well, judging by a brief search on the Internet (e.g.Cicerone, Sassaman, & Swinney, 2007; Northamptonshire County Council, 2003;Scottish Executive, 2004), although there are numerous cautionary comments aboutits use on motorways and freeways (where it should not be used), on major arterialroads and in major cities. In the United States, the roadside interview is rarely donefor larger urban areas, because of concerns with road rage and adverse publicityresulting from the inevitable delays to drivers when such an interview is conducted.

On-board transit surveys were something of an afterthought in the survey toolkitfor transport planning. This resulted from the strong focus on highway-basedsolutions that emerged from the US-led transport planning eld, which only slowlyacknowledged an ongoing role for public transport in the urban context, partly as aresult of growing interest in transport planning in Europe and partly as a result ofa growing recognition in the United States that highways could not solve all of theurban transport issues. Again, in the wake of the home interview and roadsideinterview, early on-board transit surveys focused on an interview survey either onboard the public transport vehicle, or occasionally at the stops and stations wherepeople waited to board. Subsequently, and by the late 1970s, the on-board interviewbecame a self-administered survey, often to be accomplished whilst passengers rodeon the public transport vehicle (Stopher, 1985). Household travel surveys andnational surveys often result in apparent undercounts of public transport ridership(Transmanagement, 2004), but on-board transit surveys are also subject to very lowresponse rates in many instances. Conducting an on-board transit survey involvesfacing a number of challenges. These include creating a survey form that can be lledout whilst the vehicle is moving and by passengers who may have nothing on whichto rest the form; keeping the survey short enough that most passengers can completethe form during their transit trip; providing forms in more than one language wherenecessary and the logistics of handing out forms and retrieving them, especially onovercrowded vehicles. On-board surveys, however, continue to be an important partof the travel survey toolkit.

The home interview (BPR, 1954; CATS, 1956; FHWA, 1973) was conductedby sending interviewers out to households, where, unannounced, they knocked ondoors of selected sample households and then proceeded to ask a battery of questionsabout the household, its members and recollections of travel that took place on theprevious day. The focus of these surveys was weekday travel, so interviewing wasusually conducted throughout the week, on Tuesday through Friday evenings and onSaturdays, so that the data on the previous day was collected for Monday throughFriday. In some surveys, interviewing took place only on weekday nights, and theMonday evening surveys asked people to recall their travel on the previous Friday.The sample sizes for these home interview surveys were generally dened as apercentage of the metropolitan regions population, and often ranged from 1% to3% of the population. Thus, it was not uncommon for the sample to comprise asmany as 20,000 or 30,000 households in large metropolitan areas. The Chicago Area

The Travel Survey Toolkit 17

Transportation Study Home Interview Manual (CATS, 1956) specied the design ofa survey to be conducted on a sample of 56,000 households. It was also usuallyclaimed that the response rate was upwards of 90%. However, it is unlikely that thiswas calculated using the CASRO (1982) or AAPOR (2000) formulae, but was mostprobably simply estimated as the proportion of households that were successfullyapproached and had completed the interview.

By the end of the 1970s in the United States, home interview surveys were beingreplaced by other types of surveys. It was becoming evident in the late 1970s that thehome interview survey could not be carried out safely in certain parts of most urbanareas, and that alternative methods of surveying were required. One of the last largehome interview surveys to be conducted in the United States was probably the oneundertaken in Detroit in 1981 (Parvataneni, Brown, & Stopher, 1982). In the nextfew years, the home interview survey began to be replaced by telephone recruitment,using random-digit dialling (RDD); this was followed by a mail-out, mail-backsurvey. However, response rates to the mail-back portion of the survey were oftendisappointingly low, so that the mail-back option began to be replaced by atelephone retrieval of the diary data. The telephone retrieval method largely replacedmail-back options by the early 1990s. A few regions continued to use either a mailsurvey or the telephone recruitment with mail-out and mail-back, but the homeinterview in the United States had been almost totally replaced by a telephonerecruitment, mail-out, telephone retrieval method of surveying by the mid-1990s.

At the same time that the standard methodology of conducting travel surveys inthe United States changed, the sample sizes also dropped and were no longer referredto in terms of the percentages of the population. Rather, sample sizes were moregenerally described in terms of a range from around two or three thousand upto as much as ten or fteen thousand in a few cases. Samples of the size used in theearly days of the home interview survey were no longer to be seen. During the 1970s,even before the emergence of disaggregate models and their incorporation intothe standard transport planning package, there began a process of moving awayfrom the large samples of the early home interview surveys (Foster, Benson, &Stover, 1976). The increasing costs of such surveys, coupled with improved under-standing of sampling statistics, produced pressure on the survey work to reduce thesample sizes.

The methodology for household travel surveys has undergone somewhat lesschange outside North America. The home interview survey remains the methodologyof the Sydney Continuous Household Travel Survey, for example (Battelino &Peachman, 2003). In Brazil and Chile (Ortuzar, 2006), face-to-face interviewing isalso still being used. In other cities of Australia, one of the methods currently in use isa face-to-face recruitment interview with survey instruments left with the householdand either picked up in person by the interviewer subsequently or sent back by mail.In New Zealand, the method is quite similar, except that on the return of theinterviewer, a computer-assisted personal interview (CAPI) is done, and the materialsleft with the household consist only of a memory jogger that covers the designateddays of the survey. In Canada, most household travel surveys appear either to followthe U.S. procedure or to be pure telephone surveys. However, the sample sizes are

18 Peter R. Stopher

very large comparatively, with the recent Toronto survey covering more than 100,000households. In the United Kingdom, the face-to-face interview is still the preferredmethod, using CAPI, and following a pre-notication letter and diary provided toeach household. Different methods are recommended for smaller area surveys,namely, a mail survey with CATI retrieval for smaller urban areas and regions. InFrance, a face-to-face survey is compulsory (to obtain the CERTU label and statenancing) in large conurbations, but a telephone survey (CATI) is recommended formedium-sized conurbations surveying only one person in small households and twopersons in larger ones (CERTU, 2008). In the Netherlands, surveys arepredominantly mail surveys, with telephone contact to remind households tocomplete and return the surveys, and also to correct problematic or missing data (vanEvert, Brog, & Erl, 2006). This survey has evolved from a predominantly face-to-facesurvey up to 1983, and a telephone survey from 1983 to 1998. In Germany,recruitment by telephone using RDD is the method used in the current Germanmobility panel (Kuhnimhof, Chlond, & Zumkeller, 2006) and the German long-distance panel (Chlond, Last, Manz, & Zumkeller, 2006), with some data collected inthe recruitment stage by CATI, followed by a mail survey for the transportinformation, whilst the last National Travel Survey in Germany sampled from anaddress register, but conducted the survey by a mix of CATI and self-administeredpostal survey. In South Africa, surveys are still largely conducted as face-to-faceinterviews at the household. These surveys are generally done using paper and pencil,although some are conducted using CAPI.

Irrespective of the method used, household travel surveys, in common with manyother types of surveys, have been experiencing declining response rates over the pasttwo or three decades. Some of the decline is clearly associated with changingmethods, but there is a growing reluctance from the public to get involved in surveysof any type, and reassurance that the survey is not a marketing device has onlymodest impact on response rate declines. The Sydney Continuous Household TravelSurvey provides a good example of how even the face-to-face interview isexperiencing a declining response rate (Battelino & Peachman, 2003). While thestatistics in that paper show results only through 1999, the trend towards decliningresponse rates has continued into the 21st century. Another example is provided inthe case of the Dutch Travel Survey (van Evert et al., 2006), although this surveyshows a signicant jump in response rate, due to adoption of an improved surveymethod in 1999. However, even the improved rate has showed a decline in the fouryears documented in the paper by van Evert et al. (2006).

During the past 50 or more years, the household travel survey instrument hasundergone a signicant and substantial evolution. Figure 2.1 shows an example ofthe type of trip data interview form that was used from the 1950s to early 1970s. Thisexample is similar to many that were used in this period and which were designedspecically for use in a face-to-face interview survey. The survey forms were, to alarge extent, designed by trafc engineers and transport planners, who had little orno formal training in survey techniques. Fortunately, because the surveys wereconducted as interview surveys, it was possible to train interviewers in the use of theforms.


Figure

2.1:Exampleofatrip

data

form

from

FHWA

(1973).

20PeterR.Stopher

The earliest designs of self-administered surveys used forms that were littledifferent from the interviewer forms that were developed in the early days oftransport planning. However, during the 1970s, especially as disaggregate modelsbegan to move into mainstream modelling, changes began to occur in the layoutand structure of the survey form. Probably, the most signicant of these changeswas made by Brog, Fallast, Katteler, Sammer, and Schwertner (1985a) through theintroduction in Germany of the Kontiv design of diary survey. A similar diarydesign was then introduced in the United States (Brog, Meyburg, Stopher, &Wermuth, 1985b) in the Detroit survey in 1981 and then in a number of subsequentsurveys in the United States. This began the development of travel diary booklets inthe United States, which have now become more or less the standard for mosthousehold travel surveys. Figure 2.2 shows an example of a recent booklet used in theMichigan Household Travel Survey. Differences between Figures 2.1 and 2.2 areindicative of the extent of the changes that have occurred in the design of surveyinstruments. In the early 1970s, the standard survey instrument was a set of leaetstyle sheets, usually of foolscap or legal size in landscape mode that was used by aninterviewer to enter the responses to questions asked of a householder. By the middleof the rst decade of the 21st century, the standard survey instrument is a bookletwith questions to be answered by the respondent.

Recruitment methods have certainly undergone signicant change, but this oftenremains as one of the more difcult aspects of a household-based survey. In manycountries around the world, there is no listing of household addresses, which iseither publicly available or even for purchase. This creates various problems forundertaking good random sampling. In the United States, this problem has given riseto the almost complete dependence on random-digit dialling as the method to drawthe sample, although this method is unpopular in many other countries.

Another major change that has taken place in the method of undertakinghousehold travel surveys is the denition of the day of interest for the survey. In thetraditional home interview survey, conducted in the early days of household travelsurveys, the interviewer invariably asked questions about the preceding weekday tothe day of interview, and relied on the ability of the respondent to recall behaviourfrom that day. It was recognised early on that this recall survey under-reportedtravel, as evidenced by discrepancies between screenline counts and expanded homeinterview survey data. Nevertheless, in the early days of self-administered surveys inthe United States, as well as in many surveys elsewhere, this retrospective collectionof travel data continued. By 1980, especially with the introduction of the diarysurvey, prospective diary completion was beginning to be used (Stopher, 1996). Inaddition, as Stopher (1996) notes, a secondary issue that was also coupled with thischange from retrospective to prospective was an overt attempt to reduce the level ofproxy reporting. In the conventional retrospective surveys of the 1970s and earlier, itwas customary to ask the person who answered the door to the interviewer to reporton travel not only for themselves, but also for everyone else in the household. Inintroducing the prospective diary, each person in the household (often over a certainage, such as 5 years old) was provided with a diary to be lled out by the personthemselves, if able to read and understand the questions, and by proxy only if the


Figure

2.2:Pages

from

Mich

igansta

tewideplace-b

ased

diary

(reprin

tedwith

perm

issionofMich

iganDOT).

22PeterR.Stopher

person was unable to read and understand the questions (which would bepredominantly children under the age of about 12).

Many other changes have taken place, both subtly and more obviously in thedesign and conduct of household travel surveys. These are likely too many in numberto be of interest to catalogue in this paper. However, the changes already noted hereare certainly among the signicant changes that have taken place and that are leadingto improvements in the quality of household travel surveys, and other transport-related surveys in general.

Perhaps, however, the most signicant change that has occurred in the history ofthe household travel survey has been the emerging awareness by those involved inhousehold travel surveys of a vast literature on conducting surveys of various typesfrom other elds. In the early days of the home interview survey, the surveys weredesigned almost solely by trafc engineers and transport planners who had little orno training in survey design, and were unaware of the efforts of others in surveydesign. It is, perhaps, rather amazing to see how well these engineers and plannersdid in a eld that was unknown to them and to which they could bring only theirengineering and planning training. It was only in the 1970s that the eld of householdtravel surveys began to be a little more informed of the efforts of others outside theeld of transport that could provide pointers to better survey design, and it is almostcertain that some of the changes that began to appear in the survey designs were as aresult of such inuences, albeit inuences that were largely unrecognised at that time.However, such changes as the move to booklet-style diaries, prospective diarycompletion and reduction in proxy reporting all mirror information well known inthe wider eld of survey design. Household travel survey designers have only becomeaware of some of this work in the last decade or so, and references to such works inpapers and articles dealing with household travel surveys have appeared onlyrelatively recently. Nevertheless, the eld could clearly have beneted by consultingworks such as Sudman and Bradburn (1974), Cialdini et al. (1975), Dillman (1978),Bradburn and Sudman (1979), Sudman and Bradburn (1982), and Groves et al.(1988) among others, most of which have been absent from the references of papersin travel survey methods until quite recently. It is, however, now evident that the eldis becoming increasingly well informed about research in other elds on surveydesign and execution, and that current designs are often now improved throughbringing such knowledge to the table.

2.2.1. The Recent Past

Changes in the last 10 or 15 years have, perhaps, accelerated along with so manyother things, such as technology and the sense of the passage of time. During thattime, we have seen increasing willingness to experiment with technological solutions,such as Internet surveys, GPS devices and computer-based surveys, among others. Areview of the books produced from the past three conferences (TRB, 2000; Stopher &Jones, 2003; Stopher & Stecher, 2006) and a comparison to the papers in a much


earlier conference (Ampt, Richardson, & Brog, 1985) gives something of a avour forhow far the profession has moved in this respect. Among specic areas worthmentioning is the increasing interest in behavioural processes, as evidenced by thework of Doherty, Papinski, and Lee-Gosselin (2006). In the CHASE methodologydeveloped by Doherty, Nemeth, Roorda, and Miller (2004), the way in which peopleplan for and then reschedule and change particular activities and associated travelhas been explored. There has also been an emerging focus on process data, ratherthan outcome data (e.g. Bradley, 2006; Pendyala & Bricka, 2006). The schedulingand process directions are concerned with both developing a better understandingabout how travel decisions are made (as a foundation for a better understanding ofhow behaviour can potentially be changed) and improving the ability to model travelbehaviour, and not just measuring the outputs of the decision process in the form ofthe travel that people undertake. Also, while transport data has almost always beena signicant input to modelling, it has also always been a signicant input to policyformulation and policy testing. In this respect, policy issues have changed over time,and these changes also inuence the changes in travel survey methodology. Amongthese one can note particularly concerns with greenhouse gas emissions that aredriving the need to look at more than just a typical weekday in the spring or autumn,and the increasing concern with bicycling and walking trips that are driving changesin how accurately we collect such non-motorised travel.

Also, in the last 15 years, we have seen a shift from concentrating on the trip as theunit of measurement to activities (Stopher, 1992), and then to time use (Harvey,2003). It is perhaps interesting to note here that the shift to activity diaries was madeinitially in the belief and expectation that this would improve the accuracy of datareporting by respondents, and was not made with any modelling implications inmind. Indeed, consulting the paper by Stopher (1992), one can see that some painswere taken to stress that the resulting survey would still be able to produce theconventional trip-based data. The development of activity-based models largelyfollowed this shift in the data collection mechanism, rather than driving the change indata collection methods. However, the shifts in the modelling paradigms have beeninstrumental in pushing the data collection methodology more towards an interest inboth activities and tours. Again, however, the introduction in some locations of time-use surveys was also motivated by efforts to improve the accuracy and completenessof travel reporting, rather than being driven by a modelling paradigm. Perhaps,however, because modelling paradigms that specically require time-use data havenot emerged, the time-use version of the diary has not moved so clearly into themainstream of data collection as has the activity-based diary and a close relative of it,namely the place-based diary.

Another development over the past 15 years has been the re-emergence andsubstantial improvement in the use of Stated Response measurement as an input toboth policy evaluation and modelling of travel demand (Louviere, Hensher, & Swait,2001). Again, the key element here was the development of improving methods forundertaking the surveys that would provide the stated response data and particularlythe tying together of revealed choice and stated choice measurement into a coherentpackage in which the stated choice questions clearly relate to the experience of the

24 Peter R. Stopher

respondent, as measured by revealed choice data (Stopher, 1997). Efforts to continueto improve the methods of Stated Response measurement continue (Jones & Bradley,2006). In the 6th International Conference, Lee-Gosselin (2003) put forward a usefultaxonomy of stated response methods, including stated preference/choice, statedtolerance, stated adaptation and stated prospect. At this point in time, only the statedpreference/choice subset has really seen attention in the recent past.

Two further interesting developments in the past 15 years or so have beenconcerned with an aspect of survey methodology. The rst of these was the initia-tion of several transport panel surveys, most notably the Dutch Mobility Panelwhich began in 1984 and ended in 1989 (van Wissen & Meurs, 1989), the PugetSound Transportation Panel which began in 1989 and ended in 2002 (Murakami &Watterson, 1989) and the German Mobility Panel which began in 1994 and iscontinuing (Zumkeller, Madre, Chlond, & Armoogum, 2006). For some reason,transport professionals have been reluctant to embrace panel surveys, despite the factthat panels have many advantages. In 1993, there was a conference held in LakeArrowhead, California, which was called the First US Conference on Panels forTransportation Planning (Golob, Kitamura, & Long, 1997), but after 15 years thereis yet to be a second such conference. Stopher, Swann, and FitzGerald (2007) havebeen using panels for the past three years for evaluating travel behaviour changes inAustralia. However, apart from these and a few other instances, there have, to date,been few panels in transport applications. Perhaps one of the reasons that morepanel surveys have not been undertaken is again because modelling has yet to takeadvantage of the richness of data and the dynamics of behaviour offered by suchsurveys. As is discussed later in this paper, there are clear reasons for using panelsmore extensively, but it is doubtful that a strong case will be made for this until themodelling catches up with the potential use of the data.

The other methodological development that has appeared in transport surveys isthe continuous survey. This is not to be confused with a panel. In continuous surveys,a fresh sample is drawn each year, but in such a way that the sample membersdrawn in each year should not include respondents who were included in the recentpast years of the continuous survey. The survey is continuous, in that it takes placeusually throughout the year and on all days of the week and weeks of the year, andcontinues from one year to the next. At present, apart from National Travel Surveysin the United Kingdom, which is a continuous survey that has been running since1988; Sweden, which was continuous from 1994 to 2001; Denmark from 1992 to2003; the Netherlands since 1978 and the Czech Republic since 2000 (Bonnel &Armoogum, 2005); the only examples of regional continuous household travelsurveys appear to be in the Antipodes. Possibly the rst of these Antipodeancontinuous surveys was the Victoria Activity and Travel Survey (VATS) which ranfrom 1994 to 2002, and which involved collection of data from about 5000households each year during that period. VATS has now been replaced by theVictoria Integrated Survey of Travel and Activity (VISTA), but it is not yet known ifthis will be a continuous survey like its predecessor. From 1997, the Transport DataCentre in New South Wales has collected a continuous household travel survey in theSydney region (Battelino & Peachman, 2003). This survey collects data from about


5000 households each year, throughout the entire year, and provides data based on athree-year rolling average, which therefore covers about 15,000 households onaverage. The survey is ongoing. The third continuous household travel survey iscarried out by the Land Transport Safety Authority in New Zealand and covers theentire country with a sample of about 2200 households each year. It began in 2002and is continuing at this time, and aims to have a four-year rolling average torepresent the New Zealand population. Although not a household survey, there isalso a Canadian Vehicle Survey that has been running as a continuous survey since1999. Again, continuous travel surveys have not yet made it into the mainstream,although there are many good arguments for them, especially the constancy of thebudget allocation from year to year, as opposed to having to seek a major budgetitem for the periodic household travel survey as well as the issue of measuring CO2,which requires year-round measurement and measurement from year to year todetermine the direction of changes in greenhouse gas emissions. One of the majorarguments against continuous surveys is the use of rolling averages in modelling, apractice that some transport planners and modellers regard with discomfort, if notoutright opposition.

2.3. Emerging Challenges and Directions

In many parts of the world, the paper and pencil interview survey is still the mainstayof the household travel survey. However, in some parts of the world, other surveymethods are emerging at this time that are hoped to bring improved accuracy andcompleteness to the data that are acquired. It is probably important to reiterate at theoutset that the main purpose of a household travel survey, and indeed most othertypes of survey in the eld of transport, is to collect data that may be regarded asrepresenting accurately the various travel behaviours of the subject population, be ita city, a region, a state, a nation or whatever. There are two key words in thatpurpose that should be kept very clearly in mind as we pursue other methods ofundertaking surveys accurately and representative.

We undertake sample surveys always because it is cheaper, quicker and moreexpedient to conduct a sample survey than a census. Funds to obtain a census of travelbehaviour have never been available and are unlikely ever to become so. However, thedesire would be to measure the entire population (i.e. undertake a census) if it could beafforded and could be done within the time frame required. Therefore, the requirementof the survey is that it be as representative as possible of the population from which it isdrawn. A use

Documents

227555475-1848558449