ED 237892-

AUTHORTITLE

INSTITUTION

SPONS AGENCY

PUB DATE'GRANTNOTEPUB TYPE

EDRS PRICEDESCRIPTORS

DOCUMENT RESUME

CG 017.147

Gottfredson;0,Gary D.;.And OthersThe School ltction:effectiveness Study: Second Inter'imReport. Part I. Report W). 342.Johds.Hopkins Univ.,' Baltimore, Md. Center for Social'Organization of Schools.National Inst. for Ju'venileJustice and DelinquencyPrevention (Dept. of qustice/LEAA), Washington, D.C.;National Inst. of Education (ED), Washington', DC.Jun 83 ,

80 -JN -AX -0005; 82-JS-AX-0037; NIE-G-80-0113154p.; For related document, see ED 222 835.Reports - Evaluative/Feasibility (142)

.04F(14./pco'i Plus Postage.' ,Adolescents; *Delinquency Prevention; *EvaluationMethods; Measurement Techniques;. *Nontraditional`Education; *Outcomesof Education; ProgramDevellopment; Program' Ef'fec'tiveness; ProgramEvaluation; PsYchometrica-v Secondary Education;. Youth.Problems ,

ABSTRACTThis report, prepared by the Delinquency and 'School

environments Program, Ideseribeal,further int:Grim results of .theprogram's national evaluation 6f the Alternative Education ProgFaAthe Office of Juvenile Justice and Delinquency Prevention (0JJ015).

.The,repor is directed primarily to project 'implementers and to the-0JJDP and is organized in.two.sections: the first-presents adiscussion of topics-relevant to the entire study. Chapter 1recapitulates briefly the first"chapter of the first interini reportto provikle a quick introduction for those unfamiliar' with theAlternative EducatIon\ Program. Chapter 2 reviews chapters -4 of the.

. first interim report tooutline the record'ol accomplishmen4t-10f-earlier delinquency prevention efforts, andithe program developmentevaluation. ChiPter.3 tlescribes changes made in tilt measures used:inthe school action effectiveness study. Chapter 4 describes the landsof delinquency prevention projects in theinitiative, the majorinfluences on the evaluation, and the develoment of the preventionprojects during the second year of the Alterefftive Education Program.Chapter 5 provides an overview of the school-level evaluation,results. Chapter 4.summarizes information about the; effects ofinterventions targeted at high-risk individuals for projects thathave such targeted, .'interventions as distinct, evaluatable components._Thee -final chapter draws implications and makes recommendations. Part2 of the report consists of 14 independent reports of the preveRtionproject. The appendices contain detailed statistical tables andothermaterial relating to the results summarized in part_l. (JAC)

**********************************0**********************************

* Reprbductidng supplied by EDRS are the best that can be made *; ,

.; * 4r *1- . from the' original document.

*****************W***********W*****************************************

REPORT NUMBER 342

JUNE 1983o

THE SCHOOL ACTION EFFECTIVENESSSTUDY: .SECONDiINTERIM REPORT PART I.

Gary D. Gotfredson, Denise C. Gotfredson, And,-

Michael Cook

4

U S DEPARTMENT OF EDUCATIONNATIONAL INSTITUTE OF EDUCATION

E U CATIONAL RESOURCES iNFORMATION

CENTER (ERIC)This document has been reproduced asreceived from the person or organizationoriginating itMinor changes have been made to improvereproduction quality

I

_ 1Points of view or opinions stated in this document do not necessarily represent officialNIE - TO THE EDUCATIONAL RESOURCESposition or policy INFORMATION CENTER (ERIC)"

cr

"PERMISSION TO REPRODUCE 'THISMATERIAL HAS BEEN GRANTED BY

,b.

f

STAfF

`Edward L. Mcbill, Co-Direttor

James M. McPartland, Co-Director

Karl L. Alexander.

Henry!J. Becker

Jomills H. Braddock, 1r

'-Shirley Brown:

Ruth H. CaUter

Michael Cook

.Robert L. Crain.

Doris R. Entwisle

Joyce L. Epstein

James Fennessey

Denise C. Gottfredson'

Gary D. bottfredson

Linda S. Gottfredson

Edward;J. Harith.

John H.-Hollifield

Barbara J. Hucksoll*

r°

f

Ramona M. Humphrey

Lois G. Hybl.

Helene M. Kapinos

Nancy L. Karweit.

Hazel. G. Kennedy

Marshall B..Leavey

%Gretchen M.-Luebbe

Nancy A. Madden.

Kirk Nabors

Alejandro Portes

Donald C. Rickert, Jr

Laura Hersh Salganik.

Robert 'F. Slavin

Jane St. John

Valarie Sunderland

Gail E. Thomas

William T. Trent'

1

L

THE SCHOOL ACTION EFFECTIVENESS STUDY:

SECOND INTERIM REPORT

PART I o

Gary D. Gottfredson, Device C. Gottfredson,' and Michael S....00k'

'Editors

Report No. 342

Jtne 1983

This work is supported primrily..by GOInt'sfrom the Office of Juvenile.Justice.ali4 Demei:t of Justice,. and in parlqtly'Grant!a.Institute of Education, U.S. .Department ofdo not necessarily iepre,seAt the position

No..80.4.JN-AX-01:105 and 82-JS-AM-0037

linquency Prevention,.U.S Depart-NIE-G-804f13 from the NationalEducation. 'The opinions expressedor polky of any agency.

The Johns Hopkins University3505 NorthCharles StreetBaltimore, M.:aryland 21218

./ 4'

file .Centero

The Center for Social.Organizatbin of Schools has two primary1 -

objectiVes:. 'o develop a scientific lsnowledge of how schools affect their

students, and.to use this knowledger:,to develop better school practices and. .

organization.'

The Center Works through three research programs twolachieve arts'

vbjectives. The School Organization Program investigates how school and

classroom organization affects student learning and other outcomes. Current!

studies focus on parental involvement, microeoMputers, use oV time iri.schools,

cooPerativeklearning, and other organi2Ational factors. The Education and^

Work Program examines the relationship between schooling and students' later-

life occupational and educational success. Current projects include studies. -

of the competencies required in the workplace, phe sources of'training and

experience that lead to employment, college students' major field choides,

arkd employment_of urban minority youth. The Delinqueifdy-atid-Schobl-Iniiiiron_

. ments Program researches. the. problem of crime, violence, vandalism, and

disorder in schOol.s_antrihe role that schools play in delinquency. Ongoing'

.studies address the need to develop a strong theory of delinquent behavior

while examining school effects on delinquency and evaluating delinquency'

preventibn programs in and outside of schools.

.The Center also supports a Fellowships in Education Research program

that provides opportunities for talented.young res?Archers to conduCt and

publish significant research and encourages the participation of womei and

minorities in research on education.

This report, prepared by the Delinquen4, and School Environments PI-ogram,

describes further, interim resultsPof the program's national evaluation of.

the Office of Juvenile Justice and Delinquency Prevention's (OJJDP's)

,Alternative Educati,n program. First interim results were reported in

esOS Report No. 325, April 1982.

N-04

School Action Fffectivencea Study Staff

ThetJohns Hopkins University

Gary D. GottfredsonProject D rectorDenise. C. Gottfredson,I Associat Research ientistMichael S. Cook, Aesociate.Resea ch Scientist'.Donald E.' RickereJr.,'Researdt AnalystHelene,Kapinos, Data TeChnicianLois: Hybl, Secretary.

Andrea NuzzolovResearchAgistantJdbert Kirchner, Student.Research Assistant.Abhilit Mazum4er, Student Research AssistantStuart Gavurin,,Sivant Research Assistant

. .NormRingel, Community Psychology. Intern

2

Social Action Research Center

J. Douglas Grant, Project AdvisorDeborah Daniels,'_ResearchAssociate,,Jane St. John; Research Associate

Federal Project Officers

National Institute for Juvenile Justice and .%

Delinquency Prevention

a

Barbara Tatem Kelley, Project Officer

Special Emphasis Division, Office,of Juvenile Justice\and Delinquency Prevention

Roberta Dorn, Project OfficerCatherine Sanders, Project Officer

Technical Assistance Division, OJJDP

Mary Santonastasso, Project Offider

Tedhnical Assistance

Polaris Research and Development

Noel Day, Project CoDirector'David Sheppard, Project CoDirectordonseratte Diaz, 'Project. AssbciateMichelle Magee, Project AssociateJohannei Troost, Proiect Associate

Preface

The.Schooi Action EffectivenessStudy.(SAES) -id the national'evalua-:tion of'pe Office-for Juvenile Justice and Delinquency Prevention's(0JJDP's)..Alternative Education po-

t gram..Thestddy is rooted in the.perception..that reducing''the- risk-ofyouth'criliierequires thexollabtorative effort of prkotitioners,researchers, and, Ojectwonsors,

. Together; these groups, can create'change and examine itacon0equencesin settings where answers are, neededand problems are.;eal. r

The study is apsCrooted in thenotion that .theo is an essentialingredientof bot03.04grala develop-ment and emaluationsearcb. Con-sequently; SAES aims:'40implementthe:Program Development valuation(PIE) method, collaborating withpraCtitione'rs:in spealfxing.theOry-:"bdsed research questions and desi,gn-ing evaluations ?as An aid to organi-zational self-study and thedevelopment of effeCtive programt to.preventlouth crime. 10

As evaluator's, -we are assisting

in the development of effectiveprojects;-critically assessing proj7ect effectiveness, and Contributingto knowledge about wayato-reduce.youth crime. At root, we sHarevith:OJJDP and the legislators whip, . ot,-4

created that agency the convictionthat the public deserves delinquencyprevention and educational'effbrts,whose effeCtiveness. has been'demon-strated. In a Federal demonstrationProgram such as the Alternative. EdUcation Initiative, the expenditureof public funds is justified by theevaluation of the resulting effortto learn how to develop and imple-ment similar prolapts effectively.The current evalltion, ialthough.accounting for a small fraction ofthe cost of the Alternative Educa7"

stor

1. !

Lion Program, ,has the/ importanta';'.mission of summarizing and making

'available for transfer-to others, the, 'knowledge gained In the .broader oft-,

. ,4, ',

.

We have,not assumed' this. /..,

important tasl. will. be easy to i

A..%

accomplish,,and we are gratified ...

that we have'been as successfu'aswe have bben in translating our:

\i_

.

amb.ttions4

into eal\ty. The excel- .

, : ,1.:nt..rapport and-cooperation ;6-have:, . Wi Irtfie Federalo:agenciesinvolved,

an Withmdst of tie action proj-..

ects,.ba've been critical in this

Ngram.

SUC. eSp f

\ --- ,,.. ..,

/This interim report summarizes

Some of what we have' learned in thesecond\year Of the SAES. We' ate

pleased that evaluation is becomingroutinized as an expected and well-understood part of the activities of '

md4 projects. .1.

Isr

. We are prebently in a third year*of interactionviith 14 of the 17 .

Projects with'whOm we began working.in 19'80--thbse that coht!inue to

operate. In most cases,.evaluationdesigns that are strobger than thosepossible in'the firt-- two years arenow being implemented. . Becausesounder projects and sounder evalua-tion designs were available in the

f) second year of operation than in the.

first, the urrent report is moreinformative in describing theeffects of project Interventionsthan was our firstinterim.teport.

ure reptrts should be even Moreproductive in assessing projecteffects on students andsohools.

In our first interim report read-ers:may find information' about eachdelinquency prevention' projecC:shistory,..its start-up activities,and its successes and ptoblems in.

Prefa?.e

t' implementation duxing Sits firstN.year. Alsb, in that report may be

founcL some organizational diagnoses,and some ideas abgut improving proj-ects. Oc ca Sional ly , . you wild. find

upon.,.

Micahael S. COok, Deborah Dan-.iel,s; Denise. C. Gottfredson'i Deborah

Ogawa, Donald E. Rickert, Jr. ,preliminary at t erapt s/go ',as sess,

0 effectiveness. Tftese -preliminary-.,attempts were, not--nor were .they,

intended tb be--authoritative and.... ,,conclusive statements.. They were,intended to provide informatitn''use-ful fo'r project development.

.1 \vEffectie projects develop Over

;Norm Riegel, and Jane St. John, .

worked long and hard with actionprOject,personnel in workshOps, site,visits, and on the phone to preparefor the surveys, to evolve.ProgramDevflopment .Evaluation plans4arid todraft prOj'ect narratives. Lois Hyblarranged workshOps,`qirganized docu-ments, typed 'manuscripts,. prepared

time,% incorporating feedback from-

graph's, .and provided some Muchtheir own observations and those of needed ,,order and predictability forevaluators to become stronger. Thissecond report is therefore, like the

the project., Helene Kapinos kept. .

the PDE worksheets flowing, main-first, directecf, primVrily, to projectim$ementers and to 'OJJDP'and its

tained our calendars, coordinatedthe flow4 of day-to-day &to entry

technical assistance contractor, tit() tasks performed by the researchhave a stake in fostering project ,` assistants, and helped us ,maintaindevelopment or in planning new ini- . the comply( and massive data, files

... -.' ti.Aives., This, repo t also-con..tainsa information about the rogress of . ''. .

cre.atecl, by this large project. Thisrcpoxt possj.ble becauseVof their

',i each project in ac'hitiling file tresults sought"by OJJDP 'in its ini-

a

help and ,kactical wisdom. \.

tiative. Part wo'of tbits report MinBirds eye, Doris Coax',.im, Bar-describes each- -individual project' bale Mat.tha,Stewart,geparately, and fdctises',, on the . Hilda Gutier'rez, .Glen Bader, Chest'er

" extent t6 which each individualproject is meeting its own goals;and

Wooten, Richard Snfith, Chares,Almo, .

,Hferman "StePtoe, Mary Lewis.. Prestonobjectives, and describes the pr'ob- Elract, .Pau'1 Friday,. Vanita Vactor,lems 'and achievements of each. . Bill, Harris, David Bailey, Anita

0 , Batisti, Anadia Andrews, JeanetteAcknowledgments *

0 ".

This interim report 'is the pro-'P-. duct of many people'a contributions,.

and it is possible to describe. only .3,

some of these contributions here., The thousands of youths and teacherswho shared their views about theirschools and provided informationabco.lt themselves made major contril,-buti,ons 4to the empirical basis for

' thie report. The principals of: par-r ticipat'ingVschobls, who facilitated

both project- development and evalua-' tion, were essential contribiaors,to., . the development of knowledge. We

hope thatn ail of these participants ,

will be -rewarded for theif help by -.having; their views heeded' and acted

Bass, 'fiyrone'Seals , Phyllis Betz,Ciorah Montes, Nilda ROdriguez,Chris Lopez; Philip Cano, NancyCohen MarilynalcKnight, Prentice:.Deaciri'ck, 'Dave Reisk, Joan Bellafo-nataine, Joe Natt4n, MarkCpuOr,, Nic Cooper, Sally Wizotsky,Sonny' Luster., WilliaM Kottman, Tom .Leighty, Roy Mahoney, and Pat Kennywere prevention project personnelwho contributed in .basic ways to thework reported here. They providedthe theories that guided much of theinstrument 'and scale' construction,developed .project 'plans using evalu-,ation terminology,. and made theaction projects and data corlectrongo. ,

N, .

Denise C. Gottfredson performedthe superhuman task of coordin tingan anaging all of the data; G'aottfredson and Donald E. Rickert

modified the surv'y instrument's used;in the project's 'first year'.Michael CoOk, Denise Gottfredson,Donald Rickert, and Jane:St. John,.

. put in many long days analyzing dataunder- great time pressure. Others

. who assisted with data analysis andmanagement' include': Stewart Gavu-iin, Richard D. Joffe; Robert Kir-chner., Helene ,Kapinos, AbhijitMazumder,,'Andree Nuzzolb, and Debo:-rah. K. dgawa. . Raul Romero traps-,lated some newt items for the studentquestionnaire into Spanish, and Den-nis pillon and Mary' Ellen Hartmannof Intran Corporation produced theOptically scannable instruments.

Roberta Dorn and Barbara TatemKelley of the Off ice for JuvenileJustice and. Delinquency Preventioncleared the way for this project to,"proceed, and helped to resolvenearly countless problems along theway..

Opinions expresssed are theauthc3rs' or ediLors", and do notnecessarily reflect the position orpolicy of any agency or institution.

Organization of the Report

The: remainder of this report isorganlzed into two sect4ons. The, ,

f irst of these discusses topicsrel-evant, to the entire study. Chap-ter 1 recapitulates in briefer formthe first chapter of thie firstinterim report to provide readersunfamiliar with the Alternative Edu-cation Program and the School Action

, Effectiv-eness Study with a quickintroduction. Chapter,2 recapitua-lates in br ief er form chapters 2'througk .1-§4of the first ineetlimreport to acquqint those who havenot read that document with impr-tant information about (a) the

'reface

.'record of ,accomplishment in e,arliel'delinquency ,prevent ion 'efforts, ,(b) conditions necessary to make

inferenpes about prevenLion project,effectiveness, and- (c) program,development evaluation. Chapter 3descripesbschanges made in the meas'-2

thsed in the school actioneffectiveness study, summarizing the

. psy.chometriic prqperties of instru': meats redesigned for easier compre-

tension by schoorofficials or 'improved through new research.

.'Chapeer 4. pfovides brief descrip-tiong of the kinds of delinquencyprtventicin projects in the initia-

. t lye., It. alko describes majorinfluences on\ the evaluation and thedevelopment of the prevention proj-

.ects during the second year of theAlternati,ve Education Program.Chapter 5 provides an. overview of

tthe school7level evaluation results,fcir the results sought by OJJDP.Cliapter 6 summarizes informationabout .,the effects of interventionstargeted at high -risk individuals..f or projects that have such targetedinterventions as' distinct evaluate-.ble comktonents. Chapter 7 draws,implications of the' study that seemappropriate at the present' time, and

'makes recommendations for .futurework. to, reduce youth crime.

Part II of this report consistsof . independent reports on the pre-vention 'project. Most chapters weredrafted by the field worker assignedto. the- iojeCt:' Therefore, theygenerally have the benefit of havingbeen given direct attention by themember -the valuation team most'familiar with th'at particular alter-native education project. At thesame time, "however, the involvementof multiple authors, eechNith adifferent background. and perspectiveon evaluation, leads to some 'uneven-news in prenenta4on. Some .,authorshave bluntly 'provided the, gott'd andthe bad news in a straightforwardf ash ion. Others have leaned toward

9

. Preface

presenting the projects in ways thatmake their strepgths salient. In

the editorial poc,p6 we have nottried to-eradicate the personal andstylistic differences that exist'among the authors. of the separateproject reports. The reader istherefore urged to consider each ofthese a distinct essay, and to avoi'dmakinecomliarisons across projects,on the, basis . of th e ae :*ind iv id ual 1 y

drafted accounts. Many readers mayT.

"s

-iv-

be interested in 'reading Part'Ithis report, Olectivelpr dipping Vinto Part II to learn mote aboutspecific projects..

Appendices contain detailedstati:sticalitablea and other materialrelating to the results summarized

A %in Part..

a

`a.

O

10.

. .0G

I

Prefa-ee

Table.of Contents:

9%.

..

.... Ackni/ dgments . .11',:. . .-. .-:4-

. . A eeeee e.o ....... iiOrgan tion of the'Rerlait . . . : ;.... : . . . .. iii,

1-e

Para'' I: School Action Effectieness Study Over'vi'ew, a o

. Chapter 1' ' 1

./ IntipduCtion to.the SO'ool Action Effectiveness Study ...: ... . . . . ,, 1.

..

The AleernailvvEduoation Program ..... ,. .. , . ',;,. : , ... .. . 1

Evatuationt.iimE;. 1 ' 11 : . . . . 2

Planning aril .Implementation P 084, .3

Evaluation, Che Spontor, and the Action.Projectti- :

4...

. ,

9

. . 9

.

Chapter 2 ;

- Elements df Effective ,Program Developmentand Evaluation . .

r

4P et Vforts I

Ntii' 9

'''I ferences about Project effectiveness 9

T e Current Egort- ... - '. . , 10

StIrengtheningPrevention.Projects thrOugh-Eyaluation .,. . .. : .. : 10

' PDE . e. . ' . 11\

. .. , . ..

,.,

. , 11...On oing Process .;,

, . ..

.

4 .ChapOer 3. , 15

, .

. An Update o theMeasures psedjsok the School Action,Effectiveness Study . 15,.

Meas ring Inilividuals and -Or onszatii 1an 5..

Some Essential' Psychathetric concepts : . . . . . .04 .. . ..... '.. 16Relative- Measurement ..'. . '.' . 16

... .

. . , .... . , , . . . . ... . . ...

.., Reliability . y 17. Validity . . ,. ..\.

. .:

' 18-Measures of'Students . .. . . : . 1

..o1 ,.... .. .` ..... 19:.Socia0Bacitground . . e .- . *19SOcial\Relations ,

' . 22.

... ,Attitudes and.Psychosocial Development 23

', SeltL.Reported Benvior ... , ,' > .... ..... .. . .. .... , . '26

.0

'

'., Measures 'of School Experiences .... :..,..... .' ,- ,Quality Control . 9. ... ' . .. .,,

, 3s Over T .StabilityOffStUdent Measures Time ,i

:i

Measures of TeacheTeacher'st i

- Stability of the Teacher MeasuresMeasures of ;School Climate 1.. .... .. . . ,Compositiohal and Psychosocial.Climate ScalesMeasures of psychOsocial Climate Based on Student Reports .....Climate ScaleB BaLsed on Teacher ReportsStability of the Psychasocial.Climate MeasuresInterpreting the School Climate Measures ; : .... i . . . 0

..Student Scaleb \..

.

.Teacher .&alea \ .. . : ... .. , . ...

Suggestiens forlising Climate Reports in Project,Planning%Using Measures fl r. Individual Students and Tea;cheisThe Utility of Irr formation far Project M a n a g e r s . .. . ,

.

0

.....

.

a 4

'

3. '.

.

,

.

.

.

.

.

'

.

28

29. k

29

30

34

34

44.4W

34

37'40

4040 .

-'44 ..,

41

41

42

Li

V V.

'Ohapter . . ..... ....8

'The Alternative Education 'Program: Kinds. of Projects and, the Youths::

Involved 4+3

Crossgutting Dimerisions 43

'.Individual vs. Environment 1 . . . .; ..: . . .44, . .

6 . A Classificatioh oLPreventiOn Projects- . . . ...... ic,. . . .. 44

The ClasSificatiop:4Ppliea' tO.thei4ternAtive Education Projects. , .*46. -

What Are The' Interventions? .. . : ,. ... .. ... .. ...... ., : .,.1..: 47

47'

ftWho-Are the Target Populations? .

.. ,

11

How Much -Delinquent Behaviror Occus? 60

Some Consequences of! Victimization . C. ... , . . : .... . . . ;61''.

Who EngligeS in Delinquent Behavior? : . . , . . - ::- 61..,. ,,

.

. .

.

,. .

Chapter 5' , ,1 65' e...

`Thenvironmental,Effect of'. the Alternative EdUcation Protram 65.

Oyerview, of'the'ResUlts SOUght ly-the Program -. . . .. : .. ... :. . .. 65 . ,

MesSureicbf,Program Goals"`. . 1 . .. .: , 65... . ..... . .

.. Measures ofProgram Objectives , , . ; .., 66.

Theoreticakly\Important Outcomes '-`., . :'.'. . .. ... . . - ... .. " .. ..... 67,,

Method's, .. ', '', - .. . .' ... . . .... . .. . : .. ... .. . .4, .:O.'I / '.67 '

..

Simple 'Box Scor'tt.

School-by,-Schook' lEkamination A.s .. 68'

Results . . .. . . . ., 68...

Program Box Scares) ... ' .. . .. a V . J ...... .' , :. 66

School-by-School Summary .,... . " .... .... ' ... .. - . ...., '. 71

. Discussion' .- . . ..- .. :. ., . .. z.. In;

. , ., I . .. .i, . ,. 91'. ALiMitatlons ... . ' . , ....

Differences,gpidemiological Indicators vs. Between Grout:..., .1.. , .j ,. 0%-

nV

.C7

. .

Some Interpretations .

..

102,A

. I ..,\. . 103 :Chaster 6. . .(7'. . ', .r. .

The-Effects of InterventionSTargeted at Identifiable .croup0.1 . i4 . . 103.

,

.valuation Designs , .. ... / . 104

Highlights' i. .- , : 16,6

Compton . j . 4'. .. l '. . . c '106-

Pasadena ,

/06'a.

PCD!,

t. 106 _-

Kalamazoo 0,:,

.j, . . . ..- 107',

.ii . 4'107.Puerto RicoCharleston

.\ '467

\

/

Virgin Islands ..\' (.1 f

. , S .. '.-.. . .168

Miami ...... \ ..... '. ........ . . , .... . . yr. : 10.8 ,

Plymouth ..... \. . . . .. ..... - .

Milwadkee ... : . ..: .. . .. ... .- ...

.

.. ..,. . . 10:. ,

t .' ...4.) .. .. .

A Summary109

L

.' . . . .,...112-Progress' y

\--'-ii.,...:-

Chapter 7 , \ 113

Educational' 'Interventions and the Preventjon of Delinquency: 'Scme,Clos-,

iing Observations , .

,

.'. ,... ...*: .1 . . . ,. : . . . 1175

Schooling and Delinquency .. : .... ." ...... .' .. 113., . . :

.

Developing Effective ProgramsSound Implementation. . . , . ...... '.1. , . . . . : ... .. . 114

1 .

,, Evriluation and Expertise ' ,:. i. 114

4

L ;2

The Work Ahead

Written Product's Regulting from the Evaluation'

References

Glossary

Subject Index

Of'

A

Name Index

*Additional Mate)ials Not Included in Part I:

"70

ON.

115

1'17 -.

125

131

'137 ss

. . 141

..

Appendices,

.

A: Narmg for School. Climate Measures _

B: Profile Sheets for. School Climpte Measures

CE Worksheets for Using School Assessment Results ,.

D: Detaildd C/imaterResults for 1981 and 1982 .

E:, Total Enrollment, Number ,in Sample, and Completion Rates for - Spring

-1982 SAES Student Survgy f

. ,

Part II: Interim gValuations of Specific Projects

Compton Action. Alternative School:l Second Interim ReportD. K. Daniels and P.-D. G tfredson

Project STATUS:D. K. Ogawa

Second Illterim Report

.Project RETAIN, Chicago Board of Education: Evaluation Report'

J. St. John

.The MilwoOd ALt.r.natne Project: Second Interim Report,M. S, Cook

1-

'Project PREP: Second Interim ReportD. K, Ogawa

The Jazzmobile Alternative Arts Education Project:Evaluation Rep6rt,D E. Rickert

.41

vii-

131

.

o

I

Qtro Camino: Second Interim ReportJ. St. John

Project PATHE: Second Interim ReportD. C. Gottfredson

Virgin Islands Alternative Education ProjectJ. St. John

Interim Quantitative Evaluation of the Academy for CommunityEducationD. E. Rickert

Alternative Education for Rural Indian Youths:Lac Courte QreillesM. S. Cook

Plyinouth Alternative Education Project:Second Interim ReportM. S. Cook

The Education Improvement Center-South Alternative Educationproject: Second :Intrim ReportD. E. Rickert,

Jewish Vocational ServicessAlternative Education Project..Evaluation ReportM. S. Cook

a

14Lviii-

eo

Part I.

School Action Effectiveness Study Overview

Introduction to -the School Act

The Alternative' Education Program

The Office of .Juvenile Justiceand Delinquency Prevention(0JJDP,0funded 17 demonstration projects inthe fall of 1980 and the earlymonths of 1981 as part of a Programin Delinquency Prevention through .

Alternative EduCation. This OJJDPinitiative is premised in part on'the'obserlatioxythat delinquentbehavior is. associated with a numberof scho91-related or school -based

. problemi, including disruptiveclassroom conduct, absenteeism,truancy, and dropout (Bachman,O'Malley,. & Johnston, 1978; Gott-.fredson, 1981, 1983b; OJJDP, 1980).

An educat approach to delin-quency prevent -..a is strongly sug-g\4sted by the most widely influen-cial contemporary theory ofdelinquency (Hirschi, 1969), inwhich commitment to educational orother conventional goals, artach-ments to teachers and the school,and belief in 'rules:are viewed asbonds of social control which pre-vent delinquent behavior. Learningtheory, especially social learningtheory (Bandura,. 19711, provides anexplication of the ways in Whichthese elements-of the social bondmay be -strengthened by appropriateeducational interventions. Social

learning theory also helps toexplain how the influence Of alter-native school organization, and theinfluence peers, teachers, andparents, can converge. in preventing,or failing'to prevent, delinquency.These theoretical perspectives findsubstantial support in the evidenceprovided by research; they concur in

For a more extended discussion ofthe topics presented here, see thefirst interim report

ion Effectiveness Study

implying that alternative education.programs can be structured in wayS .

that will reduce delinquent behavior(Gottfredson, 1983b; HaWkins & Wall,1979).

\

Both primary and secondary prer-vention of delinquency might beachieved in alternative educationprograms through their effects on.the academic and social developmentof the youth involved.

i.

The Aemonstrat n program isf.6rthe most part targe d at schools

.serving grades 6 thro gh 12 in rela-tively high crime communities, Withhigh rates of delinquency, dropout,suspensions, expulsions, absentee-ism, and youth unemployment. ?roj-ects funded as part of this programwere to be aimed at achieving

. (a) decreases in delinquent behaviorin and around schools,-(b) decreasesin dropouts, ,suspensions,expul-sions, and truancy, (c) increases inattendance, (d) increases in aca-demic success in school with conse-

. quent increases in graduation ;rates,and (e) improvements in the esrly,postschooling labor market experi-

. ences, or in the post-secondarytraining or education, of youthassociated with participating'schools. The .first of, these resultssought is known to be associ tedwith the results "b" through "d" inthe foregoing list, which ar gener-spy regarded as important rsk fac-.tors for subsequent delinquentbehavior. The final result soughtwouldould likely be influenced- y inter-

,

lentions that reduce the ri k of .-

delinquency, and of special/importance to the Departmenit of(Labor (which transferred funds toOJJDP to support part of this initi-

/ative).

The achievement of these objec-tives requires some reorganization

Introduction

of.school policies, practices, andenvironments. Specifically,. theOJJDP program announcement calledfor achieving the following objec-tives which were eeen,as instrumen-tal in'fostering the attainment ofthe overarching program goals:(al limiting or decreasing referralsto the juvenile justice system;(b) making 'school discipline fairand consistent while providing fordue process; (c) increasing youth,parent,, and community agency parti-cipation in school decision makingto reduce student alienation andfeelings of powerlessness;(d) decreasing the grouping of stu-dents according to inappropriatecriteria (such As social class_orrace) which, accompanied by inproved,learning environments, should pre-clude labeling effects and stigmati-zation while enhancing educationalsuccess; and (e) providing a struc-ture for learning that promotes edu-cational and social developmentbecause it is tailored to realisticlevels of performance for individual-students.

Some of these instrumental objec-tives are in accord with research onthe characteristics of schools andcommunities that are associated withvictimization (Gottfredson & Daiger,1979;'National Institute of Educe-tiom, 1978)%. Others accord withadvice offered by national adVisoryPanels (e.g., President's Commissionon Law Enforcement and Administra-.tion of Justice, 1967), or practi-tioners (McPartland & McDill, 1977).

These, OJJDP- generated projectspecifications constitute the firstof three bases.for an evaluation.The second basis igoals and objec-tives of each of the seventeendelinquency prevention projects.The third bdsis for the evaluationis the broader literate on theprevention Of youth clime, whichspecifies some intermediary objec-

tives that are important fordelinquency prevention efforts (seeGottfredson, 1981, and Empey, 1981,for reviews of this literature).

Evaluation Aims

The overarching goal of theSchool Action,Effectiveness Study isto create transferable and scien-tifically sound knowledge about.delinquency prevention theory andpractice; But a complex evaluationsuch'as the School°Action Effective-ness Study must accomplish many aimsif it isto-be effeAive. As Ogawa

01982) makes clear, previous delin-quency prevention efforts and their aevaluations have been fraught with'problems of inc2mplete implementa-tion, weak evaMtions, and lack ofintermediary and outcome measuresrequired to assess the efforts. :

Not only delinquency preventionprOgrams.suffer from'these problems.Sarason (1971)-describes the disap-pointingAegree of in7lementation ofattempted'educational innovationssuch as the "new math." Whereas the

developers of the innovationintended to-alter the ways teachersinteract with students,..the majoroutcome was the use of:some new math.

L books. Many educational evaluationsare, as Charters and Jones (1973)put' it, evaluations of "non-events."Likewise, theory. is lacking in manydelinquency prevention and correc-tional programs,Vbut is an essential-'element in the prograMs and theirevaluatibns XEmpey, 1980; Glaser,1977; Gottfredson, 1982a). The SAES

has taken steps,to'aVold evaluatingnon-events, and. also to avoid the.otlier problems from vihich.earlierprevention evaluations have Suf--fered.

RedUcingyouth crime.in Ameri ais bound tO'be at-least As diffic ltas building a space shuttle. But

those who envis$on, programs to pre-

vent 4elinquency or to. rehabilitateyouthfu? offenders all too oftenhope for effectiveness without-

-developing plausible plans or usingthe technology. needed to raise theirinert and clumsy programs from theground. Developing effective pro-grams to retuce youth crime will notbe easy: M ch worthwhile technologyhas been developed; butvusually thistechnology goes underutilized or ismisapplied in schools.

The history of previous delin-quency prevention efforts impliesthat most previous programs have .

.4P4*, been poorly implemented, implausible. from the outset, or poorly evalu-

ated. This history implies thatconcerted effort is required toimplement highly plauSible programswith strength and fidelity, and,toevaluate these programs rigorously.

The scientific literature pro-vides good,reason to believe thacythe risk of delinquent behavior canbe reduced, the evaluation litera-ture provides strong grounds for'insisting on strong, theoreticallybased, and well evaluated, programs.The School'Action EffectivenessStudy was .designed to strengthen theprojects being. implemented in theAlternative Education Program, eval-uate them rigorously, and creatp-

transferable knoWiedge about delin-quency prevention.

Planning and Implementation

The history of evaluationresearch in delinquency preventionis replete with examples of programsin which the'implementation was

' undocumentel5OrnOt carried out as .

planned (Dixon.& Wright, 1974; Kriq-berg, 1978; Ogawa; 1982). Knowingthe fidelity-with which program

.'plans are implemented; the strengthof the "treatment," and the contextwithin which the program operates isessential for three reasons. FirSt,

Introduction

any evaluatioa result--eitherpositiye:or negative--is of littlevalue unless the nature of the pro-gram is well. described. Second, .

informatiOn derived from monitoringthe activities and.the impleMenta-:tion of Plans is needed tostrengthen the. integrity of the pro-

and to detect, unforseen conse-quences or potential breakdOwn6.inproject plans or the evaluationdesign. Third, negative results of

summative evaluations have sometimesled observers to co'nclude that the.interventions intended to be imPle-mented.do not I./ark, where'as the

interventions may nOt'in fact havebeen implemented,, implying a quitedifferent conclusion (Sechrest,White, & Brown, 1979).. Knowledge ofwhat was .actually implemented isessential in drawing conclusions.from tests of .any planned interven-

. tion. . .

Strenith and integrity of plannedinterventions. Assessment of theplanning analimplementation processconsists of two distinct components

. (Sechrest, West,'Phillips,.Redner, &Yeaton, 1979). The first.relates,t6.considerations of the strength ofthe intervention plan. This isessentially a matter of the con-struct validity of the" measuresintended to be taken in an intervention: Several procedures are avail-able to assess the strength ofdelinquency prevention prograins.These include: (a) analysis of theplausibility of the plans'- theoreti-Cal premises, and determination ofhow closely the specifics of theplane are linked to delinquency-prevention theories; (b) expert judg-ments about the likelihood that theproject as specified will prodiicethe desired outcomes; sand-

(c) comparisons ofthe intendedpro-grams.with the range of current or

. past eff.ortsat delinqdency preven-tion (in 'this way a program that was°othervise'unremarkable but resembled

-3- n

Introduction

a previous ineffective effort 'flight. -be' judged a weak program). In addi-

tion to a theoretical basis, parame-ters involved in making assessments,_of strength include staff stabilityor qualifications, intensita.andduration of treatment, focus ofeffort, clarity of plans, and theextent to which the plans involve .

.different responses todifferentpersons (e.g.; individualized

\ instruction). In general, replica-'

'1 tions of previously tested or, wellengineered interventions, comprehen-sive attempts. to cope with the mil-,tiple causes of a problem, treat-

I ments with clearly spelled Outtreatment protocols or implementa-tion manuals, or primarT.prevention.efforts that affect a.substantialproportion of an environment's inha-bitants are likely to be stronger:than. those that lack these charac-

teristics.

The second aspect of assessingprogram implementation relates to .

''the integrity ACT fidelity with whichiplqns are implemented.. Clear plans

/'are more likely to be implemented/

/ with fidelity than diffuse plans,fuziy promises, or; vague projectdescriptions.. Some components ofimplemetation that must bp moni-tored or observed are (a) staffing

'Patterns (including experience,

,training, numbers, and stability),(b) methods used to select, adm4t,or reject the youth involved in eachproject and each of its components,(c) the differential assignment ofyouth to aq,ernative'programs, orthe basis for indivfualization ofinstruction, (d) the ature, dura-

tion, circumstances, nd frequencyof services to individLals or f.groups, (0) methods used to deter-mine,yylio (including ,students) isinvolvedik implementation, .(f) the

7 interventions' elements' and their.durtion, (g) the degree of projectstaff commitment, (h) project super-visory and management practiceS,and

curricular materials.;

individualized education plans, les-son plans, diagnostic protocols, ,

treatment plans,'and the like.

The .importance of this aspect ofassessing implementation can ,4

scarcely be overestimated. The

scope of the alternative educatiopaction projects, encompassing asthey do many.distinct components;makes the faithful implementation ofall plans unlikely. A failure toobtain sound evidence about the,strength and integrity of these pre-vention projects could lead to erro-neous conclusiohs about the efficacy\of the delinquency prevention ideasbehind these. projects.

Evaluation, the Sponsor, and theAction Projects A,

.The\AlEernativeEducation Program

is:sponsored by the Office of Juven-ile, Justice and. Delinquency Preven-tion,with supplemental funding pro-vid'ed through OJJDP by ;the

Department of labor. TOwee divi-sions of OJJDP are involved directlyin. this program. First., the Special

Emphagis Division has programmaticresponsibility for%the grant awardsmade to the 17 action prolectslisted in Tables land 2. Second,

the Technical Assistance and Train-ing Division has responsibility fbr'providing.assislance in 'project'development, and.wOrks through con-tractors.to'do ao.' Initialby, theWestinghouse National Issues CenterWas assigned these technical assist-ance tasks as part of its larger:contract to provide assistance forOJJDP's Delinquency Prevention -.

4 ;.

11. An lath project was "funded, too

I late to be included in this evalua-tion:. Initially denied a grantunder the Alternative-Education:Pro-.

gram., it successfully challengedthis denial and waseventuallyawarded a grant.

19

I

4

4: I

I.

Ia

II

to

N

419

r

IMNI 11 mar op mir Immo II= MB wIr MEI MO

. Orga

J

halmemmpammweidwwm

Prevent ion ofAllatiwonnIP4IP

4

a

E

di

1

r

t

4

gip. WI. 1114 104 .+Ar i- .we IIINP IPA WO lb% IN- V' 1W°' Mr IIMU 41MI IMI I

I

Organization Name

Edi Imo ma.emmak4am-W4w.aw.I

Research and Development efforts.Late in the first year of operationof the ,Alternative Education Pro-:gram, Westinghouse was replaced byoPolarisReseatch and Development inthis nole. Third, the National.Institute for Juvenile Justice andDelinquenciPrevention,is.responsi-ble for the evaluation.' The insti4tute made a grant to the Johnskins University to perfoymthisevaluation, and the Univer ity sub-

L

Introduction

contracted part of the work to itscollaborator, the Social ActionResearch Center. In ,short, a total

of 23 organizational entities aredirectly involved in this effort.The participation of each is essential to the successful conduct oftheevaluation.. The degrqe of col-laboration ands,Coop4ration'amongthese'grOups has generally beenexemplary.

a

d

. .

Elements of Effective Program Development and Evaluation'

Past Efforts

Examples of highly plausible`,well implemented, and carefullyevaluated delinquency preventionprojects are extremely rare. .Dixon

and Wright (1975) reviewed95 delin-quency prevention reports publishedafter 1965 and 'concluded that thereis a paucity of evidence about theeffectiveness of existing programs.Dixon and Wright'attributed part ofthe unThpressive record of .accoM-plishment in this area to unclearproject objectives, difficulties inimplementing rigorous deaigns. andcollecting meaningful measurements.

More recently Krisberg (1979)reviewed 16 exploratory 'delinquencyprevention' projects funded by OJJDP.After the first year Ofoperation,only one of these prOjects.ffidimplemented_even a.quasi-experimen-tal.design. Not only were most,ofthe- projects unevaluatable becauseof problems in data collection andthe lack.of comparable controlgroups, but,nOne of the 16 projectshad articulated a useful theor.yabout delinquency in their catchmentareas or spelled out how their seer-Vices would reduce the. problem.Krisberg concluded that uals wereoften too ambiguous,. not cleirlyrelated to the problems the projectssere. intended to address, and thatprOjects had engaged in. incompleteplanning.

InIshort,.most.previous evalua-tions'in'the delinqeuncy preventionarea have suffered 'from evaluation'design flaws, the use of irrelevant

For a more extended discussion ofthe topics presented here, see chaP-ters 2 througb4 of the firstinterim report.

measures or nomeasures at all,dependence on a single source ofinformatiop, a dearth of theory,' andambiguity about intent.

There are, however, a few goodexamples of delinquency prevehtiondemonstration projects. The EmpeyAnd Lubeck (1971) and Empey andExikson'(1972) reports show howtheory can be integrated with delin-quency prevention efforts. AndAlexander and Parsons (1973) illus-trate a family intervention thatinvolved. (a) a clearly describedintervention, (b) process evalua-tion, (c) careful summative evalua-tion using clearly defined and non-reactive behavioral criteria in apersuasive evaluation design; Simi

larly,' the results of interventionsdescribed by Reid and Patterson(1976) are impressive. These fewexamples illustrate that higk.qual-.ity and well evaluated projects canbe implemegted.

_The present evaluation .attemptsto build on the previous.experiencein this area'to anticipate and avoidas many pitfalls as possible. We

-aim to'clarify prevention projectgoals and theory and their linkageSwith short-term or intermediaryobjectives and the intervention'saimed at bringing theie objectiveabout. We also.aim to facilitatethe development of workable struc-tures foi managing project implemen-tation and evaluation.

Inferences about-Project Effective-ness ,

Once a project has implemented a'plausible intervention intended to ,

influence student attitudes, behav-ior, or development, assessing theconsequences of that in,:erventiOnbecomes important. Making thisassessment is not always easy.

-9-

?.3 .

Evaluation

Yoting people are growpg andchanging all.the time. Rates ofPartiCipatien-in delinquent behaviorapparently' rise and then fall withage. Scholastic competencies usn-ally grpw over time, but at differ-ent rates for different people-.Students make new friends and aban-don old fines, and every parent knowsthat his or her child's tendency toconform or rebel. is different atdifferent stages of development.:Isolatingthe influence of some spe-

. cific experiences intervention, orset of interventions is thereforedifficult.

Making inferences about the,/

causes of some difference 4n studentoutcomes--about the effects ofplanned inierventions--is, however,a major goal of evaluation. Putanother way, an aim of a thoroughevaluation.is to dethrmine whetheran observed difference in studettbehavior or attitudes (if any Ofference is observed at all) can rea-sonably be attributed to a specifiedintervention. Certain conditionsmake. the search for the effects ofan intervention easier; other condi-tions preclude making any confidentinferences. An excellent discussionof the conditions. that make'infer-ence possible is provided by. Cookand Campbell (1979)', and readers maywant to consult their book for ela-boratibn.

Creating these conditions is what0 evaluators mean by "evaluation

design." Most prOjects participat-ing-in the Alternative Education

. Pogram did /not anticipate fullytheneed-tb crente'rigorous evaluation 'designs. In the first interimreport, we described at length anumber of objections raised at one,time or andaher by project implemen-ters to the rigorous evaluation oftheir' projects.

The Current Effort

A major accomplishment of theAlternative Education Program in itsfirst two years of operation is thatit has succeeded in implementingevaluation designs for a number ofthe prevention projects that aremuch stronger than those typicallyfound in this area. In the secondysar of operation, six projects suc-cessfully implemented. true random-

,ized field trials. Several otherprojects implemented carefullythought out quasi-experimentaldesigns.

A second major accomplishment ofthe Altertative Education Program inits first two years of operation isthat it 'has succeeded in .collecting

outcome measures that are clearlyrelevant to the Program's goals.Whereas the collection of informa-tion about delinquent .behavior wasresisted by several projects in thefirst year of operation, by year twoonly one project was unable to'assist the 'evaluation in collectingthis information (the Harlem proj-ect).

Strengthening Prevention Projectsthrough EValuation /

A Program Development Evaluation(PDE) tetliod providAs'the structurefor the evaluatio.67of the various

,

projects in the Alternative Educa-tion Program. - This method is :

intended to anticipate and fosterthe development of these protects byinvolving project personnel in acycle of evaluation activities: The

method is intended to (a) make rig- \

oroua evaluatiOn possible, (b) takethe evaluatien,relevant not. only tonational concerns but also to theconcerns of project°PersOnnel andmanagers, (c) doCument peojectimplementation, (d) facilitate proj-.ect implementation, (e) tie.theevaluation explicitly to theory, and

-10-

(f) integrate research with prtdectoperations so that projects developby using the results of research inproject planning. Related struc-tures, differing mt.wht In

detail, are provided by Empey (1980)and Tharp and Gallimore (n.d.).

PDF,

The Program Development Evalua-tion method pro-rides this strategyand structure, in part through thefollowing components (for moredetails, see the first interimreport; or re Gottfredeon, 1982a;and Gottfredson, Rickert, Gottfred-son, & Advani, 1983).

Clear Goals. A project withoutclear goals is o the road tonowhere. Clear measurable goalshelp a project focus its activitiesand they provide an integratingtheme for a delinquency preventioneffort. In using the PDE method,researcher, and project implementerswoek,togetaer to design an agenda toachieve clearly articulated goals.

Exblicit Theory. Theory helps toorganize knowledge, provides a guidefor developing or selepting.an'intervention, and provides a basefor assessing the program's effec-tiveness. Behind every delinquency.prevention/ project lies a set, ofideas, or practical "theories." If,

left unarticulated, these ideas pro-vide .little guidance for...projectdevelopment. The more carefullythought through these ideas are, themore useful they are'in guiding,project decision making..

Intervention. The program compo-'nents--the actions taken by a pro-gram to move:closer to achieving itsgoals--are rooted in clear-headedthinking about goals and the pro-

,

gram.s theory of action. Interven-

tions are implemented with an exper-imenting spirit. Each element of a

Evaluation

program can be evaluated throughevidence about how well it is.beingimplemented and what it is /:corn-

, Forcefield Analysis. All actionsoccur in a dynamic program environ-,ment in which available resources'co-exist with obstacles to action.Initial analysis of Oirrorcefieldincreases the likelihood that inter-ventions and research designs willbe. implemented as intended, But

periodic further analyses are'neededbecause initial analyses may beincomplete, or. incorrect, becauseperception's change over time, anslbecause the projects actions changethe forcefield.

Plans. 'Effectiye programs deriveplans for implementation that arederived.from,the forcefield analy-ses; they use available resources toovercome obstacles to implementa-'tion. fully articulated pldnincludes vtandards for impromentingeach intervention and managing theoverall program. A careful plandetails each Of t'he following:

o Critical benchmarks--:key pointsat which a decision, agreement,action, or arrangement mustoccur to keep the prcojct movingforward.

o Implementation standards--ob-'servable quality control stan-dards that let everyone involvedwith a program.know what consti-tutes acceptable performance.

o Task statements--Details of whowill do what by when.

Ongoing Process1

The Program Development Evalua-tion method stresses the collabora-tion of researchers and projectimplementors at each and all stages .

of the change process. The develop-

c25

11.

0

i

I

i p

1

1

I

Ill )tilirik I : The 1' Ili 1

?valuation

'mental expectation is symbolized in-Figure A, which illustrates the com-..ponents of the PDE method. In

applying this method, a detailedProgram Development Evaluation Planis created to'gether with implemen-ters. of each project. The manage-ment plan for the program is compu-terized,:,..andis updated every six

weeks. Quality control checks onthe implementation of the plan aremade frequently by project implemen-ters, aryl information on the accom-

plishment of key performance stan-daFds, objectives, and goals isentered 'into the computer."' Informa-

ral of program development unwinds.-information feedback is used toimprove .the prevention programs andthe ways they .are managed.

' We have attempted'to.apply theforegoing method with all of theprojects invplved in the AlternativeEducation Program. We have, of

course; met with mixed. success. The

magnitude- of the task of workingwith ,so many projects With extremelylimited' resources has meant that

even in working with those-pr.ojectsmost eager to implement this methodwe have not implemented it nearly. as

tion flows-from the-projects-to-the_ fully_as would be 'desirable.

researchers and back again as a spi-

.,

4

-13--

2?

Ad Update on the Measures used in the School Action Effectiveness Study

Measurement is a central compo-nent.of sound program developmentefforts, and measurement is essen-tial in program evaluation. Thischapter is a gu de to using andinterpreting m asures of. school cli-mate, individu 1 psychosocial devel-

''''Opment,and d linquent behavior thatare used in the Alternative Educa-tion Evaluadon. It serves as amanualto help readers interpretresults of the School Action Effec-t ivemes-&-StUdy'

Measuring Individuals and Organiza-tions

A two-tiered set of measures areused to assess the outcomes of theAlternative Education Program, Onetier assesses the characteristics ofindividual students and individualteachers ehq are relevant to organ-izational climate, or to importantpersonal outcc,,:ies. The other con-sists of school-level climate meas-ures that directly assess someimportant dimensions on whichschools vary.

r

The psychometric. work'reportedhere was sponsored in part by agrant from the National Institute ofEducation, U.S. Department -of Educa-tion. The opinions expressed doinotnecessarily reflect the positions orpolicies of any agency. This chap-ter is abridged substantially, butcovers some material covered in thefirst interim.report, and it reportson improved Methods of presentingresults. Material presented in somedeitail in the first interim report.,is repeated here because it isnecessary far an understanding of-the results presented elsewhere inthis report. ,''

The measures are divided intothese two classifications for animportant reason. We have allexperiencyd differences in the psy-chosocial climates of different.organizations, and we can easilyappreciate that organizations differin the environments that they pro-vide. Yet we also know that differ-ent individuals often have differentviews of the characteristics of thesame organization. Therefore, inassessing a given climate,it isimportant to average across manydifferent reports - -in essence .treat-.ing individual differences as error.These differences are, however,every reason we measure individua s.Accordingly, two `distinct sets of.measures are called for. Besidesthe general climate.lassessments,individual measures are needed forpersonalizing instruction and forcomparing the effectiveness of.alternative educttional treatmentsreceived by some people in a given,school or community.

The measures described hereweredeveloped specifically for theSchool Action Effectiveness Study(SAES) because no' comprehensive andpsychometrically adequate batterywas available elsewhere. They arerooted directly in a program ofresearch on delinquency and schoqlenvirowments conducted over the\pastseveral years at the;dohns HopkinsUniversity. The development of theinstruments used was'.guided in partby an-examination of instrumentsused in the National Institute ofEducation's (1978) Safe SOlooiStudy, instruments suggested by Foxand associates (19/4), the SchoolInitiative Evaluation questionnaires(Grant, Grant, Daniels; Neto, & .

Yamasaki, 1979), and a number ofother instruments,use.4 in major

-15-

28

Measures

social surveys or for indiyidualassessment. in recent years. Rele-

vant items (with necessary modifica-

tions) from other devices are some-times used.

Decisions about useful measuresape based on a review of the goalsand objectives of the OJJDP Alterna-tive Education Program and of thevarious alternative education proj-ects being evaluated, on currentdelinquency theory (Hirschi, 1969;Gold, 1978; LeMert, 1972; Greenberg,1977) on Gottfredson's (1983b)account of some implications ofdelinquency theory and strategiesfor organizational change. Many

discussions with prevention projectpersonnel--using the Program Devel-opment Evaluation framework--of thegoals and objectives of their parti-cular delinquency prevention effortscontributed greatly to the formula-,

tion of the measurement needs.

Some Essential Psychometrit Concepts

In order to use the measuresabout to be described in an informedmanner, it is important to under-

stard several ideas: (a) the rela-tive nature of psychosocial measure-ment, (b) reliability and (c)

construct validity'.1 The following .

paragraphs review these ideas.'

Relative Measurement

We have.few absolute measures inbehavioral science. In other words,

simple counts of "units" of acgieve7

Ment. or interpersonal competency orfairness or delinquencyare impossi-

ble to obtain. Instead, we typi

cally express their levels in rela-tive terms. For example,achievement test results are often

presented in terms of percentile-

1. For more thorough discussion seeThorndike (1971).

rank or stilidatA-score form. These

forms of expressic;iiinvolve state-ments of the standing of an indivi-dual (or organization) relative tosome norm group of people (or organ-

izations). For example, a percen-tile rank of 76 on an individualtest would mean that out of 100.individuals representative of thepopulation on which the test's norms.are based, 76 persons would have a

score lower than this one. We useboth percentile ranks and raw scoremeans and standard deviations topresent results. (The mean is the

abnarithmetic average of a set of'scores, and a standard deviation is

a unit of dispersion or spread.)

In interpreting such scores it isimportant to bear in mind that theyexpress stores relative to other '

scores in the study. sample. Differ-

ent samples of people or of schools ,

will differ somewhat in their meansscores (and also in their dispersion). Therefore a score that is,for example, at the 65th percentilerelative to one norm group could beat the 30th percentile relative toanother norm group.' There is nosuch-thing as a magically "correct"or even "most appropriate" norm

group.

,Please note that the psychometricuse 'of the word "norms" has little

or nothing to dowith some everydaylanguage uses of the word. In

everyday language we sometimes use"nom" to mean an idehl or requiredstandard. It is quite posSible for

a school to have students who show

an "average" degree of satisfactionwith school but who are ratheruncomfortable7-or who are average inreading achievement according tolarge city norms, but who do, iot

read well at all. In.inteOretingany partiCular results, readersshould probably consider' both their'own "ideal" norms and the "statisti-

cal" norms presented here.

Reliability

Chance, sloppiness, ambiguity,temporal instability, and hetero-geneity of meaning or interpretationcan influence any measure. Measure-ments of the distance between Balti-more and New York made by the odome-ters in a number of different carswould tend to agree pretty well, butnot.perfectly. ,They would havehigh, but not perfect, reliability.Reliability-is_ a technical term usedto describe the relative contribu-tions of measurement error and"true" scoreTiA iability to a scale

;or other measur . Technically,reliabilitYs,the proportion of thevariance (a statistic summarizingvariability) that is not error tothe total variance in the score.Because there are many ways of .

adefining error, there are many waysof estimating reliability (Stanley,1971)..

The reliability coefficientsreported in this chapter are of twokinds.. One kind is based on theanalysis of items administered on asingle occasion and therefore L_...._,

.4 excludes temporal instability from /. the definition of error. They can/.

be interpreted as an index of .howwell the scales measure whateverthey measure at a given_point intime. This kind of reliabilitycoefficient is called 4 "homogene-ity" coefficient; we estimate itusing coefficient alpha. -' The secondkind is based on the stability ofscores over time._ We estimate it bycorrelating scores obtained by indi-viduals or schools in the Spring of1981 with scores for the same indi-viduals or 'schools obtained in theSpring of 1982. This kind of reli-ability estimate is called a."retest" reliability; it is a meas-ure of the stability over time of ascore.

Knowledge of the reliability of a

1

.

MeasuresI

.I

test or other index isiimportantbecauSe a lOw homogeneity coeffi-,cient means that the device does not/measure anything well. A:high homo-\geneity coefficient means that the- device measures something. (What

that something is, is" what construct'validity is all about.) Homogeneitycoefficients can range from 0 to.0- A reliability of 1.0 is high,'meaning that the score contains no

.'error. A high retest reliabilitymeans that a stableicharacteristicof a person or organization is beingmeasured.' High retest reliabilitiesmay mean that (a) the characteristicis resistant to Change, (b) that theenvironment is pre ehting the indi-vidual or organiza ion to change, or

-17-

that (c) nothing has been done tochange the characteristic..

Over the years practitioners havedeveloped rOes'of thumb for accep-table levels of reliability for dif-ferent purposes. / In general, it is.

not sound practice to use tests withreliabilitiesmup below -a or .8for individual, diagnosis,: personneldecisions, and so forth. This ishecautie one wound want to be reason-ably certain that a score is reason-.

k

.ably error -free when making impor-

tant decisions bout individuals.When interpretations of patterns orprofiles are .tq be made, it is espe-cially importat that reliability bethis high, or higher.

For evalua I

ion purposes, lowerLevels of relfability. of Measurementat the individual. level .are accepta-

ble and are sometimes to be prefer-red, becabsebf three related_con-siderations. First, because. the .

scores of may y individuals are usu-ally .average 'in an evaluation, /

dependable e timat6s of true- scaremeans can be obtained even withrather unre fable individual ,meas-ures (see Sianley, 1971). Second,'

the longer the scale (i.e./( the more

I

3:0

1

1

Measures

:items), the more reliable.it is,other things being equal, but it isoften Aifficult, time consuming, orcostly to administer long scales.As an alternative, using shortscales with many persons gains goodestimates of group means. Third, in

an evaluation it is necessary tomeasure many things. This isbecause pr)vention programs havemany goals and objectives, andbecause it is always wise to searchfor-unanticipated positive outcomesor side-effects of a program. Butadministering many, highly. reliable(i.e., long) scales is prohibitive.Fortunately, a large test group .again tComes to the- rescue. Usingshort 'scales with'many.people solvesthe problem and yields satisfactoryestimates of true-score means.

As a tule of thumb, scales withTtliabilities as low as .5 (or evenlower) are adequate for use in anevaluation, provided that the proj-ect being evaluated uses randomiza-tion as a selection device, or thatany selection'is absolutely indepen-dent of (i.e., unrelated to). thegoals or objectives of the program.In suchah evaluation, it is notneceSsaryto attempt to adjust forpre7existing or spurious group dif-ferences on outcomes. When it isnecessary to make'such adjustmentsby using statistical "controls,"reliabilities for the control variatles must be as high as possible.The rule of ,5 is too lax in thiscase because when the "control"Variables are unreliable they do aninadequate job of correcting forspurious.differences-between groups.Therefore, to enable a saund evalua-tion, a project which does not ran-Adamize should use more reliable(i.e.,longer) scales encompassingmeasures of all relevant character-istics in.whichithe treatment andcomparison groups may differ.

6

0

Validity

Validity has to do with the mean-ing and interpretation of an indexor score. The exploration of mean-ing is a.never-ending proceSs,because it is, so closely linked withtheory. Theory involves constructsor ideas about the causes or natureof phenomena, Often, Measurementhas meaning only in the context ofsome theory.' For example, some edu-cators have a theory thata.generalability called intelligence under-.lies much human performance, or at

. least scholastic achievement.' The. measurement of intelligence using apaper and pencil verbal ability testmay make sense in terms of thistheory. Because the theory predictsthat this test will correlate withschool grades, evidence about thevalidity of a test for measuring theconstruct of. intelligence can comefroean'examination f the empiricalrelation between test scores and .

school grades. The same evidenceprovides information aboUt the util-ity of the theory. Theories andmeasures 'are thus validated in acommon process. We speak of a testas validated when'eMpirical evidencehas in general shown, the testresults to follow the predictions of ,

a theory that has been uSeful.

-18-

In addition, when there is agree-Ment about what a construct means,some :evidence abbut validity cancome from an examination of the itemcontent of a test. For example,most of us would probably agree thata test to see how many bricks a per-son can-load on a truck in an houris a poor test of verbal ability,and that a, list of multif le-choice.vocabulary items would*provide.amore valid measure of that con- '

struct. (Similarly, the vocabularytest would be a yoor test of endu-rance.) .Therefore, deliberatelyincluding items to measure a givenconstruct in itself can provide some

31

limited degree of confidence in ascale's construct validity.

The eviden.ce is strengthened ifthe scale shows expected patterns ofcorrelations, with other scales. And

it is especially strengthened if-applicable experimental manipula-tions influence scores in predictedways. Other evidence of validity,can come from an examination of dif-ferendes in scores on the scaleamong groups known or believed todiffer in the characteristic beingmeasured. For widely used instru-ments, these kinds of evidence accu-mulate over time. Eventually, abasis for judgment about a scale'sconstruct validity emergesalthoughdifferent judges often disagree.

41/

Subsequent' sections describe theorigins, development, and some psy-chometric properties of a two-level

set of assessments of schools and. their inhabitants. These sections

. are intended to provide informationaboutreliabilitY and validity, andto describe the normative interpre-

. tation these assessments.

Measures of Students

Five sets of measures of indivi-dual students have been developedfrom di&CUssidne, with the staff ofdelinquency prevention projects.about, the problems they faced andwhat they hoped to accomplish, thedemands of evaluating'a program withoverarching'goals of school organi-zational change and delinquency pre-vention, and the. theory of .detin-

.quency prevention. Ihese measuresof students are needed to assessproject effectiveness under diffi-cult field .research conditiobf andto learn more about what works for.whom.

4:*

Measures

Social Background

Measures of social background orfamily 'Characteristics are neededfor two reasons: (a) .They provide

essential statistical controls toaid in, demonstrating project effec-tiveness when evaluation designs:**calling for. statistical adjustmentsare necessary, or when strongerdesigns fall apart. (b) In a fewcases, Projects aim tooalter familycharacteristitsusually the extentto which'parents value education andencourage their children to performwell in school.

Accordingly, the following twomeasures were deVeloped: 4

Parental Education. This tiro-

item scale 'is based on decades ofresearch that show parental educa-tion to be a powerful antecedent ofschooling outcomes, especially ofpersistence in education (Sewell,Hailer, & Portes, 1969a). The twoitems ask how much education a stu-dent's father and mother completed.The scale has a' reliability coeffi-cient of .78 overall, although thecoefficient for the smallgr'oup ofNative Americans in the sample isonly .51.2 Table 1 displays the-

2. Somewhat More detailed resultsof the examination of homogeneitycoefficients for measures employedin' the'first.year of this evaluationare presented in the first interimreport; ; -Some improvements in themeasures afe reflected in resultspretkentedvhere..: Homogeneitycoeffi7cienls reported here:Wre-Calctilatedfrom a 10% sample Af..the.whites, a.10%sample.of Blacksa 10% samplefor the combined groups, a 20%.sam-ple of mainland Hispanic youths, anda,100% sample of all other groupsmeasured in the Spring of 198.2. '

Reliability coefficients are re-es-c,1/4

-timated here because of a new scor-n

719-

32

t!.

Table 2

Reliability Coefficients (Alpha) for Individual-Level Student Scalesby Gender

ti

Scale Male FemalesTotalSample

Numberof items

Family background ..

Parental ecication1

'76 72 78 2

Parental emphasis on education,

° 57 51 .50 .4

Social relationsAttachment to parents 61 60 60 6

Negative peer influence- 63. 67 65 .9

-Nt-tttudes-and-social-development-AllenatiOn 60 44 51 ,6

Attachment to school .76 75. 76 10

Belief in rules '52 54 53 6

Interpersonal competencyInvolvement'.

4

Positive self- concept

4360'8

47 ,

62

60

4262'

61

5.12

L2

Practical knowledge, 73 75' 75 7

Rebellious autonomy 49 49 47 3

Internal control 58 56' 52 7.

Behavior .

School effort_ 62 56 59 5

School non-attendance. 61! 62 ,61 2

Self-reported delinquencY.(total) 84 85. 85 19

Self7reported drug use 78 77 /5 5

Self - reported.. serious delinquency 77 80 83 11

School experiences.School punishments 54 53 54 . 4

School rewards 63 58 . 56 4

Victimization , 58 73 69 7

ValidityInvalidity 44 45 44 5

Note: Decimals are omitted.

34

,,

measures

scale's hbmogeneiy estimates forsix ethnic groups.;.Table 2 displaysthe scale's reliability estimatesfor males and females and for thetotal (combined) sample. TWis,meas-ure may be taken as an indicator of'family .socio- economic status. It is

known to be a good Oredictor ofschOoling outcomes s'uch as persis-tence and grades (Bachman, Johnson,& O'Malley, 1978; Jencks, 1979), butit is usually only weakly related todelinquent-behavior at the.indiVi;dual' level--although perhaps it hasa stronger relation to more seriousdelinquency (Tittle& Villimez,1978; Gottfredson, 1981a).

Parental Emphasis on Education.This four-item scalb asks for infor-mation about the degree of parentalattention to the student's school .performance and parental expecta-tion for school persistence. It

was suggested by prevention projecttheories that attributed studentnon-attendance to a lack cf parentalencouragement or "value" on educa-tion. And, parental influence isdemonstrably predictive of studentpersistence in school (Otto, 1976).The scale is only moderately relia-ble--.50 overall, with homogeneitycoefficients ranging froni .45 to .57for, race-sex subgroups. The scale

has moderate negative correlationswith'self-reported delinquency (seeGottfredson et al., 1982, Table 4),'and has an expected, but small,positive correlation withstudentreports of effort 'spent on sch6o1

work.

Social Relations

Three easures of a student'ssocial relations were developedbecause of (a) empirical and theo-retical links between bonds of Y.

affection or respect for others andconforming (non-delinquent) behav-ior', (b) powerful statistical asso-ciations,between delinquent behavi9rand'delinquent peer influence,(c) the central place given to peer ,

-influence-in-the-theories-of_several_of the prevention projects, and

.

(d) the explicit assumption made byseveral projects that parentalsupervision governs student atten-dance.

Attachment'to Parents. Thisscale, intended to measure Hirschi'sconstruct of the same name, incorpo-rates several items closely relatedto items shown in earlier studies tobe, correlated with delinquent behav-ior (Hirschi, 1969;Hindelang, Hir-schi, & Weis, 1981; D. Gottfredson,1981b)A An attempt has been made'toengineer a potent scale by includingsix items related to this construct.The scalq, asking students bow closethey are to their parents, how .muchthey like them, and so forth, has anoverall reliability of .60. It,_-Feor-

relates as expected with self-reported delinquent behavior (see.Chapter 4 (this volume) and Gott-.fredson, Ogawa, Rickert, & Gottfred-App,1 1982), in accord with Hirschi's

( 19,69) theory thatattachment,,toparents creates a stake in conform-

ing behaVior. This agreement pro-Vides some evidence of the constructvalidity of this scale.

A 1

Negative Peer Influence. This

Scale measures a construct centralto the.. explanations of delinquencyand non attendance formulated' byseveral of the action,projects. It

:is.rooted'directly in earlierresearch (summarized by Empey; 1978)that shows delinquent. peeraSsocia-

c.

ing procedure implemented in yeartwo to increase, the interpretabilityof the results, and because somemeasures (Alienation and. InternalControl) were lengthened by addingnew items. /

.0

3 S. -22-

C

tions to bq powerful,predictprs ofdelinquent involvement. In addi-2,tion,.it incorporates items relatedto dropout, similar to those used inearlier studies of persistence inschooling (Bachman et al., 1978).It is au attempt to engineer a long,powerful, and broadrbased measure ofnegative peer'influence. This

nine -item scale has reliabiltiew:ranging from .55 to .70 across BO-groups and, it is a potent correlateof delinquent behavior (Chapter-4,this volume; and Gottfredson, p82).It contains items asking, whether thestudent's best friend is interestedin school, thinks getting goodgrades is'important, thinks schoolis a pain, or has been involved indelinquent activities.

Attitudes and Psychosocial Develop-ment.

Psychosocial development is amajor goal of the Alternative Educa-tion Program. In this area, therewas considerable prior work to buildon in choosing measures/;to includein the battery.

Alienation. The.six-item Aliena-

tion Scale is based in.part onSrole's (1956) Anomla Scale, butfewer items are included, and, the

wording of, items ha's been changed togive them more school-related con-tent and to make them sound a littleless bizarre. Alienation items usedin the School Initiative Evaluation(Crane 'et al., 197) and in otherprevious studies were modified foruse here, Items include, "Thesedays I get the feeling that. I'M justnot a part of/ things." And, "I feelno one really cares much about-whathappens to me." Overall, this shortscale has a reliability of .51.

(The reliability is improved overthe 4-item version used in 1981.).As expected, the scale correlates,positively with self-reported delin-quent behavior; and negatively with

-23-

36

Measures

reports of effort expended on schoolwork (see Gottfredson et al., 1982, -

Table 4; and Chapter 4, thisvolume).

Attachment to School., This is a

central construct for many projectswhose major goel-Or objective is thedevelopment of positive studentttitudesjtoward school. The con

struct iaalso central to socialcontrol theories of delinquency'Olirschi, 1969) that,view-attachmentto school as a major social, bondrestraininijndividuals, frOm_parti7cipation in delinquent behavior.Consequently; we'have constructed arelatively long and broad -basedmeasure of attachment to school.This 10-item scale has.reliabilitiesranging from .66 to .82 across sub-groups--a6 overall. Items .ask the.students if they like the school, :ifthey like the classes, how-importantg t.ting good grades is, and sofirth.' ThqRscale is; aU expeCted, apowerful correlate of delinquentbehavior (negative) and effortexpended at sChaolwork (positive)(Gottfredson et al., 1982).

Belief. The expectation `1

individuals differ in the extent to,which they believe in the moral val-'idity of. conventional social rules,.and that the degree of belief influ-ences behavior, is widely shared., Acommon goal of peer-group7based"interventions to prevent delinquencyis to strengthen belief. by -usingpeer pressure. The item content ofGough's (1964) Socialization scale(which was developed through empiri-cal effbrts to discriminate betweenadult offenders and non-offenders)lends support to this popular,notion. And, belief is a centralconstruct in social control theory,which postulates, that people differin the degree to which they have.internalized rules, and that theytherefore are constrained frominvolvement 'in delinquent behavior

Measures

to different degrees. Muchempirical evidence supports thisidea (e.g., D. Gottfredson, 1981b;Hirschi, 1969).

Consequently, in orders to measure

this aspect of psychosocial develop-ment. we have assembled a short scalefrom well-worn items used in otherresearch, whose characteristics wi.!reknown. The six-item scale contar.nsitems such as, j'It is all rightget around th{law if you can;""Taking things froth 'stores 'doesn'thurt anyone;" and "People who leavethings around deserve it if theirthings get taken." The scale has a

reliability of .53 overall; itsreliability is lower for the Span-ish-speaking and Spanish surnamedsubsamples, and'higher for the othersubsamples. The scale has a.sub-.stantial,negative correlation, withwith delinquent behavior (Gottfred-son et al., 1982; Chapter 4, thisvolume), as earlier research andtheory imply it should.

Interpersonal Competency. This

scale is composed of four items fromHolland and 'Baird's-(1968) InterpersonaL Competency Scale. It consis-tently has moderate reliability andcorrelates positively with othermeasures.of psychological health Oradjustment, and.negatively withmeasures of alienation. The fifthitem was written by Holland espe-cially for the present purpose,- togive the scale more school-relatedcontent. It has a reliability coef-ficient overall of- .42. This -mess-.

ure correlates positively with .

reported ,effort expended onwork, and it is nearly independent(bncorrelated with) self-reporteddelinquent behavior (Gottfredson et

1982). This accords with otherevidence that delinquent involvementis only modestly associated with'psychological health (Waldo'& Din-itz, 19671 cf. Quay, 1964).

Involvement. This scale' is

intended to measure a central con-struct\in social control theory thatdoes not appear to have been wellmeasured in the past. The ideathat involvement in conventionalactivities creates a stake in con-formity, because a person involvedin rewarding activities has some-thing to lose. by misconduct. This

,'scale (not to be confused with envi-ronmental measures of student-influ-ence or involvement in decision-mak--ing.)-is-CompoSed-Of-12 items-(most.-of which were adapted from therecent National Longitudinal Studyquestionnaire) asking about a stu-dent's participation in a wide vari-ety of in-school activities. It hasan overall reliability of .62, butdoes not correlate as expected withreports of delinquent behavior,casting some dOubt on its constructvalidity or on the _utility, of theinvolvement construct in theories ofdelinquency. Although this scalewas intended to serve as an impor-tant intermediary outcome measure,its utility is in doubt.

Positive Self-Concept. A number'of self7esteem scales with well-re-searched'properties are available(Robinson & Shaver, 1973, reviewmore than 30 measures). To create ashort' scale, items previously usedby Rosenberg (1965) and an itemsimilar to one used by Coopersmith(1967) were subjected to analysis .

along with another set of items con-structed to capture aspects ofself-concept specific_to schoolingand.delinquency. This scale also'isbased partly in the labelling per-spective XLemmert, 1972), whichimplies, that if people are treatedas slow'learners or delinquents,they will come to incorporate

.'aspects of those social definitionsinto their awn self-concepts. Posi-tive self-concept, therefore, is animportant intermediary outcomeaccording to labelling theory.

-24=

7

qMeasures

According to this perspective,. Peer Culture Development Project in

effective alternative education Chicago,. explamtiOnn of the problem

projects wouldindrease scores on of delinquency sometimes involved a

the positive self-concept scale, and kind of peer or. gang culture that

a program with unexpected negative resembles Miller's (1958) character-

side-effects could decrease scores. ization of subcultural socializdtion. The peer or gang culture may

Item analysis did not justify - incorporate a set of socially-shared

treating self-esteem as a separate expectations. that are different from

scale from these labelling outcomes, what might be called middle-clans

because items'' are about. equally cor- expectations. Differences may be so

related across the two'sets. Weak great that in behaving according to

items were excluded, leaving a the "lower-class" system a'perion

12 -item scale with reliabilities may violate norms of middle-class

ranging from -52 to .65 across sub- culture, and maY appear & be deli-

.groups,....61overal1.Items-include-, berately non-conforming or malicious

"My teachers think I am a-slow lear- to a "middle-class" observer. In

ner;" "Sometimes I think I am no particular, middle class concerns

good- at all;" "I am the kind of per-- with achievement may not be shared

son who will always be able to make by "lower class" youth (cf. Attach-'

it if I try;" and "I do not mind ment to School and Educational.

steeling from someone--that is just 44' Expectations). Inatead,.theae "low-

the kind of person I am.° The scale er- class" youths, according to _

correlates .48 with reported effort Miller are concerned with trouble,

on school work, and -.24 with self- toughness,.Smartness (i.e.,.manipu-

reported delinquency, and it corre- ,lative skill), excitement, fate

lates -.39 with alienation and ,39 "(explaining events by reference to.

with interpersonal competency (Gott- chande or luck), and autonomy (an

fredson et al., 1982), landing sup- ambivalent relation-4o author -'

port to its construct validity. ity--overtly desiring not'to bepushed around but covertly.desiring-

Practical Knowledge. To provide to be cared for and controlled).

a simple measure of self-reportedCompetencies'needed for coping with . Because of this recurrent' theme

everyday life, a seven-item measure in our discussions with action proj-

was created for the evaluation. ect persohnel, it seemed important

Althougll, this'self-report scale may to incorporate brief measures of

be a poor substitute for .a more com- this type of "subcultural value

prehensive or task sample approach, system. Item analyses of a largeriit seemed the only way to build a set of items implied that three of

measure of this kind of social these items formed a scale.for all

development into a multi-purpose race-sex subgroups. The deletion'of

battery. The scale has a reliabil- pooritems, however, narrowed the '

ity coefficient of .75' overall, and content of the set down to items

good item properties across. all"' that appear'to reflect a rebellious,

groups studied.- It is relatively autonomy: "Whether or not I spend

independent of the other measures,, of time on homework is my own busi-.

attitudes and behavior. 'Because it ' ness;" "I should not have to explain

has not been well-studied, it should to anyone how I spend.my money;" and

be interpreted cautiously. "I don't like anybody telling mewhat to do." The scale has a reli-

Rebellious Autonomy. In talking ability of .47 overall. The scale

with persOns running the delinquency correlates as expected with Delin-

prevention projects, especially the quentBehavior and Belief

O-25-

38

Measures

(Gottfredson et al., 1982; Chapter4, this volume).

4

Internal Control. A number ofthe prevention projects view delin-quent behavior as a result of weakinternal controls, that is of a perHon's sensor of powerlessness overthe environment. If what one doesmakes little difference for therewards, punishments, or achieve-ments one experiences, then one isfree to engage in unrestrained,self-gratifying delinquent behavior.These speculations appeared to us tobe related- to Rotter's (1966)notions of internal and external,control. 'Accordingly, we attemptedto include a small number ,of items

related to this construct in surveysconducted in the first year of theevaluation. Item analyses did notsupport the utility of.'a scale basedon those items, and renewed attemptsto develop a suitable measure weremade in the second year. The scalethat finally emerged has an overallreliability coefficient of 62, andworks reasonably well for each raceand sex subgroup. A sample item is,"Much of what happens to me is justa matter of chance."

Self- Reported Behavior

At bottom:it is the behavior ofthe young people subject to theinfluence of the Alternative Educa-tion Program that is important.The measurement of behavior istherefore essential to' the evalua-

tion. One source of informatiori

about the behavior of individuals isthe archival records that are ymin-tained in various ways by schoolsand criminal justice agencies.Those archival records are, however,'subject to many limitationp: They

vary in completeness, 'accuracy, and4 availability. Different behaviors

are recorded in different places,and they are recorded in different,ways. And, official records measurenOt only the behavior of the people

-26--

who are their subject, but also thebehavior of school and just ice sSea-tem officials who make decisionsabout what to record.

44 Accordingly, to provide for thesystematic measurement of behaviorin parallel ways for all of the pre-vention projects, we have developedseveral self- report measures. These

, self-reports are also subject tolititations, but they Ao make rapidanalysis possible, they are parallelacrossall projects, ancli)reviousresearch generally( supports theiruse (Hindelang et al., 1981).

ho .1 Effort. That 'students 'who

low grades in school tend torop out of schOol and to engage in

*delinquent behavior more than othersare two of the best documented andconsistent findings,in the*litera-ture (D. Gottfredson, 1981). Social

class and ability are modestly asso-ciated with these same outcomes butdo not completely account for theseasiociations. Therefore, it seemslikely that these outcomes are determined at'least in'part by grades--the,major, if infrequently applied, o

reward system of traditional school-ing. Grades in schoOl are not det-ermined solely by ability and socialClass, of course.' Industrial psy-chology's instrumentality theory(Porter & Lawler, 1968) suggests amechanism whereby effort is expendedif valued rewards are perceived asattainable, and in which effort is''one of the determinants of both per7formancp and rewards. :Thereforeeffort is an important intermediaryoutcome variable that should beassessed in the evaluation of a pro-gram designed toprevent delinquencyand foster persistence in-schooling.

Unable to locate existing ques7tionnaire measures of this con-struct, we developed one. This

dive -item scale has a'reliability of.59 overall. (It is somewhat lessRreliable fot Puerte44fan subsample,

39. ,

preaumbly because some of its itemcontent deals with homework, whichis rarely assigned in Puerto Rico.)The scale includes these items"Compared to other students, howhard do you work in school?"; "Iturn my homework in on time"; and "Idon't bother with homework or clapsassignm6nts." As expected, femaleasure hiaer on this scale on theav,rage than do males. It corre-.

.lates.%%19 with selfreported gradesand .34 with attachment to school(Cottfredson; et al.4982), sup-

. porting its interpretation as ameasure of effort expended on, schoolwork.

School Non-Attendance. TheAlternative Education Program isintended to demonstrate and evaluateprojects that aim .to increase atten-dance. Dependable attendance dataare not always available from schoolrecords, so a brief self-reportmeasure of attendance was incorpo-rated in the questionnaire to pro-vide back-up data. This dediSicnproved to be' wise. Attendance datafrom school records proved erratic,incomplete, error-ridden, and slowin coming., .

Two items, one asking =how often--the-sttident-cuts school all day and

one asking about class skipping,compose this brief scale, with an

.overall reliability of .61.

. Self-Reported DePinauencv(Total). One'way to'find out Whatpeople do:is to ask them. Natu-'

rally, not everyone tells the truth,perhaps especially when the quei-tions are sensitive. A commonassumption is4 that peoplewillcon-ceal information about their parti-cipation in illegal behavicir,:and sounder-report. At.the same time, therate of delinquent behavior esti-mated by the self-report method arehigher than those derived from offi-cial records (Empe' & Erikson,

Measres

1966)., There is thus, ii'-e,reat deal

of dqbate among criminologidts about,the appropriate way to measure crimeinal behavior.

Although' thereis no need to gointo the'arguments in any detailhere, a major issue is that typicalself-report measures (e.g.,, Nye,1958) tend to measure minor"offenses," some of which are not"crimes," or would not be crimes'ifcommitted by an adult. Elliot andAgeton (1980) have recently pre-sented evidence that self-report .

scales involving more seriousoffenses tend to resemble measures"-based on official' data more than doscales involving only 'trivial items.Hindelang, Hirschi, and Weis (1981)have recently, ,published a disquisi,7'

tion on the measurement of delin-quency by self-report'and officialmeasures.

The bottdm insofar as ,it

can be perceived at present, is thatfairly long, variety-type scalesinvoLving a rarigeof serious delin-quent behavior do produce resultsthat parallel official records forsome subgroups but-not for othera.Hindelang,et al. (1981) report val-idity coefficients for a numberalternative measures thatdmply verylow validity of self-reported datafor officialOr'"delinquent" blackmales,- and mach better validity forother subgroups. This is a 'cliffi-

cultY that shoUldbe kept inmind-ininterpreting these self-reporteddata.

1. 6

3. This difficulty appearstoa similar problem of di entialreliability in studies of edues-tional persistence (Bielby, Hauser,& Featherman, 197; D.' Gottfredson.1981a), and this potential problemincreases the importance of obtain,ing bffitialdata for purposes ofevaluation. Thedifferential Vali2F,

Measures

The specific self - report measures

used here are modified from thoseUsed by Elliot and Ageton (1980) andby Hindelang, Hirschil, and Weis Op(1981). Many of Elliot's items %ereused, but a "last-year variety"scale format was used because theHindelang et-al. (1981) results sug-gested the usefulness of this for-mat. These items, ask, "In the last'year haveryou...",'Respondents indi-,cate, for example, whether they have

4 "stolen or tried to steal somethingworth more than $50."

(-

A 19 -item scale constructed inthis way has very nice characteris-

°- tics-,considering,, that only a small

116foportipn of respondents answer yesto any given question. erall,

reliability is .85. The ubgroup

reia4 bilities range from 83 to.88

Self-Reported Drug_ Involvement.Prevention project personnel haveshown considerable interest in acomponent' of delinquent behaviorinvolving drug use. To provide a'measure teeet their needs, we havealso leored'a five-item subset ofthe longer (total) S-It delinquencyscale. It is composed of items ask-ing about the, use of cigarettes,

Wy-Tproblem is discussed in Part II

of the present teport (Daniels &Gottredson, 1983) but is notresolved.

4: In our first interim report we. -

estimated reliabilities for morenar -rowly defined subgroups. The

single lowest coefficient was for'''Asian-American females, who report

almost no delinquent behavior. The

reliability for that group vas .63.These reliabilities compare fay(4a-bly to those obtained by HindeLAPget al. (1981) with a 63-item last-year variety scale -.83 to .92 for

black and white males: and feM'ales:

I

liquor, marijuana,-and other drugs,and about going to school "high."(A sixth item about glue, sniffingwas left out because the analysesdid not support its inclusion forall ethnic groups.) This group of

items closely resemblesthe Hinde-lang et al. (1981) Drug Index. It

has an overall reliability of .75.

Self-Reported Serious Delin-quency. A second subscale was con-structed to measure only conductthat nearly everyone wouldjregard ascriminal. It includes 11 items

1tincluding one about sellilipdrugsthatThindelang et al. would place inhe drug cluster) and has an overall

.reliability of .83.

Measu es6 of School Experiences

It is anticipated that the proj-ects in the Alternative EducationProgram will,ex0and the range ofschool rewards, beyond those repre-sented by traditional classroomgrades. Accordingly, in an effortto assess this important but hard-to-measure set of outcomes, we havedeveloped two scales to measure'stu-dents' rewarding and punishingexperiences. School rewards andpunshments make sense intuitivelygs probable causes of school attach-ment, effort; and persistence.

-28-

One kind of.school experience isorspecial importance: victimiza-

tion. A .key measure of the successof the delihquency prevention projects under study is the level ofpersonal victimization experiencedby perspns in thoite schools.Accordingly, victimization experi-ences must be measured to assess theeffectiveness of the projects, andto learn more about the'victimiza-tionexperience itself.

School Punishments. This four-item scale ie, an index of the nega-tive sanctions an individual student

41

experiences. It asks whether thestudent was required to stay afterschool, given an extra assignment,or had his or her grade lowered as apunishment. Its reliability coeffi-cient for the total sample is .54.According to this index males exper-ience more punishment, as expectdd,and the scale correlates .30 withself reported delinquency, -.28 withpositive self-concept, -.30 withbelief, -.22 with school effort, and.24 with negative peer influence(Gottfredson et al., 1982).

School Rewards. This six-itemguile is 'an index of the positivesanctions an individual studentexperiences. It. includes reports ofincidents in which the teach d com-plimented the student's work, thestudent was given a prize or, award,or the studelt won an award for hisor her class. The reliability coef-ficient for the entire sample is.56.. The scalesis relatively inde-pendent of sex, and is correlated.25 with school attachment (Gott-freddon et al., 1982).

Victimization. A final measureof school experiences deals withpersonal victimization. It is

intended for use in assessing theamount oCtrime in the environment,and it is used in he aggregate tocharacterize the school. The Vic-timization Scale is also intendedfor use in research on the victimi-zation experience. The scale'scharacteristics at the individuallevel are therefore of interest.Containing five items, the scale hasa reliability coefficient of .69.Victimization is correlated .24 withself-reported delinquency, implyinga moderate tendenfor persons whoare victimized to engage more indelinquent behaVior themselves; Itcorrelates -.27 with school attach-ment and -48 with self-esteem; itshighest correlate among ,the varia-bles examined is punishment

Meagures

(.35)--students who report morefrequent personal victimization alsomore often report being punished inschool (Gottfredson en al., 1982).

Quality Control

There iivalways some concern thatstudents may not faithfully completetheir questionnaires, that they mayfool around or give silly answers.\As a check on this, a scale wasincluded to detect unusual pr'non-\sensical responses.

1.

Invalidity. This five-item scalei.s composed of items that a carefulrespondent would answer in only oneway. I& is keyed so that <a rareresponse earns a point. This*scaleis used as a check on the resultsand as a quality control mechanism.Invalidity scales are intended notto measurea reliable characteristicof individuals and hence usuallyhave low reliabilities. The overallreliability of this scale is_ .44.

Stability of Student Measures OverTime

One-year re-test reliabilities of ----each of the measures of studentcharacteristics described above arepresented in Table 3. Th,ese stabil-

ity coefficients provide informationabout the degree to which young peo-ple tend to retain their relativestanding on these measures from yearto year.

-29-

Re-test reliabilities for severalmeasures not already 'described arealso presented in Table '3. Thesepersonal 'characteristics were meas-ured using single. items (so it isnot possible to calculate homogene-ity coefficients. They aredescribed in the following list.

Educational Expectation. An itemasked students how far in schoolthey expected to g4. The response,

42

*

rJOne-Year Retest ReliOilities of Student Characteristics

Table 3

Student characteristic

Family background

Males

rxxN

Remales

rxxN

Parental Education .70 546 .72 626

Parental Emphasis on Education .34 373 .39 471

Social relations.,.Attachment to Parents .38 879 .47 1007

Negative Peer Influence .44 849 .39 1007.

Attitudes and psychosocial developmentAlienationa .33

lc674 .39 870

Attachment to School. .

.53 791 .46 '975

Belief in Rules .38 662 .40 888

Interpersonal Competency .32 602 .32 810

Involvement .37 747 .50 888

Positive Self-Concept .45 576 .50 798

Practic'al Knowledge .36 669 .43 893

Rebellious Autonomy .37 552 .40. 766

Educational' Expectation .48 959 .41 1081

BehaviorSchool Effort .46 851 .40 966

School Non-Atte3idance .42 969 :45 1081

Self-Reported Delinquency (total) .63 419 .55 584

Self-Reported. Substance Use .66 416 .60 583

Self-Reported Serious Delinquency .46 390 .30 . 563,

School experiencesSchool.Punishments .27 805 .32 .979

School Rewards .33 804 .32 982

Victimization .35 788 _ .23 961

Self-reported Grades .41 991 .52 1085

Validity indicatorInvalidity .32 677 .31 396

Note. Reliabilities calculated on a random half sAmple of students

who completed questionnaires in both 1981 and 1982.

aAn improved Alienation Scale was available in 1982. The correla-

tion reported is the correlation between this improved measure and

ta less reliable measure used in 1981.

-30--

43

.

Table 4,

Reliability Coefficients for the Iridiridual-LevelTeacher Scales, Number of Items in Each Scale, and

Scale Means and Standard Deviations

Scale

.

Alpha

,Prointegration Attitude

,// /

,69

Job. S'atisfact'ion/

.80

Interaction with Students .67

Type A Sanctions .47

Type B Sanctions .60

Victimization .67

Classroom Disruption .78

Low Expectations .57

Professional D4elopment .74

Nonauthoritar.ian Attitude ' .54

Note. Reliabilities and scale means and

N ofitems

4!-!

1

Mean

11.56

31

8.4;)

i

6 . 13.79i

5 .09

5 13.42

,

18 1:23

2 4.60

2 65.06

8 -.39 4

3 7.43

I-

SD's are based on

SD

2.88

1.70

4.20

2.82,

3.08

1.45

1.38n

44.47

4.67

2.17

results from the 19 &l Spring administration of these scales andare calculated on the "holdout" sample (see Cottfredgon et at.,1982), N's range from 555 to 643 due to item nonresponse.

-31- tO

Measures

which is intended to provide anindicator of commitment to, a conven-tional goal, has a re-test reliabil-ity of .48 for males and ..41 forfemales. Educational expectationsgenerally have substantial negativecorrelations with delinquent behav-ior (D. Gottfredson, 1981). The

correlation in a random half of the1982 survey data between this item-and.Self-Reported Delinquent Behav-ior was -.12 (p < .001) fO-i'boys and-.08 (p < .01) for girls.

Self-reported Grades.- We antici-

pated the potential necessity ofhaving a questionnaire-based measureof school performance to supplementdata collected from school records.Accordingly, a self- report of schtolgrades was included in the question-:

naire. This-item has a re-testreliability of Al for boys and .52for girls.

Measures of Teachers

The second largest group of inha-,bitants of a school environment arethe teachers who-workthere. Stu-

dents in the aggregate helletocreate an environment for the teach-ers, just 'as teachers, create anenvironment for the students. . Acharacterization of.the teachers istherefore important in describing aschool.

Several of the action projects'theories lead to ,interventions .

ge'aredtoward teachers. The inter-

ventions are intended to improveclassroom management, to changeteachers' attitudes, or to involvethem in new kinds of activities. L

One aspect of the evaluation there-fore involves the measurement ofteacher characteristics.

Pro-Integration Attitude. This

four-item scale is a measure ofattitudes toward integrated educa-tion. It is included because thesedelinquency and achdol-improvement

-32-

45

programs are designed to provideservices to heterogeneous groups ofstudent's. One component of several'projects is training teachers tomanage heterogeneous classrooms andto interact with a variety of kindsof students. It has a,reliabilitycoefficient of .69 (Table 4) and is'relatively independent of the otherteacher scales (see Gottfredson etal., 1982). As might be expected,nonwhites tend to score somewhathigher than white& on this scale.

Job Satisfaction. This scale is

composed of three of the four items,in Hoppock's (1935) scale of thesame name, which has been usedwidely in research. Even shortenedto three items it has a reliabilityof .80. It may confidently be takenas a measure of how well teachers

like their jobs.

Interaction with Students. This

six-item scale measures the extentof out- of'-class interaction that ateacher has with students. Items

ask about tutoring individual stu-dents before or after school and .

discussing their personal problemswith them. It his a reliabilitycoefficient of .67, and correlatespositively with Job Satisfaction,negatively with reports of classroomdisruption, and positively with theextent of recent continuing educa-.-tion activities.

Type A Sanctions. This is one oftwo scales developed in an attemptto describe the types of responsesto student conduct used by theclassroom teacher. We are unaware

of. any short questionnaire measuresof this aspect of classroom wanage-ment, but provocative evidence fromearlier research (McPartland &McDill, 1977; Gottfredson,& Daiger,1979) suggests that responses toconduct are important in preventingdisruption. Therefore we used thebest advice we could get to develop

lists of various ways classroomteachers might respond to studentbehavior. 'These lists becanie items

in the questionnaire. Through fac-tor analytic examination and inter-nal consistency item analysis, twoscales emerged.

The .first set of items ie.termed"Type A" Sanctions. A teacher whoreports lowering grades as a.punish-ment, sending misbehaving studentsout of class, and paddling or repri-manding the studentsin class isgiven a high seore. The scale has areliability of .47. Its largestcorrelate among the other teachermeasures is the amount of disruptionthe'teacher reports; it is also mod-erately negatively correlated withnonauthoritarian attitudes, "

Measures

responses to this item, whereasthose on Type A correlate only .07(n.s.)

Victimization. As one way tomeasure the amount a delinquent

ti behavior in d school, teachers areasked about their:experiences ofpersonal victimization. In the

ggregate, these reports may beItaken as an indicator of the amountof disorder in the school. An

eight-item_scale, asking about.,events'ranging from obscene remarksor gestures to physical attack, hasa reliability of .67.

Classroom Disruption. A second

way to assess the level of studentmisbehavior .experienced, by a. ,.acher

is provided by a two-item classroomdisruption scale.' It asks towhatdegree classroom disruption inter-feces with teaching, -and how much of,,the teacher's time is devoted tocoping with disruptive students.Its reliability is .78..

Type B Sanctions. This scale wasdeveloped in the same way. din con-

trast to the Type A scale, whichseems to include respousesrooted infrustration, Type B Sanctions appearto involve a wider range ofresources. To earn a high score on "-F,

this scale, a teacher reports givingextra schoolwork, awarding specialprivileges for good behavior, takingaway privileges for misconduct,calling par'ents, and referritg stu-dents to the counselor or elsewhere.This five-item scale has a reliabil:ity of .60. It correlates only .16with Type A Sanctions, even though.both scales would be-elevated if .ateacher' frequently had make somekind of response to "misconduct.Cottfiedson et al. (1982) examinedthe'construct validity of the twosanctions scales by examining theircorrelatiOns with responses to aquestion about home-based reinfor-cers., The use of home-based rein-forcers to extend the range ofrewards and punishments in theschool appears to be a highly effec-tive strategy (Barth, 1979; Atkeson& Forehand, 1979).. Scores on theType B scale correlated .35-with

Low Expectations. A labelling.

theory peippective impliesthat teacher eMpectationsfor studeni. performance maybecome incorporated into the stu-dent's self-tottept and result inmisconduct or poor academic perfor-mance. T provide a Measure of thisvariable, a two-item LOw Expects-.tions Scale asks_ teachers to judgewhat percentage of their studentsare of low ability and have "behav-ior problems:" The scale has areliability of..57. It correlates-.24 with Job Satisfaction and .43with Classroom Disruption.

Professional Development. Eightitems form a scale measuring theextensiveness. of recent continuingeducation or in-service learning.This scale, with reliabilitty' .74, isfor use in:documenting the 'implemen-tation of training component. It

also helps to lend evidence of con-

46

Measures

struct validity to othet. teacher

measures. Correlations reported by '

Cottfredson et al. (1982) suggestthe interpretation that teachersscoring high in professional deyel-opant ate more satisfied,ointerac4more with students, and are moreopen to student suggestions.

Non -Authoritarian Attitudes.Intended In part to measure sympath-etic, attitudes (as one way to get atthe "caring, competent teacher" con-stellation), a measure of punitive

'.moralism is included. To earn ahigh score on.this.scale, a teacher_rejects such.items as, "A few pupilsare just young2hoodlums and should

be treated accordingly." Thisthree -item scale has a reliability

of .54.

Stability of the Teacher Measures

We were not permitted to identifyteacher questionnaires, and there-fore cannot report on the stabilityover time of the individual-level .

teacher measures.

Measures of School Climate

The assessment of school climatesis fundamentally different from themeasurement of individuals. Whereasindividual differences are theentire point of measurement at theindividual level, these differencesare "error" or "noise" in theassessment of an environment based,on the reports of its individualinhabitants.

Compositional and. Psychosocial Cli-

mate Scales

Compositional climate, Environ-

ments are sometimes characterized byaggregated or averaged characteris-tics of individuals. We have conystructed compositional climatescales based on such aggregated per-sonal characteristics to descripb

-34-

climates using averagedcharacteristics of individuals (cf.Astin & Holland, 1961). Composi-tional climate scales are reportedfor inforthation about the studentsand the teachers who inhabitschools. In general, this type ofclimate scale describes the peoplewho inhabit the schools.

Psychosocial climate scales. An

alternative, and for some purposesmore useful way, to characterizeenvironments is to regard the inha-bitants=-teachers and students--asinformants about the environment.To construct this kind of climatemeasure, reports about the environ-ment (rather than about the'indivi-duals who inhabit it), aie used. For

psychosocial climate scales, reportsare first averaged, and then itemanalyses proceed based on schoolmeans for the items.

11-111.

Measures of Psychosocial ClimateBased on Student Reports

Community Crime. This is a ,

three-item scale based on avRragedresponses to questions about whetherthere are gangs in the student'sneighborhood, whether the gangs try.to get' the. student to join andwhether the student's parents wererobbed in, the last year. This scalemay be useful in describing the com-munity context of the school (cf.National Institute of Education,1978). It has a homogeneity coeffi-cient of .59 estimated from the 1982,data (Table 5).

Gangs in' School. This scale is

composed of averaged responses toquestions about whether there aregangs in the schpol and, if so, howmuch trouble they cause. The reli-ability (homogeneity) of this scaleis .80.

Safety. This is a 13 -item scale.asking .if students stay away from

47

I.

ti

Table 5

Reliability .Coefficients for the PsychoSocial ScalesBased on Student Report and Number of Items in Each Scale

1981 1982

Scale Alpha Alpha

Community Crime .57 .59

Gangs in School .80 .80*

Safety .92 .94

Individualized Instruction .58 .42

4

Disrespect for Students .78 .85

Student-Teacher Interaction .60 .64

Planning and Action .65 .84

Fairness .62 .76

Clarity .64 .67

Student Influence .62 .74

Grouping -. .55 .41

.1.

One-yr. N ofretest Items

.91

.82

.83 13

.80 2

.83 3

.79

.82

.76

.70 4

.84 6

.70 3

Note. Alphi-reliabilities for 1981 are generally based on asmaller number of schools than .those in 1982, which include allschools in the Initiative except those from St. Paul: We assume

retest xeliabilities are sometimes higher than the alpha coeffi-

cients because psychosocial climate scales are based on schoollevel item means which are themselves very reliable and items arenot strictly parallel as assumed by classical true score theory.

48

any of a list of places in theschool. It also asks if studentsfeel safe at school, or if they fearsomeone will hurt them at school oron the way to school. It resembles

what was called ."School Climate" in

the Schools fnitlaGive Evaluation(Grant et al., 1979). its reliabil-

ity coefficient is .94.

Individualized Instruction. This.

scale is an' attempt to use studentrepdrts as evidenCe about the levelof individualized instruction char-acterizing the school as a whole.Individualized instruction, as usu-ally construed, involves the devel-opment of individgal learning plans,rewards for improvement averpastlevels of performance, and a pace ofinstruction suited to thd

dual. Two aspects of this concep-tion are incorporated in this.meas-ure--students' reports that theyhave individual learning plans, andreports that they can work at theirown speed in class. The homogeneity

coefficient is .42.

Disrespect for Students. One

theoretical perspective (Greenberg,

1977) assumes that Aelinquency is inpart a result of a special statusaccorded youth, one, which isolatesthem from meaningful adult roles andsubjects them to de grading interper-sonal exchanges to which adultswould not be subjected. This scale

is intended to assess the degree towhich students feel that a schoolenvironment-as a whole eitherdegrades them or treats them with

*dignity. A low score could indicatethat students feel ,they are.treated

with dignity. Items include, "Stu-

den, are treated like dhildren

here;' " e hers treat students with

.respect;" and "Teachers do things tomake students feel put down." Its

reliability coefficient is .85.

Student-Teacher Interaction.This scale aims to assess the degreeof out-of-glass .positive social

Measures

interaction with teachers, from thestudents.:* point of view. It is

based on' the averaged e'sponses totwo items: "I talk to some of myteachers about things other, thanschoolwork;" and, "Teachers help mewith schoolwork outside of ,class:"Its homogeneity coefficient.is :64.

Planning and Action. This scale

is intended to assess, from thepoint of view of the students, the,degree to which schools engage inexperimenting and problem-solving,or the degree to which they resist

. change. It is composed of the fol-lowing three aggregated items: "It

is hard to change the way, things are'done in this school "; fthe teachersand principal in this school makeplans to solve probleMs"; and "Thisschool hardly ever tries anythingw." It has a homogeneity coeffi-

ci nt of .84.

Fairness. Evidence, is accumulat-ing that the degree to= which.stu-

',dents perceive a school's rules asfair and clear is associated withthe degree of orderliness of theschool (National Institgte,o Educa-tion, 1978; Gottfredson & Daiger,'1979). Consequently, scalesdesigned to assess these constructswere developed. ,Fairness is athree-dtem aggregate-level scale-based-on student reports that the .

rules are fair, that the punishmeAt.for breaking rules is the same foreveryone,' and that the principal is.fair. It has a reliability of .76.

.Intended to measure theclarity of school rules from the.point of, view of the school's stu-'dentsi this scale is compoSed ofquestions asking whether everyoneknoWs what the, rules are, whethereadhers let the students know whatis expected, whether the principal.is firm. This four-item:scale has a .

reliability 'coefficient of .67.

-36-

ef9

'Table 6

Reliabi ities of School Psychosocial Climate ScalesBased on Teacher Report and Number of Items in Each Scale

Scale'

Involvement of ParentsCommunity

Individualized,Instructioand Grading

Resources or InstructionIntegration vs. Segregation by.Ability or Conduct_

School Race RelationsTeaching Staff CommitmentPse of Grades as a Sanction

. Staff Morale (vs. alienation)Planning and ActionStudent InfluenceSmooth AdministrationSafetya

1981

Alpha

.80

1982

Alpha

.81

One-yr. N ofretest Items

.77

.60 .36 .70 4'

.86 .81 .81 4

.55 .59 .82 6

.77 .74 .53 2

.82 .91 .73 2

.84 .65 .56 2

.90 .94 .84 11

.87 .89 .84 10

,81 .85 .83 5'

.92 .93 .80 12

.94 .75 10

Not . Number of schools ranges from 48 to 50 schools for .49814ndran es froM,47 to 49 for 1982. An 'outlier 'school wap deleted inthe 81-analyses, and all schools with fewer thri 10 teacherswere deleted in the 1982 analyges. N'$ for the re-test reliabil-ities range fibM 33 to,37,acbools. We assume retest reriabili-ties are sometimes higher than the alplia coefficients becausepsychosocial climate scales are based on school -level item Meanswhich-are themvelves.very reliable and items are not strictlyparallel as assumed by classical true score theory.

4aReliabilty not Calculated in 1981. Two ,item's relating to

iclassroom disruption included in'this scale in 1981 were deleted 04in 1982, and in results presented for-1981 in this report the .

1981 scale is re-scored to, correspond with the 1982 scores.

-37-

\ 50

Student Influence. It is often

assumed that student influence onthe way a school is run may lead toa'number of positive outcomes, andan increase in student participationin planning and decision making issought- by OJJDP in the Alternative.Education initiative. This six-itemscale is intended to assess how muchinfluence students have in theirschools. Sample items include:.."Students have little say inhow theschdol is run"; "Students havehelped to make the school rules";and "Students are seldom asked tohelp solve a problem the school ishaving." The scale's homogeneity'coefficient of .74. ,

, Grouping. This scale assessesthe students' perceptions of group-'ing, or Segregation of students withspecial characteristics within the

school. It is composed of the fol-lowing three items:= "Students ofdifferent races usually end up indifferent classes"; and, "Thisschool has special classes for slowlearners "; and, "There are specialclasses for trouble makers." It has

4 a homogeneity coefficient of only

.41..

Climate Scales Based-on Teacher

Reports

An alternative perspective on theclimate of a school is provided bythe reports of teachers. Accord-ingly, 11 climate scales were con-stru ted from the teacher question-naire, using averaged teacherrespo ses about their school. Their

names and.reliabilities are shown in

Table 6.

Involvement of Parents and Commu-

nity. A goaLof, the AlternativeEducation!Program is, to. increase the

'use of community and familyresources by schools as a tructuralschool improvement. This scale

Measures

seeks to assess parent and community,involvement according to aggregate

teacher reports. It asks aboutparent influence on policies orpractices, direct parent assistance,relations between parents and teach-ers, and community receptivenessThe six-item scale has a homogeneitycoefficient of .81.

Individualized Instruction andGrading. The Alternative EducationProgram seeks to create structuralchanges in schools to increase indi-vidualized instruction, and thisintervention is planned by severalof the action pr9jects. Accord-.

ingly, this foUr-it n scale, aims to

measure individut. '-struction,

by asking if indiv.). learning

plans are used, and if grading is

based on improvement versus "thecurve." The scale has a homogeneitycoefficient of .36.

Resources for Instruction. This

scale is intended to measure rela-tive levels of resources (equipment,materials, learning opportunities)available' in the school. It con-

tains items asking about teachingsupplies,.space,extra-school set-tings used for instruction, andtimeliness of availability ofresources. This four-item scale has

a reliability of .81..

Integration vs. Segregltion by

Ability or Conduct. This scale isaiso included to measure an aspectof project implementation sought-bythe Alternative Education Program:the avoidance of tracking or isola-

'tion. The 0.x-item scale containsitems such as: "Students of mixedability work together in smallgroups in my class;" "This school

has special classes for slow lear-ners ;" and "In this school there are

special, classes for students whorepeatedly misbehave."' ,Its reli-ability is .59, and the appropria.teinterpretation of the scale is

-38-

. 5:1

unclear. Opinions differ about thewisdom of homogeneous vs. heteroge-wnedus grouping according to studentconduct or academic performance,although the current climate, andsome evidence (Slavin, 1980),implies that heterogeneous groupingcan have some virtue.

Schdol Race Relations. This

brief two-item measure asks aboutrace relations from the teacherpoint of view.. It asks hoW welldifferent groups get along. Its

reliability is .74.

Teaching Staff Commitment. Anec-

dotal and correlational evidencesuggests that the commitment of anorganiiation's staff is related to.project. implementation (Grant etal., 1979; Berman & McLaughlin,1976). Accordingly, a two-itemscale toassess staff commitment wasincluded. Its reliqbility is .91.

Use of Grades as a Sanction-. The

use of.-grades as a response to mis-conduct.is correlated with schooldisruption rates (Gottfredson &Daiger, 1979). On the face of it,this also appears to be a poor prac-tice ,because it makes the gradingand sanctioning process ambiguous..A two-item indexluses teacherreports to chardcterize the extentof this practice in schools. It has

a reliability of

Staff Morale. As with commit-.ment., morale is sometimes suggestedas a concomitant'of success inimplementing innovations, and it is.an important characteristicof anorganization in its own right, An

11 -item scale containing items suchas, "Our problems in this school areso big that 'it is unrealistic toexpect teachers to make much of a

'dent in them;" and "(Is the teachingfaculty) frustrated?" Its relia41:-ity is .94.

52

Measures

Planning and Action. .Presumably,organizations engaging in, systematicplanning and that are open to change,are most likely to successfully '

implement:innovations. Based onthis assumption, we constructed anine-item scale to assess planningand action. It asks, "How often doyou work on a planning committeewith other teachers?" "(Is theprincipal) progressive?" "(Is theteaching faculty). open to change?" '

Its reliability is .89.

Student Influence. Student par-

ticipation in school decision makingis one ot the major structural ele-ments thTvAlternative Education Pro-gram wants to create through theaction projects. The assumptionapparently is that student influencewill help to create other beneficialstructural changes, or it'may,con-tribute to decreased alienation.orsense of powerlessness. Measures of

student influence used in previousstudies (National Institute of-Edu-.cation, 1978; Gottfredson & Daiger,1979) assessed a limited range ofinfluence, and certainly do notassess the kinds of student influ-ence possible. Therefore, altholigh

based on the scale used earlier byGottfredson'& Daiger (1979), thisscale is expanded somewhat (to fiveitems). Sample questions are "I

often change my lesson plans,basedon student suggestions;" and "Teach-ers and their students work togetherto make rules governing behavior inthe classroom.". The scale has areliability coefficient of .85.

Smooth Administration. Our.ear-lier research (Gottfiedson & Daiger,1979) suggests that the-way a schoolis run is important in understandingits climate and in preventing schooldisruption. To the best of ourknowledge, detailed studies of 'Xschool administration tend to focuson the°personal characteristics ofadministrators (e.g., Miner, 1967),

Measurers

or are ethnographic accounts of thetypical activities of administra-tors. Here we wished to assess theperceptions of administrative styleand procedures from the point ofview of the body of teachers whoexperience them. Accordingly, weconstructed a 12-item scale. 'Typi-cal items are: "Simple, non-timeconsuming procedures exist foy theacquisition and use of resources;""There is little teacher-adminiStra-tion tension in this school." "(Theprincipal is) open." In a sense -

this scale represents a global rat-ini,kf the positiveness with whichteachers view the schools's adminis-tration, although the item contentfocuseson both principal bthaviorand some probable practical conse-quences.of that behavior. Its reli-

ability is .93.

F

Safety. This 10-item scale meas-ures teachers' perceptiohs of thesafety of their schools., It asks,

for example, how safe, the class-

rooms, halls; restroomsoetc. are.Its homogeneity coefficient is.94.

Stability of the Psvchosocial Cli-mate Measures I

-Cte-year stability coefficientsfor the psychosocial climate meas-ures derived from student reportsare presented in Table 5, and thecorresponding information for psy-chosocial climate measures derivedfrom teacher reports are presented

2 in Table 6. With the exception ofthe measure of school Race Rela-tions, these climate measures arefairly stablkover time.

Interpreting the School ClimateMeasures

In August 1982 we prepared fourkinds of feedback about the schoolsin which the delinquency preventionprojects are operating. This feed-back, based on surveys.conducted

53. -40-

with students and teachers, took the,.following forms:

o Average characteristics ofeach school's students.

o Reports by students about 10

each school's psychosocialclimate.

o Average charteristics ofeach school's teachers.

o Reports by teachers' abouteach achool's psychosocialclimate.

\ 1

'Formative evaluation informationof this kind is most useful when theprojects have developed clear ideasabout what they expect to see.Accordingly, base-line informationfrom surveys conducted in Oe Spring" \of 1981 was presented, and projectswere asked to make predictions aboutthe Spring 198rresults based' on theprojects' goals and objectives, andthei.r knowledge of the degree ofimplementation of their variousinterventions.

Student Scales

Individual-level ptudent scalesreport the average item score forall items in,the scale: Scale.

scores are computed such chat if aperson gave the keyed reaponse to 6items in a 12-item scale, his or herscore would equal,..'50'.4 As with

5. For items with more than tworesponse options (e.g., "yes," "no")'item responses were dichotomized.This ,differs from the scoridg methodused to report scores in our previous reports.' In previous reportsall of the variability in mulit-op-tion items was utilized by adding'together standard scores for itemsto compose scales of equally

0

L

001.4 kinds of psychological meas-urement, norms are useful in inter-preting scores because they,tellwhether a given score is high or lowin reference to an identifiablepopulation. Norms for '!.c school

compositional and psychosocial cli-mate scales based on the sample of.schools in the Alternative EducitionProgram are provided in AppendiX A,for this purpose.

Profiling scores. Using thesenorms a school's climate scalescores can be plotted on a profilesheet for easy interpretation. In

August, 1982, such profiles wereprovided for each school. The space

required to profile each school pre-vents us from presenting .the information in full in that' fbrm here,the profile sheets shown in AppendixB can be used to plot any schoolspro ile given the norms and thedetailed'school-by-school resultsenumerated elsewhere in this report.

The profile sheets provide a ver-bal interpretation of the climatemeasure results for a school. This

interpretation is based on thetranslation of percentile ranks intowords. The translation tablepreceding the illustrative profile 7.

sheets shows how percentiles mapinto verbal interpretations.

weighted items. The modificationwas introduced to enhance the inter-pretability the scales for practi-tioners not accustomed to use ofstandard scores. An examination ofthe psychometric properties of bothkinds of scales implies that thecurrent procedure is almost as effi-'cient as our original scoringmethod.

-41-

5.4

Measures

T 1,1.c. :;,,!101.

Teachor stales are scored byadding together the items; that com-.pose a'scale with the item responsescale constructed so that a highnumber always corresponds with the"high" end of the scale. The abso-lute (raw) scores on these scaleshave no intrinsic meaning, as do thestudent scales, and they can only beinterpreted by reference to norms.Norms are provided in Appendix A forthis purpose.

Profiling scores. Teacher,cli-mate scales can be ptofileB in the .

same way as the studenrscales usingthe illustrative profile sheets.

,Suggestions for .Using Climate

Reports in Project:Planninik

The worksheets,providedto proj-ects to facilitate the constructiveutilization of. the school climatemeasures are shown in Appendix C. .'

These worksheets are used to make 'discrepancies' between what implemen-ters expect to soe and what the

.actuallviobserve salient. Someprojects found it more useful ,toexamine this information in alterna-tive formats.

Using Measures for Individual Stu-, .

dents andTeachers

The confidential inclividual-/evel

measures are presented in the formof statistical summaries for treat-ment and Comparison group youths foreach of the delinquency pieventipnprojects. InterpretatiOn of theseindividual-level meadureis madewithin the context,ofAtie evaluatkeidesign for each of the various proj-ects, and these results are discussed in the projeq:-specific eval-uation reports in Part II of thisreport for each project that hasprojeCt components targeted atwell-defined groups and'an evalua-

SI

Measures

tion design that enables anassessment :of these project compo-nents.

The Utility of Information for Proj-ect Managers

In workshops conducted in August,11981, and again in August, 1982,school rbfiles were made availableto project'directors. These pro-files provided assessments ofschools useful.for-diagnostic andprescriptive purposes. The effortsof thousands of students and teach-ers in completing these surveyswould go partly to waste if thisinformation were not used in projectplanning -end ,continued project-development. We earnestly hopedthat this inforMation would be used, .

and are gratified that several proj-ects have made extensive use of thisinformation in renewed project plan-ning.

-42-

55

Similarly, interim feedback wehave provided to project directorson the characteridtics of theirclientele (in summary form), andabout the effectiveness of theirinterventions based on the statisti-cal analyses of individual scales isintend5d to be used in refininginterventions. No one expects tosee dramatic effects of projects intheir developmental stages, 'but pro-gress in at .least home areas is tobe expected. Projects will increasein effectiveness largely by usingthe information provided by thisinterim feedback. A subsequentchapter provides an overview of thechanges in school characteristicsacross years, and Part II a thisinterim report describes interimevaluation results for those por-..tions of each project targeted atidentifiable grbups of youths.

The Alternative Education Program: Kinds of Projects and the Youths Involved

The seventeen delinquency preven-tion projects that are the focus ofthe School Action EffectivenessStudy are diverse.. Providing a

, brief account of the similaritiesamong them and their major differ-ences is therefpre a difficult task:They differ in size, goals, theoret-ical rationales, stage of develop-ment, and in many other ways aswell. Nevertheless, it is possibleto characterize each FToject interms of some crosscutting dimen-sions.

Crosscutting Dimensions

The first conceptal dimensionalong whicheany delinquency preven-tion project may be placed is adimension of primary prevention vs.secondary or tertiary prevention.

Primary prevention. Primary pre-vention is activity directed toreduce the risk of delinquent behav-ior in a population. Making provi-sion for safe water supplies andenvironmental sanitation is an exam-ple of a primary-prevention-activityin the health area: It is intendedto reduce the population's risk ofdiseases transmitted by water. Theefficacy of this approach to healthpromotiOn is unquestionable. MakingJules in a school or community__snorer and more widely understood--is an example of a primary preven-tion activity in the delinquencyprevention area. Such an'activitywould be intended to reduceiroithe riskthat young people in the school's orcommunity's pcpulation will engdgein. delinquent behavior. In primaryprevention, the emphasis is onreducing the incidence or severityof some target problem in' the popu-lation at large.

-43-

Secondary -)revention. .Secondary

prevention is activity directed atreducing the occurrence of some neg-ative outcome for persons believedto be at especially high risk ofexhibiting that negative outcome.The administration of drugs and therestriction of sodium- intake to man=age blood pressure for hypertensiveindividulals who are at high risk ofsubsequent cardiovascular disease isan example of a secondary preventionactivity in the health area. Theseinterventions are intended to reducethe risk of strokes and heartattacks for persons with high bloodpressure. Prior research shows thatdrbgs and restricted sodium intakecontrol blood pressure, but that itis difficult to get people to adhereto the prescribed regimens. Provid-ing young people who are performingpoorly in school and who are in dis-ciplinary difficulty in school withacademic curriculum and learningstructures tailored to their needand using the techniques of applied

-behavior analysis to assist themoinLanaging their conduct is an exampleof a secondary prevention activityin the delinquency preVention area:Thes"iatervihtions are intended to

_ .

reduce the risk of future delinquentbehavior_for'youths already at sub-stantially elevated risk of display-ing such behOvior. The emphasis insecondary prevention is on reducing

--the-incidente or severity of some,target problem in a selected subset

. of the population that is considetedto be at tnusuaf.risk.

Tertiary prevention. Tertiaryprevention is for the most partsynonomous with the colloquial useof the term "rehabilitation." ,,Sur-

gery to remove an inflaitbd apOndixis an example, of tertiary prevention'in a medicalcontext. A rehabilita-

56

Projects

tion program for incarceratedoffenders might be an example of atertiary prevention program in thedelinquency area. Tertiary preven-tion is usually not regarded as aform of prevention, but rather as a'form of remediation.

Individual vs. Environment

A second conceptual dimensionalong which any delinquency preven-tion project may he placed isrSdimension of a focus on the adapta-tion of'the individual vs. a focuson altering the environment.

Promoting individual adaptationor resistance. The promotion ofindividual adaptation or resistanceto negative sources of influence isactivity directed to "innoculate"the individual against sources ofharm or to enhance the ability ofthe person to adjust to or cope withan environment. An example of anintervention aimed at preventingdisease through this approach is theinnoculation of individuals against

smallpox. After introducing a non-virulent strain of a micro-organism,the immune system develops defensesagainst subsequent invasion by viru-lent strains of the same organism.For some well understood diseasesthis approach is of unquestionable.utility. An example of an interven-tion aimed at preventing delinquencythrough this approach is one aimedat strengthening a person's attach-ments to prosocial others so that he

or she develops greater stakes in

conformity. In tertiary prevention,the emphasis is on reducing the fur-ther incidence or severity of a tar-get problem in the subset ofthe,population that has already exhi-bited the problem.

Altering theenvironment. Inter-

ventions adopting an approach ofaltering the environment aim to eli-

-44-

57

minite or attenuate sources ofinfluence in the environment thacontribute to problems. An exa le

of an .intervention in the healtarea using this ,approach is the useof dust extractors in grain storagesilos to reduce the risk of deathresulting from the explosion of thesilos. Such interventions haveproven to be of considerable value.Examples of interventions adoptingan approach of altering the environ-ment in the delinquency preventionarea include projects that widelydisperse immigrants and familieswith low socioeconomic statusthroughout urban areas rather thanallowing them to concentrate inurban slums, and interventions thatalter the reward structures ofschools so that individuals experi-encing difficulty in academic workwill not experience only failure inschool.

A Classification of Prevention Rrojects

The foregoing two dimensions can

be used to create the classification /of Alternative Education Preventionprevention projects illustrated inFigure 1.1 Quadrant 1 (the upper

bleft quadrant) includes prima .y pre-vention projects focused mai' lonenvironmental factors that contri-bute to delinquency. They aim to

reduce the risk of youth crime foi at6tal population. The effectivenessof Quadrant 1 interventions shouldbe reflected in epidemiologicalindicators of youth characteristicsand behavior. Their interventionsshould be such that everyone, ornearly everyone, in the target popu-lation is affected by them.

1. For a related classificationwith different content see Associ-ates for Youth Development, 1980.

Figure 1

A Classification of Preventive Interventions

Alter theenvironment

o Target is thepopulation.

o Focus is change instructures, policiesprocedures.

Example: Changingdisciplinary pro-cedures.

Primaryprevention

o Target is thepopulation..

o Focus is altering44individuals' be-haviors, attitudes,or competencies.

Examples: law-related education,career development.assistance.

o Target is a selected group

of high risk indivduals.o Focus is change in struc--tures, policies, procedures.

ExaMple: Reducing availabilityof school area for. gangactivity or conflict.

. _

Secondary,tertiaryprevention

o Target is a selected groupof high risk individuals.

o Focus is altering individuals'behaviors, attitudes,or- competencies.

Adapt theIndividual

-45-

Example": Counseling orpsychotherapy for 'high-riskyouths or offenders.

58

Quadrant-2 (the upper rightquadrant) includes secondary andtertiary prevention and remediationprojects focused mainly on environ-mental factors that contribute todelinquency. They aim to reduce therisk of youth crime for a group ofindividuals who are at high risk of'displaying deliniluent behavior.Their effectiveness should bereflected in measures of the per-sonal characteristics and behaviorof youths in the selected, high -riskgroup who receive the direct ser-vices ox treatments of the project.Quadrant 2 interventions should besuch that the treatments are deliv-.ered to members of the target groupwith 'sufficient strength and integ-rity to prevent further exposure toenvironmental conditions promotingdelinquent behaVior, to create envi-ronmental conditions that restrainthe individual from delinquentbehavior, and (if a tertiary preven-tion project) remediate existingproblems or rehabilitate an offen-

der.

Quadrant 3 (the lower left quad-rant) includes primary preventionprojects focused mainly on adaptingpeople to their environments. They

aim to reduce the .risk of youthcrime.in a total population byenhancing the ability of people inan environment to adjust to or copewith the environment. Quadrant 3'interventions should be such thateveryone-, or nearly everyone, in thepopulation is affected by the inter-ventions in ways that foster adjust:::

went or adaptation. Theirpeffectiveness should be reflected inepidemiological indicators of theattitudes, behavior, or personalcharacteristics of the population.

Quadrant .4 (the lOwer right quad-rant) includes projects focused..-mainly on adapting individuals tothe situations in which they findthemselves. They aim to reduce the

5 9 -46-

Projects

risk of subsequent delinquentbehavior by individuals in a targetgroup of high risk individuals.Quadrant 4 interventions should besuch that the treatments are deliv-ered to members of the target groupwith sufficient strength and integ-rity to develop. personal restraintsfrom engaging in delinquent behav-ior, prevent further development ofpersonal characteristics promotingdelinquent behavior, or (if a terti-ary prevention project) remediateexisting problems. Interventioneffectiveness should be reflected in

measures of the personal character:::istics and behavior of youths in theselected, high -risk group who

receive the direct services ortreatments of the project.

Like any typology, the presentone is an abstraction--these areideal types. No project'is likelyto resemble one and only one ofthese ideal types. In actuality,most projects will have characteris-tics in common with two or more ofthese 'approaches. The typology is . ,

nevertheless useful in providing ageneral characterization of preven-`tion projects.

The Classification Applied to theAlternative Education Projects

In Table 1. the foregoing classi-fication is applied to Ihe 17 proj-ectsbeing evaluated in the SchoolAction Effectiveness Study. We haveconsidered what these preventionprojects are doing and what they aimto accomplish, and then used ourbest judgment in preparing thetable. The classification of a par-ticular project in one category oranother does not imply that it is 'apure type, or even that the classi-fication is partictilarly apt. The

primary .cla'ssification means that,in our judgment, the project mostresembles that type.

Table 1

A Classification of the Alternative Education Project

Project 'Primary type Secondary type

Compton, CACYD

Pasadena, STATUS

ChitagC), PCD

Chicago, RETAIN

Kalamazoo, AEP

South Bronx, PREP

East Harlem, AAEP

Puerto Rico, OC

Charleston, PATHE

Houston, GIS

Virgin Islands, AEP

Hayward, LCO

Miami,' ACE

New' Jersey, EIC-S.

Plymouth, AEP

Milwaukee, JVS

St. Paul, Together

Secondary/Individual

Primary/Individual

Secondary/Individual

Secondary/IndiVidual

Primary/Environment

Secondary/Environment

PridarY/IndiviBual

Primary/Environment

Primary/Environment

Secondary/Environment

Secondary/Individual

Primary/Environment

Secondary/Environment

Secondary/Environment

Secondary/Individual

Secondary/Individual

Primary/Environment

Primary/Environment

Primary/EnvironMent

PrimarytEnvironment

Secondary/Environment

Secondayy/Individual

Secondary/Individual

Secondary/Individual

Secondary/Individual

Primary/EnVironment

Secondary/Individual

Primary/Environment.

Secondary /Individual

-47-

60

For Atk

fuller description of eachproject-see the detailed projectdescriptions,in our first interimreport, and in the second interimreport, part II. Those fullerdescriptions make clear that simpleclassifications do not portray thefull complexity of any of these.projects.

What Are The Interventions?

The individual interventionsbeing implemented by the projectsspan a wide range,of educationalapproache9.and techniques. Table 2categorizes a majority of the inter-ventions being implemented by theprojects. An "X" indicates onlythat we have reason'to believe thatsome version of that particularintervention has actually occurred.No attempt is made here to judge thestrength, fidelity, integrity, theo-retical reasonableness, or effec-tiveness of a particular interven-tion within a projeat. Many of theinterventions are, in fact, only,weakly implemented. The main pur-pose of the table is to give someidea of the scope and diversity ofinterventions across the entireAlternative Education Program, andwithin any particular project.

Who Are the Target Populations?

The evaluation's data base shows,that a cumulative total of 6,548youths were the targets of direct . .

interventions betWeen September 1980and April 1982. The number ofyouths subject to indirect influenceby the projects is larger: A total

of 23,934 youths were indirect tar-gets of interventions of projectswith primary prevention componentsin the schools in which they oper-ate. Detailed information .on bhenumbers of youths involved in eachproject is provided in Table 3. The

first column shows_the.cumulativenumber of youths' receiving,-dixet7---

61

Projects

services as part of the AlternativeEducation Program. This numberranges from 88 to 1,151 across theprojects. The second column showsthe number of youths receiving

i direct services as of April ,1982.The third column shows the number ofstudents indirectly served by proj-.ects with appreciable preventiveintervention aspects in the 1980-81academic year, and the fourth columnshows the corresponding information

-147ifi the 1981-82 academic. year.

Amore detailed description ofthe youths potentially_ affected bythe Alternative Eduaation Program isprovided in Tables 4, through 6.These tables show estimated ethnicgroup and sex breakdowns. Theseestimates are made using ethnic and,gender self-identification on theSpring 1982 School Action Effective-ness Study surveys.

The Alternative Education Programmainly involves ethnic minorities.Table 4 shows the ethnic' composition.of public schools in which,delin.;..quency prevention projects are oper-ating.2 Table 4 shows that only32.7% of the schools' population arewhite, 36.6% are Black, 24.7% areSpanish-speaking or Spanish gur-named,.1.5% are Native American,1.7% are Asian-American, and 2.7%gave some other ethnic self-identi-fication. The ethnic composition of

2. The V's shown in the table arenot equal to_the N's in Table 3because of survey or item nonres-ponse, and because the indirect ser-vice totals of Table 3 are derivedfrom principal questionnaires about .school enrollment, while the samplesfor Tables 4 to 6 were drawn fromactual school rosters. The tablealso shows the ethnic composition ofseveral non-intervention schools

__ which serve as "control" , l-cols in

the evaluation.

. -487'

Table 2

4Characteristics and Interventions of -the

Alternative Education Projects

1 2

Project Number5 6 7 8 9 10 11 12 13 14 15 16 17

Project school characteristicsProgram operates in: /

V/Regular chool

Alterna L e schoolOrganizatio of targetschool(s):

Elementary schoolJunior/middle schoolHigh school

XXXXXXXX4

X X X XX X X

X X x,X X XXXXXXXXXXXX X X X X X

Project interventionsCurriculum development X

Individualized instructionor tutoringTeacher.del!Wed X

Peer deliveredComputer delivered

Vocational/career educationAdaptive/affective educationTedcher training/developmentIndividual counseling X X X X X X X X

Group/peer counseling X' -X X X X'

Change classroom manage- X' X X X X

X X X

X X X

X x x

X X X

X

X

X

X

X X

X

X

X X X X XX X X X X XX X X X X

X X X X X XX X X X

X X X X.

X X X X

X X X X XX X X XX . X X X -X X

X X

X X X X X

X XXXXX X X

X X X

X

xxxXxxlx'

X X X

X X

X X X X

ment/organizationIncrease extra-curricular X

activitiesChange school sanctionprocedures

. DisciplineSuspensions

Increased.student participa-ti.on incdecision making

Improve school climateInvolve community in school:

Channel resources into X

school'Parental involvement X

Involvement of persons X

other than parentsImprove parent-studentrelations.

Diversion ffom juvenilejustice system

X X

X' XX X X

X X X

XXXXXX

X X

X XX X X X X

X

X

XXXXX'X X

X X X X

1=Compton-CACYD2=Pasadena-STATUS3=Chicago-PCD.4=Chicago-RETAIN5=Kalamazoo-AEP6=South Bronx-PREP

7=East Harlem-AAEP8=Ptierto Rico-OC

9=Charleston-PATHE10=Houston-GIS11=Virgin Islands-AEP12=Hayward-LCO

-49-

62

13=Miami-ACE14=New Jersey-EIC-S15=Plymouth-AEP16=Milwaukee-JVS17=St. Paul-Together

1

Table 3

Cumulative' and Current NUmber of

Clients Receiving ServicesI

4 -- - - --

1 Total Receiving Total Receiving

Direct Services Indirect Services

.0 Cumulative

Project to Apr. '82

Compton 0 4 132.

Confftiiutional Rights 421d

Foundation, PasadenaPeer Culture Development, 946

Chicago 1

Chicago Board of Education 205

Kalamazoo 115

Bronx 329

Jazzmobile, Harlem 781a

Puerto Rico 976d

Charleston 1,151 ,,'

Houston 119 .-

Virgin Islands 88

Lac Courte Oreilles,. Hayward 124b

Miami 114

Plymouth-Canton 213

New Jersey 154

-Jewish Vocatiohal_Services, 329c

MilwaukeeSt. Paul 351

Current inApril '82 1980-81'

65 0

1981-82

0

250d 3,445 3;069

432 5,531 5,712

''

'128 0 0

115 657 665

150 0 0

251a '0 0

491d 22,245 1,608

,630 4,597 4,078

. 75 0. 0

53 0 : 1,356

100b 100 95

85 , 0 0

161 0 0

154 0 2,812

329c________!__

318 3,722

_ __0___

4,539

.Note. Counts are based on the number of clients who received at least

some direct program services, according to information provided to the

National Evaluation Mfi'nagement -Information- System. _Clients not named or

identified with an ID number are not entered into the MIS. Direct ser-

vice recipients include all students enrolled in or receiving services

through a program component. Indirect or preventative services are

recorded only for prOjects involving a substantial school change or pri-

mary prevention component.

aFigures do not include students currently enrolled in the elementary

school program.

bFigures do not include 1981-82 youth center clients.

c These figures do not include Returh Center clients and MYEC clients who

entered the proiram after the "evaluation phase" which ended in January,

1982.

dThese figures reflect the number served through May, 1982, because no

data were available in April.

-50-

63

0'

Table 4

Ethnic Composition of Schools in the Alternative Education Program

Spring, 1982

City, School,and Project

Row Percentage

Native Asian Spanish Weighted

American American American Black White Other N

Papadena, CaliforniaSchool 70 l.,66 '6.21 19.45 43.98 23.28 5.43

at876

School 82' 0.87 2.69 14.38 45.03 31.95 5.08 1377

Project subtotal 1.18 4.06 16.35.: 44.62 28.58 5.22 2253

Peer Culture Development,Chicago

School 1370 0.27 0.74 68.41 27.71 2.12 0.74. .1097

School 14)0 0.55 6.05 61.26 7.86 21.31 2.94 732

School 1820 0.30 1.21 26.02 22.35 49.21 0.91 3298

School 3200' 0.00 0.94 30.19 4.72 59.43 4.72 106

School 4720 0.67 1.33 94.00 0.67 3.33 0.00 150

School 5070 1.35 5.'41 28.38 43.24 20.27 1.35 . 74 .

School 5550 3.26'' 7.61 56.52 5.43 25.00 2.17 92

School 6010 1.08 1.08 89.17 3.25 4.33 1.08 277

Project subtotal 0.43 1.88 43.77 19.75 32.96 1.22 5826

Chicago Board of Education,,

...

ChicagoSchool 1240 0.00 0.14 56.32 38.91 0.00 4.63 720

School 1340 0.00 0.59 13.40 50.81 32.99 2.21 679

School 2300 2.82 7.91 59.89 -2.26 23.73 3.39 177

School 4440 0.00 2.86 43.81 33.33 17.14 2.86, 105

School 4550 0.87 0.00 0.87 98.26 0.00 0.00 115

School 5090 0.55 0.00 32.97 0.55 59.34 6.59 182

School 5750 0.61 0.61 0.61 97.58 0.00 0.61 165

School 5880 -1.06, 0.24 82.77 11.25 2.08 2.61 410

School 6180 0.79 0.00 15.75 77.95 0.79 4.72 127

Project subtotal 0.50 0.90 39.92 40.45 14.98 3.25 2680

Kalamazoo, MichiganSchool 318 1.27 1.09 3.27 21.82 69.09 3.45 550

School 327 1.69 1.06 1.91 35.55 52.88 1.91 466

Project subtotal 1.47 1.08 2.65 28.11 63.95 2.75 1016

-51-

64

continued

S

Table 4 (continued)

Ethnic Composition of Schools in the Alternative Education ProgramSpring, 1482

City, School,and Project

South Bronx, New York

NativeAmerican

' Asian

American

Row Percentage

SpanishAmerican Black White Other

II

Weighted

School 22 2.21 0.91 46.28 44.98 2.91 2.59 309

School 55 0.00 1.64 22.95 70.49 3.28 1.64 61

School 63 3.68 0.00 39.71 52.94 0.74 2,94 136

School 64 3.33 0.00 69.17 20.00 4.17 3.33 120

ScluEol 82 1.23 0.41 45.08 50.00 2.46 0.82 244

School 117 2.73 0.78 59.77 31.25 0.39 5.08 256

School 132 0.00 0.00 35.37 60.98 1.22 2.44 82

School 145 1.60 0.40 43.90 51.80 0.80 1.60 378

School 141 3.80 0.80 ' 29.50 63.60 0.80 1.50 200

School 148 2.71 0.68 26.78 66.78 1.02 2.03 295.

School 166 1.60 0.00 40.80 54.50 0.60 2.50 629

School 229 2.50 0.50 21.50 72.50 0.00 3.00 200

Project subtotal 2.20 , 0.40 ''',.. 40.10 52.90 1.20 2.50 2941

East Harlem, New YorkProject subtotal 3.31 0.83 11.57 82.64 0.00 1.65 121

Playa de Ponce,Puerto Rico

School 1 0.83 2.22 87.50 1.39 7.78 0.28 360

'School 2 1.84 0.00 87.56 0.92 8.76 0.92 217

School 3 0.39 0,399

. 0.59 5.71 0.39 508

Project subtotal 1.00 0.84 89.79 9 1.00 6% 0.42 1195

Charleston,South Carolina

School 242 0.40 0.40 0.80 41.50 54.80 2.00 451

School 741 1.40 0.76 1.66 95.40 0.00 0.77 393,

School 742 0.64 0.51 1.02 96.42 0.26 1.15 .392

School 743 1.12 0.00 0.26 97.41 0.43 0.78 382

School 751 0.32 0.50 0.50 98.36 0.00 0.32 313

School 754 0.00 0.47 0.23 99.06 0.23 0.00 426

School 755 0.96 0.14 0.00 98.48 0.00. 0.42 717

School 944 1.47 0.29 1.49 66.67 27.57 2,48 339

School 951 1.02 0.00 , 0.43 78.53 19.00 1.02 623

Project subtotal 0.80 0.10 0.60 85.80 '11.50 1.00 4036

St. Croix,Virgin Islands

Project subtotal .29 0.00 30.99 '65.43 1.19 .2.10 1041

continued

:

-

Table4 (continued)

Ethnic Composition of Schools in the Alternative Education ProgramSpring, 1982

City, School,and Project

Plymouth, Michigan

NativeAmericah

AsianAmerican

Row Percentage

SpanishAmerican Black White Other

Weighted

School 31 1.68 1.51 1.35 0.17 88.89 6.41 604

School 41 1.68 0.12 1.51 0.00 92.0511, '4.59 861

School 42 0.00 3.57 0.89 0.00 92.86 2.68 112

School 43 0.00 2.80 0.00 1.87 88.79 6.54 107

Project subtotal 1.44 1.00 1.38 0.18 90.78 5.22 1710

New JerseySchool 1 1.65 0.00 30.34 15%17 44.94 7.87 178

School 2 1.03 0.69 3.10: 14.25 74.84 6.0E 924

School 3 4.64 0.00 8.21 67.14 17.14 . 2.86 280

School 4 1.69 0.56 16.01 12.92 65.17 3.65 356

Project subtotal 1.82 0.48. 9.36 22.59 60.50 - 5.25 1738

St. Paul...

School 210 1.60 3.95 11.68 29.31 59.09 4.37 1207

School 230 1.73 8.63 3.88 3.02 80.58 2.16 1323

School 342 2.98 0.72 0.93 14.94 76.09 4.33 478

School 352 - 6.82 7.82 5.02 6.69 65.45 8.21 353

Prod, t subtotal 2.39 5.58 2.77 14.75 70.48 4.02 3456

Total 1.50 1.70 24.70 36.60 32.70 2.70 28378

-53-

66

1.

Table 5

Ethuie Self- Identification of Students Receiving Direct Solvicvs.Alternative Education Program, Spring 1982

City, School,and Project

Compton, California

Net ive

AmericanAsianAmerican

Row Percentage

SpanishAmerican Black Whips Other

CACYD 3.51 5.26 8.77 77.19 1.75 3.51 57

Pasadena, CaliforniaSchool 70 2:43 2.43 24.31 44.98 18.55 7.29 82

SciloOl 82 0.96 5.77 11.54 61.54 16.35 3.85 104

Project subtotal 1.61 4.30 17.18 54.22 17.32 -5.37 186

Peer eulture Development,Chicago

School 1370 0.90 0.00 42.34 50.45 6.31 0.00 111

School 1430 3.16 1.05 60.00 18.95 13.68 3.16 95

School 1820 0.00 0.00 12.82 37.18 50.00 0.00 78

School 3200 0.00 0.00 20.83 0.00 75.00 4.17 24

School 4720 0.00 3.85 88.46 0.00 7.69 0.00 26

School 5070 \ 4.55 0.00 27,27 50.00 18.18 0.00 22

School 5550 3.33 6.67 43.33 13.33 33.33 0.00 30

School 6010 2.38 0.00 83.33 4.76 9.52' 0.00 42

Project subtotal 1.64 0.93 45.99 28.04 22.66 0.93 428

Chicago Board of Education,Chicago

School' 1240 0.00 0.00 66.67 33.33 0.00 0.00 A

School 1340 0.00 0.00 . 0.00 0.00 100,00 0.00 1

School 2300 7.14 0.00 57.14 7.14 28.57 0.00 14

School 4440 0.00 0.00 25.00 66.67 8.33 0.00 . 12 '\'

School 5090 0.00 0.00 57.14 7.14 21.43 14.29 14

School 5750 0.00 0.00 0.00 100.00 0.00 0.00 11

School 5880 0.00 6.30 93.70 0.00 0.00 0.00 16

School 6180 7.69 0.00 23.08 69.23 0.00 0.00 13

Project subtotal 2.30 1.15 47.05 35.84 10.36 2.30 87

Kalamazoo: MichiganSchool 318 1.03 1.03 2.06 28.87 63.92 3.09 97

67-54-

continued

T (coot tote d

Ethnic Sell-Identification of SindentiOteceiving Dileit !;CIVItCel

Alternative Edocation Program, Spring 198/

Now Percentage

City, School.and Project

S,nith Bronx, New York

School 51School 55SC1101)1 63

School .64

School fi2

School 111School 132School 147School 148

.Project subtotal

East Harlem, New YorkSchool 88

Playa de Ponce,Puerto Rfco

School 1

School 2School 3

. Project subtotalCharleston, SouthCarolina (PATHE)

School 741School 742School 743School 754Scliool 755

School 944School 951

Project subtotal

Wouston, TexasAProject pu opal

St. Croix,. i,.

Virgin IslandsProject; subtotal

NativeAmerican

6.450.000.000.000.000.00'0.000.00,

5.562.29

.

4.88.:;',

.

0.562.220.701.00

3.03

2.131.67

0.000.000.000.001.01

0.00

3.70

..

A61110

American

0.00

0.000.000.000.00

0.000.000.000.000.00

1.22

3.390.000.351.17

1.52

2.130.001.72

1.59.0.000.00 11.01 1

.

0.00

.0.00

American

25.8127.27

33.1490.0033.1340,0021.4314,295.56

29.77

12.20

84.7588.1592.3189.15

7.580.001.67

0.000.000.000.001.51

94..64

40.74

.

Black

61.2912./3

6160T0

66.61

1;01:04

85.71

88.8965.65

79.27

2.821.481.05

1.67

87.8895.7495.00'96.5598.4166.0087.0489.95

0.00

48.15

A

:

White

0.000.000.00o.000,000.00

0.000.000.000.00

0:00

7.91

6.675.59

6..51

0.000.000.001.720.0032.0012.966,03

1.79

.

7:41

other

6,450,00-0.00

.00

0.007.14

Q.000.00.

2:29

2,44

'0.56

1.480.000

..

0.000.001.670.000.002.00

0.000.50

.

3.57,

"

0.00 \

, s, r

=;-

contind ''.ue

N

31

II

15

lo15

10

14

7

18

131

8i

,

177

135

286 .

599 99

66

47

60

58.

63

50

54

398

56

27

\

.

.

0

`

1,

.I

fi

-55-

68

Table 5 (continued)

Ethnic Self-Identification of Students Receiving Direct Services,Alternative Education Program, Spring 1982

Row Percentage

City, School,And Project

Heyv,ird (LCO)

Native

American

AsianAmerican

SpanishAmerican Black White Other

Project subtotalMiami

98.63 0.00, 0.00 0.00 1.37 0.00 73

Project subtotal 0:00 0.00 16.18 51.47 32.35 0.00 68

Plym,- h, MichiganCudth Worl?s 0.00 0.00 3.20 0.00 93.50 3.20 31

School 31 0.00 0.00 0.00 0.00 92.90 7.10 14

School 41 5.60 0.00 5.60 0.00 88.90 0.00 18

fichool 42 3.40 1.70 0.00 0.00 87.90 6.90 58

School 43 2.30 0.00 0.00 2.30 88.40 7.00 43

Project subtotal 2.40 0.60 1.20 0.60 89.60 5.50 164

New JerseySchool 1 0.00 0.00 63.64 13.64 22.73 0.00 22

School 2 0.00 0.00 0.00 5.26 84.21 10.53 19

School 3 4.17 0.00 8.33 75.00 8.33 4.17 24

School 4i 0.00 0.00 4.55 13.64 54.55 27.27 22

Project subtotal 1.15 0.00 19.54 28.74 40.23 10.34 87

St. PaiT1

School 210 0.00 0:73 4.06 ,31.94 61.08 2.18 275

School 230 0.00 0.00 0.00 0.00 100.00 0.00 2

School 342 0.86 0.00 3.83 18.70 74.88 1.72 116

School 352 0.00 19,0/. E.22 8.11 62.91 8.11 55

Project subtotal 0..12 2.78 )..71 25.44 65.06 2.78 448

a

- 5 1)-

Table 6

Gender Self-Identification of Students in Participating Sdhoolsand of Students Receiving Direct Services, Spring 1982'

City and School

Total School Directly Served

Female

ComptonCACYD

Pasadena, CaliforniaSchool 70 48

School 82 46

"Project subtotal 47

Peer Culture Development.,Chicago

School 1370 44

S 4chool1430' 55

School 1820 '50

School 3200 53

School 4720 51

School 5070 53

School 5550 46

School 6010 53

Project subtotal 50

Chicago Board of Education,ChicagbSchool 1240 48-School 1340 '52 1

47/

School 2300School 4440 48

School 4550 (control) 52/

School 5090 51

School 5750 56School 5880 53,

School 6180 ,.0

Male

\52

54

53

56,

5

50/47

49/47/ 54

4750

52

4853

52

48494447

40

Project subtotal / 51 49

Kalamazoo, MichiganSchool 318 : 5? 48

School 327 46 54

Project subtotal 49 51

WeightedN. Female Male

985

1431 .

: "2416

38

4855

51

62

52

10.

49

60

96

108

205

1120 .s 57 43 113

742 49 51 95

3308 63 37 78

110 52 48 25

149 50 50 26

77 64 36 22

94 53 47 30

287 49 51 43

5887 55 45 432

756 43 57 7

720 100 0 1

177 29 71 14

110 31 69 13

124

192. 38 62 16

169 55 45 11

427 47 53 18

138 67. 33 15(

2813 45 55 95

i 45 55 100, -,

., 1 .+5 55 100

Note: A dash signifies not applicable. continued

70

A

Table 6 (continued)

Gender SelfIdentification of Students in Participating Schoolsand of Students Receiving Direct Services, Spring 1982

City.and School

South Bronx, New York

Female

Total School

MaleWeighted

N

Directly Served

% .

Female Male N.

School 22 51 49 333 --

School 53 (mini unit) 32 68 34

School 55 56 44 66 .38 62 13

School 63 48 52 143 43 57 14

School 64 52 : 48 131 15 85. 13

School 82 ' 48 52 260 33 67 15,

School 117 52 48 291 18 82 11

School 132 55 45 87 47 530 17

School 145 56 44 413 --

School 147 52 48 233 50 50 8

School 148_

54 46 326 22 78 18

SChool 166ScSchool 229

50.'

53

50

.47

667212

--

--

--

Project subtotal 52 48 3196 33 67 143

East Harlem, New YorkProf?ct subt'otal 53 47. 137 59 41 92

Puerto RicoSchool 1' 51 49 384 59 41 189

School 2 46 '54 237 56 44 152

.School 3 57 43 526 57 43 293

Project Subtotal 5? 48 1269 57. 43 635

Charleston, SouthCarolina (PATHS)

School 242- (control) 47 53 468

School. 741 54 46 423 47 53 75

School 742 53 47 427 31'.. 69 52

School 743 44 56 402 48 52 66

School 751 (control) ,65 35 323 ---

School 754 50 50 444 '52. 48 60

School 755- 53 47 734 47.. 3 64

School 944 49 51 363 38 62. .

School 951 47 53 650 38' 62 55

Project subtotal 51 49 4235 44 56 425

Note: A dash signifies not applicable. continued

Table 6 (continued)

Gender Self- Identification of Students in Participating Schoolsand of Students Inceiving Direct Services, Spring-1982---

Total School Diredtly Served

City and School

Houston (GIS)

%

Female MaleWeighted

N Female Male

Project subtotal 52 48 ' 61

Virgin IslandsProject subtotal 51 49 1266 28 72 29

Heyward (LCO) .

Project subtotal 46 v 54 78

MiamiProject subtotal '46 54 69

Plymouth, MichiganGrowth Works -- 23 77 31

School 31 59 __AI__ 669 18 82 17

School 41 48 52 919. 56 44 18

School 42 53 47 115 48 52 64

School 43 49 51 110 3.8 62 -47

-Project subtotal '52 48 1837 39_ 61 177

New JerseySchool 1 44 56 192 25 75 24

School 2 52 .48 1001 30_ 70. 23

School 3 50 --. 50 316 47 53 34

School 4 55 45 379 30 70 23.

Project subtotal 51 49. 1888 35 65 104

St. PaulSchool 210 46 54 1209 48 52 277

School 230 47 53 1311 0 100 , 2

School 342 46 54 '488 58 42 117

School 352 48 52 357 56 44 55

Project subtotal 47 53 '3463 51 49 451

Total 50 50 29851

Note: A dash signifies not applicable.

72

-

Projects

the schools involved differsmarkedly by city and sometimeswithin city.

The ethnic bieakdown of personsreceiving direct services throughthe Alternative Education Program ispresented in Table 5. This tableincludes all projects providingdirect services, regardless of thelocation -of those services. (The

project operating in Compton, for

example, is included in Table 5-but .

not Table 4 because it primarilydevotes its efforts to a group ofhigh risk youths directly served byits alternative school.) For the

most part the ethnicity of youthswho receive direct services resem-bles the ethnic composition of theschools served by the projects.

The gender composition of theschool populations and of groups

-receiving direct services aredescribed in Table, 6. As expected,

'about half of the school populationsare male and half female: There aresometimes slightly more males thanfemales among direct service reci-

pients. Males are, of course, morelikely to have disciplinary diffi-culties in school and to engage indelinquent behavioi than arefemales, so this slight predominanceof males is to be expected.

1

How Much Delinquent Behavior. Occurs?

To provide some perspective onthe youth population involved in theAlternative Education Initiative,' itis useful to characterize it in 'terms of the amount of delinquentbehavior these youths engage in.There is no foolproof way to esti-mate the amount,of delinquent behav-ior anygroup.engages in (:se Chap-ter 3), but one method to use

information .derived-,fxam voluntary

self- report. Adctirdingly, Table 7

shows the proportion of youths

-60-

admitting to have committed each ofseveral kinds of delinquent behaviorin the past year. These tabula-tions, which have been statisticallyadjusted to reflect the populations.of the schools involved, imply thatthese youths (especially the males)have committed a large number ofcrimes. Note that the table showsonly the percentage who admit toeach crime at least once. Undoubt-

edly the total number of crimes com-mitted is much larger.

According to Table 7, the popula-tions of the public and alternativeschools involved in the AlternativeEducation Program engage in a sub -

scantial amount of delinquent behav-ior. Of the males, 13% damaged ordestroyed school property at leastonce, 17% damaged or destroyed other

property, 19% carried a concealedweapon, 13% were involved in gang'

fights, and 10% hit or threatened tohit a teacher. The absolute numbersof males in these schools whoengaged in these behaviors are:2210 vandalized school property,2852 vandalized other prope'rty, 3291

carried a conceaed weapon, 3134stole or tried t6"-steal somethingworth more than $50, 1035 usedstrong-arm methods to rob someone.Females engage in each of the fore-going behaviors much less often thanmales--half as often-or less.

%

The percentages of males andfemales smoking, drinking and usingdrugs are more nearly equal. Forty

six percent of the males and 44% ofthe females report drinking, 23% ofmales and 20% of females reportusing marijuana.

Only 24% of the males and 36% ofthe females report engaging in none

of the behaviors listed. A rela-

tively small percentage of youthsreport engaging in a great variety

of delinquent behavior: 7% of the

males and 2% of the females...reported

Table 7

Percentage. o ales and Females Reporting They CoMmittedEach of a Variety of Delinquent Behaviors n Past Year'

and Estimated Numberof Youths Committing, 1982

Behavior

Males

N

II

Females

Damaged or destroyed school 12.9 2210 6.2 1132

propertyDamaged or destroyed other 16.7 2852 6.1 1112

propertyStolen or tried to steal 7.7 1317 1.8 322

something worth less than $50Carried a concealed weapon 19.3 3291 5.8 1050

Been involved in gang fight 13.0 2205 5.2 936

Sold marijuana or other drugs 7.9 1336 4.0 714

Hit or threatened to hit .a 10.5 1785 4.7 851

teacherHit or threatened to hit a 50.4. 8511 31.5 , 5618

studentTaken.a car without owner's 8.4 1429 3.1 553

permissionUsed force or strong arm methods 6.1 1035 1.9 338

to robStcilen or tried to steal ( 18.6 3134 . 10.4 1858

something worth more than $50Stolen or tried to steal some- 13.0 2189 5.0 895

thing from locker orelsewhere at school .

Broken or tried to break into a 7.5 1253 1.7 309

building or carSmoked cigarettes 24.3 4062 33.1 5884

Consumed alcohol 45.7 7566 43.6 7634

Smoked marijuana 23.4 3898 20.2 3586

Taken other drugs 7.7 1285 7.5 1323

Cone to school drunk or high 12.0 2006 8.7 1540

Used inhalants 6.1 1015 5.3 935

None of the above 24.0 3957 35.8 . 6316

One of the above or 39.9 6593 54.2 9571

fewerTwo of the above or 43.2 7128 56.8 10037

fewer .

Half br more of the above 6.8 1123 2.2 389

Note. Based on weighted tabulations from the Spring, 1982, SchoolAction Effectiveness Study survey. The total weighted N is 19167males and 19274 females. Percentages-exclUde non-respondents. N's

shown in the table are the estimated number, of persons performingeach type of behavior in past year but do not include survey or itemnon-respondents.

761-

74

engaging in half or more of thebehaviors listed in the past year.

Some Consequences of Victimization

Most work on crime neglects thevictim, focusing exclusive attentionon the offender. Yet it is the vic-

tim of crime who most directly..experiences the impact of theoffense, and the recent report ofthe President's Task Force on Victims of Crime (1982) begins the taskof focusing greater attention on thevictim.

Reducing victimization is animportant 'goal of the AlternativeEducation Program, and results

thein Chapter 5 discuss the_.achievement in this area through theProgram's second year Data col-lected as part of the, evaluationillustrate hoW Crime may affect the

victim. Students who report moreextensive victimization in surveysconducted-as part of the SchoolAction Effectiveness Study are sig-nificantly more alienated (feel theyare less connected to the social

order), like school) significantlyless, and have significantly lowerself-esteem. When the psychological

health of students is studied over

time.we find evidence of negativeeffects of victimization on aliena-tion and the amount of effortexpended at school work for female

students. (A technical account of

our preliminary correlationalresearch in this area may be found

in .Gottfredson (1983a).

Much remains to -be done in the

Area of documenting and measuring

the effects of victimization on the

victim. We interpret these prelimi-nary results as suggesting negativeeffects that interfere with the aca-demic work of the victim as well as'having serious direct harmful psy-chological, effects, especially forfemale students.

Projects

In short, delinquent behavior andvictimization are without questionserious problems' in the public andalternative schools involved in theAlternative Education Program.

_ Who Engages in Delinquent Behavior?

Naturally, nott all youths engagein.delinquent behl-Vior to the same

degree. Table 8 shows correlationsbetween the number of differentkinds of .delinquent behavior admit-ted and various personal character-istics. The results shown in Table.8 generally accord with the resultsof previous research, and imply that -

the youths who engage in more delin-quent behavior are characterized by:

Weak attachment to parents.

o Associaticn with delinquent peers.

o Al.Netion, or a feeling of notbeing connected to the social

order.

o Weak attachment or dislike for

amr school.

o Lack of belief in the validity of

. rules.

o Low self-esteem or a delinquent

self-concept.

o Premature and rebellious expres-sions 'of autonomy.

o Little effort expended at school-

work.

o Truancy.

Youths engaging in much delinquentbehavior are not much different interms of parental education (a meas-

ure of socioeconomic.status) thanthose engaging in little delinquentbehavior. The more delinquent.youths are punished more in school

zdnd also are victimized somewhat7 more than other students in school..

-62-

Table 8

Correlations between Selected Personal Characteristicsand Variety of. Delinquent Behavior Reported

Characteristic

Males

r N

Females

r N

Parental Education .06 328 .12* 448

Attachment to Parents -.26* 398 -.34* 4550

Negative Peer Influence .52* 446 .42 614

Alienation .19* 399 .25* 564

Attachment to School -.36* 427 -.43* 600

Belief in 'Rules -.35* 386 , 549

Positive Self-Concept -.22* 361 -.281v 520.

Rebellious Autonomy .22* 307 .32* 460

School Effort -.31* 387 -.32* 529

School Nonattendance .30* 459 .27* 618

School Punishments .23* 437 .33* 605

Victimization .16* 432 .23* 605

Note. These correlations are calculated using a random'.half of the students who completed SAES surveys in theSpring of 1981 that included the self- report delin-'qUency meaeure. Correlations are computed on this sub-sample to save the cost of processing a much larger,file; the pattern of results would be substantially thesame were calculation performed on the entire sample.

*p. < .01. .

The foregoing results are not newto researchers in the delinquencyarea. They once agEin-suggest theappropriateness of testing interVen-Lions to prevent delinquency usingalternative education approaches.Interventions in school to alterstructural relations in th'e environ-ment or to enhance the ability ofyouths. to adapt to schooling aresuggested by these data. Activitiesby educators tc decrease'the nega-tive influence of delinquent peers,to create greater feelings of con-nectedness to the social order ofthe school, to increase,attachment.

0

7 7

Projects

to or liking for school, to fosterbelief in the validity of rules, andto develop (or at least not ravage)the students' self concepts may verywell reduce youth crime.

In the next chapter we begin toaddress two questions: "Are theprojects participating in the Alter-native Education Program providingevidence that they are preventingdelinquency?" And, "Are theseects providing evidence that theyare influencing the known concomi-tants of (arid presumed risk factorsfor) delinquent behavior?"

1

The Environmental Effect of the Alternative Education Program

and Population Result's for Delinquent Behavior

The Program Announcement for theProgram in Delinquency, Preventionthrough Alternative Education( OJJDP, 1980) makes clear teat theOffice aimed to demonstrate delin-quency preventionprograms. thatcreated structural changes in theorganization a schooling to bringabout changes in the behavior andpsychOsocial dpvelopment of studentsand teachers in those schools. In

other words,'the Program aims atdelinquency prevention throughchanges in school climate andchanges in the attitudes and behav-iors of students and teachers in theschools. These outcomes involve theentire populations of the schools

involved. Outcome measures are epi-demiological indiCes of behavior andpersonal characteristics for schoolsand measures of scho9,1 environments.The 14-esent chapter 4-eports on pro-

.gress\in these areas.

Overyiew of the Results Sought bythe Program

The results.'sought by the OJJDPare recounted in Chapter 1. Chapter

3 describes in detail some of themeasures we have developed to .meas-

ure these desired outcoms. The

'forlowing paragraphS collate meas-ures with results sought.

Measures of Program Goals

Reduce delinquent behavior in andarouad schools./ This goal is ofcentral importance and is measuredin several ways. (a) -A Self-Re-

ported Delinquent Behavior scale andtwo scales composed of subsets ofitems from this scale: Self-Re-

ported Drtig Involvement and Self-Re-ported Serious Delinqubnt.Behavior.The first of these subscales con-

-4_

tains only items related tosubstance use and the second is res-tricted to the illegal behaviors ofgreatest seriousness (excludingdrug-related items). (b) A StudentVictimization scale. (c) Student

reports of Gangs in School.(d) Student reports of Safety.(e) Student reports of CommunityCrime. (f) Teacher reports ofSafety. (g) Teacher reports of Vic-,timization. (h),Teacher reports ofClassroom Disruption. Additional

measures of delinquent behavior werecollected from official records fora number of projects, but they werenot.-pollpcted in parallel ways fromProject to project,, and no results

based on official,rbcords areareported in this chapter. Someresults based on official recordsare included in _Part II" of this

report.

Decrease suspension. This goal

is measured in a parallel way forall projects through students'reports of suspension from school.The definition of "suspension" dif-fers greatly from project to project(and school system to school sys-um). School systems'often'adoptdefinitional changes to alter theappearance of high suspension rates.Accordingly, although informs an

was collected from school reco s on

suspension, that information may notbe regarded as parallel across proj-ects. Those non-uniform data will12e used elsewhere, but are" not

reported here.

Increase attendance. This goal

is measured in a parallel way forall projects thrOugh students'reports of School Nonattendance.The definition of "attendance" andthe.methods used to maintain these

:6 57 8

Environmentt; and Delinquncy

data differ greatly from schoolsystm to school system. Accord-ingly, information collected fromschool records is put to use'else-where in examining project compo-nent.s within school systems but itis not reported here.

Increase acacemic success. This

goal, is measured in a parallel wayfor all projects through students',reports of their school grades.School systems differ in their grad-ing practices and reporting formats,and they differ in the standardizedachievement tests administered andtheir test administration practices.Accordingly analyses of indlvidualproject grade and qrst 'score infor-

' mation derived from school recordsare presented elsewhere.,

Improve transition to work andpost-secondary education. Our uni-form measure of this goal is stu-dents' educational expectations.Extensive research shows this to bea useful predictor of subsequentcareer and educational attainment.Educational expectations are there-forefore n excellent proxy, for actualfolloW-ups of career and educationalbehavior. Those data, which wouldbe coscly to collect and which wouldrequir waits of several years are,not available now.

Measures of Program Objectives

Fair and consistent school disci-pline. Two scales measure thisobjective: (a). Rule Clarity', and

(b) Fairness of Rules.

. Youth, parent and community par-ticipation in school decision makingand reduced student alienation.This objective is multi-faceted, andis measured by the following:(a) student reports of StudentInfluenet,(b) teacher reports of

Parent and Community Influence,(c) teacher reports of Student.Influence, (d) a student AlienationScale.

'Preclude labeling effects.Labeling theory hypothesizes thatwhen people are treated as delin-quent, stupid, or bad that they cometo see themselves as delinquent,stupid or bad. That is they developnegative self- concepts which contri-bute to future delinquent, stupid,or bad behavior (called "secondarydeviance"). Consequently this ,

objective is measured by the Posi-4tive Self- Concept Scale (students) '

and the Low ExpectationS Scale(teachers).

Ysovide_a learning t.ructure real-i_stically tailored to promote educa-tional and social development. This

multi-faceted objective includesobjective& related to psychosocialdevelopment, educational develop-ment, and educational itructuralarrangements. Educational develop-ment is redundant -with one of the

Prdgram's goals. The following listare the measures of the psychosocialand structural objectives:(a) student Rebellious Autonomy,(b) student Prattical Knowledge,(c) student Interpersonal Compe-tency, (d) Pal4ental tmphasis on Edu-cation, (e) student reports ofSchool Punishment, (f) studentreports of School Rewards,(g) teacher renortO.of Individual-ized Instruction, (h) teacherreports of the Use of Grades as aSanction, (i) student reportsIndividualized Instruction,(j) teacher reports of the use ofType A Sanctions, (k) teacherreports of the, use of Type B San,c1-.

tiOns, (I) teacher Non-AuthoritarianAttitudes, and (m) teacher reportsof Interaction viith Students.

Theor erica! ly_ Important Out comet;

In addition to the objectivesexplicitly mentioned in the OJJDPProgram Announcement, theory andresearch in delinquency preventionand organizational change and thetheories of action underlying one ormore of the 17 prevention projectssuggest several other important out-comes.' It is importantlthat aneV, Alultion of a dplinquency(TrZwen-

.

Lion program attend to these impor-tmit intermediary outcomes because,they should help 'explain the .succesS-

or failure of a project. Theseadditional outcomes fall into twogroups: outcomes related to organi-

. zational health that may be relatedto.the ability of the organizationto implement strong intervdhtions,and outcomes known or believed to beimportant risk factOrs for delin-quent behavior.

'Organizational health. We reporton the following five(measures oforganizational health: (a) Studentreports of school Planning and

. Action, (b) teacher reports, ofschool.Planning and Action,(c) teacher reports of smooth SchoolAdministration, (d),Teaching StAffCommitment, (e) TeaChing StaffMorale; jf) teacher Job Satisfac-

j tion, and (0. teacher ProfessionalDevelopment.

Delinquency. risklfactors. Thefollowing five additional measuresof risk factors for delinquentbehavior are important(a) Attachment to Parents,(b) Attachment to soloed, (c) Belief .

in Rules, (d) Negative Peer Influ-ence, (0 Disrespect for Students.

Methods

Results presented in this chapterare based on surveys 06 students andteachers conducted in the Spring of1981 and 1982. We requested allparticipating schools to survey al'l

Enviionments and Delinquency

ni hoe lvachvr:i who Leach atI cm; t. one student in gradesthrough twelve in both years, and werequested all participating schools.Lo cooperate with the assessment ofa pr obabi 1 i ty sample of studentsselected to make possible estimatvsfor the school's population ofschools. This request was met inmost cases.'

In assessing changes in measuresof Program goals, objectives, andadditional outcomes over time, threedifferent methods were used. These

methods are described in the follow-ing paragraphs.

Simple Box Scores

Firgt, simple "box scores" areused to obtain an overall picture ofthe pattern of progress towardsgoals and objectives for the Programas a whole. In tallying thesescores, we co4ared'the,school- com-positional and psychosocial climate

T.. The New Tersey and Jazzmobileprojects did not survey teachers in1981 and there were irregularitiesin the administration of the ,studentsurveys in those projects. TheMiami project was not yet running'its alternative school in 1981. TheMilwaukee project was not operatingin schools at, the time of eithersurvey. The Plymouth project didnot use a probability student sam-pling procedure in 1982. Short

forms of the student questionnairewere made available to the Comptonand Virgin Islands projects in 1981where difficulties with reading ley-.el s/ were anticipated. Items were,-ceAsored from the 1981' questionnaireby the St. Paul, Plymouth, ChicagoBoard of Education (and thereforePCD), Pasadena, Harlem, New Jersey,and Charleston projects (listed indecreasing order of number of itemscensored). Items were censored from

so

Env t t (mulct! t n and Is. I inquency

meattores .htr each tit hool accoldi

h) asnensments made in the flitting (d

19H1 and again it the f;pling of

1982. Any given school'couldmnprove on a measure, ItTri'hb ()I

worse on the MCIISIICC or stay the!=0. In making these tallies weexcluded nchools which did notadministet the relevant hurt ion'; of

th !;chool Action Filcctivenchh `.;tit

vey for either year, schools forwhich the sampling sttatogy changedin major ways from one year to thenext, and scdlools in which thosur-vey response rates for the two yearsdiffered by more than 25;l.. Siveral

!;chool fly st ems or project di,ec t ors

censored items from the survey in1981, and two projects ceuSoreditems from the survey iu )982. For

a variety of reason: ;, we were unable.to prevent the sampling/Strategyfrom differing in some Schools inNew Jersey, the Bronx,: Plymouth, andPCD from one year to the next (saefootnote 1) . Finally, poor stuveyadministration in one or anotheryear made measures non-comparablefor the two years in some schools inCBE and the Bronx, as well as in the

LCO and Houston projects.

The sign test (Siegel, 1956) wasused to estimate the probabilitythat the number of positive or nega-tive changes observed would arise bychance were there no true Wference)from year to year.

School-e-School ExamiAt ion

The detailed results of the cli-mate assessments for each schoolinvolved in the Program were alsoexamined. I

the 1982 questionnaire by the Miamiand St.. Paul projects. No items arebeing censored from the 1983 ques-

tionnaire!

-68-

4114)40 C1110/1411iQ0. Ot1 hoot COW icnitio1I11 CI lathe

h+INI'd .ht odent repottn, t it I tc-

t jct. for the difference between 1981and 1982 means `on measures 01 hliv-

dont chatacteristics were computedbased on the observed moans andstandard deviations.for each schoolfor each year. For these schoolromposition measures, difieiencesmay be regarded as dependdble if the

-statistic exceeds 1.96.

Ichopi_pity_chosocWslimato. For

comparisons of 1981 and 1982 psycho-social climeete measures, a differentkind of "t-sratistic" was compul.ed.This statistic uses informationabout the psychometric properties ofthe measures to compare each differ- -

ence in scores to an index of the. .4

margin of error for that difference.,Specifically, the "L.-statistic"reported is the ratio of the differ-ence between 1982 and 1981 scores tothe standard error of measurement: ofthe 19$2 scores. (In calculatingthe standard error of measurement,the St. Paul schools' data were notinclulled, because these data became

available much later than all therest of the data.) As a rule ofthumb, differences that exceed twicethe standard error of measurementmay be regarded as dependable. That

is, diatferences for which theN.,

"t- statistic shown in the tablesequal to or greater than 2.0 may beregarded as dependable.

Results

Program Box Scores

Goals. An overview of the number

of schools that improved Of

regressed on each Program goal isprovided by Table 1; The firstcolumn in the table shows the numberof schools that improved on eachmeasure, the second column shows thenumber of schools that regressed(got worse), and the third column

Table

Number 4.t Alternative Education Program Schoolsthat Improved and Regressed from 1981 to 1982 on

Results Sought Goals

Number of AlternativeEducation Sites

Measure

!)ec tease Delinquent BehaviorIn and Around School

Improved Regressed No Data

H.udents' Total Delinquency 11 8 19Stndents' Drug floe - 11 8 ) 19

Students' Serious Delinquency 11 5 t, 22

Students' Victimization - 17 . 16 5

Students' Gangs in School 14 10 14Students' Safety 18* 4 i 16

Students' Reports of Community Crime 10 15 '. 13Teachers' Safety '- 21* 4 : . 13

Teachers' Victimiation 2*. 5 11

Classroom Disrupt iona : 14 13 10

Decrease SuspensionsStudents Susperv;ions 5 16 7

Increase AttendanceStudents' School Attendance 20 13 5

Increase Academic SuccessStudents' Grades 14 19 5

Improve Transition tO Work andPost-secondary EducationStudents' Educational Expectation g 21 12 5

Note. Twenty-seven Alternative Education program schqols are excluded .

from this table because of significant differences in the-samplingfrom 198] to 1982.

* p<.01

a0ne school had.no change.

769-

-fable 2

Number of Alternative Education Program SchoolsThat Improved and Regressed. from 1981 to 1982 on

. l'esuls Sought: Objectives

--,...N.Number of Alternative

Education Sites

Measure Improved Regressed No Da;,d

Making School DiscipUne'Fair and Consistent,While Providing for Due Process

Students' Reports of Clarity-of - Rriles '16 17

Students' Reports of Fairness of Rules 12 --' 21 5

Increasing Youth, Parent and Community AgencyParticipatiop in School Decision Making to (

Reduce Studeht AlienationStudents' Reports of Student Influence 9 18

Teachers' Reports of Parent and Community/'Influence i 17 11 10

Teachers' Reports of Student Influence . 15 13 11-

Students" Alienation, 23* '10

Preclude Labeling EffectsStudents' Positive Self-Coll 28** 5

Teachers' Low Expectati'ons 17 11 1

Provide a Learning Structure Tailored toRealistic Levels to'Promote Educationaland Social Development

Students' Rebellious Autonomy 23*'k 4 11

Students Practical Knowledge 11 22 5.F.,

Students' Interpersonal Competency 15 11 1'

Students' Reports of Parental Emphasis.)

od Education 2

.

12** 24

Students' Self-reported School Punishments 17 16 5

Students Self-reported School Rewards 19 14 5

Teachers Reports of Individualized .

Instruction 10 18 10

Teachers Use of Grades

as Sanctiona 12 15 10

Students" Reports of IndividualizedInstruction 10 23* 5

Note. Twenty-seven.Alternative Education program schools are excluded fromthis table.because of significant differences in the sampling from 1981 to

1982.

aOne school had no change. * p<.05 ** <.01

-70-

shows the number ,ols excluded. from the tally (!- y anavail-ablity of inform r one yearor the other, TI _le /shows thatfor nine of the tell measures ofdelinquency in and around .schoolsthe measures show less delinquencyin 1982 than in 1981. The diffcr-ences for measures of school Safetyand Teacher Victimization reach sta-tistical.significance.. The numberof school.s increasing in safety isstatistically significant accordingto both the student and teachermeasures.

Evidence in Table 1 about theother goals, sought is not clearcut,although th;,,, number of schools with

higher Attendance and students withhigh Educational.Expectations. isnOnsignificontly greater than the.number of /schools which declined onthese. measures.

ObIgctives. Box scores for Pro-gram objectives that parallel theresults for goals are shown in Table2. The evidence 'frca this tablesuggests that the objective ofincreased fairness and consistencyof the school rules is not generallybeing met. Although not signifi-cant, the pattern of results, is thatmore schools decreased on Fairnessof Rules than increased. The numberof schools that increased on t:he twomeasures of Student Influence andthe measure of Community Influenceis not significantly different fromthe number that regressed. StudentAlienation, however, decreased insigniiicantly scio than it

':early six times as many schoolsImproved as regressed on the measureof st.Idents' Positive Self-Concept.This pattern is signifi ant. Teach-ers' expectations for studentsinere,.-Ied in mere-schools than itdecreased, but this pattern is notstntistically significant.

Environments and Delinquency

Changes in student psychosocialdevelopment as measured by Rebelli-ous Autonomy and Interpersonal Com-petency. are favorable: 'Higher forInterpresonal Competency and lowerfor Rebellious Autonomy in moreschoOls in 1982 than in 1981. Thenumber of improved schools is signi-ficant for Rebellious Autonomy. By

and large,. the 'schools in the Pro-gram regressed on measures of thepresence of learning structures pre-sumed to promote social development.In particular, students' reports of.Individualized Instruction was lowerin more schools in 1982 than it wasin 1981 fon more schools than wouldbe expected by chance.

Additional outcomes, TeacherCommitment and Morale increased inmore schools than it decreased, asshown in Table 3. For Morale thispattern is significant. No "distinct

pattern of change was observed fo'rany other measure of organizationalhealth.

Table 3 also shows that moreschools improved than regressed oneach of the five theoretical riskfactors or delinquent behavior,although one of the.patternsobserved fb these changes were sta-tistically. significant.

School --by- :school Summary_

The foregoing overall talliesignore changes from one year

and

thenext for specific projects andschools. In the paragraphs thatfellow, a' detailed summary of these

--to-year changes is provided.this section., only changes that

are no,.7nally statistically Signifcant: r. described. A complete

of every school's resultsfor all measures is.provided inAppendix I).. In preparing the tablespresented in this section, a great'many significance tests were per-.:formed. When many such tests are

Measuct:

Table. 3

Number of Alternative Education Program Schoolsthat Improved and Regressed from 1981 to 1982. on

Re3ults Sought: Learnin -g Structures and Additional Outcomes

198] to 1982 Change on Measures ofLearning Structos

Number AlternativeEducation Sites-

Improved Regressed No Data

Teachers' use -4 Type A,Sanctionz,

16 13 9-

Teacheis' use of Type B Sanctions 13 15 10

Teachers' Non-Authoritarian Attitudes 16 12 10

Teachers' Interaction with Students 13 16 9

1981 to 1982 Change on Measures ofOrganizational Health

Students Rc;ports of Planning and Action 12 15 11

Teachers' Report's of Planning and Action .13 15 10

Teachers' Reports of Smooth SchoolAdminislration 15 13 10

Teachers' Commitment 17 11 10

Teachers' Morale 22* 6 10

Teachers' Job Satisfaction 16 13

Teachers' Professional 'Development 12 17

1981 to 1982 School Changes on AdditionalTheoretical Predictors of Delinquency

Students' Attachment to Parents 14- 13 11

Students' Attachment to School 17 16 5

Students' Belief in Rules 21 12 5

Students' Negative Peer Influence 16 i.3 9

t.udents' Reports of Disrespect for 16 12 t10

Students

Note. Twenty-seven Alternative Education icTogram schools are excluded from

this table because 'of significant differences in the sampling from f1981 tb

1982,

Environments and Delinqnency

performe&,Q some of them are almostserely "significant" by chancealone. For this reason,cance tests should be regarded asnominal, and interpreted with-can-

In summarizing the .sehool-by-school results, we will not beleborthe reader with a line-by-lineaccount of the tables. The moti-vated reader can sift through thedetails without our guidance.'instead, we shall-highlight selected-

' results that appear to us to suggestmeaningful patterns.

Delirvilaenc behavior. At least

one school in Charleston, VirginIslands, and Compton projects showeda significant decrease on one orMore measures of self-reporteddeliwinency. These results are pre-sented in Table 4. Specifically,

St. Johns High School in Charlestonand Elena Christian Junior HighSchool in ,the Virgin Islands 'hadsignificantly lower mean scores onthe'Self-Reported Delinquency Scale(and on the Drug Involvement sub-,-scale) in 1982'thanein 1981. BrdwnHigh School in Charleston and .theAlternative School in Compton hadsignifi-c-nntly lower scores on-. the

Serious 'Delinquent Behavior sub-Plymonth Central Middle

`ctii,ul 'and thiecontrol school forthe Kalamazoo project (South Junior

-2. In all, 1,377 tests were per-formed.- If each of these tests wereindependent, 69 "significant" dif-*Ierences at r,Iie would heexpected by .-.nance-alone, and 14

wouldbe expected by chance alone atthe .01.I-evel. Differences attribu-table to chance should be roughlyequally -divided between positive andnegative:outcomes. We observed _179differences that reached "signifi-cance" at the .05 level,-102.posi-tev,e and 77 negative.

High School) both increased signifi-cantly in Self-Reported Delinquency(and the Drug Involvement sub-scale). Finally, one of the BronxElementary Schools (No. 63)increased significantly in -Self-Re-ported Drug Involvement. .

Other measures of delinquency inand arenid schools. On other meas-ures of delinquency in and aroundschools, Charleston, Puerto-Rico,Pasadena, and 21ymouth stand oat asprojects with most significantchanges from 1981 to 1982 with someCharleston, Pasadena, and PuertoRico schools showing. decreases in

=the other measures: of .1elinquent

behavior in and-around s,hcols, orincreases in school Safety, and aPlymouth school, showing the oppositepattern (see Table 4). Table 4 alsoshows that of the schools showingsignificant changes in the variousmeasures of eelirquency, there weremore than twice rs many instances ofsignificant ime:avement as decl ir.e(27 instances of improvement, and 12of decline).

Suspenions. Results for suspen-sions are shown in Table 5. Thetable shows that in seven scheolsthe number of suspensions reportedby students decreased significantlyand in two schools the number ofs!ispensiols incteased s;gnificantly.Suspensions decreased in at leastone school in the Charleston, PuertoRico, Chicago Board of Education,and Plymouth projects; suspensionsincreased significantly in one Cdmp-

,ton and St. Paul school.

Academie and career outcomes.Table 6 shows significant decreasesin selfLreported grades in sevenschools. Table 7 shows that onlytwo schools changed significantly inthe- level of educational expectations-Tboth increased.

Table 4

1981 to 1982 School Chnniy, on Alternative Education Goal:

Decrease Delinquent ';:ehavior In and Around School

Project and Scale Score t Degrees of

School 1981 1982 statistic freedom

Measure: Students' Self-reported Delinquency- Total.

Y.alamazooSouth JHS (327)(control)

CharlestonSt. John's HS (951)

Virgin IslandsEleAa Christian JHS--

PlymouthCentral MS (41)

at,

.14 .20 3.11 423

.12 .10 -2.05 549

.09 .06 -2.92 440

.14 .22 2.66 212

Measure: S, ,:nts' Self-reported Drag Use

KalamazooSouth JHS (327)(control)

Bronx63

Charleston'St. John's HS (951)

Virgin Islands'Elena Christian JHS

PlymouthCentral MS (41).

.17 .27 3.Y 428

.05 .16 3. 193

.24 .19 -2.27 557

.11 .07 -2.81 445

.17 .32 3.48 216

Measure; Students' Self-reported Seriou,s Delinquency

Compton Action Cente,rCACYD

CharlestonBrown HS (754)

.26 -.15 -2.72 78

.10 .07 -2.01 573

O

-

Table 4 (continued)

Project dnd Scale Score t

statistic

Victimization

Degrees offreedomSchool

Measure: Students'

Constitution:',1 Rights '-

Foundation

1981 1982

Self-reported

Ell iot ..111S( 70) .20 .16 -1.99 573

KalamazooSouth,JIIS (327)(control) .13 -.21 3.79 420

Chicago Board of EducationBowen US (1240) .09 .05 -2.68 284

LeMoyne Cl (4440) .28 .17 -3.65 185

Bontemps El (5750) .15 .11 -2.27 307

Puerto RicoSantiago Gonzales (1) .14 .07 -5.31 745

CharlestonBoyke HS ,(7)5) .15 . 9 -4.01 666

PlymouthCentral MS (41) .12 .19 2.22 217'

St. Paul

JohnsouC230) .04 ..07 2.32 418Wahington(352) .10 .15 2.80 311

Measure: Students' Self-reported Gangs in School

CharlestonBrown HS (754) . .49 .74 2.07

Mease:e: Students'.Self-reported Community Crime

Virgin islandsElena Christian JUS .40 .22 -2.06

..)':!

Table 4 (continued)

Project andSchool

Measure: Teachers.

Puertollico

Scale Score t

statisticDegrees offreedom.1981 1982

Victimization

. Santiago Gonzales JHS .14 .08 34

Constitutional Right's Fndn.Muir HS (82) .16 .10 -2%28 153

Chicago board of EducationBowen HS '(1240) .23 .16 -1.96 84

Measure: Classroom Disruption

Constitutional Rights Fndn.Elliot JHS (70) 2.66 2.38 -2.14 79

Peer Culture DevelopmentLake View HS (1430) 2.10 1.82 -2.26 65

.Puerto RicoDr. Agnayo. HS 2.31 1.65 -3.25 57

-76-

Table 4 (continued)

.

.

Project and Scale Score t Degrees of

School 1981 1982 statistic freedom

Measure: Students' Self-reported Safety

Bronx53 .72 5.65

Jazzmobile88 .73 .79 2.45

CharlestonBurke HS (755) /;) .82 2.21

Measure:

Constitutional RightsFoundation

Teachrs". Salety4

Elliot JHS (70) 3..19 3.60 2.77

.KalamazooMilwood JHS (318) 3.46 3.75 2.01

Puerto RicoSantiago Gonzales JHS (1) 3.52 .).17

Dr. Aguayo hS (3) 2.81 397 7.81

CharlestonCourtenay MS (741) 3.90 4.24 2.28

A.B. Rhett MS (742) 3.35 3.67 2.17

,Cent: 1 MS (41) 3.78 3.38 -2.72

Note. Oui. schools where do change from 1981 to 1982 is regarded as depen-

dable are included on this tabe. Reports of change on teacher survey measuresare exclude,: frniii this taMe when the numbv of. twcher surveys on which the 1982

mean is based, is fewer :1, t- .statistics fo compositional measures are

based on the means and standard deviations for each school. For psychosocial cli-

mate measai-e.s, the "t statistic",. is, the ratio.. of the difference between 1982 and\

1981 .scor2s to the standard error of measurement of the 1982 scores. As a rule of

thu.ib,'psychosociat climate measure differences that are twice the Standard error

of measurement may be regarded as dependable. Dashesin the colUmn for degrees of

freedOm indicate that the measure is a psychosocCal climate measure.

c

Tablv 5

1981 to 1987 School Change ,ori, Alternative Education Goal:Decrease SUgpensions

Project and Scale Score Degrees of

School 1981 1982 statistic freedom

Measure: Students'

Compton Action Center

Self-reported Suspensions

CACYD 5 .34 -2.02 92

Chicago Board of EducationLeMoyne El (4440) / .32 .18 -2.27 185

Puerto RicoSantiago E4Onzales JliS (1) .22 .13 -3.23 743

Charl-eston

A.B. Rhett MS (742) .21 .1.3 -2.62 541

Brown HS (754) .49 .,28' 575.

Burke,HS (755) .26 .17 -3.08 655

PlymodthGrowth Works (1) .74 .20 -4.45 50

East MS (31) ..2,2 .11 -2.13 231

St. Paul .

Washington MS (352) .13 .22 2.08 301

.

Note. Only those schools where the change from 1981 to l98 ix regarded as depen-;

dable are included on this table. Reports of change on tea\q er. survey measures

are excluded from this table when the number of teacher sdrveys on which the.1982

mean is based, is fewer than 11. t.-statistics for compositional measures are

based 'on thelmeans.and standard deviations for eachs-chodl. For. psychosocial cli-

mate measures, the "t-statistic" is the ratip'of the-difference beftween 1982 and

1981 scores to the standard error of measurement of the 19'82 scores. As a rule of

thumb, psychosocial climate measure differences-that ..._e tWicethe'standard error. ,

of measurement miy b.e'regarded a dependable. Dashes Ph the column for degrees of

freedom indicate that the measure is a psychosocird clim.'at measnre.

/ -78-

I

lab It (

I ° B I A 1 t orn:, t i ve Educat i on Goa 1

I a, case Academic Success

Proiect and Se Scory t Degrees of

School 1981 I 982 statist i freedom

Coi Itt i ()[ta I Rights.

i on

Muir HS (82)

Puerto RicoRuis Belvis El (2)

CharlestonR-ivers MS (743)Brown HS 47541

VirginElena Christian JHS

r,t11,Thilt Self-reported Grades

2.40 2.18

2,89 2.652.45 2.32

529

-2.31 38'

583

-2.76 619

2,36 2.16, -2.52 496

St. PaulJohnson HS (230) 2.83 2.67 -2.02 422

Washington MS (352) 2.68 2.46 .-2.02 322

Only "those schools where the Lhaage from 1981 tc 1982 is regarded as depen-dable are included on this i!able. 2prts of chat on teacher survey measuresare e,:luded from this table when ttx number of teacher surveys on which, the 082'mean is based, is fewer than 11. L-sti;.tistics for compositional measures are

based on the means, and stzwdar eviations for each school. For psychosocial

mate measures, the "t-statistic" is-the ratio of the' difference between 1982 and

1981 scores to th7 standard error of measuremc:_gt of the- 1982 scores. As a rule of

thumb, pschA'ocial climate measure differences that are twice the standard error,-c-)fmeas.uroient may- boregarded as dependable. Dashes- in tho column for ,iegrees'of

freedom indicate that monsur:-e is :1 psychosocial climate measure.

5..

-Cir)Li 4,

-$

-1..1 1) 1 y 1

1 9 8 1 t 1 9 t i 2 51w , 1 ( " 1 3 a t l , , , t i n A l t. rna t vi. Educ a t ion Goa 1 ;

Improve Trans it ;,on to Work. and Dos t -se condary EdtiA t i on

_ , _ - . - _Pro :leer and :lcale Score t Degrees of

School 19 81 1982 stntisric r;cdom

Measure: Stu,Iyut. Educational Expettati9nf:

Chicago Board of Educationflonteml),: El ( 5750) 3.44 4..06 3 1

Br9nx61

---------------3.01 3.53 2.01 290

Note . 0U1y those cliools where the chine from 1981 to 1982 is regarded as depen-

dable are included on this tattle. Reports of change on teacher survey measuresare excluded from this table when the number of teacher surveys on which the 1982

mean is based, is :ew t than 11. t-staz.istics for compo'si...iJnAl 41easures are

based on thImeans and stLndard deviations for c:ach school. For psychosocial tli-

mate meastie'.,;; the "t_-statistic" is the rotio of the difference between 1982'and

1981 scores to the stzvidard error of measurrmont of the 1982 scores. As a,rule of

thumb, psychosocial mate measure differences that are twice the standard error

of measurement may by l',.,.;arded as dependable. Dashes in the column for :ices of

freedom indicate that the meat:er is a rychosocial climate measure.

40

1

p.

ElIV i FOIIII.,111 !, mid !4.1 1 mi ,,,, ,

1

.

Tli r:vIr y . T.! h I e 8 show!, t 1;it

' in t II i ei- Prole( t s had ?;i rii iI i cant ly less se l 1- repo. t ed 'icho,,1Nona t 1 (.110 a qcp ; one von t ioi !;(' Imo I

had Si }',1,1 1 iCall I I y moi c si:11001 N.loo.t ,ndan cc; . .

School d isc ip 1 ine . Result:: formeasures of the f a i riles s and claritye t .chool rules at e pre!,timed inTa o 1 e 9. The results in this tablemi r cur the resin t s shown earl ier forthe Pr 01,,,r am box' score on these meas-ures. As many schools sign if ican t 1 yincreased in the CI at it y -of Rules as ,decreased ( two each) , and the onlytwo school s that chin e d F; i gn i 1 i

can tly on the measure of Fa irness ofRu 1 e S .M.6: reared.

Parent, column i4.y and studentpar t ic iria t ion . Table 10 shows thatPlymouth Central Middle School si.:,

I nif icantly declined on both measuresof Student Inf thence, while one /school each in Charleston and PcD

-showed significant increases in -...1 .

Par ent-Comthun i ty Invol vemeft , andPuerto ri et(1-'sli Ted an increase in

..-Student lnf luence according toteacher reports.- Q

Ot

eienat ion. Table 4 0 also/ (di(x...;s.

that 1 0 school s significantlyt+ changed in the 'measure -n'if. student

Alienation (down in five and up illNye) . In the five schaol s withsi gni t icant increas4 in Alienation,two are control schools. Itrojett .--

school s in Plymouth. ( 2)- kind St :,.Paul(1) also showed significant..'increases in Alienation. Two ,.

.

school s in Charleston and one schooleach in Pasadena, PCD, and Ilarlem3h :d lower Alienation scores in 19 82than in. 1 981.

3. Al tlioilgh the Harlem school 'sresponse rate din 1982 was wi this 25'

its 1981"response rate, it wa'svi'; loci both y c=*a r s .

I I II tic! , I live hoots in the'

cas Posit 1 ve 1 I -conceptt ive

iurluded two ,pro ject and out controlschool Char lest on , Milwood Juni elII gh 'ic.ltool in Ka 1:1111:11',Ou d I hehaul Project -111/0 ( ()Ill 110t

).

It

eac het expect at jons . (uly oneschool changed significant ly on themeasure Of teacher expect a t ions .Expectations of 'students were moreposit ill .a middle. school iii t'

mouth in 1982 than they we re in1 981 . (Th i t. outcome is counter' to"most results for this school.)

Other measures of psychosociadeve lopment Six schools changeds f icant ly on me zistir (`.s of rebel-1 ions autonomy. Table .12 ,shows thatof these six, five program schoolshad lower scores in 1982 and one

control school had higher ';cores in1 981 . The snme table shows thatonly two progiun `schools showedi nc reas es on the measure of st u-swot s' Tactical. Knowle4ge (one eachin Ital.-A em and PI y u . Five. pr 0-grail) schools and'One control school`..ad si gni f i cant 1 y lower scores ont his 'measure. Th.esc five includeone cent re: school in Kal arnr....no , an

. e Ica:lent:any school in CBE, and threeprogram school s in St. Paul . Twoprogram school s ir e,-,,ased in In ter-pc r s ouali Comp:, t ss:loT , one *inCBE and one in ' Cons id-

r'inR these pt levelopmentarC as

n any pr og r cande C

t i011,1 S t LICi r(..`: ',Thereare 27 sigitif icant differences ontu'asitres; of educational structures

h "wn Ta b 1 e 1 2 Of.f these. 27,,. 9.r n a favo abte di. r ect i 611, andwere in the rect ion ppcsitqt

that sought cl r or systera,,t_ iic 7

a t t ern acrosc schools or pre_ject

toi'e d asp

94

1q81 to.4-1982 S(hool Chatwe on AlCernative Education Goal:Incleaue Attendance

Project and Scale Score begreeti of_

.S(hool 1981 1982 aLirdic freedom

Melpi ). Studentt;"

Veer Culture Development

Self-reported School Non-attendanCe

Harritwil US (1370) .62 .53 -2.02 628

Kafmna.Zoo

South JHS (3.27)(control) .19 .28 2.Q4 450

Jazzmobile88, .31 .18 -2.73 381

PlymouthGrow't.h Works (1) ..89 .65 -2.12 57

Note"; Only those schools where the change from 1981 to 1982 is,regarded as depen-dable are,included on this table. Reports of Change on teacher survey measures

-4r cluded,from this table when the number of teacher surveys on which the 1982mean is bAed, is r.ewer than 11, t-stutistics for compositional measiires-are

based 00 the means and standard deviations foi each school. For psychosocial cli-mate measures, the "t-statistic" is .thr rat.c of the difference between 1982 and

1981 ,scores. to the standard error ofmeasu L?ment of the 1982 scores. As a rule of

'thumb, psychosocial climate measure differences that are tw;ce the standard errorof.measurement may be regarded_as dependable. Dashes in the column for degrees of

freedom indicate that (the maasurc, fs a'psy,Aosocial cliMate measure.

s

4

;'

-

0,

.!

Tahlv 9

1981 to 1982 Slmel Ch;,,,,,,e on Alternativv Education Objective:Making Iiuu1 Discipline Fair and

Gonsitent While Providing fec Doe Process

t .111(1

heel

Scale Syore t Degrees of1981 19,82 statistic freedom

Measere: Students'

Huard of EducationRontrmps El (5750)

-r 'Thlture Development

Reports of

.77 .67

Clarity of

-2.12

Rules

Harrison HS (1370) 762 2.24

PlymouthGrowth Works (1) .62 .73 L.20

Central MS (41) .72 .60 -2.51 -s-

Measure; Students' Reports of Fairness of Rules-r

Kalama.zoo

South JHS (327)(.control)

PlymouthCentral MS (41)

.67 .54

756 .39

-2.35

-3 .34

Note, Only those schools where the,changefrpm 1981 to 1982 is regarded as depe:nda-.:.

.

ble are included on this table. Reports of change do teacher survey measures' areexcluded from this table when the number of teacher surveys on which the 1982 mean

is h;i:ted, is fewer than 11. t- statistics for compositional measures are based on::Ile means and standard deviations for each school. For psychosocial: climate mea-,

,ure,

"t-statistic" is the ratio of the difference between 1982 and-1981 scores'4to the sta dard error of measurement of the 1982 scores. As a rule-of thumb5 psy-

chosocial limate measure differences that are twice the standard error of measure -

ment may es.regarded as dependable.. Dashes in the column for degrees of freedomindicate that t1,2 Measure is a psychosocial climate measure.

.

Tdble 10

1981 to 1982 School Change on Alternative'Etication Objective:Increasing 'Muth, Parent, and Community Agency -...Participation in School Decision Making. to

Reduce Student AlienationCt.

Project andSchool

ti 2%Measure:

PlymouthCenral MS-(41)

Scale Score- t Degrees of1981. 1982 statistic fieedom

- ,r .

r

Students'- Reported StudentInfluence"r" 7".r.

.46. .35 -2.17\N

Measure: Teachers'-Repo .k of Parent and ComMunity Involvement

Peer Culture DevelopmentHarrison HS (1370)

.ChariestOnCourtenay MS (741) 1.31 1.45

1.09 1.30 . 2.9

-44;a?

.16.

Measuie: Teachers' Reports of Student. Influence in School

Puerto Rico -

Dr. Aguayo HS (3)

PlymouthCentral MS (41),

'1.52W 1.71

1.54' 1.35

G

9 7 8 4

45,

(cont.

Table 10 (continued)

'Project' and

SchoolScale Score t ' Degrees bf1981 1982 statistic freeslop

Measure:

Constitutional Right)s

FoundationEllioe'JHS (70)

Peer Culture DevelopmentHarrison HS (1370)

Students' Alienation

.42 .35 -2.99

1 .

.39 .32 -3.040

.

522

542 %-.7Th'.

*

KalamaZoo.South MIS (327)(aontrol) .37 .43 r

2.25 384

Jazzmobiledeg .31 .20 -2.04 135

-CharlestonLaing MS -(242)(control) .38 .44 2.98

F" 573;

A.B. Rhett MS(742) .33 .25 -2.97 511

Burke HS (755) .27 .22 ,2.51 608

PlymouthEast MS (31) .33 .42 _ 2.04 224

Central MS (41) .35 .44 2.03. 220

St. Paul,Murray MS (342) .27 .36 3.04 383.

Note. -Only those schools where the change from 1981 to 1982 is regarded as depen-dable are. included on this table. Reports of change on teacher survey measuresare excluded from thig table when the number Of teacher surveys on which the 1982mean' is based, is fewer than 11. t-statistics for compositional measures arebased on the means and 'Standard deviatiOns for each School. For psychosocial cli-filate measures, the "t-statistic" is the ratio of the difference between 1982 and1981 scores to the standSrd error 'of measiurement of the _1982 scores. As a.rule of

r thumb, psychosbcial climate measure difOrences that are twice the standard errorof measurement may be regarded as dependable. Dashes in the column for degrees offreedom indicate that the me&sure is a psychosocial climite meabure.'

4

.

-85- 98

Table 1144.

1981 to 1982 School Change on Alternative Edutation Objective:Preclude Labeling Effects

L

Proj'ect.and Scale Score 4 Degreesfreedot

o.

School 1981 1982 statistic

Measure: Students' Positive.Self-Concept

KalamazooMilwood MIS (318)'., .69 .74 2.69 576

JazzmObile%

88 .71 .80 3.1,7 133I 0

CharlestonCharleston HS (751)Ccontrol) .77 .80 2.08 -, 421

Burke HS (755) :79 .82 2.97 550

Haut Gap MS (944) .i3 .77 ' 2.18 444'* .

Measure: Teachers' Low Expectations

Plymouth .

Central MS (41) 35.36 23.29 -2.10 51

Note. Only those schools where the change from 1181 to 1982 is regapled as

dependable are included on this table. Reports of change on teacher'survey

measures are excluded from this.table when the number of teacher surveys on

which the'1982 mean is based, is fewer than 11. t-statistics for composi-

tional measures are based on the means and standard deViations for each.

school. For psychosocial climate' measures, the "t-statistic" is the ratio of .

the difference between 1982 and 1981 scores to the standard error of measure-

ment of the 1982 scores. As a ruleOf thumb, psychosocial(climatemeasuredifferences that are twice the standard error of measurement, may be regarded

as dependable. Dashes'in the column for degrees of freedom indicate that the

measure is a psychosocial climate measure.

-86-

I

93

4

Table 12

el1981 to 1982 School Change on Alternative Eddca.tion#3jective:eromote Educational and Social Developmentby Providing Appropriate Learning Structures

Project andoSchool

Scale Score t, Degrees of1981 1982 statistic freedom

Measure: Students' Rebellious Autonomy

Chicago Boaig Of EducationBowen HS. (1240) .62 .53 -1.98

Luella (4550)(Contro0 .55 .68 0, 2.09

KalamazooMilwood JHS

fazzmobile88

(310- ,.73

'.74

.63

.56

-3.14

-2.78

Puerto Rico..Santiago Gonzales JHS%(1) .47 .39 -3.13

PlymouthEast MS (31) .79 .69 -2.28

Measure:

, 264144

49A

116

688:

215

r

Students' Practical Knowiedge1.

.

Chicago Board of Education .

Blaine El (2300) .

KalamizooSouth JHS (327)(control).

Jazzmobile .

88. .

Plymouth..Central MS (41)

St. Paul

1..37.1.26 .315

.391.25 X2.92 375-,

.97 1.35 4;33. 126

Ce;,.tral HS (210) 1.57

Murray MS.(342) 1.40WaphingtoA MS (35.2) 1.19

-87-

.444 2.10 221

1.36 -4.03 2581-,415 -4.82 3661.03 / -2.58 261

Project andSchooltit-,

Table 12 (continued)

Scale Score t Degrees' of

1981982 statistic freedom

7,

4Measure: Students' Interpersonal Competency

Chicago Board of Education'.Bont6mps El r(5759) .80 .86

z,

2.50 277

Charleston7

4A.B. Rhett MS (742) .75, .80 2768 .503

Measure:. StudeAs-Reported Parental Emphasis .on Education,

Peer Culture DevelopmentLake View HS (1430) -2.253 . 480

Chicago Board of EducationLeMoyne El (4440)-

' KalamazooRilwood JITS1(318)

IoutOHS.(327)(contrp0

Bronx22

63.

Puerto RicoRuis ,E1 (2)

Dr. Aguayo. HS' (3)

.68 .60. L2.:03 188.

.68 .62 -2%56 623

.68. .61 -2.45 425

. 66 .56 416

. 75 .61 -3:57 169

. 71 .64 , -2.35 . .336

. 63 .5g -2.18 805.

Measure: Students' Reports of School Punishments

Bronx22 .27 .21 -2.30

463 .31 .22 -2.66

CharlestonLaing MS (242)(contrOl) :16 .22 2.45

grown HS (754) .24 .19 -2.401.10

PlymouthCentral MS (41) .16 .27 2.t4

437187

613580

221

.Project andSchool

Table 12 (continued) "'"".:

Stale'Score1981 1982

Degrees of

statistic-- freedOm.

Measure:, Student's' Self-reported School Rewards

Chicago Board of EducationBlaine,E1 (2300)

Bronx22

4 63

Puerto Rico I. 4' ,

Santi'ago Gonzales MIS, (1)

Ruis Belvis El (2)

Virgin 'Islands. ...',

.43 .28 -6.52 742

--.47 .35 -3.55 336

.28 .33 . 2.00 - 473

.

.26 .16 -2.68 22..

.231 3.0.3

.33 .24

.52 .32

.29 .19

.15 .22

-3.16-3.91

-3.322.37

319

4\

438187

402309

Elena ChristianJaS ,.

PlymouthCentral MS (41)

St. Paul

0 Murray MS (342)Washington ,MS (352)

Measure: Teachers' Reports of IndividualiiedInstruction,

Charleston .

Charleston HS-4451)(controi) 1.82 1.09 -2.08

Bronx63 1.94 1.18 -2.80

-89- 102IL

-ra

A-able 12 (dontinued),

. 4 ..44

-

-PrOj4t and - Scale'Score t Pegrdes of

School .. 19tlf 1982 ' - statistic m./ freedo.

,.-u-q

, . .7..:

Measure: Tea,chers U. se of Type BSanctipns.,'

, , f. .

Charleston .

s.

Charleston HS (751)(control). 2,47 1.98TO '

-2.31, - 27 . A,

,.

.. 4%...p.,_

Measure:

-

TeaOlers' Non-Authorftar,lan Attitudes

-,Virgin Islands .

4,Elena Christian JX5(0) 2.31 2.61 : 118

4

Note. Only those schools where 'Elie change:from\198I'to 1982 is regarded as-

dependable aye includ on this table. !Reports of change on teacher survey

measuresare excluded from this table when the number of teacher surveys on

which the 1982 mean is based is fewer than 11: t-statistics,.ftv composi-

measgreaarebased.on the means and standard deviations for each t

school.. For psychosocal climate measures, the "t-statistic" is th*e. ratio,

of the CA:ference'between 1982 ansd 1981 scores to the standard error of

measurement'of the 1982 scores ,Asa rule ofthucb, psychosocial climate

measure differences that are twice the standard error of measurement may

be regarded as dependable. Dashes in-thecOlumn for deKliees oijreedom.

indicate that,the measure is a psychosocial climate measuie.5

1

\.;

1

t

CP

Environments and Delinquency ,

c..,appears in the results, al.though itis striking that'no School increasedon Parental,Emphasis on Education' _r

k; Individualized Lnstrction, And %

eightodecreased on'Parental Emphasisandtiwe'deCreased in the measureofIndividualized Instruction..

.

Organizational'health. Thlteopes%

for the measures . presumed to "berelated to prospects for organiza-

. . .

tiontl del/elopmedt e'sfiown in

Y Table03,, In all; 25' ignificantoceurred, of whi h 17 are in

'the positive direction. Of theeight negative changes, three werefor'control schools, The measUres.showing most positiIe changes wereTeacher Commitmentand Teaches

Morale.. Three Charleston schools,'ahigh school in Puerto Rico, and EastMiddle.gool'in.plymPuth signifi-7:Cantly improved on Commitment. orMorale'. Two schools in PuertO Rico,one...in-Charleston, And the program '

school in Kalamdzoo.shpw.ed signitfi-"cant increases in teachers' reports''Of Smooth -Administration:

Other delinquency risk"faciors.A' summary, of significant differenceSbetween 1981 and 1982 for other the-oretical risk factors for deliniivetbehavior is 'presented in Table 14.Four program schools showed m

.decreases in Attachment tp Parenlp,and a control school showed 'anincreape. Attachment to Schoolincrealped in two program schools and'one cpptrol school and, decreased infour programachools and, two controlschools. Four schools increased instudent Belief in Rules and oneschool dei'feased (all are programschools). The 'program schools show-ing'in renses!wer6)inSt. Paul,Char]. ston, Harlem Ssee footnote 3),and C E.

o.

1

Negative PeeruIpfluence decreased. inlour,Trogram schools in Charles-,ton (2), Harlem, and Pasadena; andit - increased bn qne program and onecontrol.. school in Charleston.-

. '

Discussion

Co-occurring events. The majorlimitation of the results presentedhere:is that they all.detcribe'changes in die pOpulation le"trels ofdeliAquefitibehavior, other, behavior,1

other pdycho415cialcharacteristies,and2gichoel environment betieen 1981and 19821 Without referent to the

source -of the change. HAny'ma'or.\event that occurred in the environ.men , structure staffing, aeader--s .p, .or' financial resources availa7

:ble to a school--or even economic.conditions. and school system

potdtial explanations of4S'OCII changes from year' to year. The

AlfernativeEducttion proj4cts oper7.'acing ix these schools are generally.anesucll matior' occurxance.

s'Sometimes it-is'difficult to havemuch confiderfee in an interprqationthat a project operating in a schoolcaused ttkchanges observed. For j.

example, ajunior high sqbool in.'Ply.lhouth/shawepl signifitant

increases 0- Delinquent BehaVior,student Aliendti3On, School Punish-niehts student Victimization; 'itshowed significant Aecrease6 in .

teacher Safety,. Fairness of Rules,Clarity of Rules, Student Influence

. as reported by teache4g and;stu-degts, School' Rewards, Attachment toParents, And.studpnt'reports ofschool Planning and Action. Butthese outcomes are,hard.to interpretas effects bf the alternative educa-1,tion project operating in thesChOol. The project primarily provides direct services to a selectedgroup of students had no majorproject component 'directed at Schoolorganizational change, and thereforeis not a plausible explanation ofthis substantial deterioration in,school climate. A more plausibleexplanation' of the results is thechange in grade-level Organization

-91- EN/

A

4.

Table134P'

1981 to 1982 School Change ck'Measures (4 Organizational-Health

Project and' Scale Score,

t .Degrees, of

qtatistiC freedom

-11r

planning and Action

School" 1981 1982

.

Measure: Students' Reports of School

Peer Culture Development ,° Harrison HS(1370) .46. .56 .2.05

, 1. .

.

Chicago Board of Educationc

LeMoyne El .39 .49 2.08,(4440).

.,

°CharlestemBtorke HS (755) .66 .55, -2.19

.Plymouthf

Central MS (41) %47 .37. -2.04

Measure: Teachers' Reports of Planning and Action

Puerto RichDr. Aguayo HS (3) 1.68 1.83 .26

. . Measute: Teachers' Reports of Smooth AdmOnistration

Peer Culture DevelopmentCurie HS (1870) 1.63 1.50 -2.17

KalamazooMilwood JHS (318) . 1.58 1.80 3:60'

Puerto Rico(1.' Santiago Gonzales J8S (1) 1.63 1.75 2.09

-Dr. Aguayo HS (3) 1.69 1.83 2.41

CharlestonCh4fleston HS

(751)(oontrol) 1.70 1.53 -2.88 -7'

Haut Gap MS (944) 1.66 1.80 2.46

co

10592,

Project andSchool

ti

Table 13 (continued)

Scale Score19810982 StAtigteiC

Degrees of'

freedom

Measure: Teachers' CoRmitment

T

Constitdtional' Rights'FOundationMuir HS (82) 1.68 1.50

Puerto RicoDr. Aguayo HS (3) 1.63:1.87.

Charlestonk

A.B. Rhett MS (742) 1.41 1..62

Haut Gap MS (944) 1.66 1.91St. John's HS (95117 '1.49 1.67

PlymouthEast` MS (3 1.40 1.76

Measure: .Teacher Morale

Kalamazoo A

South JHS (327)(control

Puertp RicoAguayq H$ (3)

CharlestonHaut Gap MS (944)St. John's HS (951)

1.62 1.49

1.57 1.76

1.61 1.761.50 1:62

2.85

.' 2.49

. 2.91

. 2.18

3.54

3:60

6

2.782.12 ,

>

-2.40

.0

r

4.11

los

A

.

k4,

a5

Table 13 (continued)

\

Project and Scale Score t. Degrees of,

School , t '19.61 1981 statistic freedomv., _ , , . r

'. hl--- '' Jobgc pis $.atiSfaction IIP7-

8

- .,7.

, ---, n

Measure: "ie

» v.

.. .... ).

Puerto Rico ,

°tic>. Camino\ , .3.30 2.:94 .-2.13 - 18

. ,

''

Virgin Islancis8.

... ,

. .Eleta thrittiadd.JHS . 2.;2 2.72 ;!97128 ..

A ..0

I

.

, Meaturk:. Tgchers' Professional Development...

If . / ,r- 7-7.

48

-.t1.. r

ihicap Boa'ra of Educatton-f- 0 $ %

Bowen HS 0.140) --:\ J.31 1.44 2.01. 88.

.

Kalamazoo ' . .

A

South ..HIS (327)(control) 1.54. 1.36 72:45 30

<4,

% \ .

Note: Only thote schoolt where the change from 1981 to 1982 is regarded:as .depen-

dable are included on flqs table, Reports.of,chtnge on'' teacher survey measures

,are vxcluded fOmIthis table when,the.nO0.mber of ,teacher surveys on which the 1982

mean is based, .ks fewer than 11.. t-stastics for compositional measures are

abased 'on the means and standard deviatiOntjor each school. For psychosocial cll-.

%mate measures, the "t- statistic?" is the ratio of the difference between 1982 and

.1981 scores .eo the standard error of measurement of the 1982 scores. As a rule of

thumb, psychosocial climate measure differences tHat are twice the standard error

of measuretent may be 'reg4rded as dependable. iDavhet in the column for degrees of

freedomindice that the measure is a 'psychosocial climate measure.

e

V

-96-7

.

107

ti

Table 44

1981 to 1982 School Change onAdditional Theoretical Prediciora of Delinquency'

Project and.School

.

. Scale Score t Degrees of4981 1982 statistic - .freedom

Measure: Students' Attachment .to Parents

00*

Constitutional RightsFoundation . 1

Elliot JHS (70)1,-:.

.64 ,..58 -2.39Muir HS (82) :. .0, .56 ' . -2.66

--, . ',.

Chicago Board of EducationLuella El (4550)(control) .57 .65 . 2.114

Btonx\63 .71 .62 -2.48

PlymouthCentral. MS (41) .72 -.56

J'Ileasure: Students' attachment to SChool

Peer Culture 6evelopment.Hdrrison HS (1370)

Chicago Board of EducationLuella El (4550)(control)Bontemps El (5750)

. .

KalamazooSouth JHS (327)(control)\ .65 .57 -3.08

I

r

.71 .75 2.01

,

.57 .65 2.13

.71 .64 ., -2.78

4E

-957

108

601

508

1 '

225

4

-'5904

...,

II"\

176308

(1

'r

4

Table 14 (continued)

'Proje'ct and

School ,

Scale Score t Degreds of1981 1982 statistic freedom

Measure: Students' Attachment to School (continuO)

Jazzmobile88

CharlestonLaing MS (242)(cintrol)

Si. Paul.

Murray MS (342).

.§8 .78 2.78 183

.69 .63 611

.71 .59 4.09 402

Measure: Students' Belief in Validity of Rules

L,Chicago Board of Education

Bowen HS .(1246),Bontempi El (5750)

jazzonobile

.66

.77

.60

2.162.17

88 .62 .73 3.01

Charleston

t;4 ,Haut Gap MS (944)- .64 .69 2.32

St. PaulCentral HS (210) .67 '.74 2.33

96

270282

257

Table 14 (continued)

l'rbject andSchool

Scale Score t. Degrees of\1981 1982 statistic freedom

Measure: Students' Reports of Negative Peer Influence

Constitutional RightsFoundationMuir HS (82)

Jazzmobile88 .21 .16

arlestonA.B. Rhettq4S (742) .17 .14

. 19 .15

Rivers MS (743) .18 .22

Charleston HS ,(751)(cOntrol) .13 .16

Burke HS (755) .22 .18

Measure: Disrespect' for Students

KalamazooSouth JHS (327)(control) . 91 1.08

/-2.26 /499

-2.13 312

-2.34 585

2.31 5352.08 489

-2.95 699

2.20

Note. Only those schools where the change from 1981 to 1982 is regarded asdeperdAble are included on this table. Reports of charge on teacher survey measures areexcluded from this table when the number of teacher surveys on which the 1982 meanis based, is fewer th en 11. t-statistics for compositional measures are based onthe means and standard deviatioils for each school'. For psychosoci'al climate mea-sures,the "t-stati,flic" isthe'ratio of the'difference between 1982 and 1981 scoresto the standard errtk4of measurement of the 1982 scores. As a rule of thumb,/psy-chosocial climate measure differences that are twice the standaid error of measure-ment may be xegaided as dependable. Dashes in the column for degTees of freedom _

"indicate that the measure is a psychosocial climate measure.

-97-,

Environments and Delinquency

that occurred in the Plymouthschools during the 1981-82 schoolyear, Both middle schools were con-verted to.'junior high schools, thusreturning more troublesome 9th gradestudents to the school rather thanmoving,' them onto the high schoolcajbtpus..

A similar grade reorganizationoccurred in the Puerto Rico schools.Structural changes in the PuertoRico schools between 1981 and 1982involved decreases in the school.population and a move from split.tosingle sessions. Results for boththe Plymouth and Puerto Rico proj'-ects should be regarded as tricky to

interpret.

A similarly difficult to under-.stand set of results occured for aPCD high school where School Nonat-tendance, Alienation, and studentreports of school Planning andAction decreased; Significantly andwhere Clarity of Rules,, Parent and -

Community Involvement, Attachment toSchool increased significantly.This project operated a closelysimilar intervention in two otherhigh schools in the same city, andthe results do not resemble thehighly positive change in schoolclimate registered for the school

just described. In this case it ispossible that the PCD project con-tributed to the positive changes,but neither the experimental compar-ison of the- project's interventionwithin this school (St. John, 1983)nor convergent evidence from other

schools supports this interpreta-tion. Other co-occurring events areexplanations that rival the inter-pretation that this felicitous out-come was due to the project's inter-

vention.

A third example involves resultsfor the LCO reservation school forwhich results were not tabled in thethis chapter (they are shown in

Appendix D) because thestudent-survey response rate dif-,feredbmarkedly for the 1981 and 1982administrations. Political changeson the reservation that resulted inthe firing of a leap proportion ofthe teachers in thelschobl are pro-bably major contributors to thedrastic negative changes reflectedin the teacher survey results.. In

this school, teachers reported sig-nificantly lower Safety, less Stu-..dent Influence, less Parent and Com-munity Involvement, and very muchlower Morale in the second year.This outcome lends support to thevalidity of the teacher-survey mess-,nres, but it would be absurd toattribuie the results-to-the opera-tion of a 'relatively ow -key project

(Cook, 1983b).

R ders can gain_greater insightinto the nature,' amplitude, anddirection of other influences oneach project by consulting the firstand second interim evaluatiqpreports specific to each.

Changes in sampling. The results

summarized in this chapter excluderesults where we knew of changes inthe ways students were sampled.These changes, which we have takengreat pains to avoid, neverthelessoccurred on occasion. -For example,

the results do not include reportsfor the Plymouth high schools wherewe were unable to obtain a suitablerandom sample of the studentry in1981, and where English classes weresampled in 1982 as an expedientapproach to the assessment of schoolclimate.

Despite our-efforts to excludeobVicusly.less meaningful compari-'sons, there may remain some compari-sons where 'the quality efesurveyadainistratibn differed to a prob-lematic degree for the two years.For example, one school in.St. Pauladministered,surveys to 56Z of the

.

sample in 1981; the report of surVeT.administration for this school indi-

, cated that surveys had been adminis-tered to 78% of the sample in 1982.Many of the St. Paul surveys weresent to us with the..identification -)numbers removed, however, and thenumber of booklets identifiable asbeing .from ,this school ,resulted in

An effecti4R. resPonse rate of 59%.Neither the effective 1982 responserate (59%) nor,,the reported responserate (78%) were sufficiently differ-ent from the 1981 response rate(56%) to warrant excluding tschool by the criterion we use .

Tables 4 through 14 show sever 1puzzling significantly negative'changes between 1981 to 1982 forthis schor' The hypothesis thatsample dii. ,nces'explain theresults is at least as plausible asthe hypOthesis that Project Togetherbrought about negative changes inthe school. There is no way ofknowing how much the group for whomquestionnaires were available in1981, resemble the group for whomquestionnaires were available in1982, or how much either groupreally represents the schools stu-.dentry.

An example where the responserate decreased, from 1981 to 1982 may.illustrate the obverse of .the prob-

lem just described. One Pasadenanschool surveyed 79% of the sample. ofstudents in 1981 but only 63% of thesample the following year. Thisschool's tesults\show lower StudentVictimization, Classroom Disruption,and student Alienation. Responserate differences are explanations of.these results that are at least asplausible'as the interpretation thatProject STATUS induced thesechanges.

These response rate problems donot, of course, affect the interpre-

Environments'ind Delinquency

tatipn of results for schools withhigh response rates for both years.Response rates' for the 1981 - survey

are presentedby Gottfredson, 1982,and a detailed listing of 1982response rates is presented inAppendix E. -

Chance. Some nominally signifi-cant results may be attributable tochance.- See the method section fora description of thi ssue.

Secular trends. cientists refer

,to shifts in the general cultureover time as secular trends. Cur-rently% there appears to be adecreasing interest among educatorsin individualized instruction. Thigsecular trend appears to be a plau-sible interpretation of the statis-tically.significant,tendency. towarddecline in individUalized educationfor schools studied here. The puzz-ling tendency_for school averages on.Parental EmphaSis on Education maY------

also reflect a secular change due tothe recession or other unknowninfluences especially because sofew. of the-prevention projects .

systematically engaged in (activities:likely to.influence this outcome:

Changes do not reflect absolutelevels. The results reported hereare for changes-in th'e level of thevariables, examinedrather than the 1

levels of the school characteristicor student outcomes themselves. A

school that is already extremelyhigh in 'Attachment to School, forexample, may remain high from yearto yeari but show no change in thidmeasure/. Perhaps no improvement isneeded1, Readers interested in this

issue )ahould consult the normativqinformation provided in an appendix,and see the second interim report;for he Charleston project (Gott-1fredson, 1983) 19here this -issue ifsexamined

1 1 2-99-

Envirdnn9ts and Delinquencyti

Some Intei\pretations

k

Despite the foregoing worries,the results.presented in this chap-ter are based on reasonably sounddata-ghd, represent an unusually tho-roughgoing scrutiny of the changesover time in the school climates,and in the behavior, and psychoso-cial status of the student popula-tions involved in the Program. The

information presented here is limited to population outcomes. Per-

spectives on the effects of project'components targeted at defined sub-populations are provided in theindividual interim evaluations ofeach of the prev,ntion projects. A'subsequent chapter provides a terse,summary of highlights from thoseindividual interim evhluations.

The most important observationsappear to be the following:

1. Schools involved in the Pro-gram are safer in 1982 than theywere in 1981. Both teachers and

. students report more safety in the

second year of the Program than theydid in the first, and the improve-ment is statistically significant.

2. Teachers in Program schoolswere victimized less in the secondyear of the Program 'than they werein the first, and the improvement is.statistically significant. Teacher

Victimization was down in 22 Programschools, and was nonsignificantlyhigher .in only 5' Program schools.

'There is a tendency for measures ofdelinquency th and around schools toreflect less delinquency in'progr.amschools in 1981 than in 1981. For

nine of ten Measures of delinquency,the measures,ahow less delinquencyin 1982 than in 1981, although onlythe difference for the Safet,vandTeacher Victimization measures reachstatistical significance. /A,school -by- school analysis of changesshow that when all measures ofdelinquency are taken together, more

than twice as many schools showedevidence of significnatly lessdelinquency in 1982 than showed evi-dence of significantly more delin-quency.

3. The number of schools withimproved attendance is greater thanthe number with worse attendanc,but this difference is not statisti-'cally significant. Schools in three

projects had-Significantly4betterattendance in 102.than in 1981; noprogram school had significantlywors%ottendance in th-Second year.

4. Student Alienation decreasedin significantly more schools thanit increased. Schoo4 in four ofthe delinquency prevention.proiectssignificantly improved on the meas-ure of Aliefiation.

5. Nearly six times as many:schools improved as regressed in

,measures of student Self- Concept,and this Pattern is statisticallysignificant.. Schools in four of thedelinquency prevention. projects sig-nificantly improved on the measure

of student Self-Concept.

6'. Students' in significantlyMore.schools reported less Rebelli-.ous Autonomy in 1982 than in 1981.

Five program. schools decreased sig-nificantly on the measure of Rebel-lious Autonomy, and no?, program.

school increased.significantly onthis measure.

7. On the various measures ofpsychosocial development, more thantwice as many program schoolsimproved as showed a decline. The

pattern pf. results suggests that \

get

modest positive results wereachieved indesired areas, but thil;\

pattern does not,provide much sup-port for a conclusion that theseresults were brought about throughthe specific structuial alternativessought in the OJJDP Program design.

-100-

113

For example; the Program Announce- .

ment c:alled for "providing learningstructures tailored to realialclevels to promote educational andsocial. development." For the pro-gram overall, measures ot individu-alized instruction went Yignifi-cantly down between 1981 and 1982,and most of the other measures Ofalternative educational structuresshowed a tendencyto decline, butmost measures of social'developmentwent up. As a second example, theProgram Announcement called for."increasing_youth,'parent and commu-nity agency participation in schooldecision making, to reduce studentalienation." Although alienationwas reduced, students report (non-significantly) less influence inmore schdols.Chan they report more,and there is only slight suggestionof increases in parent or communityinfluence. Furthermore, an examina-tion of the schoo17by=school resultsdoes not reveal many instances ofco- occurance of desired changes in .

the measures of educational struc-tures and the measures of alienationor psychosocial development.4

8. ,Teacher tommitment andteacher Morale,intreased in.moreschools than it decreased, and forMorale this pattern is statisticallysignificant. This appears to be animportant outcome because of

4. We explored this issue furtherby examining the school-level corre-lations between measures of the edu-cational structures, ant student psy-chosoCial developmat.' In general,these correlations provide littlesupport for an interpretation thatindividualized instruction isrelate4 in the way anticipated withfavorable psychosocial outcomes. In

contrast, the correlations do sup-port an interpretation that studentinfluence leis negatively related toalienation.

Environments and Delinquency

evidence.from. other research thatteacher morale and staff Lommitmentare important correlates of programdevelopment.

,

9. More schools improved thanregressed on each of five theoreti-cal risk, factors for delinquentbehavior, although none of the pat-terns were statistically.signifi.-cant.

10. Positive changes do notoccur. with equal frequency acrossall prevention projects., The mostconsistent evidence of. positivechanges in school-level outcomesoccurs for the Charleston, PuertoRico, and Kalamazooprojects. Allthree of these projects are p0.maryprevention projects that focus'onchangesin the environment,. that isthey are Quadrant 1 projects interms of the classification pre-sentedin Chapter 4. They wouldtherefore be expected.to have, largereffects on school climates thanwould projects focused primarily on.a subgroup of the population. In'

our judgment, theegharleston projectis untertaking thorough efforts tosystematically implement well -de-fined. interventions aimed at alter-ing the broad school environment;and the Kalamazoo project is alsoclearly focused on broad-rangingschool climate improvement. The.

grade structures of the schools inwhich the Puerto Rico project oper-ates were changed between 1981 and1982.' This reorganization resultedin the elimination of split sessions.and adecreate in school popula- '

tions.. These structural changes,are the most plausible explanation,for' the significant positive find-igs for the Puerto Rico schools.;The school that experienced the mostimprovement is. the school that con-.vetted from a 9-12_to a 10-12-grade.

mstructure. Also', most of the signi-ficaht inproveMents are on teachermeasures. The focus of the projectis on students, not teachers.

-101-*

114'"

Environments and Delinquency

Creating changec in the climateof schools of sufficient magnitudeto have substantial effects on the .

incidence of delinquent behavior isbound to be difficult. We interpretthe evidence presented in this chap-ter as implying that postive changesin school climates, includingchanges for known risk factors fordelinquent behavior and known corre-lates of program development, haveocOrred. Significant increases inschool safety are already,apparent.Provided that the implementation ofprevention projects continues toimprove over time, these encouragingresults suggest that future resultsmay be more impressive.

11. The overall results mask_impressive instances of progresswithin some projects. The evidencereported in this chapter, takentogether with evidence presented inmore detail elsewhere (Coo,k, 1983a;D. Gottfredson, 1983) strongly'sug-gests the interpretation that theCharleston and Kalamazoo projectsare developing as promising delin-quency prevention interventions thattake a primary prevention, environ-

,

1

mental change approach. Plans tocontinue to develop, evaluate, dis-semifiate information about, andreplicate those projects. should be'made. .

Epidemiological Indicators vs.Between Group Differences

All of the results presented in ,

this chapter are for school popula-tions. Sothe projects have no sub-

stantial interventions intended toinfluence the enviropment of theschool as'a whole. Specifically,

some 'of the :projects direct theiractivities to preventive or remedialinterventions with selected groupsof students For such projects,,effects are most likely to be evi-dent in comparisons of youths whoreceive direct services,with controlgroups. This chapter does not beardirectly on the efficacy of inter-ventions, targeted at selected groupsof students. Evaluations' of such

interventions are discussed indetail in the reports of the evalua-tions of those individual projects.Highlights from those interim evalu-ations are described iinChapter 6.

115

eJ

ti

The Effects of Intervehtions Targeted at Identifiable

Groups of Youths: Some Highlights

All of the delinquency preventionprojects in the Alternative Educa-tion Program had at least o

component directed identifiable target group ofindividuals.These interventions were averse.Some were directed vt a highlyselected.grotuo of high riskduals, some' were directed, at 'youths',

more representative of the generalpopulation.. Some involved alteringthe' environment to which the targetgroup was exPosed, others involvedefforts to enhancethe.coping skirlsor ability of the individuals toadapt to tile environment.

We took great paihs to work withprevention project implementers todevelop evaluation .deigns thatwould enable' us to gage the efficacyof the interventions targeted atidentifiable groups. Our aim was toCreate circumstances that would makepossible confiderlE cOn9lusiona aboutthe effects of specified interven-tions, while bearing in mind that in'the preliminary stages of a proj-,ect's development rigorous outcomeevaluation may be an egregious exer-cise. Put another way, there mustexist a reasonably well developedand specifiable interverCdon to

4 'evaluate. Furthermore, implementingcareful experimental evaluations offield trials is a very difficultundertaking. Some of the difficul-ties encountered in. `convincing proj-eqt implementers of the importanceof bearing the burden Of rigorousoutcome evaluation.are described:inour first interim report (Gq.ttfred-

so 1982c, Chap.

thatIt is evident that good many, of

the interventions being implementddiwthe Alternative Education Programremain., in incompletely developed

or

form. We perceive steady progress.in many projects in refining theirprograms _over time as they gainexperienCe 'benefit'frcut evaluative'information and from technicalassistance, and put rudimentaryideas to test. It is alsO evidentthat a good many. of the attempts toimplement interventions in this Pro-,gram have been thwarted by exigen-cies beybnd the control of. theiTplementersor not foreseenby any-one.- In other cases, the projectimplementers do. not appear 'to aimsysteMatically to dev.O.op specificinterlientions but rather to take

.advantage.of opportunities thatexist.' in the project's environmentto achieve the adoption of any inno-vation that appears to hold promisefor Moving the school in a desireddirection. And in 'some .cases,. the

resources--time, talent, money,technologies - - required to implemett

what was intended are not availablein a project: .

In short, the prevention projectsare not' only variable in terms ofthe _ocus'of the interventions tar-geted at identifiable groups, butthey are also variable in terms ofthe integrity and developmentalstage of those interventions, and interms of how stable or well definedthe interventions are.

Commentators (Farrar t House,.1983) on the evaluation of Push /.Ex -.cell, Jesse Jackson's highly pub-licized effort to.insPire youths tostay in school and perform well .

,here, have Made an interesting sug-estion. Push/Excell may have been

a movement rather than-a 'program.According to Farrar and House, theMOenientaspect of Filsh/Excell vas'expressed by a'compelling message:

116.

Targeted Interventions

"That hard work, self-discipline,delayed gratification, and persis-tence were qualities that youth .

needed in order to succeed" (p. 37).The movement had a catalyst--JessieJackson. The program aspect of

/Push/Excell. was expressed in a set4.,rof 10 guiding principles, in some

suggestions for impleMenting.theprogram, and a skeleton staff that.provided some (but probably notenough) assistance to the'implement-ing sites. An implication we drawfrom this commentary on Push/Excelland. its evaluation is that programevaluations may%be best suited tothe evaluation of programs, and '.not

all activities' are really programs..

None of the Alternative EducationProgram's projects.are movements,but some of them do.not_resembleprograms very much, and they aredifficult to evaluate. as prOgraMs.Such projects squirm like.reluctantcats when attempts are made to cramthem into a box. We have judged'. it

futile to attempt to encase certainaspects of some projects--the effortis beyond our resourc s, and, theevaluation would likel 'miss thepoint.

Evaluation Designs .

For all the foregoing reasons theevaluation designs for the compo-nents of the 17 prevention projectsdider. We had expected that theevaluatiOn designs implemented inthe second year of the AlternativeEducatiOn Prpg;am would beo muchstronger than they were in the st

year. They are much stronger. The

timing was better, experience hadrbeen gained in the first Year,' Andeveryone had a better idealof whatwas required. A

The designs as implemented arebriefly described in Table 1.. Trueexperiments involving the randomassignment of yol9is to treatment

and control groups are generally tobe preferred. Seven of these pre-vention projects implemented random-ization for at least one of theirinterventions.' When randomizationis impossible or not feasible, theuse of a comparison group created' insome other way. is necessary. The

more "equivalent" this 'comparison

group the better: That is, when acomparison is markedly differentfrom the group receiving treatment,a host of potential explanations fol./

differences obserVed in outcomemeasures are possible. The availa-

b il ity of pre-intervention informa-tion is often useful either toassist in ruling out a hypothesisthat outcomes observed are due to .1

pre-existing differences between the.treatment and Comp rison groups, orto increase the efficiency (static-tical power) of an evaluationdesign. Finally, sound and compre-hensive outcome measures are,required to assess the effects ofany intervention.

Taking all of these considera-tions together, the projects thatwere most mnewble to the interimoutcome evaluation of their targetedinterventionsin fhe second year (dill

the Program were those in Compton,Chicago (PC4), Charleston, and Mil-waukee. Each of these projectsinvolved the' random assignment ofyouths from a pool of eligibles fortheir Major interventions targetedat identifiable,groups; each wasable to develop comprehensive out-.dome measures. The projects locatedin 'Pasadena, Kalamazdol Puerto Rido,\the Virgin-Islands, Mi i, and Ply-7

mouth either implemented reasonablystrong quasi-experiments for theirmajor interventions, or implementedtrue experiments for some project

components. The highlights pre-sented below describe soMe of theresults for these projects whichwere most'Imenable to interim'

assessment of effects.

-104- .117. .

Table 1

Evaluation Designs for Project Components Targetedat Identified Croups

Project randoMControl group Pretreatment` 04come"equivalent" nonequivalent . measures measures

Compton, CACYD yes NA NA yes yes

Pasadena, STATUS no yes yes partial YA8

Chicago, PCD yes' NA NA no .ieso

Chicago, CBE partial . no not

yes . partial

Kalamazoo, AEP ' yesa no yes no . yes

Bronx, PREP yes NA . NA part , partial

,Harlem, AAEP no no yes no partial

Puerto Rico, OC no no yes yes yes

Charleston, PATIIE yes NA NA yes yes

Houston, dIS no no 'yes ne no

.

Virgin Islands, AEP no no yes yes yes ,.,

Hayward, LCO no ':- no -yes partial partial

Miami, ACE 'no yes NA yes yes

Plymouth, AEP yesb no yes oyes yes

New Jersey, EIC . no no yes partial partial

Milwaukee, MYEC Yes. NA NA .. yes yes

St. Paul no no yes no yes

4

aStudents who partiCipateil.in tHe student council were randomly assigned to participdte

within homerooms. The design for other project components is a non-equivalent control groupidesign.

bNo control group was available for the Learning Options (Crowthworks) component. The

design'for the middle school treatments was a non-equivalent Control groupdesign.

Eighlight

In the following abstracts wesummarize.same'of the major resultsof the interim assessments of theeffects of project components tar-7-geted at specified groups. This

'account is not comprehensive,' andreaders are encouraged to consultthe,more extensive accounts pre-sented in Part II of this report.

Compton

The Compton Action Alternative'School (formerly the Compton ActionCenter for Youth Development, CACYD)has evolved over the course of its

first two years in ways that appearto have strengthened it as a'delin-quency prevention project. Interim

results based on student self-reportsuggest that the project has beenremarkably effective in altering, a

number of student characteristicsthat delinquency prevention theory.implies must be altereeto preventdelinquent behavior, and studentself-reports of delinquent behaviorare significantly-lower than the

self-reports oa control group.The self=report data must be inter,-preted with caution, however, .

because of some- evidence of differ -.'ential validity for treatment andcontrol group members.. 'Probamswith the retrieval of some archi4a1data on official delinquency andother outcomes limit the assessmentreported here.- New data have veryrecently become available-tostrengthen the analyses performed to-date, .and the results'curientlyavailable should be regarded as ten-

tative.

Pasadena

'Project STATUS (Student Training

Alternatives Through Urban°Strate-gies) involves 5 interventions: (a)

the Options class;. (b) the YouthCommittee and Leadership TrainingClass,_(c) project: training; (d)

Targeted Interventions

parent,, involvement, and (c) theAction/Advisory Committee, aredesigned to provide students with-a--meaningful educational program. Theevaluation designs for the Optionsclass and a. CommitteeCommttee in one'

sof the two chals involved in thisproject were strong enough to meritdescription of the .interim resultshere. Results show Eliot's Optionsclass' to significantly decreaseAlienation, increase students'self-ratings of reading ability,decrease withdrawals from school,increase Interpersonal Competency,and increase student reports of theFairness and Clarity of schoolrules. It appearsto.havethe unan-,ticipated consequence of increasingabsenteeism for participants in theclass. No'statistical evidence ofeffectiveness was found for.Muir''sOptions class nor for the Youth Com-mittees. Numerous problems inimplementing the program may par-

.tially account 'for th/ null results.

PCD

Peer Culture Development (PCD),operating in the.Chicago PublicSchools, runs a peer counselingintervention as regularly meetingclasses. The interim evaluationresults suggest that the project hasproduced positive effects on beliefin conventional rules, delinquentbehavior, and school grates for somesubgroups. At the same time, nodependable eildence was adduced that

. other important project objectiveswere achieved, and the effectsobserved are not obstrved consis-tently in each-semester and acrossthe several categories of youthsinvolved. Because some interimresults suggest that the interven-

.,tion Can be strengthened, the ?rej-ect has actively engaged in clarify-ing its implementation standards anddeveloping procedures for monitoringthese standards. The PCD .project.

illustrates a serious approach toproject development over time.

- 106- 113

Kalamazoo

The Milwood Alternative Projectis primarily a school improvement

t project operating in Milwood JuniorHigh Schodl in Kalamazoo, Michigan.Accordingly the results described inChapter 5 for'the\school-wide out-comesare most relevant to the eval-uation of this project. The projectdid; however, haVe several compo-nents targeted at subgroups of theschool's population. These includean attendance monitoring.procedure,a Skills Lab class for low-achievingstudents, a schoor-within-a-schoolfor eighth grade students (the Mil-wood Alternative Program), a studentcouncil, and a project advisory.One of these components,.the studentcouncil, was amenable to outcomeevaluation 'through the fortuitoususe of a lottery in homerooms toassign students.to participate.Results suggest that involvement inthe student council may haveincreased Negative Peer Influence,lowered Self-Concept, loweredAttachment to Parents, reduced stu-dents' perceptions of ParentalEmphasis on Education, and lowerededucational expectations. A numberof alternative hypotheses to explainthese results also exist.

A second component, attendancemonitoring, resulted in,a signifi-cant reduction in the proportion ofstudents who could be considered"chronic non-attenders" when com-pared to a similar group of studentsin the Kalamazoo comparison school.

Puerto Rico

The Puerto Rico project, run by acommunity organization, functions asan extension of the school day.Project staff work cooperativelywith school administrators andfaculty to assess student academicneeds and schedule project activi-ties. The project includes an aca-demic component and activities aimed

Targeted Interventions

at student social development andcommunity participation. The evalu-

ation uses a non-equivalent controlgroup design with good pre-interven-tion data available for use as sta-tistical controls. Interim evalua-tion results suggest that the

.project has modest positive effectson school grades, students' educa-tional expectations, standardizedachievement teat scores, students'Involvement in extracurriculaiactivities, and Student Influence.Some negative results are suggestiveof negative outcomes for studentsreferred to the project for academicdifficulties, but these results maymost plausibly be regarded as due toweaknesses in t4u evalpation designfor this particular project compo-nent.

Charleston

Project PATHE"operates in sevenCharleston County Public Schools.It aims to,alter broad aspects ofschool climate and to assist in theadaptation of a group of approxi-mately 100 high risk youths in eachschool. Implementation data showsubstantial variability in thetrength of implementation of thedirect service components across theseven schools. Schools that wereimplementing the school-level inter-ventions in strongest form tended tobe weakei'in implementing the inter-ventions targeted at high riskyouth . The evaluation involves alarge sample true experiment.Interi results suggest that thePATHE program increased academicperformance for targeted individualsat both the middle and high schoollevels, and increased school attendance, promoted attachment to schooland enhanced self-concepts for tar-geted middle school students. In

one or more of the' middle schoolssignificant positive, effects wereaiao found for the following out-comes:' serious delinquency, rebel-

-107- 1 20

Targeted Interventions

lions autonomy, individual studentperceptions of the fairness andclarity of the school rules, scbolpunishments, and employment. That

is, treatment group studentsreported leas serious delinquentbehavior and rebellions autonomy

, than did control group students,reported the rules to be fairer andclearer, received less punishment,and were employed more often. In

one of the middle schools treatmentstudents scored significantly loweron the measure of Practical Know-

.ledge than did control group stu-dents, .and in a high school severalnegative effects were observed forthe target group students: School

Attachment, educational expects-.tions, and employment were lower fortreatment.than for control groupstudents.

Virgin Islands

ThE Virgin Islands AlternativeEducation ,Project is seeking to

implemet two interventions dissemi-nated by thQ National Diffusion Net-work (NDF)--Focus and PAM. Interim

evaluation results suggest thatFocus is being implemented largelyas intended, but 'with some modifica-

tions to the Focus model. PAM wasnot implemented,as Intendedin thesecond year of project operation.Despite some weaknesses in the eval-uation design and measures, resultssuggest that the modifiedFocus'intervention resulted in studentsreceiving higher grades that theyotherwise may. have received. No

other consequences of the Focusinitervention were detected by the

evaluation. Program development is.continuing, as are efforts tostrengthen the evaluation.

I .411i:

The Academy for Community Educe-'ion is a small alternative schoolthat uses a token economy system,academic education, professional/vo-cational curriculum, and otherinterventions in providing servicesto youths at high risk of delinquentbehavior drawn from the Dade County(Miami? Public Schools. The limited

data currently available suggestthat participation in the Academyrgsults in significantly less absen-teeism, fewer suspensions, less tar-diness to school, and more academiccredit earned than participation inthe public schools. Academy parti-cipants, however, withdrew fromschool involuntarily mare often thansimilar students remaining in thepublic schools. Despite some impor-..'tant limitations of the data, theseinterim results are encouraging.

P lymout'h

The Plymouth Alternative Educa-tion Project operated Student Ser-vice Centers, Student ActivitiesCenters, and an out-of-school Learn-img Options program primarily fortwo high school's and two middle

schools. These interventions pro-vide educational services, counseling, and recreation for studentswith disciplinary and attendancedifficulties in this predominantlyworking and middle-class white com-munity. Interim evaluation resultsraise some questions about the Stu-dent Service Center implementation,and suggest that some unexpectednegative effects of this counselingintervention may be occurring. Spe-

cificalliy, treatment students ascompared to controls reported fessInterpersonal Competency, lowerSelf-esteem, more. Rebellious Auto-nomy, less Involvement in convet-tional activities, lower Attachment .#,1

'to School, and less Parental Emplus=

sis on Education,..

-108-

121

Interim evaluation resultl implythat the high school Student Activi-ties Center is being implemelftedwith care, and has some promisingpositive effects on particivants.Specifically, the SAC'btudentsscored significantly higher thantheir controls on tests of writingskills and Practical Knowledge, andreport,higher school grades and lessalienation.

The Learning.Oftions program didnot participate in an outcome evalu-ation.

Milwaukee

The Jewish Vocational ServicesAlternative Education Project devel-oped and implemented three interven-tions. The Milwaukee Youth Employ-ment Center (MYEC) counseled dropouty,uth and attempted to place them inemployment. The Return Center,operated in cooperation with theMilwaukee Public schools, assessedand referred to alternative educa-tional programming youth who werecontemplating dropping out, or whohad already dropped out and wishedto re-enroll in formal education.The Job Score class was a regularhigh school course developed by MYECstaff to teach employment skills toyouth at' risk for dropout.

An experimental evaluation of theMYEC program indicates that it wasnot successful in increasing theemployment opportunities of its

clients. No evidence exists thatthe Psychological Health, Interper-sonal Competency, or RebelliousAutonomy of the youths involved wereaffected by. the program Subsidiaryanalyses suggest, that in general,clients did not receive many ser-vices, although the extent to whichclients participated in the remedialeducation provided through the proj-ect was associated\with positiveoutcomes. Evaluations of two.proj-

Targeted Interventionb

componenttithe Return Centerand Job Score classen--were not com-pleted due to the early terminationof the project.

A Summary

The foregoing list of highlightsmay be difficult.for the reader tointegrate. Accordingly, in Tables 2and 3 we summarize these highlightsin tabular form to provide a sort of"box score" for the program as awhole. This summary, for .nterven-tions targeted .at specified 6 oupsof individuals, parallels t e! sum-

maries provided in Chapter (Tables

1 through 3) of overall r, ts for

interventions aimed at en )opu -

lations.

These tables summarize the evi-dence about the effects of targetedinterventions on the characteristicsof individuals--their psychosocial .

characteristics and their.behavior.Only the goals and objectives men-tioned in the OJJDP programannouncement and selected delin-quency risk.factors are included inthis tally. Other project-specificgoals and objectives are sometimesomitted. In a few cases, however,.where a specific hypothesis thatindividuals targeted by the projectwould have different perceptions oftheir environments, some of theenvironmental objectives have beenaddressed by measurement of the grr-ceptions of treatment and comparisongroups individual perceptions; someof these comparisons are included inthese tables.

A box score for the AlternativeEducation Program's goals is pre-sented in Table 2.- The preponder-ance of the significant effects ofthese targeted interventiions ispositive. Of 23 significant differ-ences, 20 were in the positive

. direction. Two of `the three nega--tive outcomes were for subcomponents

-109- 122

Table 2

-e

Number of Alternative Education Program Projects Showing

c' Experimental or QuasiExperimental Evidence, of Effects ofInterventions Targeted at Specified Gr&ups: Program Goals

Measure

Decrease Delinquent Behaviorand Around School

Number of Alternative-Education Sites,

Positive Negative

.Seribus Delinquency 3 0

Drug Involvement 1 -0

.Deerease.Suspensions 2

Increase Attendance 3 . 0

Increase Academic SuccessrGrades 6

Standardized test scores 2

ImproVe Transition to Work andPostsecondary EducationliEducAtional Expectations 1 "2

Working faepay 1

Having a job

,Note. Only the ten project's having sufficiently rigorous outcomeeyaluatians,Of targeted interventions to warrant interpretationare included in this table. Measures or goals =he primarilyfrom the School Action Effectiveness Surveys, but are also taken

ftbm official school or police records when available.

123

Table 3.

Number of Alternative Education Program Projects ShowingExperimental or Quasi-Experimental Evidence of Effects of.'

Interventions Targeted at Specified Groups: Program Objectives

Number of AlternativeEducation Bites

Meastipe

Making'School Discipline Fair and ConsistentWhile Providing for Due Process

Students' Reports,of Clarity of Rules. 'Students' RepOrts of Fairness of Rules

Positive

2

NegatiAle

0

0

Increasing Youth, Parent and Community AgencyParticipati.bk in School Decision Making toReduce Student Alienation

Students' Reports'of Student Influence 1

Alienation 2

Preclude Labeling EffectsPositive Self-Concept 1 2

Provide a Learning Structure Tailored toRealistic Levels to Promote Educationaland Social Development

Rebellious Autonomy 2 `, .^1

Practical Knowledge 2 1

Interpersonal Competency 1 1

Parental Emphasis on Education 0 . 2

School Punishments 4 1 2

School Rewards 2 .0

Additonal delinquency riskrisk factorsAttachment to Parents 1 1

Attachment to School 2 1

Belief in Rules 2j 0

Negative Peer Influence 0-e

1

Note. Only the ten projects having sufficiently rigorous outcomeevaluations to warrant interpretation are included-in this table.Measures of objectives arle.taken from the School Action Effective7ness Surveys.

1 4

Targeted!Interventions.

of Projects with largely positiveeffects. Because so many statisti-cal tests were perforthed, some ofthese nominally significant differ-ences could occur by chance, but bychance half would be expected to bepositive and half negative. That

result was not observed:

A box score for significant dif-ferences for measures of AlternativeEducation Program objectives for thetargeted interventions is providedin Table 3. Of, 34 significant dif-ferences, 21 are positive and 13 are

negative. Once again, 5 of the 13negative effects are for subcompo-nents of projects with largely posi-tive findings. Of,the negativefindingS, six are for a single 'proj-ect, reflecting what appears to bean unexpected negative influence ofa counseling intervention run bythat project (Cook, .1983).

Progress

The results presented here and inthe previous chapter imply that pro--gress has been made not only inimplementing'the interventions con-ducted by the projects in the Alter-,native Education Program, 'but alsoin implementing the evaluation ofthose projects. The experimentaland quasi-experimental evaluation of

these projects-tp_date provide evi-dence that we interPret_as suppor-tive of this general appi:Olth-,toreducing youth crime. These areinterim results for projects thatcontinue to develop, however. Most

of these projects are being imple-mented in improved, form in the thirdyear, and evaluation designs anddata collection arrangements areimproved. As these projectsdev,plop, use information about theireffects and about the strength andintegrity of their interventions,benefit from technical assistance;and gain in experience and expertisethey may .be expected to producestronger and more consistenteffects.'

The evaluation has turned agreater' portion of its effort todocumenting the implementation ofinterventions and to working with

project to clarify the standardsfor the implementation of theirprojects. This effort, too, maycontribute not only to strongerinterventions for some projects inthe third year, but also to theefforts of others who follow theseprojects in efforts to implementeffective delinquency preventionprograms- by using the program modelsbeing developed.

-11125

Educational Interventions and the Prevention of Delinquency': Some Closing

Observations

Some observations on the Alterna-tive Education Program at the end oftwo years of operation are apt. In

this final brief chapter we offersome of our less technical observa-tions on the operation of the Pro-gram, and some observations on theimplications of'what_we have learned'so far for the futtireof delinquencypreventiotf. A shortage of time, andthe pressure to get on with the task.of continuing the evaluation in the'third year limit our ability tocarefully document and justify theopinions expressed here. Inbtead,

we assert our opinions Ind willleave to a later date.a fuller expo-sition of these opinions and Specu-lations and of the, reasons webelieve as we do.

Schooling and Delinquency

:,The Alternative Education Pro-gram's inception was based on abackground of research and carefulthought about the causes of youthcrime. The background paper thatspelled out the rationale for anAlternative Education Program(OJJDP, 1980) was a careful doeApentthat built a good case for educe-tional,,and school structural,.approaches to reducing the. risk ofyouth crime. A President's., TaskForce:(President's Commission on LawEnforcement and Administration ofJustice, 1967) had pointed to causesof delinquency in the organizationof schooling in America. .A NatiOnal

Academy of Sciences panel (Martin,Sechrest, & Redner,, 1981) called forresearch and development of school-based'interventions.for both primaryprevention and remediation at aboutthe time the Program was initiated.We have elaborated elsewhere astrong case for interventions in

schools to prevent delinquency(Gottfredson & Daiger, 1979; Gott-fredson, 1981; Gottfredson, 1983b).The evidence of the present evalua-tion provides no reason to 'questionthe scientific, theoretical, andpractical premises'upon which theAlternative Education PrOgram wasbased. To the contrary,Ipatterns wesee in the data provide every reason'to'try harder to implement and eval-uate preventive interventions basedon this approach., Oncelagain, forexample, we find the same school-re-lated risk factors, associated withdelinquent behaviqr (see Chapter 4).Wire impressive,. the evalu

ation has

produced experimental evidence that..interventions in schools can altersome of these risk factors and-evenevidence suggesting that delinquentbehavior has been reduced.

Developing Effective Programs

The most pressinglproblem in thedelinquency prevention'field is theproblem of developing sound, theory-based interventions that can beimplemented in strong enough formthat they will make a substantialcontribution to)the reduction'ofdelinquent behavior The develop-,ment and implementation of strongprograms is not a need of alterna-tive educational approaches todelinquency prevention alone. Such

development and caeful implementa-tion is required in every approachto delinquency prevention. It is

required in the eritire criminal anjuvenile justice area. Programs torehabilitate offenders have beepalleged not to work (Martinson/1974). But the lictual evicl9nde sug-gests that few rehabilitativeefforti of suffitient strength andintegrity have b en implemented and

-113- 12,6;I

Preventing Delinquency

carefully evaluated, that the task .

of creating such interventions willbe difficult, and that we shoUld tryharder (Sechrest, White, & Brown,1979). The development, implementa-tion, and evaluation of such moreeffective programs will require theattention of talented people in aconcerted effort over a period of

years. A more careful, long-term,technologically and scientifically'based prorammatic effort.mnst bemade to realize the-patential toreduce youth crime that there is nowevery scientific reason to believe

exists. A two- or three-year pro-gram with limited scientific andtechnical assistance is not enough

time to demonstrate that'this poten-tial can be-realized.

Sound Implementation

Throughout our earlier report,and in the present one; we haveemphasized the importance ofstrength and fidelity in theimple-mentation of prevention projects.Indeed, we attempted to structurethe evaluation in 'ways that would

foster the developu4nt of stronginterventions with high probability

of iMplementation. We are gratified

by what we 'perceive to be greatstrides in strengthening the interventions implemented by many of

these project's. In our judgment,however, every one of these' projects

can improve greatly in the strengthand care of implementation of it's.,interventions... This is'not a,con -demnation, far from it, for we note;with favor the progress that hasbeen made.

The point is that a number ofavailable technologies that appearto fit with the goals, objectives,

and rationales behind these'projectsare used far, far less than theycould be. Classroom reward strut-.

tures that have been experimentallydemonstrated to alter known delin-quency risk factors (attachment to

-114-

school, performance in school) arebeing used in only one project, andthere notin thorough or strongform. Home based reinforcers arenot being `systematically applied,despite their demonstrated efficacyin altering another delinquency riskfactor (disciplinary' difficulties inschool). Technologies that involvebehavioral contracting and the carefel consequation of behavior areunderutilized in all but perhaps one

of these projects.

One reason, no doubt, that someof the existing technology goesunderutilized are diffictities increating'productive organizationalchange that will lead to their.adoption. Another reason, no doubt, is

that insuffiCient attention,has yetbeen directed to ensuring that thesetechnologies are available fo proj-ect, implementers and that they havethe skills to implement them. Athird reason, probably, is that we

have not taken seriously enough as anation the development of a cadre ofprofeasiOnals expert 'in the applica,-

'tion of such technologies, and thata human resources development effort

will be required to implement educa-tional and.delinqUency preventionprojects in strong form. And-afourth reason, we are certain, isthat the jolting Wray in which pro-

gtaMs are' begun and ended, to be

started and stopped again at somelater time; n some other place,with some other personnel, is notconducive to the development ofstrong programs.

We must pursue our goals syste-matically if we are to achieve them.

Evaluation and Zxpertise

We will make 'more progress

towards the development of sounddelinquency prevention programs wheneveryone involved adopts an experi-menting approach to the .enterprise.

127

Preventing Delinquency'6 04

Fcir too many yeard, and there isonly the slightest sign that thisattitude is eroding, an attitude ofj"anything goes" has pervaded the '

administration of most programs inthe crime prevention area. Propo-

nents of correctional reform, ofschool reform, of reform in thejuvenile and criminal justice systemmore broadly, have always felt free-to condemn the existing system andoffer their ,alternatives. Seldom,

indeed, do proponents of reform rre"sume that their reforms may alsointroduce uhdesirabie effects notanticipated in advance, that theirreforms may founder on unforeseenobstacles, that they will one day bethe target of future reformers whowill condemn them.

Virtually every innovation in thecriminal and juvenile justice system, and- in the educational system,should be approached as an.experi-ment. The reasoncsimply put,that the innovations may, not help,May not help as well as expected; ormay cause harm.

We have encountered resistance toevaluation of interventions in oneform or another from many sources inthe conduct of our work. &projectimplementer may be sure his or her .

intervention works and eschew theburdensome activity of studying the'intervention's effects. .A Rroject

director may wish to avoid rockingthe .boat in his or her system andavoid steps to make available die ,necessary information or arrange-ments to make for a maximally usefulevaluation. A project officer maysee a need 'to rapidly meet service.quotas and create an environment notconducive to an orderly evaluation.The pages torn from the calendareach day create a pressure to getonwith the work. now, rather than toplan and systematically carry out:the most careful possible implemen-tation'and research.

We interpret some of the evidenceand experience generated by thisevaluatiOn as.implying A need forgreater expertise and for a tho-roughgoing experimental approach tbdelinquency prevention.' We aredisheartened when we see a projectrapidly staffed with workers notfully qualified by experience ortraining to implemept their parts:ofa project. - We waste precious timeand resources when we must wranglewith recalcitrant .implementers overwhether or not a project component,should be evaluated.

=115-

The heed for expertise and anexperimenting approach to the devel-opment of delinquency preventionprojects is obvious, and should'beinsisted upon. For example, the*negative or null results for some ofr-the counseling, interventionsincluded in the Alternative Educa-tion Program strongly imply that nosuch program should be undertakenwithout the firm guarantee inadvance that the intervention willbe undertaken as an experiment and.with highly qualified staff.

The Work Ahead

We. have only begun to explore thewealth of information about theeffectiveness of the AlternativeEducation Program and its componentprojects. Nested withiln each ofthese'projects lie unmined treasuresof information.about:sthe relativeefficacy of each of the interven-tions implemented. .In the thirdyear of the PrograM projects are-.implementing interventions ofgreater strength than those of thesecond year,'-and the arrangements toevaluate them are sounder. We willcontinue our efforts to unearththese treasures as the evaluationcontinues. This report is aninterim evaluation. More powerfuland more thoroughgoing analysesremain. to be performed; there is-

128

Preventing Delinquency

much more to learn about this.Program. The final chapter will be

written oyly after much remainingwork is completed.

.129-116-

Written Products Resulting from the Evaluation

Here r provide a'partial listing of written products produced in the

course of this evaluation.' the most important previous written product is our

first interim report (G. D. Gottfredson (ed.), The School Action EffectivenessStudy: First`Interim Report. Baltimore: The Johns Hopkins University, CSOS,

1982). That report contained an overview of the Alternative Education Pro-gram, theSchOol.Action Effectiveness Study, and a description of each of the

17 projects in,

the Prbgram.,7,

.,

The following table contains a partial listing of other written productsproduced during the course,,,of the evaluation.

130-117-

1).

Products

Table 1

Select l Written- Products Resulting from the.Evaluation )

Project topic or Title Date

Compton Report oniattendance and,behavior outcomes forFall Semester', 1981

School climate assess-ment

February, 1982

August, 1982

Constitutional Rights Report on Fall 1981 pre- October, 1981

Foundation test survey

School climate assess- August, 1982ment

Report on achievement, ,

behavior,, and attendanceoutco4s for 1981-82school year

October, 1982

Report on pre-treatment March, 1983

data for 1982 -83 experi-mental groups

Peer Culture Development Report` on types of stu-, 'dents in peer groups

PCD effectiveness as:s-

essment for Fall Semes-ter, 1981

May, 1982

May, '1982

1

131-118-

c Table 1 (continued)

A .

Project TOpic or Title

Produqts

.Peet Culture Development School climate assess-(cont.) mknt

Presentation on PCp(

effectiveness at recep:tion for CBE administra-tors, Chicago juvenilejudges, and school prin-cipals

Date

sAugust, 1982

March, 1983

Peer Culture Development: May, 1983Second/Interim- Report

Kalamazoo School climate assess- August, 1982ment

Bronx

Puerto Rico

Repprt on expendi-ture of pioject manager

Report on Parent Questi -,onnaire results

Report on, behavior andattendance outcomes forFall Semester, 19 81

Report on Fall 1981 pre-test survey

Degember,

February,

February,

February,

1982

1982

1982

1982

School climate assess- . August, 1982

ment

School climate assess- August, 1982ment

Presentation 11 OtroCamino implementationand effectiveness pre-sented to Puerto RicoDepartment of Education

November, 1982

-119- 132

Productt;

Table 1 (continued)

Project Topic or Title Date

Charleston Report on reliability of June, 1981 .

CTBS subscales

7)Reports on attendance,achievement and behavioroutcomes for Fall Semes-ter, 1981

March, 1982May, 1982

Report on intensity of March, 1982 ,

PAWE interventions May, 1982

Bar charts showingschool level change andtarget-control studentcomparisons for allgoals and objectives

October, 1982

Report on results of August, 1982

Parent Questionnaire

School climate assess- August, 1982

ment

Report on behavior out- December, 1982

comes for first quarter,1982-83 school year

133.-120-

Project

Table 1 (continued)

ProduvIN

Topic or Title Date

Houston

Lac Courte Oreilles

Miami

School climate assess- August, 1982

ment

The George I. SanchezAlternative EducationProject: Second InterimReport (Unpublishimanuscript. San Raphael:Social Action ResearchCenter)

Report on effect ofPLATO on psycho-socialattitudes from Fall 1981survey and data

March, 1983

March, 1982

School climate assess- August, 1982

ment

School climate assess- August, 1982

ment

Report on achieveMent,_ February, 1983

attendance,' and behavioroutcomes for Fall Semes-ter, 1982

Academy for CommunityEducation: SecondInterim Report (Unpub-lished manuscript. SanRaphael: Social ActionResearch Center)

March, 19 83

a -121-

134

Product,1

Project

Plymouth

New Jersey

St. Paul

Milwaukee

t 1, ( con nue(' )

Topic oe Title

Report on attendance,achievmnent and behavioroutcomes for Fall Semes-ter, 1981

.o

10,

January, 1982

School climate assess- 'August, 1982

ment

Report on utility for April, 1982diagnosis of studentBehavioral EvaluationScales

.

Presentation of effec- October* 1982

tiveness of programotosuperintendents andprincipals of South Jer-sey Schools

Report stir attendance,

achieveant, and behav-ior outcomes for FallSemester, 1981

1May, 1982

School climate assess- August, 1982ment

School climate assess- 'Fall, 1982 ,

ment

Project ToRether: Sec- March, 1981ond Interim Report(Unpublished manuscript.San Raphael: Social

Action Research Center)

Report ofi effectivenessof Job Score Class

-122- 135

June, 1982

Project

Alternative EducationInitiative

Table 1 (continued)

Produati

Topic or Title Date

Standards for programdeylopment evaluationplans (Unpublished manu-script. Baltimore: Cen-ter for Social Organiza-tion of Schools)

1983

The Schooj Action Effec- 1982tiveness Study: Over-view (Paper presented atthe annual meeting ofthe American EducationalResearch Association,New York)

The School Action Effec- 1982.

tiveness Study: Prelim-inary Results (Paperpresented at the annualmeeting of the AmericanEducational ResearchAssociation, New York)

The School Action Effec- 1982tiveness Study: Devel-oping and evaluating

*prevention efforts.(aapex-presented- at-theannual meeting of theAmerican Society of Cri-.minology,Torontd)

The School Action Effec- 1982

tiveness Study: First di-

Interim Report (ReportNo. 325). Baltimore:Centei for Social Organ-ization of Schools. ,

,

- References

Alexander, J. F., & Parsons, D. V. Shorterm behavioral intervention withdelinquent families: Impact on procas and recidivism. Journal of Abnor-ma). Psychology, 1973, 81,219-225.

Associates for Youth Development. Community development workshop. Unpub-lished manuscript, Tucson, Arizona, 1980.

Astin, A.. W., & J. L. The environmental assessment technique: A wayto measure college environments. ,Journal of Educational Psychology, 1961,52, 308-316.

Atkeson, B. M., & Eorehand, R. Ho e-based reinforcement programs designed tomodify classroom behavior: A r view and methodological evaluation. Psy-chological Bulletin, 1979, 86, 1 98-1308.

Bachman, I% G., 01Malley, P. M., & Johnston, J. Adolescence to adulthood:Change and stability in the liwIThof..young men. Ann Arbor, Mich.: Insti-tute for ,Survey Research, 1978.

,P1Bandura, A. Social learning theorT.- Morristown, .N. J.: General Learning

Press, 197E

Barth,"R. Home -based reinforcement of school behavior: A review and analy-pis. Review of.Educatioiitl Research, 1979, 49, 436-458.

Berman, & McLaughlin, M.. W. Implementation of educational innovation,Educational 'Forum-, 1976, 40, 345-370.

/r)BielbY, W. T., Hauser, R. M., & Featherman, D. L. Response errors of nonblack

males-in models of the stratification process. American Journal of Sociol-ogy, 1977 82,.I.242-1288.1

O

Carnegie.Council-on Policy Stbdies in Higher Education. Giving youth a betterg_hanc'e:, Options' for education work and service., San Francisco: Jossey-'Bass,' 1979.

Chatters, W. "W., Jr., & Jones, E. On the risk of appraising non-events inprogtaff evaluation. Educational Researcher, 1973, 2, 7.

Cook,:M. S. Y.The>Kalamazoo Alternative Education Project:' Second interimreport,'Ill G. D.. G6tffredson, D. C. Gottfredson, & M. S. Cook (eds.): TheSchoOl..ACtion Effectiveness Study: Second interim report. Baltimore: The

6 -

.Johns Hopkins University, CSOS, 1983. (a)

Cook, M. S. The Lac Cod to Oreilles Alternative Education Project. In G. D.

Gottfredson, D. C. Gott Win, & M.-S:. Cook (eds.), The School ActionEffectiveness Study: Second interim report. Baltimore: The Johns HopkinsUniversity, CSOS, 1983. (b)

4 '.

(

--125-431

References

Cook, M. S. Plymouth Alternative Education Project: Second interim report.

In G. D. Gottfredson, D. C. Gottfredson, & M. S. Cook'.' (eds.), `The School

Action Effectiveness Study: Second interim report. Baltimore: The Johns

Hopkins University, CSOS, 1983. (c)

Cook, T. D.,& Campbell, D. T. Quasi-experimentation: Design and analysis

issues for field settings. Chicago: Rand McNally, 19)7.9.

Coopersmith, S. The antecedents of self-esteem. San Francisco:' Freeman,

1967.

Daniels, D.,. & Gottfredson, G: D. Compton Action Alternative School: Second

interim report. In G. D. Gottfredson, D. C.\Gottfredson, & M. S. Cook

(eds.), The School Action Effectiveness StudY: Second interim report:

Baltimore: The Johns Hopkins University, CSgS, 1983.

Dixon, M. C., & Wright, W. E. Juvenile delinquency prevention prog_ratns: An

evaluation of policy related research on the effectiveness of prevention, .

'.programs. Nashville, Tenn.: Office of Educational Services, Box 60, Pea-

body College for Teachers, 1974.

, Duncan, O. D. A socioeconomic index for all occupations. In A. J. Reiss; Jr.

(Ed.), Occupations and social status. New York: ;Free Press, 1961.

Elliott, D. S., & Ageton, S. Reconciling differences in estimates of delin-

quency. -American Sociological Review; 1980, 45, 95-110.

Empey, L. T. 'American delinquency. Homewood,'Ill.: Dorsey, 1978.

Empey, L. 'T. Field experimentation in criminal justice: Rationale and

.design. In M. W. Klein & K. S. Teilmann (eds.), Handbook of criminal jus-

tice evaluation. Peverly Hills, Calif.: Sage, 1980.

Empey', Constructing crime: Evolution and implications of sociological

theory. In S.' E. Martin; L. B. Sechrest, & R. Redner (eds.), New directions

in the rehabilitation of criminal offenders. Washington, D.C.: National

Academy Press, 1981.

Empey, L. -T., & Erickson, FL; L. Hidden delinquency and soCial status. Social

Forces, 1966, 44, 546-554.

Empey, L. T., & Erikson, M. L. The Provo experiment: tvaluating community

control of delinquency. Lexington: Lexington Books, 1972.

Empey, L. T.., & Lubeck, S. C. The Silverlake experiment: Testing delinquency

theory and community intervention. Chicago: Aldine-Altherton Press, 1971.

Farrar, E., & Holise, E. R. The evaluation of Push/Excell: A case study. In

A. S. Bryk (ed.), Stakeholder-based evaluation. San'Francisco: Jossey-

Bass, 1983.

Fox, R., and associates. School climateimprovement. Tulsa: CADRE, 1974.

-126- 138

References

Glaser, D. Concern with theory in correctional evaluation research. 'Crimeand Delinquency, 1977, 23, 173-179.

Gold, M. Scholastic experiences, self-esteem, and delinquent behavior: Atheory for alternative schools. Crime and Delinquency; 1978, 24290-308.

Gottfredson, D. C. Black-white differences in educational attainment. Ameri-can Sociological Review, 1981, 46, 542-557. (a)

Gottfredson, D. C. Modeling delinquency: More inconclusive evidence. Paperpresented at the annual meeting of the American Educational .Research Asso-ciation, 1981. (b)

Gottfredson, D. C. Project PATHE.: Second interim report. In G. D. Gottfred-son, D. C. Gottfredson, & S. Cook (eds.), The School Action. Effective-

. ness Study: Second interim report. Baltimore: The Johns 11opkins Univer-sity, CSOS, 1983.

Gottfredson, G. D. Schooling and delinquency. In S. g, Martin, L. B.

Sechrest, R. Redner (Eds.), New directions in the rehabilitation of crim-'inal offenders. Washington, D.C.: National Academy Press, 1981.

Gottfredson, G. D. A theory ridden approach to program evaluation: A methodfor stimulating researcher-implementer collaboration (Report No7i330).Baltimore: The Johns Hopkins University, CSOS, 1982. (a)

Gottfredson, G. D. Role models, bonding, and delinquency: An examination ofcompeting theoretical perspectives (Report No. 331). Baltimore: The JohnsHopkins University, CSOS, 1982. (b)

Gottfredson, G.D. The School-Action Effectiveness Study: First interimreport (Report No. 325). Baltimore: The Johns Hopkins University, CSOS,1982. (c)

Gottfredson, G. D. An examination of the consequences of personal victimiza-tion in school: A longitudinal study in an unselected sample. Unpublishedmanuscript, The Johns Hopkins University, CSOS, 1983. (a)

Gottfredson, G: D. Schooling and delinquency prevention: Some practicalideas for educators, parent's, program developers, and researchers. Journal'of Child Care, 1983, 1 (3),.1983. (b)

Gottfredson, G. D., & Daiger, D. C. Disruption in six-hundred schools:. Thesocial ecology of personal victimization in the nation's public schools(Report No. 281). Baltimore: Johns Hopkins University, CSOS, 1979.

Gottfredson, G. D., Joffe, R. D., & Gottfredslon, D. C.' Measuring victimiza-tion and the explanation of school disruption (Report No. 306). 'Baltimore:Johns Hopkins University, CSOS, 1981.

Gottfredson, G. D., Ogawa, D. 0., Rickert, D. E., Jr., & Gottfredson, D. C.Measures used in the School Action Effectiveness Study. In G. D. Gottfred-son (ed.), The School Action Effectiveness Study: First interim report(Report No. 325). Baltimore: The Johns Hopkins University, CSOS, 1982.

-127- 139'

References

Gottfredson, G. D., Rickert, D. E., Jr., 1G7'ttfredson,' D. C., & Advani, N.

Standards for program development evaluation. Baltimore: The Johns Hop-

kins University, CSOS, 1983. / .:

Gough, H. G. ( Manual for the California Psychological Inventory. Palo Alto,

Calif.: donsulting Psychologists/ Press, 1964. 1

Grant, J. Grant, J. D., Daniels,//D., Nem, V., & Yamasaki,C. The school

team approach phase .1 evaluation (Prepared underl'grants no. 77-NI-99-6b12

and 78- JN4AX- 0016 from the National Institute for Juvenile Justice and

Delinquenpy Prevention). ',San Rafael, Calif.: Social Action Research Cen-

ter,. 1979. -.

/

Greenberg', D. F. Delinquency and the age structure of society. Contemporary

Crisis, 1977, 1, 189 -223. I .

Hawkins, J. D., & W411, J. S. Alternative education: Exploring the delin-,

quency prevention potential. Seattle: University.'of Washington, Center

,for Law and Justice, 1979.

Hindelang, M. J., Hirschi, T., & Weis, J. G. Measuring delinquency. Beverly-

Hills: 'Sage, 1981.

Hirschi, T. Causes of delinquency. Berkeley: University of-/California

Press, 1969.

Holladd, J. L., & Baird, L. L. An interpersonal competency scale. Educa-

tional and Psychological Measurement- 1968, 28, 503-510.

Hoppock, R. Job satisfaction. New York: Harper, 1935.,

Howard, E. R. School discipline desk book. West Nyack, N.Y.: Parker, 1978.

Jencks, C. Who gets ahead? The determinants of economic success in America'.

New York: Basic Books, 1977.

Krisberg, B. Executive summary of the national evaluation of prevention (Pre-

liminary report). San Francisco: National Council on Crime and Delin-

quency, 1979.

Lemert, E. M. Human deviance, social problems, and social control.- Englewood

Cliffs, N.J.: Prentice -Hall, 1972.

Martin, S. E., Sechrest, L. B., & Redner, R. (eds.), New directions in the

rehabilitation of criminal offenders. Washington, D.C.: National Academy

Press, 1981.

Martinson, R. What works? Questions and answers', about prison reform. Public

Interest, 1974, 35, 22-54.

McPartland, J. M., & McDill, E. L. (Eds.) Violence in schools. Lexington,

Mass.: Lexington, 1977.

-1.4

-128-

References

Miller, W. B. Lower class culture as a generating milieu of .gang delinquency.Journal Of Social Issues, 1958, 14, 5-19.

Miner, J. B. The school administrator and organizational character. Eugene,Oregon: Center for Advanced Study of Educational Administration, 1967.

National Institute of Education. Violent schools--Safe schools: The safeschool study report to Congress.. Washington, D.C.: Author, 1978.

Nye, I. F. Family -relationships and delinquent behavior. New York: Wiley,1958. .

4.

Ogawa, D.. K. Delinquency prevention: The record of accomplishment, In G. D.

Gottfredson (ed.), The School Action Effectiveness Study: First interimreport (Report No. 325). Baltimore: The Johns Hopkins University, CSOS,1982.

Otto, L. B. Social integration and the status-attainment process, AmericanJournal of Sociology, 1976, 81,-1360-1383.

Office of Juvenile Justice and Delinquency Prevention. Program announcement:Prevention of delinquency through alternative education. Washington, 1Y.C.:Author, 1980.

Porter, & Lawler, E. E. Managerial attitudes and performance. Home--wood, Irwin-Dorsey, 1968.

President's Commission on Law Enfoicement and Administration of Justice. Taskforce report: Juvenile delinquency and youth crime. Washington, D.C.:U.S. Government Printing Office,1967.

President's Task Forte on Victims of Crime. Final report. Washington, D.C.:The Task Force, 1982.

° , Reid,,J. G., & Patterson, G. R. The modification of aggression and stealingbehavior of boys in the home setting. In A. Bandura & E. Ribes (eds.),Behavior modification: Experimental analysis of aggression and delin-quency. New. Jersey: Lawrence Erlbaum Associates, 1976.

Robinson, J. P., Athan'asiou, R., & Head, K. B. Measues of occupational atti-tudes and occupational characteristics. Ann Arbor: Institute for SocialResearch, 1969.

Robinson, J. P., & Shaver, P. R. Measures of social:psychologlical attitudes(Rev.. ed.). Ann Arbor: Insitute.for Social Research, 1973:

Rosenberg, M. Society and the adolescent self-image. Princeton: 'PrincetonUniversity Press, 1965.

Rotter, J.B.' Generalized expectancies for internal versus external controlof reinforcement. Psychological Monographs, 1966, 80 (1, Whole No. 609).

0

-129-

141

Referedces

St. John, J. PCD, Chicago: Second interim report. In . D. Gottfredson, D.

C. Gottfredson, & M..S. Cook (eds.), The School Act on Effectiveness Study:

Second interim report. Baltimore: The Johns Hopkins University, {SOS,

1983.

Sarason, S. B. The culture of the school and the problem of change. Boston:

Allyn and Bacon, 1971.

Siegal, S. Nonparametric statistics for the behavioral sciences. New York:

McGraw-Hill, 1956.

Sechrest, L. , West, S. G. , Phillips, M. A. Redner, R. , & Yeaton,. W. Intro-

duction. In L. Sechrest, s§. G. West, M. A. Phillips, R. Redner, & W.-Yea-ton (Eds.), Evaluation studies review annual (Vol. 4). Beverly Hills,

Calif.: Sage, 1979.

Sechrest, L., White, S. 0., & Brown, E. D. (eds.). The rehabilitation of

criminal offenders: Problems and prospects. Washington, D.C.: National

Academy of Sciences, 1979. ..

Sewell, W., Haller, A., & Portes, A. Tkie.educatOnal,and early occupational

'attainment process. American.Sociological Review, 1969, 34,'82 -92.

Slavin, R. E. Cooperative learning in teams: State ofthe art. Educational

Psychologist, 080, 15, 93-111.

Srole, L. Social integration and certaint.7corollaries. American Sociological

Review, 1956; 21, 709-716.

Stanley, J. C. Reliability. In R. L. Thorndike (Ed.), Educational measure-

ment. Washington, American-Council on Education, 1971.

Tharp, R. G., & Gallimore, R. Theecology of program research and develop-

ment: A model of evaluation succession. Unpublished manuscript, undated.

(Available from Roland. Tharp) Department of Psychology, University of

Hawaii, Honolulu 96822.)-

Thorndike, R. L. (Ed.). Educational measurement (2nd ed.). Washington, D.C.:

American douncil on Education, 1971.

Tittle, C. R., & Villemez, W. J. Social class and criminality. Social

Forces, 1977, 56, 474-502.

Waldo, G. P., & Dinitz., S. Personality characteristics of the criminal: An

analysis of research studies. Journal of Research in Crime and Delin-

quency, 1967, 4, 185-202.

-130-

142

Glossary

Glossary

Administrative Removal from school for a period of one or moreremoval days as a result of an administrative decision, without

being dropped from the school rolls. In most ,cases thisis equivalent to the colloquial use of the word'suspen- Asion.

Adoption The acceptance and use of a developed program, interven-i tion, or set of. interventions.

Educational practices, forms, or structures thatdiffer from traditional practices, forms or structures.

Alternativeeducation

Analysis ofvariance

A statistical technique used in the .analysis ofexperimental data. Its application in quasi- or non-ex--perimental data is often misleading.

Behavioral A form of intervention involving an.agreementcontracting among parties about the aim of the intervention and the

consequences of achieving or not achieving the aim.

Comparison group A group. of individuals, schools,- etc., with which a groupreceiving some intervention. are compared to help learnabout the effects of the intervention. Equivalent compar-ison groups (i.e., groups where no pre-existing -differ-

, ences are present) are preferred to. comparison groups knownto be non- equivalent, and equivalence is best achieved

. through randomization (see randomization).

COntrol group A group of individuals, schools, etc.,.with which a treat-.

ment group is compared, and which is known to be equiva-lent. Control groups should be created through randomiza-tion when possible.

Control theory

Core data

A theory of delinquency that assumes people will engage inunsocialized behavior unless restrained. It specifiessome ways to restrain youths from delinquent behavior.

Information about the results sought as outlined in theAlterna.tive EducatiOn Program Announcement/and the inter-ventions specified in the program announcement.

Critical benchmark A key decision, agreement, action, or arrangement neces-sary.to move forward-,with a strategy or plan. If a bench-mark is not met, progress in executing the strategy isblocked. When a benchmark is met, the forcefield changes.

' A benchmark statement tells what 'change in the forcefieldmust occur by when.

Delinquent Behavior which is illegal. Includes some behavior

-131-

1.43

Glossary

behavior (such as alcohol consumption) that is not illegal foradults.

Design decision The choice of interventions that occurs at ,or near the endof the initial planning phase of project development.Design decisions should be reconsidered perio4ically using .

information about the decision's outcomes.

Deterrence theory A theory that assumes crime can be reduced by the threat

' of punishment.-

'Differentialassociation

A theory of delinquency that assumes people.engagin delinquent behavior because they learn definitio sfavorable to law violation from those with whom they 'asso-.

ciate.10

Diffusion The spread of knowledge or information.

Dissemination

Evaluation

Experiment

Goal(s)

Forcefield

Pao

A set of activities consciously designed to encourage theutilization of knowledge or techniques in the development

or redesign of programs.

Activity to determine what happened, why, and with what

effect. Evaluation determines whether project activitiesproduced any outcomes of importance; whether unintended aswell as intended outcomes were produced. Evaluation sub-

sumes both formative and summative evaluation.

An experiment is activity undertaken deliberately to exa-mine the consequences of the activity. The term experi-

ment is often used to refer to true experiments, involvingrandomization (see randomization, true experiment).

What an organization is trying to achieve. A goal gener-

ally the obverse of a problem; it specifies how the goal(or the level of the problem) may be measured. Goals are

not broad or general aims. Such broad or general aims may

be called missions. .

The social-psychological field that immediately surrounds

a decision or action. It includes the forces that compel

or restrain against alternative actions as they are per-

ceived by an individual or corporate actor. Organizations

are held in place (do not change) because forces are in

equilibrium. To cre to change,the balance of perceivedor actual forces mus be changed.

-132-144

Glossary

Formative /Vtivity undertaken during the course of a projectevaluation to foster project development by determining what is being

implemented, by whom, with what effect, and'how effective-ness may be enhanced.

Implementation The execution of an intervention. Interventions vary inthe extent to which they are implemented as anticipated orplanned.

Implementation A blueprint for the implementation or replicationmanual of an intervention or set of interventions (i.e., program

model). SuCh a manual includes a statement of (a) what,the intervention is intended to achieve, (b).the theoryunderlying the intervention, (c) the resources requiredfor imiire-tentatiod, (d) the training and personnelrequired, (e) detailed specification of the intervention,(f) implementation standards, (g) all forms and recordkeeping procedures required to operate and evaluate theintervention.

Implementation . A clear statement of indicators of faithful implementa-standards tion (fidelity and completeness) of an intervention. Spe-

cifications or blueprints for an intervention define the/implementation standards for the intervention.

Incapacitation

Institutional:.

ization

A.theory that assumes crime can be reduced by locking uppeople whohave engaged in delinquent behavior.

Institutionalization occurs when anactivity becomes routinized and part of the status quo inan organization. When an activity is institutionalized',more effort is required to.terminate it or substantiallymodify it than 4 required to continue,it.

Intervention Activity undertaken to achieve an objective. Interventionis often synonymous with .the word "treatment."

Involuntary Removal from the school rollswithdraWal as a result of an administrative decision. In most cases

this is equivalent to the colloquial use of the wordexpulsion.

Labelling theory A theory oi'delinquency that assumes that treating anindividual asothough he or she were a delinquent resultsin the development of a delinquent self-concept and subse-quent delinquent" behavior.-

Managementinformationsystem

A tool used in formative and summatiVe evaluationto provide information about plans,'strategies,resources, obstacles, adoption, implementation, and oucomes.

145

Glossary

Management plan

Mean

Needs assessment

plan for implementing an intervention or set of inter-

ventions. A management plan is composed of strategies foradoption of innovations (see strategy) and of standards

tfor e implementation (see implementation standards).

An arithmetic average.

Activitytives.

tended to specify or clarify goals or'objec-

Post- randomization Activity to termine whether, randomization achieved

check the equivalenc of a treatment and control group by com-paring characte istics of the two groups that existed

prior to the ini iation -of the intervention, or which the

intervention coula not plausibly influence.

Objective(s)

Obstacle(s)

Organizationaldiagnosis

An outcome that a p oject's theory of action implies must

occur to achieve a goal. Objectives cintermediary out-

comes) are stated in measurable terms. Ideally, a state-

ment of an objective will specify when an objective will

be achieved and how much improvement should occur as well

as specifying how it is to be measured.

Forces which hold the project back, impede the progress of

a plan, or move the organization or individual in a direc-

tion opposite the intended direction. Obstacles may be

'perceived when none exist, or, obstacles may exist where

none are perceived.

Activity designed to assess 'the current

status of an organization and the relations among its ele-

ments,. Organizational diagnosis may include any of the

following activities: (a) climate assessment,

(b) assessment of goal Confluence, (c) assessment of

authority and decision structures, (d) assessment of com-

munication and interpersonal relations, and (e) assessment

of the match between goals and activities. Diagnosis

attempts.tdAinterpret the interaction among the above ele-

ments at a point in time.

Powerful An evaluation with sufficiently sensitive measures,

evaluation adequfte sample size, and with a design making the detec-

tion of intervention effects likely. Evaluations differ

in power, and an evaluation lacking in power has a low

probability of demonstrating anything conclusively.

-1347 14 6

Glossary

Program A theory- ridden method of action research involving

Development goal specification, theory elaboration, objectiveEvaluation development, intervention,definition, forcefield analysis,

the development of management plans, and evaluationresearch. PDE is intended to result in an upward spiralof activity leading to greater organizational effective-ness in 'accomplishing its goals.

Resource Any tool or force that furthers the adoption of an innova-tion, implementation of an intervention, or the achieve-ment of a goal or objective. A resource may be a person,institution, physical or psychological force, information,'money, or expertise. Both perceived and unperceivedresources may exist.

Rigorous An evaluation in which one may have confidence in theevaluation inferences drawn about the consequences of a ,demonstrably

iipplementedand well-described set of ins-erventlons.Rigorous evaluation is a major goal of qrhool Action

Effectiveness Study.

Risk factor A characteristic known to be associated with an outcome(e.g., delinquent behavior). Poor school performance isone of the risk factors for delinquent behavior.

'Significance A technical term meaning thatthe outcome was unlikely tohave arisen by, chance.

"Site-specific Information about goals, objectives, and interventionsdata ,

in each prevention project, whether or not these resultsor interventions are suggested by the OJJDPtoprogramannouncement or delinquency theory. It includes informa-tion about needs, goals, forcefields, strategies, .inter-ventions, and outcomes'.

Social learning A theory that assumes people behave as they do becausethey have learned about the consequences of behaviorthrough their awn. experiences and observations of others.

(7.Standard deviation. A measure of the extent, to which individuals, schools, or

other units are dispersed around the mean. A measure ofdispersion usefuin statistical analyses.

Strategies Plans. Strategies are developed from a forcefield analy-sis: 'Ail executable strategy will appear workable to thosewho must execute it, and will make use of an organiza-.tion's resources to'avercome the obstacles to adoption andimplementation. Strategidsare composed of two kinds ofelements: critical benchmarks and tasks.'

Randomization .A procedure employed to ensure that treatment and controlgroups are equivalent except insofaras differences arise.by'chance. Randomization serves to rule out rival

147 ,-135-

Glossary

Regression

Task(s)

Theory

True experiment

t-test

hypotheses about the sources of differences observedbetween treatment and control groups, and so leads to morerigorous evaluations. Because the technical meaning ofrandomization is not widely understood, randomization ishest accomplished by experienced research personnel.

(a) A statistical technique useful in the analysis ofexperimental And quasi-experimental data. (b) Getting

worse. (c) Movement froni an extreme position in a distri-

bution to a more central position.

The part of a strategy that specifies who will do what bywhen.

A statement of why a problem exists or of how an organiza-

- tion may achieve a goal. A projects theory of actionserves as a template for choosing and assessing interven-

tions.

An experiment involving the random assignment of units(people, schools, classrooms, etc.) to two or more treat-

ments (one of which is often a non-intervention treatment,

or control condition).

A test for the significance of differences in means (see

means, significance).

Victimization .Suffering personal harm, threat, or loss as a result of a

crime.

1

143-136-

Subject. Index

Aability grouping, 38absenteeism, 1, 27

acknowledgments, iiadaptation, 44administration, 39, 82

survey, 81administrative removal, 131administrators, 39adoption, 131alcohol, 60Alienation, 19, 23, 81, 100

alpha, 17.

alternative education, 1, 131AlternativeEdncation Program,

43, 60'

, analysis of variance, 131assault, 60Attachment to Parents, 22Attachment to School, 23, 91attendance, 100

B

baseline, J40behavioral contracting, 114, 13Belief in Rules, 23, 91blood pressure, 43boX score, 67-68, 109Bronx, 68

C Lcensoring

items, 67chance, 72, 99, 110Charleston, 67, 72-74, 81- 82,.91,

99, 101-102, 104, 107Chicago Board

of Education, 67-68, 73, 81,91

Clarity, 36classification, 46Classroom Dirsuption, 33classroom management, 32climate

compositional, 34, 67-68interpretation, 40psychosocial, 34, 67-68student reports, 34teacher reports, 37

coefficient alpha, 1'7

commitment, 39, 82, 101

ti

D

E

Community Crime, 34comparison group, 131competency, 25Compton, 49, 67, 72-73, 104,,106conformity

stakes in, 44control group, 131\core dAta, 131counseling, 108crime

community, 34critical benchtharks, 11, 131

delinquencyand school, 113correlates, 62.official, 106serious, 28, 108

delinquent behavior, 28, 60, 72,131

delinquent friends, 22.Department of Labor, 4design decision, 132development, 11, 112'

need for, 114diffusion, 132discipline,74, 114Disrespect for Students, 36dissemination, 132drop out, 1, 26 -

drugs, 28, 60, 72

educational expectation*, 29, 73educational structure, 81, 100effects

treatment, 105 '

unexpected negative, 108, 115effort, 26employment, 109enviornmental change, 44environment, 44epidemiology, 44, 102ethnicity, 48evaluatability, 9, 103-104evaluation, 132

design, 10, 103-104formative, 133need for, 114-115negative, 3

- 137-149-

w.

Subject Index

outcome, 9powerful, 134rigorous, 135summintive, 10value of , 3

expectation's, 81

experiment, .104, 107,' 132

true, 136expertise, 115

F.Fairness, 36-family, 19fidelity, 3-4, 114Focus, 108forcefield, 132forcefield analysis, 11

gang culture, 25gangs, 34,.60'Gangs in School, 34gender.,48.-49, 60

glossary, 131goal, 132goals, 11

evaluation, 2-3OJJDP, 1, 65-66, 68

grade structure, ..101.___

grades, 26, 30,'73,

Grades as Sanction, 39Grouping, 37

t

HHarlem, 67, 81, 91highlights, 100, 105history, 3home-bgted reinforcers, 33, 114homogeneity, 17human-resources, 114

I

L

implementatioh, 2-3, 133fidelity, 3manual-,'133standards', 11, 133

strength, 3individual, 44individual differences, 34Individualized Instruction, 35-,

37, 82,.99,'101inferenCes, 9'inhalants,'28institutionalization,.133

Integration vs. Segregation, 38integrity, 4, 103Interaction With Students, 32Internal Control, 19, 26internal validity, *91Interpersonal Competency, 24,' 81intervention, 11, 47; 133,

choice of, 132fidelity, 47integrity, 3,.47strength, 3, 47

interventions, 103targeted, 109

Invalidity, 29involuntary yithdrawal, 133 ,

Involvement, 24Parents and Community, 37

J

.Job Satisfaction, 32Johns Hopkins University, 5, 16

KKalamazoo, 72, 81-82, 101-102,

104, 107LLac Courtes Oreilles, 68, 98limitations, 91-_Low Expectations; 33lower-clabs culture, 25

M .

management plan, 133marijuana,.28' .

maturation artifaCts, 10mean, 16, 68, 134measurement, 16

criminal behavior, 27standard error, 68

measures;interpretation, 40

'Miami, 67, 104, 108'MilWaukee, 67, 104, 109MIS, 133morale, 39, 101

teacher, 71, 82movement, 104

N

National Diffusion. Network, 108Negative Peer Influence,,22New Jersey, 67-68.NIE, 15, 36

150

NIMP, 4Non-Authoritarian Attitudes, 34norm group, 16norms, 41numbers served, 47

0

objective, 134objectives

OJJDP, 66,, 69obstacle, 134OJJDP, 4, 9, 100opinion; 113, 115

.

.orgsnizational 'diagnosis, 134organizational healb,67, 71, 82

P

Parental Education, 19Parental Emphasis pn .Edudation,

22

parents, 74, 82participation, 74, 101Pasadena, 67, 73, sr, 91, 99,

104, 106PATL, 108PDE, 10 12Peer Culture Development,,25,

. 67-68, 74, 81, 982 104, 106peer. influence, 22peers, 63, 91percentile, 16planning, 3 -

Planning and Action, 36, 39plans, 11plausibility, 3Plymouth, 67-68, 72-73, 81-82,

. 91-92, 98, 104, 108population, 102Positive Self-Concept, 24, 100post-randomization check, 134

. power, 104Practical Knowledge, 25,pre-intervention data, 104President's Commission, 113prevention, 1

delinquency, 106;113primary, 43, 45, 101secondary, 43tertiary, 43

principal, 39Pro-Integration

Attitude, 30procedures, 40Professional Development, 33

15119-

Q

Subject Index

profiles, 41program, 104

integrity, 3models, 112

Pro,,ram Development Evaluation,10-12, 135

projecteffectiveness, 102

psychologidal health, 24psychosocial development, 81, 100Puerto Rico, 73, 82, 98, 101,

104, 107punishment, 28, 33Push/Excell, 103

quadrants; 44

R.race relations, 39randomization, 18, 104, 135Rebellious Autonomy 25regr.ession,. 136

rehabil'itation, 43, 113reliability; 16 -18

differential, 27,retest, 17, 29, 34, 40

remediation, 44, 102resource, 135'Resources for Instruction, 37reward structures, 114rewards, '29risk, 43, 45, 64, 67; 71, 91,

101, 103, 108;114, 135rules, 36, 108;

S

Safety, 34, 40, 100, 102samples, 67sampling, 68, 98sanctions, 28, 39scales

interpretation, 40school climate, 34, 65, 68, 101School.Effort, 26School Non-Attendance, 27School NonatteridanCe, 73School Punishments, 28School Race Relations, 39School Rewards, 29scoring

changed prodedures, 40secular trends, 99self-concept, 24, 81

Subject Index

self-esteem, 24-25self-report, 26, 106Self-Reported Delinquency, 27Self-Reported Drug Involvment, 28Self-Reported SeriouS Delinquency,

28

significdbce, 72, 135'smallpox, 44Smooth Administration, 39Social Action Research Center, 5socialization, 23socioeconomic .status, 62Special Emphasis Division, 4StPa61,'67-68, 73, 81, 91, 98Staff Morale, 39standard deviation, 16, 68, 135strategy, 135strength, 3-4, 107', 114

Student Influence, 36, 39subcultural values, 25surgery, 43

surveyadministration, 68,

surveys, 67-68suspensions, 73

T

98 -99

t-statistic, 68t-test, 136target groups, 102task, 136tasks, 11Teacher Victimization, 100teachers, 30Technical Assistance Division, 4

technologyunderutilization of, 114

theft, 60theory, 9, 11, 18, 13f,

control, 131deterrence, 132differential association, 132incapacitation, 133instrumentality, 26labelling, 124, 133

social control, 1, 23social learning, 1, 135

training, 115treatment, 45, 102-103truancy, 1, 73Type A SanctionS, 32Type 8 Sanctions, 33typology, 46

-140-

Vvalidity, 16, 18-19

differential, 27, 106internal, 91self-report, 27

vandalism, 60Victimization, 29, 33, 65, 69,

99-100, 136effects of, 62

Virgin Islands, 67, 72, 104, 108

Westinghouse, 4

,152

AAdvani, N., 11Ageton, S., 27-28Alexander, J. F., 9Almo, C., iiAndrews, A., iiAssociates. for

Youth Development, 44Astin, A. W., 34Atkeson, B. M., 33

B

Bachman, J. G., 1, 23

ader, G. , ii

Bailey, D., iiBaird, L. L. , 24

Bandura,.

Barth, R., 33Bass,.J., iiBatistil"A., ii'Bellafonataine,Berman, P., 39Betz, P., iiBielby, W. T., 27Birdseye, A., ii

Name Index

E

Elliot, D. S.,.27-28Elrod, P., iiEmpey, L. T. , 2,, 9, 11, 22, 27Erickson, M. L., 27Erikson, M. L., 9

F

Farrar,i,E., 103..

Featherman, D. L., 27Forehand, R. , 33

Friday, P., ii

G

Gallimore, R., 11Gavurin, S. , iii

Gilbert-Cougar, M., iiGlaser, D., 2Gold, M., 16Gottfredson, D. C., ii,

, 22, 24, 26-27, 30, 99,

Brown, E., 3Brown, D., 114

C

Campbell, D. T., 10Cano, P., iiCharters, W. W., Jr.,Coaxum, D., ii.Cohen, N., iiCook, M. S., ii, iii,Cook, T. D., 10Cooper, N., iiCoopersmith, S., 24

D

Daier, D. C., 2, 32,Daniels, D.; ii, 16,Deadrick; P., iiDilligard, B., iiDillon, D., iiiDinitz, S., 24,

Dixon, M. C., 3, 9Dorn, R., iii

2

98,

36,

27

102

39, 113

'-141-

113

102

Gottfredson, G. D., 1-2, 11, 16,22-25, 27, 29-30, 32, 34,36, 39, 62, 103, 113

Gough", H. G. , 23

Grant, J., 16, 23, 35, 39Grant, J. D. , 16

Greenberg, D. F., 16, 36Gutierrez, H., ii

H,

Harris, B., iiHartmann, M. E.1iiiHauser, R. M., 2Hawkins, D. 1

Hindelang, M. J., 22, 26-28Hirschi, T., 1, 16, 22-24, 27-28Holland, J. L., 24, 34Hoppock, R., 32House, E. R. , 103

Hybl, L., ii

3.

Jackson, J., 103Joffe, R. D. , iii

Johnston, L. , 1

Jones, J. E. , 2

153

Name Index

K Q

Kapinos, H., ii, iii Quay, H., 24Kelley, B. T., iiiKenny, P., ii

Kirchner-, R., iii Redner, R., 3, 113Kotuman, W. , ii Reid, J. G. , 9

Krisberg B., T, 9 Reiss, D., iiRickert, D. E. , ii, 11, 22

L Rickert D. E., iii

Lawler, E. E. , 26 Ringel, N., ii

Leighty, T., ii Robinson, J. P., 24

Lemert, E. M., 16 Rodriguez,-N., ii

Lemmert, E. M., 24 Romero, R., iiiLewis, M., ii Rosenberg, M.; 24

Lopez,.C.,Lubeck, Sr. C. ,

Luster, G. ii Sarason, S. B. , 2

Seals, T., iiSechrest, L. , 3

Mahoney, R. , ii Sechrest-, JZ. B. , 113-114

Martin, S. E.,113 Shaver, P. R., 24

Martinson, R., 113 Siegel, S., 68

Mazumder, A., iii Slavin, R. E., 39

McDill, E. L., 2, 32 Smith, R., ii

McKnight, M., ii Srole, L., 23

McLaughlin, M. W. , 39 St.John, J. , ii, iii, 98

McPartland, J. M. 2, 32 Stanley, J. C. , 17

Miller, W. B. , 25 Steptoe, H. , ii

Miner, J. B., 40 Steward, M., ii

Montes, C., iiT

N Tharp, R. G., 11

Nathan, J., ii Thorndike, R. L. 16

National Academy( of Sciences, 113 V

Neto, V. , 16 Vactor, V. ii

NIE, 2, 16, 39 -.

NIJJDP, 5Nuzzolo, A., iii Waldo, G. P., 24

Nye, I. F., 27 Weis, J., 22Weis, J. G., 27-28

O West, S. G., 3

O'Malley, P., 1 N Westinghouse, 5

Ogawa, -D. K., ii, iii, 2-3, 22 White, S., 3

OJJDP, 1, 4, 113 White, S. 0., 114Wizotsky, S., ii-

Wooten, C., ii

Parsons, B. V. 9 Wright, W. E. , 3

Patterson, G. R., 9

Polaris, 5Porter, L. -W., 26 Yamasaki, C., 16

President's Commission, 2 Yep.ton, W., 3

President's Task Forceon Victims of Crime, 61

-142-.

154