Professional Documents
Culture Documents
Board of Directors
LAWRENCE W. SHERMAN, President
KATHLEEN HALL JAMIESON, Chair
ELIJAH ANDERSON SARA MILLER MCCUNE
STEPHEN B. BURBANK MARY ANN MEYERS
HENRY LOUIS GATES, JR. KLAUS NAUD
FREDERICK HELDRING LOUIS H. POLLAK
RICHARD D. LAMBERT JAROSLAV PELIKAN
JANICE FANNING MADDEN
Origin and Purpose. The Academy was organized December 14, 1889, to promote the progress of po-
litical and social science, especially through publications and meetings. The Academy does not take
sides in controverted questions, but seeks to gather and present reliable information to assist the public
in forming an intelligent and accurate judgment.
Meetings. The Academy occasionally holds a meeting in the spring extending over two days.
Publications. THE ANNALS of the American Academy of Political and Social Science is the bimonthly
publication of The Academy. Each issue contains articles on some prominent social or political problem,
written at the invitation of the editors. Also, monographs are published from time to time, numbers of
which are distributed to pertinent professional organizations. These volumes constitute important ref-
erence works on the topics with which they deal, and they are extensively cited by authorities through-
out the United States and abroad. The papers presented at the meetings of The Academy are included
in THE ANNALS.
Membership. Each member of The Academy receives THE ANNALS and may attend the meetings of
The Academy. Membership is open only to individuals. Annual dues: $65.00 for the regular paperbound
edition (clothbound, $100.00). For members outside the U.S.A., add $24.00 for shipping of your subscrip-
tion. Members may also purchase single issues of THE ANNALS for $20.00 each (clothbound, $28.00).
Subscriptions. THE ANNALS of the American Academy of Political and Social Science (ISSN
0002-7162) is published six times annuallyin January, March, May, July, September, and November
by Sage Publications, 2455 Teller Road, Thousand Oaks, CA 91320. Telephone: (800) 818-SAGE (7243)
and (805) 499-9774; FAX/Order line: (805) 499-0871. Copyright 2002 by the American Academy of Po-
litical and Social Science. Institutions may subscribe to THE ANNALS at the annual rate: $420.00
(clothbound, $475.00). Add $24.00 per year for subscriptions outside the U.S.A. Institutional rates for
single issues: $81.00 each (clothbound, $91.00).
Periodicals postage paid at Thousand Oaks, California, and at additional mailing offices.
Single issues of THE ANNALS may be obtained by individuals who are not members of The Academy
for $32.00 each (clothbound, $42.00). Single issues of THE ANNALS have proven to be excellent supple-
mentary texts for classroom use. Direct inquiries regarding adoptions to THE ANNALS c/o Sage Publi-
cations (address below).
All correspondence concerning membership in The Academy, dues renewals, inquiries about member-
ship status, and/or purchase of single issues of THE ANNALS should be sent to THE ANNALS c/o Sage
Publications, 2455 Teller Road, Thousand Oaks, CA 91320. Telephone: (800) 818-SAGE (7243) and
(805) 499-9774; FAX/Order line: (805) 499-0871. Please note that orders under $30 must be prepaid.
Sage affiliates in London and India will assist institutional subscribers abroad with regard to orders,
claims, and inquiries for both subscriptions and single issues.
All rights reserved. No part of this volume may be reproduced or utilized in any form or by any
means, electronic or mechanical, including photocopying, recording, or by any information
storage and retrieval system, without permission in writing from the publisher. All inquiries
for reproduction or permission should be sent to Sage Publications, 2455 Teller Road, Thousand
Oaks, CA 91320.
Editorial Office: Fels Center of Government, University of Pennsylvania, 3814 Walnut
Street, Philadelphia, PA 19104-6197.
For information about membership* (individuals only) and subscriptions (institutions),
address:
SAGE PUBLICATIONS
2455 Teller Road
Thousand Oaks, CA 91320
Sage Production Staff: BARBARA CORRIGAN, SCOTT SPRINGER, and ROSE TYLAK
From India and South Asia, From Europe, the Middle East,
write to: and Africa, write to:
SAGE PUBLICATIONS INDIA Pvt. Ltd SAGE PUBLICATIONS LTD
P.O. Box 4215 6 Bonhill Street
New Delhi 110 048 London EC2A 4PU
INDIA UNITED KINGDOM
*Please note that members of The Academy receive THE ANNALS with their membership.
International Standard Serial Number ISSN 0002-7162
International Standard Book Number ISBN 0-7619-2744-1 (Vol. 582, 2002 paper)
International Standard Book Number ISBN 0-7619-2743-3 (Vol. 582, 2002 cloth)
Manufactured in the United States of America. First printing, July 2002.
The articles appearing in THE ANNALS are abstracted or indexed in Academic Abstracts,
Academic Search, America: History and Life, Asia Pacific Database, Book Review Index, CAB
Abstracts Database, Central Asia: Abstracts & Index, Communication Abstracts, Corporate
ResourceNET, Criminal Justice Abstracts, Current Citations Express, Current Contents:
Social & Behavioral Sciences, e-JEL, EconLit, Expanded Academic Index, Guide to Social
Science & Religion in Periodical Literature, Health Business FullTEXT, HealthSTAR
FullTEXT, Historical Abstracts, International Bibliography of the Social Sciences,
International Political Science Abstracts, ISI Basic Social Sciences Index, Journal of
Economic Literature on CD, LEXIS-NEXIS, MasterFILE FullTEXT, Middle East: Abstracts &
Index, North Africa: Abstracts & Index, PAIS International, Periodical Abstracts, Political
Science Abstracts, Sage Public Administration Abstracts, Social Science Source, Social
Sciences Citation Index, Social Sciences Index Full Text, Social Services Abstracts, Social
Work Abstracts, Sociological Abstracts, Southeast Asia: Abstracts & Index, Standard
Periodical Directory (SPD), TOPICsearch, Wilson OmniFile V, and Wilson Social Sciences
Index/Abstracts, and are available on microfilm from University Microfilms, Ann Arbor,
Michigan.
Information about membership rates, institutional subscriptions, and back issue prices may
be found on the facing page.
Advertising. Current rates and specifications may be obtained by writing to THE ANNALS
Advertising and Promotion Manager at the Thousand Oaks office (address above).
Claims. Claims for undelivered copies must be made no later than six months following
month of publication. The publisher will supply missing copies when losses have been sus-
tained in transit and when the reserve stock will permit.
Change of Address. Six weeks advance notice must be given when notifying of change of ad-
dress to ensure proper identification. Please specify name of journal. POSTMASTER: Send
address changes to: THE ANNALS of the American Academy of Political and Social Science,
c/o Sage Publications, 2455 Teller Road, Thousand Oaks, CA 91320.
CONTENTS
EUROPE
GLOVER, JONATHAN. Humanity:
A Moral History of the Twentieth
Century. Brian VanDeMark . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
GONEN, JAY Y. The Roots of Nazi Psychology:
Hitlers Utopian Barbarism. Vamik D. Volkan. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
QUINN, FREDERICK. The French
Overseas Empire. G. Wesley Johnson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
UNITED STATES
ALTSCHULER, GLENN and STUART BLUMIN.
Rude Republic: Americans and Their Politics in
the Nineteenth Century. Jean H. Baker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
DIGGENS, JOHN PATRICK. On Hallowed Ground:
Abraham Lincoln and the Foundations of
American History. Barry Schwartz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
FISHER, LOUIS. Congressional Abdication on
War & Spending. Charles Tiefer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
STARK, ANDREW. Conflict of Interest in American
Public Life. Glenn Harlan Reynolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
STEPHAN, ALEXANDER. Communazis:
FBI Surveillance of German Emigre
Writers. Bernard F. Dick . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
SUMMERS, MARK WAHLGREN. Rum,
Romanism, & Rebellion: The Making
of a President, 1884. Samuel T. McSeveney . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
WEBBER, MICHAEL J. New Deal Fat Cats:
Campaign Finances and the Democratic
Party in 1936. Marc Dollinger. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
SOCIOLOGY
BAYOR, RONALD H. Race and the
Shaping of Twentieth Century
Atlanta. Clarence N. Stone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
BROWDER, LAURA. Slippery Characters:
Ethnic Impersonators and American
Identities. S. Elizabeth Bird . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
COSTA, DORA L. The Evolution of Retirement:
An American Economic History,
1880-1990. Jon R. Moen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
ECONOMICS
FISKE, EWARD B. and HELEN F. LADD.
When Schools Compete: A Cautionary
Tale. John F. Witte . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
THE ANNALS OF THE AMERICAN ACADEMY
PREFACE
PREFACE
The main aims of this special issue of The Annals are to examine the
systematic review method and to report on some of its contributions to evi-
dence-based crime prevention. The main title of this issue, What Works in
Preventing Crime? signals our primary interest in identifying those inter-
ventions that are effective in preventing crime and offending and that ulti-
mately may lead to more effective crime prevention policy and practice. The
issues subtitle, Systematic Reviews of Experimental and Quasi-Experimen-
tal Research, signals our interest in using the most rigorous methods of
research synthesis and only the highest-quality research designs to evaluate
the effectiveness of criminological interventions.
Systematic reviews have received increased attention in recent years in
the social sciences generally and in criminology and criminal justice specifi-
cally. This is part of the broader interest in evidence-based policy and practice
in public services (Davies, Nutley, and Smith 2000) and evidence-based crime
prevention (Sherman et al. 1997, forthcoming).
At the forefront of the development of systematic reviews is the newly
formed Campbell Collaboration. Named after the influential experimental
psychologist Donald T. Campbell (1969), this was set up for the purpose of pre-
paring, maintaining, and disseminating evidence-based research on the
effects of interventions in the fields of education, social welfare, and crime
and justice. Its Crime and Justice Group aims to prepare and maintain sys-
tematic reviews of criminological interventions and to make them accessible
electronically to scholars, practitioners, policy makers, and the general pub-
lic. The present work, although not officially carried out under the auspices of
the Campbell Collaboration, represents an important contribution to its
Crime and Justice Group, and the four systematic reviews reported here are
undergoing review by the Campbell Collaboration; it is hoped that they will
be approved and disseminated as Campbell reviews in due course.
This issue of The Annals introduces the path-breaking work of the Camp-
bell Collaboration and its Crime and Justice Group, examines key method-
ological issues facing systematic reviews of criminological interventions,
reports on four original systematic reviews of the effects of different interven-
tions on crime and offending, and makes progress toward an evidence-based
approach to preventing crime and offending. Throughout this issue, crime
prevention is defined as any program or policy that causes a lower number of
crimes to occur in the future than would have occurred without that program
or policy.
This special issue originated with the 2001 Jerry Lee Crime Prevention
Symposium, a 2-day conference on systematic reviews of criminological inter-
ventions, held in early April at the University of Maryland, College Park, and
at the U.S. capitol building in Washington, D.C. Convened by the University of
Marylands Department of Criminology and Criminal Justice and sponsored
8
PREFACE 9
1
by the Jerry Lee Foundation, the conference brought together leading
researchers in the fields of crime prevention, experimental criminology, and
research synthesis. The conference also benefited from the participation of
directors of governmental and nongovernmental research units, criminal jus-
tice policy makers, and journalists specializing in crime and justice. Five
countries were represented: Australia, Canada, Israel, the United Kingdom,
and the United States. Papers on substantive issues relating to methods and
findings of systematic reviews of criminological interventions were presented
and discussed. Subsequently the papers were revised in light of editorial com-
ments, and they are now presented here.
The systematic review and the meta-analytic review are the most rigorous
methods for assessing the effectiveness of criminological interventions and
2
have the most to offer to evidence-based crime prevention.
PREFACE 11
Systematic reviews use rigorous methods for locating, appraising, and syn-
thesizing evidence from prior evaluation studies, and they are reported with
the same level of detail that characterizes high-quality reports of original
research. They have explicit objectives, explicit criteria for including or
excluding studies, extensive searches for eligible evaluation studies from
around the world, careful extraction and coding of key features of studies, and
a structured and detailed report of the methods and conclusions of the review.
All of this contributes greatly to the ease of their replication by other
researchers.
Four articles report on systematic reviews of the effects of different inter-
ventions in preventing crime or offending: parent training and support pro-
grams, hot spots policing, correctional boot camps, and cognitive-behavioral
programs. Each follows as closely as possible the methodology for conducting
systematic reviews that has been specified by the Campbell Collaboration.
In the first article of this section, Odette Bernazzani, Catherine Ct, and
Richard Tremblay report on a systematic review of parent training and sup-
port before age 3. Seven studies are reviewed, and effectiveness is assessed on
the outcome measures of disruptive behavior (for example, opposition to
adults, truancy, aggression) and delinquency. The authors find that the evi-
dence on effectiveness is mixed: three studies report some beneficial effects
(with one reporting some harmful effects), and the other four studies report
no evidence of effectiveness. The authors call for caution in interpreting the
results (for example, due to modest effect sizes of the beneficial studies) and
recommend further intervention studies in these areas.
In the second article, Anthony Braga reviews the effects of hot spots polic-
ing to reduce crime. This form of policing involves the targeting of police
enforcement measures in high-crime areas. Nine studies are reviewed, and
effectiveness is assessed on the outcome measures of crime and disorder.
Braga also examines the effects of the studies on the displacement of crime
(an unintended increase in crime, for example, in a different location) and the
diffusion of crime control benefits (an unintended decrease in crime in a dif-
ferent location). He finds evidence that targeted police actions can prevent
crime and disorder in hot spots, that displacement is rare, and that some pro-
grams produce unintended crime prevention benefits.
The third article, by Doris MacKenzie, David Wilson, and Suzanne Kider,
reports on a systematic reviewincorporating meta-analytic techniquesof
the effects of correctional boot camps (otherwise known as shock or intensive
incarceration) on offending. Forty-four studies are reviewed, and effective-
ness is assessed according to recidivism. The systematic review reveals var-
ied effects on recidivism: 9 studies report beneficial effects, 8 report harmful
effects, and 27 report no effects. Overall the meta-analysis finds evidence of
no effect. MacKenzie and her colleagues conclude that boot camps are an inef-
fective correctional treatment to reduce future offending.
In the final article in this section, Mark Lipsey, Gabrielle Chapman, and
Nana Landenberger report on a systematic reviewincorporating
12 THE ANNALS OF THE AMERICAN ACADEMY
FUTURE DIRECTIONS
In the final article in this collection, we bring together the main conclu-
sions from the individual articles and identify priorities for moving toward an
evidence-based approach to preventing crime.
The Campbell Collaboration Crime and Justice Group has begun the
important task of preparing systematic reviews of the effectiveness of a wide
range of criminological interventions and will soon be in a position to make
them accessible to researchers, policy makers, practitioners, and the general
public. We argue that alongside the Campbell effort, a program of research of
new crime prevention and intervention experiments and quasi-experiments
must be initiated. These need to be evaluated using the most rigorous
research designs, involve large samples, have long-term follow-up periods,
and include cost-benefit or cost-effectiveness analyses. These new studies
should be initiated in many Western countries.
It is a well-known fact that having convincing research evidence and hav-
ing it influence policy and practice are two very different matters. How to
overcome some of the misconceived political and policy barriers to get more of
what works in preventing crime into policy and practice is by no means an
easy task, but fortunately it has received some attention in various academic
disciplines, criminology included.
In the final analysis, a great deal of work needs to be doneby researchers,
policy makers, practitioners, and politicians (in no order of importance)to
achieve the well-intentioned yet lofty goal of using the highest-quality scien-
tific evidence in the development of public policy and practice for the preven-
tion of crime. We view this special issue of The Annals as an important step
toward this goal. Of course, should it spur academic interest, encourage more
systematic reviews, inspire further innovation among policy makers and
practitioners, and ignite the interest of politicians, these too will be important
effects.
DAVID P. FARRINGTON
BRANDON C. WELSH
Notes
1. We wish to personally thank Jerry Lee not only for his sponsorship of this important event
but also for his unwavering support of an evidence-based approach to preventing crime and
building safer communities. We are grateful to Lawrence Sherman for helping to organize this
event (with David Farrington) and to Charles Wellford and his staff (especially Cynthia
Mewborn) for their top-rate coordination and administration of the conference. Special thanks
PREFACE 13
also to Senator Barbara Mikulski and the Consortium of Social Science Associations for their as-
sistance.
2. We discuss here only the systematic review method because systematic reviews employ
meta-analysis when appropriate and possible.
References
Campbell, Donald T. 1969. Reforms as Experiments. American Psychologist 24:409-
29.
Davies, Huw T. O., Sandra M. Nutley, and Peter C. Smith. 1999. Editorial: What
Works? The Role of Evidence in Public Sector Policy and Practice. Public Money &
Management 19:3-5.
Davies, Huw T. O., Sandra M. Nutley, and Peter C. Smith, eds. 2000. What Works?
Evidence-Based Policy and Practice in Public Services. Bristol, UK: Policy Press.
Farrington, David P. 1983. Randomized Experiments on Crime and Justice. In Crime
and Justice: An Annual Review of Research. Vol. 4, ed. Michael Tonry and Norval
Morris. Chicago: University of Chicago Press.
Farrington, David P. and Anthony Petrosino. 2000. Systematic Reviews of Criminolog-
ical Interventions: The Campbell Collaboration Crime and Justice Group. Interna-
tional Annals of Criminology 38:49-66.
Farrington, David P., Anthony Petrosino, and Brandon C. Welsh. 2001. Systematic Re-
views and Cost-Benefit Analyses of Correctional Interventions. The Prison Journal
81:338-58.
Sherman, Lawrence W., David P. Farrington, Brandon C. Welsh, and Doris Layton
MacKenzie, eds. Forthcoming. Evidence-Based Crime Prevention. London:
Routledge.
Sherman, Lawrence W., Denise C. Gottfredson, Doris Layton MacKenzie, John E. Eck,
Peter Reuter, and Shawn D. Bushway. 1997. Preventing Crime: What Works, What
Doesnt, Whats Promising. Washington, DC: U.S. Department of Justice, National
Institute of Justice.
ANNALS, AAPSS, 578, November 2001
THE ANNALS OF
CHALLENGES OFTHE
EVIDENCE-BASED
AMERICAN ACADEMY
POLICY
NOTE: We appreciate the assistance of Maureen Matkovich and colleagues at the National
Criminal Justice Reference Service who supplied the data in Figure 1. The first authors work
was supported by grants from the Mellon Foundation to the Center for Evaluation and from the
Home Office to Cambridge University. Thanks to Brandon C. Welsh for his comments on earlier
drafts of the article.
14
CHALLENGES OF EVIDENCE-BASED POLICY 15
FIGURE 1
CUMULATIVE GROWTH OF EVALUATION STUDIES:
NATIONAL CRIMINAL JUSTICE REFERENCE SERVICE DATABASE
20000
18000
16000
14000
12000
10000
8000
6000
4000
2000
0
1970- 1975- 1978- 1980- 1985- 1990- 1995-
1974 1977 1979 1984 1989 1994 2000
SOURCE: The National Criminal Justice Reference Service (www.ncjrs.org).
One major problem noted with effects have often been interpreted as
regard to traditional reviews was statistically insignificant and there-
their lack of explicitness about the fore as treatment failures.
methods used, such as why certain
studies were included, the search Systematic reviews
methods used, and how the studies
were analyzed. This includes the cri- Evidence-based policy requires
teria used to judge whether an inter- overcoming these and other prob-
vention was effective or not. Consider lems with the evaluation studies and
the debate over the conclusions in the methods for reviewing them. There is
Lipton, Martinson, and Wilks (1975) consensus among those who advo-
summary of more than 200 correc- cate evidence-based policy that sys-
tional program evaluations, briskly tematic reviews are an important
reported first by Martinson (1974). tool in this process (Davies 1999;
Despite finding that nearly half of Nutley, Davies, and Tilley 2000).
the evaluations reported in Martin- In systematic reviews, researchers
sons article had at least one statisti- attempt to gather relevant evalua-
cally significant finding in favor of tive studies, critically appraise them,
treatment, his overall conclusions and come to judgments about what
were gloomy about the prospects of works using explicit, transparent,
correctional intervention. The crite- state-of-the-art methods. Systematic
rion for success was not readily reviews will include detail about
apparent, but it must have been each stage of the decision process,
strict (Palmer 1975). including the question that guided
Thes e earlier revie w s, l i k e the review, the criteria for studies to
Martinsons (1974), were also prob- be included, and the methods used to
lematic because they seemed to rely search for and screen evaluation
on statistical significance as the cri- reports. It will also detail how analy-
terion for judging whether an inter- ses were done and how conclusions
vention was successful. This later were reached.
proved to be problematic, as research The foremost advantage of sys-
showed that statistical significance tematic reviews is that when done
is the function not only of the size of well and with full integrity, they pro-
the treatment effect but of method- vide the most reliable and compre-
ological factors such as sample size hensive statement about what
(for example, Lipsey 1990). For exam- works. Such a final statement, after
ple, large and meaningful effects s i f t i n g t h rou g h t h e av ai l abl e
reported in studies with small sam- research, may be, We know little or
ples would be statistically insignifi- nothingproceed with caution.
cant; the investigator and traditional This can guide funding agencies and
reviewer would consider the finding researchers toward an agenda for a
evidence that treatment did not suc- new generation of evaluation studies.
ceed. Given that most social science This can also include feedback to
research uses small samples, moder- funding agencies where additional
ate and important intervention process, implementation, and theory-
CHALLENGES OF EVIDENCE-BASED POLICY 21
covered (1977-1991). Only the Some journals also have lengthy lag
Review of Educational Research pub- times between submission and publi-
lished an average of one review or cation, delaying the dissemination of
more per year. Unless researchers evidence even further.
were using other unknown mecha-
nisms such as oral briefings and
THE COCHRANE
internal memos to communicate to COLLABORATION
decision makers, it seems very
unlikely that this evidence got into Are there ways of overcoming
the hands of anyone other than challenges to using systematic
research specialists working in nar- reviews in evidence-based policy? A
row areas. precedent for doing so was estab-
Most systematic reviews also tend lished in the health care field. Archie
to be one-off exercises, conducted Cochrane was a noted epidemiologist
only as funding, interest, or time per- who wrote persuasively about the
mits. Rarely are they updated to take need for medical practitioners to take
into account new studies that are rel- scientific evidence into account in
evant to the review, a challenge that their practice. Cochrane (1972)
is more significant given the cumula- lamented the fact that although ran-
tive growth of evaluation reports domized trials had shown some prac-
highlighted in Figure 1. Yet years tices to be effective and others harm-
may go by before an investigator pur- ful, clinical practitioners and medical
sues funding to update an existing schools were ignoring the informa-
review. The methodology and statis- tion. He later (Cochrane 1979) won-
tical foundation for meta-analysis is dered why the medical sciences had
still rapidly evolving, with improved not yet organized all relevant trials
techniques and new software being into subspecialties so that decision
developed to solve data problems. It makers could take such evidence into
is rare to find reviews that take into account. A protg of Cochrane, an
account these new techniques, con- obstetrician turned researcher
ducting analyses to determine if named Iain Chalmers, soon identi-
results using different methods fied and reviewed randomized trials
converge. relevant to childbirth and prenatal
Some reviewers publish in print interventions (see www.cochrane.
journals, an inefficient method for org).
disseminating reviews. Because In the early 1990s, the U.K.
print journals find them too costly, National Health Service (NHS),
cases in which reviewers take into under the direction of Sir Michael
account cogent criticisms by others Peckham, initiated the Research and
and conduct reanalysis are rarely Development Programme with the
reported. Unlike medical journals, goal of establishing an evidence-
criminological journals do not have a based resource for health care.
strong tradition in routinely printing Because of the success of their earlier
letters to the editor that respond to project on childbirth and pregnancy
criticisms with additional analyses. studies, Chalmers and his colleagues
24 THE ANNALS OF THE AMERICAN ACADEMY
were asked to extend this effort to all which the intended topic area seems
areas of health care intervention. appropriate. Once a title for the pro-
The U.K. Cochrane Centre was estab- posed review is agreed on, it is circu-
lished, with core funding from the lated to ensure that no other similar
NHS, to begin the work. It soon reviews are being prepared by
became clear that the amount of reviewers from other CRGs. Reduc-
work far surpassed the capacity of ing overlap and duplication is a cru-
one center or one nation to take into cial goal for the Cochrane Collabora-
account. In 1993, in honor of his men- tion, as scarce resources must be
tor, Chalmers and his colleagues used judiciously. Reviews are needed
launched the international Cochrane in so many areas of health care that
Collaboration to help people make wide coverage is a priority. Once the
well-informed decisions about title is agreed on, the reviewers must
healthcare by preparing, maintain- then submit a protocol for the review.
ing and promoting the accessibility of The protocol is a detailed plan that
systematic reviews of the effects of spells out a priori the question to be
healthcare interventions. In just 8 answered, the background to the
years, the Cochrane Collaboration i s s u e, an d t h e met h ods t o be
has been able to organize thousands employed. The protocol then goes
of individuals worldwide to contrib- through a round or two of criticism by
ute to its work. Much more informa- the CRG editors. It must conform to a
tion about the Cochrane Collabora- certain template to facilitate elec-
tion can be found at its Web site, t ron i c pu bl i cat i on u s i n g t h e
w w w.cochrane. org. B u t t h e Cochrane Collaborations software,
Cochrane Collaboration, in a very Review Manager, or RevMan. Once
brief time, established a number of the protocol is approved, it is pub-
mechanisms to address challenges to lished in the next edition of the quar-
using systematic reviews in evi- terly electronic publication, the
dence-based health care policy. Cochrane Library, and made avail-
For example, collaborative review able to all subscribers for comment
groups (CRGs) are responsible for and criticism. The editorial board
the core work of systematic review- must decide which criticisms should
ing. CRGs are international net- be taken into account.
works of individuals interested in The reviewers then prepare the
particular health areas such as review according to the protocol.
breast cancer, epilepsy, injuries, and Although deviation from the plan is
stroke. Each CRG has an editorial sometimes necessary, the protocol
board, generally comprising persons forces a prospective, transparent pro-
with scientific or practical expertise cess. Post hoc changes are readily
in the area, who are responsible for detected, and analyses can be done to
quality control of protocols (plans) determine if they altered findings.
and completed drafts of reviews. After the reviewers conduct the
It is useful to examine how a review and write up a draft, it too is
Cochrane review is prepared. First, submitted to the CRG editorial
individuals approach the CRG in board. Once the review draft is
CHALLENGES OF EVIDENCE-BASED POLICY 25
lished in medical journals. For exam- ing an unbiased, single source for evi-
ple, Jadad and his colleagues (1998) dence and producing reviews,
found that Cochrane reviews pro- abstracts, and synopses for different
vided more detail, were more likely to audiences, they facilitate utilization.
test for methodological effects, were
less likely to be restricted by lan-
THE CAMPBELL
guage barriers, and were updated COLLABORATION
more than print journal reviews. The
Cochrane Library is quickly becom- With the success of the Cochrane
ing recognized as the best single Collaboration, the same type of orga-
source of evidence on the effective- nization was soon suggested for re-
ness of health care interventions viewing social and educational eval-
(Egger and Davey-Smith 1998). uations. Adrian Smith (1996),
Reviews by the Cochrane Collabora- president of the Royal Statistical So-
tion are frequently used to generate ciety, issued a challenge when he
and support guidelines by govern- said,
ment agencies such as the National
Institute for Clinical Excellence (for As ordinary citizens . . . we are, through
example, see Chalmers, Hedges, and the media, confronted daily with contro-
Cooper in press). In 1999, the U.S. versy and debate across a whole spec-
National Institutes of Health made trum of public policy issues. Obvious topi-
the Cochrane Library available to all cal examples include educationwhat
16000 of its employees. It is now does work in the classroom?and penal
accessible by all doctors in Brazil, the policywhat is effective in reducing
reoffending? Perhaps there is an opportu-
U.K. NHS, and all U.K. universities
nity . . . to launch a campaign directed at
(Mark Starr, personal communica-
developing analogues to the Cochrane
tion, 2001). Finally the queen recog- Collaboration, to provide suitable evi-
nized Iain Chalmers for his efforts dence bases in other areas besides medi-
with the United Kingdoms greatest cine [emphasis added]. (378)
honor: knighthood!
Thus the Cochrane Collaboration A number of individuals across
has been able to meet many of the different fields and professions orga-
challenges posed by evidence-based nized and met to determine how best
policy in health care. By requiring to meet this challenge. Several ex-
detailed protocols, the Cochrane Col- ploratory meetings were held during
laboration addresses the lack of 1999, including two headed by the
transparency in most systematic School of Public Policy at University
reviews of research. Through rigor- CollegeLondon, and one organized
ous quality control, they produce in Stockholm by the National Board
commendable reviews. By publishing of Health and Welfare. These meet-
electronically, dissemination is ings, which included researchers and
quickened, and the ability to update members of the policy and practice
and correct the reviews in light of communities, provided evidence that
new evidence is realized. By provid- the development of an infrastructure
CHALLENGES OF EVIDENCE-BASED POLICY 27
FIGURE 2
C2-SPECTR
6000
Criminology
5000 Education
Psychology
Social
No of reports of trials
4000
3000
2000
1000
0
1950 1955 1960 1965 1970 1975 1980 1985 1990 1995
Year
Louis, and Frederick Mosteller, eds. Gawande, Atul. 2001. Under Suspicion.
1992. Meta-Analysis for Explanation. The Fugitive Science of Criminal Jus-
New York: Russell Sage. tice. New Yorker, 8 Jan., 50-53.
Cooper, Harris C. and Larry V. Hedges, Glass, Gene V., Barry McGaw, and Mary L.
eds. 1994. The Handbook of Research Smith. 1981. Meta-Analysis in Social
Synthesis. New York: Russell Sage. Research. London: Sage.
C2 Steering Group. 2001. C2: Concept, Greenberg, David and Mark Shroder.
1997. The Digest of Social Experi-
Status, Plans. Overarching Grant Pro-
ments. 2d ed. Washington, DC: Urban
posal. Philadelphia: Campbell Collab-
Institute Press.
oration Secretariat.
Harlen, Wynne. 1997. Educational Re-
Cullen, Francis T. and Paul Gendreau.
search and Educational Reform. In
2000. Assessing Correctional Rehabil- The Role of Research in Mature Educa-
itation: Policy, Practice, and Prospects. tional Systems, ed. Seamus Hegarty.
In Policies, Processes, and Decisions of Slough, UK: National Foundation for
the Criminal Justice System, Criminal Educational Research.
Justice 2000, vol. 3, ed. Julie Horney. Hedges, Larry V. and Ingram Olkin. 1985.
Washington, DC: National Institute of Statistical Methods for Meta-Analysis.
Justice. New York: Academic Press.
Davies, Huw T. O., Sandra M. Nutley, and Hunt, Morton. 1997. The Story of Meta-
Peter C. Smith, eds. 2000. What Analysis. New York: Russell Sage.
Works? Evidence-Based Policy in Pub- Jadad, Alejandor R., Deborah J. Cook,
lic Services. Bristol, UK: Policy Press. Alison Jones, Terry P. Klassen, Peter
Davies, Phillip. 1999. What Is Evidence- Tugwell, Michael Moher, and David
Based Education? British Journal of Moher. 1998. Methodology and Re-
Educational Studies 47:108-21. ports of Systematic Reviews and
Davies, Philip, Anthony Petrosino, and Meta-Analyses: A Comparison of
Iain Chalmers, eds. 1999. Proceedings Cochrane Reviews with Articles Pub-
of the International Meeting on Sys- lished in Paper-Based Journals. Jour-
tematic Reviews of the Effects of Social nal of the American Medical Associa-
and Educational Interventions, July tion 280:278-80.
15-16. London: University College Kirby, Bernard C. 1954. Measuring Ef-
London, School of Public Policy. fects of Criminals and Delinquents.
Sociology and Social Research 38:368-
Egger, Matthias and George Davey-
74.
Smith. 1998. Bias in Location and Se-
Lipsey, Mark W. 1990. Design Sensitivity:
lection of Studies. British Medical
Statistical Power for Experimental Re-
Journal 316:61-66.
search. Newbury Park, CA: Sage.
Farrington, David P. and Anthony
. 1992. Juvenile Delinquency
Petrosino. 2001. The Campbell Collab-
Treatment: A Meta-Analytic Inquiry
oration Crime and Justice Group. An-
into the Variability of Effects. In Meta-
nals of the American Academy of Polit- Analysis for Explanation, ed. Thomas D.
ical and Social Science 578:35-49. Cook, Harris Cooper, David S. Cordray,
Fischer, Joel. 1978. Does Anything Work? Heidi Hartmann, Larry V. Hedges,
Journal of Social Service Research Richard J. Light, Thomas A. Louis, and
1:215-43. Frederick Mosteller. New York:
Fitz-Gibbon, Carol. 1999. Education: The Russell Sage.
High Potential Not Yet Realized. Pub- Lipsey, Mark W. and David B. Wilson.
lic Money & Management 19:33-40. 1993. The Efficacy of Psychological,
CHALLENGES OF EVIDENCE-BASED POLICY 33
THE ANNALS
CAMPBELL COLLABORATION
OF THE AMERICAN
CRIME
ACADEMY
AND JUSTICE GROUP
NOTE: We are very grateful to the Home Office and to Paul Wiles, its director of research and
statistics, for supporting the work of the Campbell Crime and Justice Coordinating Group. An-
thony Petrosino was also supported by a Mellon Foundation grant to the Center for Evaluation,
Initiative for Children Program.
35
36 THE ANNALS OF THE AMERICAN ACADEMY
ration must agree to the following re- 8. A commitment to make the re-
quirements: view available to broader audiences
than readerships of peer-reviewed
academic journals through electronic
1. A commitment to conduct up-
publication and dissemination into
dates of the systematic review to in-
policy, practice, and media outlets.
corporate new evidence, respond to
criticisms, or use more advanced
methods, on a regular basis (for ex- The CJCG consists of 14 members
ample, every 2 years). from 10 countries: David P. Farring-
2. A commitment to undergo a ton (United Kingdom, chair), Ulla V.
rigorous editorial review process B on des on ( D en mark ) , Vi cen t e
from not only researchers but also Garrido (Spain), Peter Grabosky
policy makers, practitioners, and citi- (Australia), Jerry Lee (United
zens to ensure that the review meets States), Mark W. Lipsey (United
high scientific standards and is also States), Friedrich Lsel (Germany),
written to be understandable to Joan McCord (United States), An-
nonacademic audiences. thony Petrosino (United States),
3. A commitment to maintain Lawrence W. Sherman (United
transparent and open review pro- States), Chuen-Jim Sheu (Taiwan),
cesses so that users can comment Richard E. Tremblay (Canada),
and criticize each stage of the review, Hiroshi Tsutomi (Japan), and David
from its proposal through to its com- L. Weisburd (Israel). Anthony
pletion. Petrosino was appointed part-time
4. A commitment to use the most coordinator for the CJCG, and Joan
rigorous search methods available to McCord was appointed to liaise with
ensure that all relevant studies are the Campbell Collaboration Steering
considered for inclusion or exclusion Group. It was also decided that the
and not just those reported in easily institutional base for the Crime and
accessible journals and books. Justice Group should be the Jerry
5. A commitment to cover litera- Lee Center of Criminology at the Fels
ture from around the world and not Center of Government, University of
just the English-speaking world. Pennsylvania, under the direction of
6. A commitment to code and Lawrence Sherman. Since the Brit-
computerize key features of each ish government in general and the
evaluation study reviewed (so that Home Office in particular are in-
anyone accessing the review can or- creasingly committed to evidence-
ganize the studies according to such based policy and practice, David
features as sample size, design, and Farrington applied to the Home Of-
effect size). fice for funding and obtained support
7. A commitment to explicitly re- for the first 2 years of the CJCG, be-
port the final review so that readers ginning in April 2000.
can understand decisions made at To encourage international partic-
each stage, justifications for those de- ipation, the first meeting of the
cisions, and how conclusions were CJCG was held in Paris in May 2000,
reached. coinciding with meetings of the
CAMPBELL COLLABORATION CRIME AND JUSTICE GROUP 41
knows what about the intervention proposals. These people are being
(since double blind trials were ideal); kept informed about all develop-
measurement of dependent variables ments. An informal advisory board of
(for example, official records of crime 30 persons has also been created. The
or self-reports of offending); before members of the board have been
and after measures of crime or asked to suggest topics for review
offending; and follow-up period after and to identify funding opportuni-
the intervention. ties. The Jerry Lee Crime Prevention
Symposium in April 2001, titled Sys-
tematic Reviews of Criminological
SUBSEQUENT DEVELOPMENTS
Interventions, can be regarded as the
official launching meeting of the
The CJCG met for the second time
Crime and Justice Group.
in San Francisco in November 2000,
The next meeting of the CJCG was
coinciding with the American Society
in Paris in May 2001. The work of the
of Criminology meetings. At this
Crime and Justice Group will be pub-
meeting, more topics for review were
licized at the American Society of
proposed, including repeat victimiza-
Criminology meeting in Atlanta,
tion programs; treatment of sex
Georgia, in November 2001. Plans
offenders; domestic violence inter-
include having a panel session on
ventions; reentry programs for incar-
systematic reviews and also a work-
cerated offenders; sports/recreation
shop (sponsored by the Division of
based programs; aftercare treatment
Sentencing and Corrections) provid-
for juvenile offenders; drug courts;
ing training on systematic reviews
prison-based drug treatment; inter-
and meta-analyses. There are also
ventions for serious, violent youth;
plans to have the next Jerry Lee
and interventions for gun violence.
Crime Prevention Symposium, titled
Anthony Petrosino constructed a
Systematic Reviews of Criminologi-
brochure for the Crime and Justice
cal Interventions, in April 2002 and
Group, which is being used to
to publish the proceedings of this
respond to enquiries and to circulate
conference in the first 2003 issue of
at meetings. A Web site for the Crime
The Annals.
and Justice Group has been con-
structed by John Myrtle and Peter
Levan of the Australian Institute of KEY CHALLENGES
Criminology. It has been decided to
house a central registry of studies at Several major challenges were
the Fels Center of Government in identified. The first (perhaps most
Philadelphia. This requires the important and controversial) is what
development of a core coding system criterion of methodological quality
for every study. should be set for including evalua-
More than 120 people from 16 tion studies in systematic reviews.
countries have written to express Several CJCG members argued that
interest in the Crime and Justice only randomized experiments should
Group, many of whom wish to help in be reviewed since these were able to
conducting reviews or reviewing demonstrate effects most convincingly
CAMPBELL COLLABORATION CRIME AND JUSTICE GROUP 43
group has even less internal validity internal validity because of its fail-
since it fails to address many threats ure to tackle the threats mentioned
to internal validity (for example, his- above. An interrupted time series
tory, that is, the idea that changes in study is more convincing than a one-
the outcome might be attributable to group pre-post design, especially if
events other than the intervention; an experimental time series is com-
maturation or the continuation of pared with a control time series.
preexisting trends; testing or instru- It is expected that persons con-
mentation effects; and regression to ducting systematic reviews of inter-
the mean). ventions will select studies with high
Different considerations apply to internal validity for detailed consid-
intervention studies based on areas eration and inclusion in systematic
compared to intervention studies reviews. However, studies with lower
based on individuals. Randomized internal validity may also be listed in
experiments can rarely be carried less detail in the review, together
out to evaluate area-based interven- with summary information about
tions. It is rarely feasible to allocate them (for example, sample size,
at random a sufficiently large num- design, effect size). This will permit
ber of areas (for example, at least others to argue that these studies
100) to experimental or control con- should have been included in the
ditions in order to equate these con- expected dialogue between reviewers
ditions on all possible extraneous and others that will be facilitated by
variables before the intervention the electronic publication of system-
(within the limits of statistical atic reviews. The plan is that reviews
fluctuation). should be continually updated in
In area-based studies, internal light of new studies, relevant criti-
validity is usually maximized by hav- cisms, and methodological develop-
ing before and after measures of an ments. This updating may well
outcome (for example, crime) in com- change the studies that are included
parable experimental and control or excluded.
areas. Even better, the effect of an External validity will be mea-
intervention on crime can be investi- sured according to the extent to
gated after controlling (for example, which the selected studies report
in a regression equation) not only for similar results. It is important to
prior crime but also for other factors determine how well interventions do
that influence crime. Another possi- or do not work within different con-
bility is to match two areas and then texts or boundary conditions. Other
to choose one at random to be the types of validity (for example, con-
experimental area. Of course, several struct validity, statistical conclusion
pairs of areas would be better than validity) will also be summarized by
only one pair. An intervention study the reviewers for each study, who
with no comparable control area also may try to draw conclusions
(merely measuring crime before and about what is the active ingredient
after an intervention, as in a one- of an intervention or about causal
group pre-post design) has low pathways between interventions and
CAMPBELL COLLABORATION CRIME AND JUSTICE GROUP 45
THE ANNALS
RESEARCH DESIGN
OF THE
AND
AMERICAN
STUDY OUTCOMES
ACADEMY
ABSTRACT: Does the type of research design used in a crime and jus-
tice study influence its conclusions? Scholars agree in theory that
randomized experimental studies have higher internal validity than
do nonrandomized studies. But there is not consensus regarding the
costs of using nonrandomized studies in coming to conclusions re-
garding criminal justice interventions. To examine these issues, the
authors look at the relationship between research design and study
outcomes in a broad review of research evidence on crime and justice
commissioned by the National Institute of Justice. Their findings
suggest that design does have a systematic effect on outcomes in
criminal justice studies. The weaker a design, indicated by internal
validity, the more likely a study is to report a result in favor of treat-
ment and the less likely it is to report a harmful effect of treatment.
Even when comparing randomized studies with strong quasi-experi-
mental research designs, systematic and statistically significant dif-
ferences are observed.
NOTE: We are indebted to a number of colleagues for helpful comments in preparing this arti-
cle. We especially want to thank Iain Chalmers, John Eck, David Farrington, Denise Gottfredson,
Doris MacKenzie, Joan McCord, Lawrence Sherman, Brandon Welsh, Charles Wellford, and
David Wilson.
50
RESEARCH DESIGN AND STUDY OUTCOMES 51
programs and their outcomes (for would be relevant to only a small fraction
example, see Boruch, Snyder, and of the key questions for policy and prac-
DeMoya 2000; Campbell and Boruch tice in criminology. Where there are few
randomized experiments, it is expected
1975; Farrington 1983; Feder, Jolin,
that reviewers will select both random-
and Feyerherm 2000). Indeed, a task
ized and non-randomized studies for in-
force convened by the Board of Scien- clusion in detailed reviews. (3)
tific Affairs of the American Psycho-
logical Association to look into statis-
tical methods concluded that for In this article we examine a cen-
research involving causal inferences, tral question relevant both to the
the assignments of units to levels of Campbell Collaboration crime and
the causal variable is critical. Ran- justice effort and to the more general
dom assignment (not to be confused emphasis on developing evidence-
with random selection) allows for the based practice in criminal justice:
strongest possible causal inferences Does the type of research design used
free of extraneous assumptions in a crime and justice study influence
(Wilkinson and Task Force on Statis- the conclusions that are reached? As-
tical Inference 1999). suming that experimental designs
While reliance on experimental are the gold standard for evaluating
studies in drawing conclusions about practices and policies, it is important
treatment outcomes has become to ask what price we pay in including
common in the development of evi- other types of studies in our reviews
dence-based medicine, the Campbell of what works in crime and justice.
Collaboration Crime and Justice Co- Are we likely to overestimate or un-
ordinating Group has concluded that derestimate the positive effects of
it is unrealistic at this time to restrict treatment? Or conversely, might we
systematic reviews on the effects of expect that the use of well-designed
interventions relevant to crime and nonrandomized studies will lead to
justice to experimental studies. In about the same conclusions as we
developing its Standards for Inclu- would gain from randomized experi-
sion of Studies in Systematic Reviews mental evaluations?
(Farrington 2000), the group notes To examine these issues, we look
that it does not require that review- at the relationship between research
ers select only randomized experi- design and study outcomes in a broad
ments: review of research evidence on crime
and justice commissioned by the
National Institute of Justice. Gen-
This might possibly be the case for an in-
erally referred to as the Maryland
tervention where there are many ran-
domized experiments (e.g. cognitive-be-
Report because it was developed in
havioral skills training). However, the Department of Criminology and
randomized experiments to evaluate Criminal Justice at the University of
criminological interventions are rela- Maryland at College Park, the study
tively uncommon. If reviews were re- was published under the title Pre-
stricted to randomized experiments, they venting Crime: What Works, What
RESEARCH DESIGN AND STUDY OUTCOMES 53
1
the researcher can argue with confi- comparisons made. However, in non-
dence that if a difference has been experimental studies, statistical con-
observed between treatment and trols are the primary method applied
comparison groups, it is likely the in attempts to increase the level of a
result of the treatment itself (since studys internal validity. In this case,
randomization has isolated the treat- multivariate statistical methods are
ment effect from other possible used to isolate the effects of treat-
causes). ment from that of other causes. This
In nonrandomized studies, two deman ds of cou rs e t h at t h e
methods may be used for isolating researcher clearly identify and mea-
treatment or program effects. Quasi- sure all other factors that may
experiments, like randomized exper- threaten the internal validity of the
iments, rely on the design of a study outcomes. Only if all such fac-
research study to isolate the effects of tors are included in the multivariate
treatment. Using matching or other models estimated can the researcher
methods in an attempt to establish be confident that the effects of treat-
equivalence between groups, quasi- ment that have been reported are not
experiments mimic experimental confounded with other causes.
designs in that they attempt to rule In theory, the three methods
out competing causes by identifying described here are equally valid for
groups that are similar except in the solving the problem of isolating
nature of the treatment that they treatment or program effects. Each
receive in the study. Importantly, can ensure high internal validity
however, quasi-experiments do not when applied correctly. In practice,
randomize out the effects of other however, as Feder and Boruch (2000)
causes as is the case in randomized note, there is little disagreement
experimental designs; rather they that experiments provide a superior
seek to maximize the equivalence method for assessing the effective-
between the units studied through ness of a given intervention (292).
matching or other methods. Threats Randomization, according to Kunz
to internal validity in quasi-experi- and Oxman (1998), is the only
mental studies derive from the fact means of controlling for unknown and
that it is seldom possible to find or to unmeasured differences between
create treatment and control groups comparison groups as well as those
that are not systematically different that are known and measured
in one respect or another. (1185). While random allocation
Nonexperimental studies rely pri- itself ensures high internal validity
marily on statistical techniques to in experimental research, for quasi-
distinguish the effects of the inter- experimental and nonexperimental
vention or treatment from other con- research designs, unknown and
founding causes. In practice, quasi- unmeasured causes are generally
experimental studies often rely as seen as representing significant
well on statistical approaches to potential threats to the internal
2
increase the equivalence of the validity of the comparisons made.
RESEARCH DESIGN AND STUDY OUTCOMES 55
TABLE 1 TABLE 2
STUDIES CATEGORIZED BY SMS STUDIES CATEGORIZED BY THE IRR
Studies Studies
SMS n Percentage IRR n Percentage
1 10 3 1 34 11
2 94 31 0 76 25
3 130 42 1 198 64
4 28 9
5 46 15 Total 308 100
TABLE 4
CROSS-TABULATION OF SMS AND IRR
SMS
1 2 3 4 5
IRR n Percentage n Percentage n Percentage n Percentage n Percentage
1 0 0 8 9 13 10 6 21 7 15
0 2 20 16 17 31 24 5 18 22 48
1 8 80 70 74 86 66 17 61 17 37
TABLE 5 TABLE 6
COMPARING QUASI-EXPERIMENTAL COMPARING HIGH-QUALITY QUASI-
STUDIES (SMS = 3 OR 4) WITH EXPERIMENTAL DESIGNS (SMS = 4)
RANDOMIZED EXPERIMENTS (SMS = 5) WITH RANDOMIZED DESIGNS (SMS = 5)
SMS SMS
3 or 4 5 4 5
IRR n Percentage n Percentage IRR n Percentage n Percentage
1 19 12 7 15 1 6 21 7 15
0 36 23 22 48 0 5 18 22 48
1 103 65 17 37 1 17 61 17 37
compare these to the randomized this was true for only 37 percent of
studies included in category 5, the the randomized studies in category 5.
relationship between study out- Accordingly, even when comparing
comes and study design remains sta- those nonrandomized studies with
tistically significant at the .05 level the highest internal validity with
(see Table 6). There is little difference randomized experiments, we find
between the two groups in the pro- significant differences in terms of
portion of backfire o u t comes reported study outcomes.
reported; however, there remains a
very large gap between the propor- Taking into account tests
tion of SMS category 4 and SMS cate- of statistical significance
gory 5 studies that report an outcome It might be argued that had we
in the direction of treatment effec- used a criterion of statistical signifi-
tiveness. While 61 percent of the cat- cance, the overall findings would not
egory 4 SMS studies reported a posi- have been consistent with the analy-
tive treatment or intervention effect, ses reported above. While we cannot
RESEARCH DESIGN AND STUDY OUTCOMES 63
simply through the process of ran- argue just the opposite. The inflexi-
domization. However, nonran- bility of randomized experimental
domized studies demand much designs has sometimes been seen as
insight and knowledge in the devel- a barrier to development of effective
opment of comparable groups of sub- theory and practice in criminology
jects. Not only must the researcher (for example, see Clarke and Cornish
understand the factors that influ- 1972; Eck 2001; Pawson and Tilley,
ence treatment so that he or she can 1997). Here it is argued that in a field
prevent confounding in the study in which we still know little about the
results, but such factors must be root causes and processes that
measured and then controlled for underlie phenomena we seek to
through some statistical or practical influence, randomized studies may
procedure. not allow investigators the freedom
It may be that such manipulation to carefully explore how treatments
is particularly difficult in criminal or programs influence their intended
justice study. Criminal justice practi- subjects. While this argument has
tioners may not be as strongly social- merit in specific circumstances, espe-
ized to the idea of experimentation as cially in exploratory analyses of
are practitioners in other fields like problems and treatments, we think
medicine. And in this context, it may our data suggest that it can lead in
be that a subtle form of creaming in more developed areas of our field to
which the cases considered most significant misinterpretation and
amenable to intervention are placed confusion.
in the intervention group is common.
In specific areas of criminal justice, CONCLUSION
such creaming may be exacerbated
by self-selection of subjects who are We asked at the outset of our arti-
motivated toward rehabilitation. cle whether the type of research
Nonrandomized designs, even in rel- design used in criminal justice influ-
atively rigorous quasi-experimental ences the conclusions that are
studies, may be unable to compen- reached. Our findings, based on the
sate or control for why a person is Maryland Report, suggest that
considered amenable and placed in design does matter and that its effect
the intervention group. Matching on in criminal justice study is system-
traditional control variables like age atic. The weaker a design, as indi-
and race, in turn, might not identify cated by internal validity, the more
the subtle components that make likely was a study to report a result in
individuals amenable to treatment favor of treatment and the less likely
and thus more likely to be placed in it was to report a harmful effect of
intervention or treatment categories. treatment. Even when comparing
Of course, we have so far assumed studies defined as randomized
that nonrandomized studies are designs in the Maryland Report with
biased in their overestimation of pro- strong quasi-experimental research
gram effects. Some scholars might designs, systematic and statistically
RESEARCH DESIGN AND STUDY OUTCOMES 67
significance tests showing effectiveness and significant results were present) but no
the preponderance of all available evidence statistically significant results supporting the
supporting the same conclusion (Sherman effectiveness of treatment.
et al. 1997, 2-20).
11. It is the case that many of the studies in
this area would have been excluded anyway References
since they often did not have a crime or delin-
quency outcome measure (but rather exam- Bailey, Walter C. 1966. Correctional Out-
ined early risk factors for crime and delin- come: An Evaluation of 100 Reports.
quency). Journal of Criminal Law, Criminology
12. While the Maryland Report is consis- and Police Science 57:153-60.
tent with other recent reviews that also point Boruch, Robert F., Anthony Petrosino,
to greater success in criminal justice interven- and Iain Chalmers. 1999. The Camp-
tions during the past 20 years (for example, see
bell Collaboration: A Proposal for Sys-
Poyner 1993; Visher and Weisburd 1998;
tematic, Multi-National, and Continu-
Weisburd 1997), we think the very high per-
centage of studies showing a treatment impact ous Reviews of Evidence. Background
is likely influenced by publication bias. The paper for the meeting at University
high rate of positive findings is also likely in- CollegeLondon, School of Public Pol-
fluenced by the general weaknesses of the icy, July.
study designs employed. This is suggested by Boruch, Robert F., Brook Snyder, and
our findings reported later: that the weaker a Dorothy DeMoya. 2000. The Impor-
research design in terms of internal validity, tance of Randomized Field Trials.
the more likely is the study to report a positive Crime & Delinquency 46:156-80.
treatment outcome.
Burtless, Gary. 1995. The Case for Ran-
13. The coding scheme for scale A was as
follows. A value of 1 indicates that the study
domized Field Trials in Economic and
had any statistically significant findings sup- Policy Research. Journal of Economic
porting a positive treatment effect, even if Perspectives 9:63-84.
findings included results that were not signifi- Campbell, Donald P. and Robert F.
cant or had negative or backfire findings. A Boruch. 1975. Making the Case for
value of 0 indicates that the study had only Randomized Assignment to Treatments
nonsignificant findings. A value of 1 indicates by Considering the Alternatives: Six
that the study had only statistically signifi- Ways in Which Quasi-Experimental
cant negative or backfire findings or statisti-
Evaluations in Compensatory Educa-
cally significant negative findings with other
tion Tend to Underestimate Effects. In
nonsignificant results.
14. Scale B was created according to the fol- Evaluation and Experiment: Some
lowing rules. A value of 2 indicates that the Critical Issues in Assessing Social
study had only or mostly statistically signifi- Programs, ed. Carl Bennett and
cant findings supporting a treatment effect Arthur Lumsdaine. New York: Aca-
(more than 50 percent) when including all re- demic Press.
sults, even nonsignificant ones. A value of 1 in- Chalmers, Iain and Douglas G. Altman.
dicates that the study had some statistically 1995. Systematic Reviews. London:
significant findings supporting a treatment ef- British Medical Journal Press.
fect (50 percent or less, counting both positive
Clarke, Ronald V. and Derek B. Cornish.
significant and nonsignificant results) even if
the nonsignificant results outnumbered the
1972. The Control Trial in Institu-
positive statistically significant results. A tional Research: Paradigm or Pitfall
value of 0 indicates that no statistically signifi- for Penal Evaluators? London: HMSO.
cant findings were reported. A value of 1 indi- Cox, Stephen M., William S. Davidson,
cates that the study evidenced statistically and Timothy S. Bynum. 1995. A Meta-
significant backfire effects (even if non- Analytic Assessment of Delinquency-
RESEARCH DESIGN AND STUDY OUTCOMES 69
Meta-Analytic Methods
for Criminology
By DAVID B. WILSON
71
72 THE ANNALS OF THE AMERICAN ACADEMY
studies of interest to the reviewer has been argued that the correlation
must examine the same basic rela- coefficient is the ideal effect size
tionship, even if at a broad level of (Rosenthal 1991). However, the stan-
abstraction. At the broad end of the dardized mean difference and odds
continuum would be a group of stud- ratio effect sizes have distinct statis-
ies examining the effects of school- tical advantages over the correlation
based prevention programs on delin- coefficient for intervention research
quent behavior. At the narrow end of and are more natural indices of pro-
the continuum would be a set of repli- gram effects.
cations of a study on the effects of the
drug DepoProvea on the perpetra- Standardized
tion of sexual offenses. The research mean difference
designs of a collection of studies
would all need to be sufficiently simi- The standardized mean differ-
lar such that a comparable effect size ence, d, represents the effect of an
could be computed from each. Thus intervention as the difference
between the intervention and com-
most meta-analyses of intervention
parison group means on the depend-
studies will stipulate that eligible
ent variable of interest, standardized
studies use a comparison group
by the pooled within-groups stan-
design.
dard deviation. Thus findings based
The specific effect size index used on different operationalizations of
in a given meta-analysis will depend the dependent variable of interest
on the nature of the research being (for example, delinquency) are stan-
synthesized. Commonly used effect dardized to a common metric: stan-
size indices for intervention research dard deviation units for the popula-
are the standardized mean differ- tion. An advantage of d is that it can
ence, odds ratio, and correlation coef- be computed from a wide range of
ficient. The standardized mean dif- statistical data, including means and
ferencetype effect size is well suited standard deviations, t tests, F tests,
to two group comparison studies (for correlation coefficients, and 2 2 con-
example, a treatment versus a com- tingency tables (see Lipsey and Wil-
parison condition) with continuous son 2001). Although conceptualized
or dichotomous dependent measures. as the difference between two groups
The odds ratio is well suited to these on a continuous dependent variable,
same research domains with the d can also be computed from dichoto-
exception that the dependent mea- mous data.
sures must be dichotomous, such as
whether the participants recidivated
Odds ratio
within 12 months of leaving the pro-
gram. The correlation coefficient can The odds ratio, o, represents the
be applied to the broadest range of effect of an intervention as the odds
research designs, including all of a favorable (or unfavorable) out-
designs for which standardized mean come for the intervention group rela-
difference and odds ratio effect sizes tive to the comparison group. It is
can be computed. Because of this, it used when the outcome is measured
76 THE ANNALS OF THE AMERICAN ACADEMY
very modest for all but very small studies, such as reliability and valid-
sample sizes, but the adjustment is ity coefficients. The logic of these
easy to perform and routinely done adjustments is to estimate what
when using d as the effect size index would have been observed under
(for formulas, see the appendix). more ideal research conditions.
When using the odds ratio, one These adjustments, while common in
encounters a complication that is meta-analyses of measurement
also easily rectified. The odds ratio is generalizability studies, are rarely
asymmetric, with negative relation- used in meta-analyses of interven-
ships represented as values between tion research. If they are used, it is
0 and 1 and positive relationships recommended that a sensitivity
represented as values between 1 and analysis be performed to assess the
infinity. This complicates analysis. effect the adjustments have on the
Fortunately, the natural logarithm of results.
the odds ratio is symmetric about 0
with a well-defined standard error. Statistical independence
The importance of the latter is dis- among effect sizes
cussed below. Thus, for purposes of A complication with effect size
analysis, the odds ratio is trans- data is the often numerous effect
formed into the logged odds ratio. sizes of interest available from each
Results can be transformed back into study. Effect sizes that are based on
odds ratios for purposes of interpre- the same sample of individuals (or
tation using the antilogarithm. other units of analysis, such as city
Similarly the correlation coeffi- blocks and so forth) are statistically
cient has a distributional shape that dependent, that is, correlated with
is less than ideal for purposes of com- each other. Meta-analytic analysis
puting averages. Furthermore the assumes that each data point (effect
standard error is asymmetric, partic- size in this case) is statistically inde-
1
ularly as the correlation approaches pendent of all other data points.
1 or +1. This is easily solved by Thus we can include only one effect
applying Fishers Zr transformation, size per sample in any given analysis.
which normalizes the correlation and An independent set of effect sizes can
results in a standard error that is be obtained through several strate-
remarkably simple. As with the odds gies. First, each major outcome con-
ratio, final results can be trans- struct of interest can, and should, be
formed back into correlation coeffi- analyzed separately. For example,
cients for interpretative purposes. effect sizes representing employ-
Hunter and Schmidt (1990) pro- ment success should be analyzed sep-
posed adjusting effect sizes for mea- arately from those representing
surement unreliability and invalid- criminal behavior. Second, multiple
ity, range restriction, and artificial effect sizes within each outcome con-
dichotomization. These adjustments, struct can be averaged to produce one
however, depend on information that effect size per study or sample within
is rarely reported for outcome mea- a study. Alternatively, a meta-ana-
sures in crime and justice evaluation lyst may choose a single effect size
78 THE ANNALS OF THE AMERICAN ACADEMY
based on an explicit criterion. That is, the overall mean effect size, com-
the meta-analyst may prefer rearrest puted as a weighted mean, weighting
data over reinstitutionalization data by the inverse variance weight. A z
if the former are available. Finally, test can be performed to assess
the meta-analyst may randomly whether the mean effect size is sta-
select among those effect sizes that tistically greater than (or less than)
are of interest to a given analysis. 0, and a confidence interval can be
Note that several analyses can be constructed around the mean effect
performed, each with a different set size. Both statistics rely on the stan-
of independent effect sizes. dard error of the mean effect size,
computed from the sum of the
The inverse variance weight weights. Thus both the precision and
number of the individual effect sizes
An additional complication of influence the precision of the mean
meta-analytic data is the differential effect size. (For equations, see the
precision in effect sizes across stud- appendix.)
ies. Effect sizes based on large sam-
The mean effect size is meaningful
ples, all other things being equal, are
only if the effects are consistent
more precise than effect sizes based
across studies, that is, statistically
on small samples. A simple solution
homogeneous. If the effects are
to this problem would be to weight
highly heterogeneous, then a single
each effect size by its sample size.
overall mean effect size does not ade-
Hedges (1982) showed, however, that
qu at el y repres en t t h e ef f ect s
the optimal weight is based on the
observed by the collection of studies.
variance (squared standard error) of
In meta-analysis, consistency in
each effect size. This is intuitively
effects is assessed with the homoge-
appealing as well, for the standard
neity statistic Q. A statistically sig-
error is a statistical expression of the
n i f i can t Q i n di cat es t h at t h e
precision of parameter, such as an
observed variability in effect sizes
effect size. The smaller the standard
exceeds statistical expectations
error, the more precise is the effect
regarding the variability that would
size. Thus, in all meta-analytic anal-
be observed across pure replications,
yses, weights are computed from the
that is, if the collection of studies
inverse of the squared standard error
were indeed estimating a common
of the effect size. This is called the
population effect size. A statistically
inverse variance weight method.
nonsignificant Q suggests that the
Equations for the inverse variance
variability in effects across studies is
weight for each of the three effect size
no greater than expected due to sam-
indices discussed above are pre-
pling error.
sented in the appendix.
A heterogeneous distribution (a
The mean effect size significant Q) is often the desired
and related statistics outcome of a homogeneity analysis.
Heterogeneity justifies the explora-
A starting point for the analysis of tion of the relationship between study
effect size data is the computation of features and effects, an important
META-ANALYTIC METHODS FOR CRIMINOLOGY 79
statistical terms, this is a random- effect size per study for any given
effects model. analysis may also affect the meta-
Methods for estimating random- analytic findings. For example, in the
effects models in meta-analysis are boot camp systematic review by Mac-
well developed. The basic method Kenzie, Wilson, and Kider (2001), the
involves modifying the definition of analyses were performed on a single
the inverse variance weight such effect size selected from each study
that it incorporates both the subject- based on a set of decision rules. A sen-
and study-level estimates of instabil- sitivity analysis showed that using a
ity. The inverse variance weight is composite of all recidivism effect
thus based on both the standard sizes produced the same results, bol-
error of the effect size and an esti- stering the authors confidence in the
mate of the variability in the distri- findings. Third, if the meta-analysis
bution of population effects. The lat- has included methodologically weak
ter is computed from the observed studies, analyses examining the rela-
distribution of effects. Random- tionship between method features
effects models are more conservative and observed effects are essential.
than fixed-effects models. Confi-
dence intervals will be larger, and Illustration: Cognitive-
regression coefficients that were sta- behavioral programs
tistically significant under a fixed- for sex offenders
effects model may no longer be signif-
To illustrate the methods outlined
icant under a random-effects model.
above, I have selected a subset of
It is recommended that meta-analy-
studies included in a meta-analysis
ses of criminological literatures use a
of sex offender programs (Gallagher,
random-effects model of analysis
Wilson, and MacKenzie no date).
unless a clear justification to do oth-
Presented below are the programs
erwise exists.
based on cognitive-behavioral princi-
ples. Studies were included if they
Sensitivity analysis
used a comparison group design and
A final analytic issue is the sensi- the comparison received either no
tivity of the results to unusual study treatment or non-sex-offender-spe-
effects and decisions made by the cific treatment. Studies also had to
meta-analyst. First, it is wise to report a measure of sex offense recid-
examine the influence of outliers in ivism at some point following termi-
the distribution of effect sizes and nation of the program.
the distribution of inverse variance A total of 13 studies met the eligi-
weights. A modest effect size outlier bility criteria for this meta-analysis.
with a large weight can drive an The recidivism data were dichoto-
analysis. Rerunning an important mous and as such, the odds ratio was
analysis with and without highly selected as the effect size index. The
influential studies can help verify odds ratio and 95 percent confidence
that the observed result is not solely interval for these 13 studies are pre-
a function of a single unusual study. sented in Figure 1. Visual inspection
Second, the method of selecting one of these odds ratios shows a distinct
82 THE ANNALS OF THE AMERICAN ACADEMY
FIGURE 1
ODDS RATIO AND 95 PERCENT CONFIDENCE INTERVAL FOR EACH OF
THE 13 COGNITIVE-BEHAVIORAL SEX OFFENDER EVALUATION STUDIES
APPENDIX
EQUATIONS FOR THE CALCULATION OF EFFECT
SIZES AND META-ANALYTIC SUMMARY STATISTICS
(13) ES = (ES w ) Weighted mean effect size, where ES is the effect size
w index (equations 4, 5, or 6) and w is the inverse
variance weight (equation 12)
META-ANALYTIC METHODS FOR CRIMINOLOGY 87
APPENDIX Continued
1
(14) se ES = The standard error of the mean effect size
w
ES
(15) z = A z test; tests whether ES is statistically greater than or
se ES less than 0
(16) LowerCI = ES 1.96se ES Lower bound of the 95 percent confidence interval
Homogeneity test Q
( (ES )
2
w)
(18) Q = (ES 2
w) Homogeneity test Q; distributed as a chi-square,
w degrees of freedom equals the number of effect
sizes less 1
( (ES )
2
wj
(ES
j
(21) Q j = 2
w ) Q between groups; where j is 1 to the number of
w
j j
j categories for the independent variable; distributed
as a chi-square with j 1 degrees of freedom
(22) QW = Q QB Q within groups; where Q is the overall homogeneity
statistics defined in equation 18 and QB is defined in
equation 21; distributed as a chi-square with the
number of effect sizes minus the number of categories
in the independent variable as the degrees of freedom
(23) Use specialized software For example, SAS, SPSS, or Stata macros by Lipsey
and Wilson (2001); SAS macros by Wang and
Bushman (1998)
88 THE ANNALS OF THE AMERICAN ACADEMY
THE ANNALS
EARLY PARENT
OFTRAINING
THE AMERICAN ACADEMY
NOTE: We thank the following agencies for financial support: Canadian Institute for Ad-
vanced Research, FCAR, Fonds de la Recherche en Sant de Qubec, Molson Foundation, Social
Sciences and Humanities Research Council of Canada, and St-Justine Hospital Research Centre.
90
EARLY PARENT TRAINING 91
two years following birth (Olds et al. effects from other effects including
1998). These home visitation pro- developmental effects.
grams aim at a wide range of out-
comes including maternal physical Types of participants
and psychosocial health, parenting
The review was limited to families
skills, and childrens psychosocial
with a child under age 3 at the start
development and physical health.
of the intervention to ensure that the
The long-term impact on delin-
interventions were provided early in
quency of intensive home visitation
the childs life. However, no limits
during a period of more than two
were set concerning the childs age at
years supports the hypothesis that
the end of the intervention. In addi-
quality of family environment during
tion, selected interventions could tar-
the early years is a key to delin-
get either the general population
quency prevention (Patterson, Reid,
(universal intervention) or a high-
and Dishion 1992; Yoshikawa 1994;
risk group (selective intervention).
McCord, Widom, and Crowell 2001;
Nagin and Tremblay 2001). Early
Types of intervention
parenting interventions generally
postulate that quality of parent-child Studies were eligible for this
relations will facilitate learning of review when parent training or sup-
control over impulsive, oppositional, port was a major component of the
and aggressive behavior, thus intervention, although not necessar-
reducing disruptive behavior and its ily the only one.
long-term negative impact on social
integration. Types of outcomes
This review aims to address The original aim of the review was
whether early parenting and home to assess the impact of the interven-
visitation programs are effective in tions on the childrens delinquent
preventing behavior problems and behavior. However, since we found
delinquency in children. only one study assessing delin-
quency, we used a broader scope in
our review and selected studies with
CRITERIA FOR CONSIDERING
outcome measures of disruptive
STUDIES FOR THIS REVIEW
behaviors. These assessments
included self-reported delinquency;
Types of studies self-, parent-, or teacher-rated mea-
sures of disruptive behavior; and
Only studies employing random observer-rated assessments of dis-
assignment or quasi-experimental ruptive behavior in the classroom.
(preintervention and postinter-
vention assessments and adequate SEARCH STRATEGY FOR
control groups ) des ign s w ere IDENTIFICATION OF STUDIES
included. Studies lacking control
groups were excluded as they cannot Our starting point for searching
help differentiate intervention through the literature was two
EARLY PARENT TRAINING 93
with preschool children, but only 34 interventions (that is, they targeted
met the 4- or 5-star classification. Of high-risk groups, mostly socially dis-
the 34 studies, a total of 6 trials met advantaged families or, in one study,
our inclusion criteria. Three addi- premature babies). Boys and girls
tional trials were identified in were included in all studies. Two
Tremblay, LeMarquand, and Vitaro studies targeted minority groups:
(1999), but they were not kept in our African Americans and Mexican
review as they did not meet the 4- Americans. The latter study was the
star design criteria of Mrazek and only one that did not attempt to
Brown. obtain a representative population
The PsychINFO search yielded sample due to major recruitment
151 new abstracts, none of which challenges. While it can be argued
were included in the review. Most of that nearly all studies tried to
them were excluded because they involve families, in practice, most
targeted older children. Others were studies intervened mainly with
excluded for methodological reasons, mothers.
mostly because of the absence of a In total, 7917 families were ran-
control group. Search i n g t h e domized to receive parent training or
Cochrane Library and the Future of to be in a control group. One study
Children publications generated an had more than 4000 participants
additional four reviews that provided involving 21 sites, two had more than
information about one trial that had 1000 participants, three had more
not already been identified and met than 300, and one had 125. Attrition
our criteria. rates varied greatly from one study
Thus seven trials met our criteria. to another, ranging from 20 to 67 per-
The data were summarized using cent. Sample numbers relevant to
effect sizes but were not combined in our review varied from 117 to more
a meta-analysis due to the small than 2000 (exact number was not
number of studies and the presence available for St-Pierre and Layzer
of substantial heterogeneity among 1999).
them.
Intervention characteristics
RESULTS Four interventions began when
the child was 12 months old or youn-
Sample characteristics ger (see Table 2). All four continued
beyond age 2, up to age 3, 5, or 6. Two
All seven trials were randomized trials began during the prenatal
controlled experiments (see Table 1). period, and both continued up to 2
All but two were conducted in the years. Finally, one trial began when
United States; one was done in Aus- children were 24 months old and
tralia and another in Bermuda. Two ended when they were about 4 years.
interventions targeted the general Overall duration of interventions
population (universal preventive ranged from more than 2 to 6 years.
interventions) while the remaining Length of follow-up ranged from
five were selective preventive immediate end of intervention to 13
EARLY PARENT TRAINING 95
TABLE 1
SAMPLE CHARACTERISTICS OF STUDIES INCLUDED IN THE REVIEW
a
Study Target Population Country Final N
TABLE 2
INTERVENTION CHARACTERISTICS OF STUDIES INCLUDED IN THE REVIEW
TABLE 3
OUTCOME FINDINGS OF THE STUDIES INCLUDED IN THE REVIEW
Direction of
Study Outcome Effect Sizea p Value Outcome
(continued)
98 THE ANNALS OF THE AMERICAN ACADEMY
TABLE 3 Continued
Direction of
Study Outcome Effect Sizea p Value Outcome
program that started early during available for this review. In addition,
the pregnancy of high-risk women overall results were mixed: four stud-
and continued during the 2 years ies reported no evidence of effective-
after birth. The nurses promoted sev- ness, two reported beneficial effects,
eral aspects of maternal functioning and one reported mainly beneficial
and well-being including competent effects with some harmful effects.
care of the children. The nurses com- The latter effects, however, con-
pleted an average of 9 visits during cerned one specific item only: late for
pregnancy and 23 visits from birth to school. Studies varied greatly from
the childs second year (Olds et al. one another on various aspects,
1997). including outcome measures, childs
age at evaluation, the nature and
DISCUSSION duration of the intervention, and
sample size. Studies reporting bene-
A very limited number of well- ficial effects showed no specific pat-
designed studies including both terns allowing distinction from the
early interventions and outcomes other studies. In this context, it is
related to disruptive behaviors were impossible to make a definitive
EARLY PARENT TRAINING 99
THE ANNALS
EFFECTS OF HOT
OF THE
SPOTS
AMERICAN
POLICING
ACADEMY
ON CRIME
By ANTHONY A. BRAGA
NOTE: The author would like to thank Phyllis Schultze at Rutgers Universitys Criminal Jus-
tice Library for her valuable assistance in completing this review.
104
EFFECTS OF HOT SPOTS POLICING ON CRIME 105
TABLE 1
HOT SPOTS POLICING EXPERIMENTS AND QUASI-EXPERIMENTS
a
Study Treatment Hot Spot Definition Research Design
TABLE 1 Continued
a
Study Treatment Hot Spot Definition Research Design
(continued)
112 THE ANNALS OF THE AMERICAN ACADEMY
TABLE 1 Continued
a
Study Treatment Hot Spot Definition Research Design
Houston (TX) Patrol initiative designed Seven highest crime Quasi-experiment with
Targeted to reduce index crimes beats were nonequivalent control
Beat in seven beats selected for this groups; target beats
Program Three beats used high- program were matched to
(Caeti 1999) visibility patrol at hot Enforcement actions noncontiguous
spots targeted at hot comparison beats
Three beats used zero spots in beats through cluster analysis
tolerance policing at identified by and correlations of
hot spots computer analyses census data
One beat used a POP Difference of means in
approach comprising reported crime were
mostly traditional used to evaluate
tactics to control hot program effects for
spots 3-year preintervention
Two-year intervention and 2-year intervention
period period
Three high-visibility
patrol beats managed
by one substation
Experienced police
resistance to the
program
EFFECTS OF HOT SPOTS POLICING ON CRIME 113
TABLE 1 Continued
a
Study Treatment Hot Spot Definition Research Design
TABLE 2
RESULTS OF HOT SPOTS POLICING EVALUATIONS
Displacement/
Study Crime Outcomes Other Outcomes Diffusion
(continued)
TABLE 2 Continued
Displacement/
Study Crime Outcomes Other Outcomes Diffusion
St. Louis (MO) All three drug locations expe- None Compared trends in
POP in rienced varying reductions calls at targeted
Three Drug in total calls addresses to
Market Regression analysis suggests trends in calls at
Locations that reductions on blocks other addresses
Study (Hope where drug locations were on same block
1994) located were greater than Location 1signifi-
other blocks and intersec- cant displacement
tions in surrounding areas into surrounding
addresses; loca-
tion 2no dis-
placement or dif-
fusion; location
3no displace-
ment or diffusion
EFFECTS OF HOT SPOTS POLICING ON CRIME 117
TABLE 2 Continued
Displacement/
Study Crime Outcomes Other Outcomes Diffusion
TABLE 2 Continued
Displacement/
Study Crime Outcomes Other Outcomes Diffusion
program did not work in treatment was implemented to control hot spots
beats with reported significant crime did not experience noteworthy
reductions relative to control beats decreases relative to a control beat.
with significant crime reductions The limits of the analytic framework
were not justified. It is completely preclude conclusions that certain
possible that the observed significant types of policing strategies may be
reductions in the treatment beats more effective in preventing crime in
were significantly greater than the hot spots. Nevertheless, the results of
significant reductions in control this study can be broadly taken to
beats. support the position that focused
Given these caveats, the Houston police enforcement efforts can be
Targeted Beat quasi-experiment sug- effective in reducing crime at hot
gests that the aggregated treatment spots.
beats experienced significant reduc- The Beenleigh Calls for Service
tions in auto theft, total part 1 index quasi-experiment found no notewor-
crimes, and total part 1 patrol-sup- thy differences in the total number of
pressible crimes (robbery, burglary, calls in the town of Beenleigh rela-
and auto theft) relative to aggregated tive to the matched town of Brown
control beats. The three treatment Plains (Criminal Justice Commis-
beats where zero-tolerance aggres- sion 1998, 25). However, simple
sive disorder policing was used to nonexperimental pre-post compari-
control hot spots experienced mixed sons found noteworthy reductions in
reductions in part 1 crimes relative total citizen calls for service in 16 of
to control beats; the three treatment 19 case studies included in the
beats where high-visibility directed report. The research team concluded
patrol was used to control hot spots that the POP strategy enjoyed some
experienced reductions in a wide success in reducing calls for service
variety of part 1 crimes relative to at the targeted locations, but due to
control beats; the one treatment beat the small scale of the project and lim-
where an enforcement POP strategy itations of the research design, these
EFFECTS OF HOT SPOTS POLICING ON CRIME 119
crime prevention gains were not this resulted in low levels of treat-
large enough to be detected at the ment during the early months of both
aggregate town level (Criminal Jus- experiments. In the Jersey City
tice Commission 1998, 28). DMAP experiment, this situation
The Minneapolis RECAP experi- was remedied by providing a detailed
ment showed no statistically signifi- crackdown schedule to the narcotics
cant differences in the prevalence of squad commander and extending the
citizen calls for service at addresses experiment from 12 to 15 months.
that received the POP treatment as This problem was remedied in the
compared to control addresses Jersey City POP experiment by
(Sherman, Buerger, and Gartin 1989, changing the leadership of the POP
21). These results were probably due unit, developing an implementation
to the assignment of too many cases accountability system, and providing
to the RECAP unit, thus outstripping additional training in the POP
the amount of resources and atten- approach as well as through other
tion the police officers provided to smaller adjustments.
each address (Buerger 1993). More- The patrol treatment in the Min-
over, the simple randomization pro- neapolis Hot Spots experiment
cedure led to the placing of some of (Sherman and Weisburd 1995, 638-
the highest-event addresses into the 39) was disrupted during summer
treatment group; this led to high months due to a peak in the overall
variability between the treatment calls for service received by the Min-
and control groups and low statisti- neapolis Police Department and a
cal power. Although the overall find- shortage of officers due to vacations;
ings suggest that the RECAP pro- this situation was further compli-
gram was not effective in preventing cated by changes in the computerized
crime, a case study analysis revealed calls for service system implemented
that several addresses experienced in the fall. The changes in the calls
dramatic reductions in total calls for for service system and the disappear-
service (Buerger 1992). ance of differences in patrol dosage
Beyond the RECAP experiment, between treatment and control hot
only three other studies reported spots during summer months were
potential threats to the internal addressed by conducting separate
validity of the research designs. The outcome analyses using different
Jersey City DMAP experiment intervention time periods; there were
(Weisburd and Green 1995, 721) and no substantive differences in the out-
Jersey City POP at Violent Places comes of the experiment across the
experiment (Braga 1997, 107-42) different time periods. Of course,
reported instances where the treat- these implementation problems are
ments were threatened by subver- not unique to these experiments;
sion by the participants. The officers many well-known criminal justice
charged with preventing crime at the field experiments have experienced
treatment hot spots were resistant to and successfully dealt with method-
participating in the programs, and ological difficulties.6
120 THE ANNALS OF THE AMERICAN ACADEMY
from the Minneapolis Spouse Abuse Studies. Albany, NY: Harrow and
Experiment. Journal of Quantitative Heston.
Criminology 4:209-23. Clarke, Ronald V. and David L. Weisburd.
Braga, Anthony A. 1997. Solving Violent 1994. Diffusion of Crime Control Ben-
Crime Problems: An Evaluation of the efits: Observations on the Reverse of
Jersey City Police Departments Pilot Displacement. In Crime Prevention
Program to Control Violent Places. Studies. Vol. 2, ed. Ronald V. Clarke.
Ph.D. diss., Rutgers University, NJ. Monsey, NY: Criminal Justice Press.
Braga, Anthony A., David L. Weisburd, Cohen, Lawrence E. and Marcus Felson.
El i n Wa r i ng, L or r a i ne G r e e n 1979. Social Change and Crime Rate
Mazerolle, William Spelman, and Trends: A Routine Activity Approach.
Francis Gajewski. 1999. Problem-Ori- American Sociological Review 44:588-
ented Policing in Violent Crime 605.
Places: A Randomized Controlled Ex- Cook, Thomas D. and Donald T. Camp-
periment. Criminology 37:541-80. bell. 1979. Quasi-Experimentation:
B r a nt i ng ha m , Pa ul a nd Pa t r i c i a Design and Analysis Issues for Field
Brantingham, eds. 1991. Environmen- Settings. Boston: Houghton Mifflin.
tal Criminology. 2nd ed. Prospect Cornish, Derek and Ronald V. Clarke,
Heights, IL: Waveland Press. eds. 1986. The Reasoning Criminal:
Buerger, Michael, ed. 1992. The Crime Rational Choice Perspectives on Of-
Prevention Casebook: Securing High fending. New York: Springer-Verlag.
Crime Locations. Washington, DC: Criminal Justice Commission. 1998.
Crime Control Institute. Beenleigh Calls for Service Project:
Evaluation Report. Brisbane, Austra-
. 1993. Convincing the Recalci-
lia: Author.
trant: Reexamining the Minneapolis
Eck, John E. 1992. Alternative Futures
RECAP Experiment. Ph.D. diss.,
for Policing. In Police Innovation and
Rutgers University, NJ.
Control of the Police, ed. David L.
. 1994. The Problems of Problem Weisburd and Craig Uchida. New
Solving: Resistance, Interdepen- York: Springer-Verlag.
dencies, and Conflicting Interests.
. 1993. The Threat of Crime Dis-
American Journal of Police 13:1-36.
placement. Criminal Justice Abstracts
Caeti, Tory. 1999. Houstons Targeted 25:527-46.
Beat Program: A Quasi-Experimental . 1997. Preventing Crime at Places.
Test of Police Patrol Strategies. Ph.D. In Preventing Crime: What Works,
diss., Sam Houston State University, What Doesnt, Whats Promising, ed.
TX. Lawrence W. Sherman, Denise C.
Campbell, Donald T. and Julian Stanley. Gottfredson, Doris Layton MacKen-
1966. Experimental and Quasi- zie, John E. Eck, Peter Reuter, and
Experimental Designs for Research. Shawn D. Bushway. Washington, DC:
Chicago: Rand McNally. U.S. Department of Justice, National
Caulkins, Jonathan P., Richard C. Institute of Justice.
Larson, and Thomas F. Rich. 1993. Ge- Eck, John E. and Julie Wartell. 1996. Re-
ographys Impact on the Success of Fo- ducing Crime and Drug Dealing by
cused Local Drug Enforcement Opera- Improving Place Management: A Ran-
tions. Socio-Economic Planning domized Experiment (Final report to
Sciences 27:119-30. the San Diego Police Department).
Clarke, Ronald V., ed. 1992. Situational Washington, DC: Crime Control Insti-
Crime Prevention: Successful Case tute.
124 THE ANNALS OF THE AMERICAN ACADEMY
Eck, John E. and David L. Weisburd. trol. Washington, DC: Brookings Insti-
1995. Crime Places in Crime Theory. tution.
In Crime and Place, ed. John E. Eck Moore, Mark H. 1980. The Police and
and David L. Weisburd. Monsey, NY: Weapons Offenses. Annals of the
Criminal Justice Press. American Academy of Political and
Fagan, Jeffrey, Franklin E. Zimring, and Social Science 452:22-32.
June Kim. 1998. Declining Homicide Novak, Kenneth J., Jennifer Hartman,
in New York City: A Tale of Two Alexander Holsinger, and Michael
Trends. Journal of Criminal Law and Turner. 1999. The Effects of Aggres-
Criminology 88:1277-324. sive Policing of Disorder on Serious
Goldstein, Herman. 1990. Problem- Crime. Policing: An International
Oriented Policing. Philadelphia: Tem- Journal of Police Strategies and Man-
ple University Press. agement 22:171-90.
Green Mazerolle, Lorraine, James F. Petrosino, Anthony, Robert Boruch, Cath
Price, and Jan Roehl. 2000. Civil Rem- Rounding, Steve McDonald, and Iain
edies and Drug Control: A Random- Chalmers. In press. Assembling a So-
ized Field Trial in Oakland, Califor- cial, Psychological, Educational, and
nia. Evaluation Review 24:212-41. Criminological Trials Register
(SPECTR). Evaluation Research in
Greene, Judith A. 1999. Zero Tolerance: A
Education.
Case Study of Police Practices and Pol-
Pierce, Glenn L., Susan Spaar, and
icies in New York City. Crime & Delin-
LeBaron Briggs. 1988. The Character
quency 45:171-81.
of Police Work: Strategic and Tactical
Greenwood, Peter, Jan Chaiken, and Joan
Implications. Boston: Northeastern
Petersilia. 1977. The Investigation
University, Center for Applied Social
Process. Lexington, MA: Lexington
Research.
Books.
Reppetto, Thomas. 1976. Crime Preven-
Hesseling, Rene. 1994. Displacement: A tion and the Displacement Phenome-
Review of the Empirical Literature. In non. Crime & Delinquency 22:166-77.
Crime Prevention Studies. Vol. 3, ed. Shaw, James. 1995. Community Policing
Ronald V. Clarke. Monsey, NY: Crimi- Against Guns: Public Opinion of the
nal Justice Press. Kansas City Gun Experiment. Justice
Hope, Timothy. 1994. Problem-Oriented Quarterly 12:695-710.
Policing and Drug Market Locations: Sherman, Lawrence W. 1995. The Police.
Three Case Studies. In Crime Preven- In Crime, ed. James Q. Wilson and
tion Studies. Vol. 2, ed. Ronald V. Joan Petersilia. San Francisco: ICS
Clarke. Monsey, NY: Criminal Justice Press.
Press. . 1997. Policing for Crime Preven-
Kelling, George, Anthony Pate, Duane tion. In Preventing Crime: What
Dickman, and Charles Brown. 1974. Works, What Doesnt, Whats Prom-
The Kansas City Preventive Patrol Ex- ising, ed. Lawrence W. Sherman,
periment: A Technical Report. Wash- Denise C. Gottfredson, Doris Layton
ington, DC: Police Foundation. MacKenzie, John E. Eck, Peter Reuter,
Kleck, Gary. 1991. Point Blank: Guns and and Shawn D. Bushway. Washington,
Violence in America. New York: Aldine DC: U.S. Department of Justice, Na-
de Gruyter. tional Institute of Justice.
M c G a r r el l , Ed m und a n d S t e v e n Sherman, Lawrence W., Michael Buerger,
Chermak. 2000. Targeting Firearms and Patrick Gartin. 1989. Repeat Call
Violence Through Directed Police Pa- Address Policing: The Minneapolis
EFFECTS OF HOT SPOTS POLICING ON CRIME 125
THE ANNALS
EFFECTS OF CORRECTIONAL
OF THE AMERICAN
BOOT
ACADEMY
CAMPS ON OFFENDING
Effects of Correctional
Boot Camps on Offending
Doris Layton MacKenzie is the director of the Evaluation Research Group and a pro-
fessor in the Department of Criminology and Criminal Justice, University of Maryland.
Her research interests include evaluation of correctional programs, correctional alter-
natives, correctional boot camps, self-report criminal activities, and offender behavior.
David B. Wilson is an assistant professor of the administration of justice at George
Mason University. His research interests include program evaluation research method-
ology, meta-analysis, crime and general problem behavior prevention programs, and ju-
venile delinquency intervention effectiveness.
Suzanne B. Kider is a masters student in the Department of Criminology and Crimi-
nal Justice at the University of Maryland. Her main research interests include correc-
tions and probation and parole reentry.
126
EFFECTS OF CORRECTIONAL BOOT CAMPS ON OFFENDING 127
ivism data were reported dichoto- recidivism are reported in the follow-
mously across all studies and were ing analyses.
based on official records, generally
reflected as arrest, reconviction, or
RESULTS
reinstitutionalization. As such, the
natural index of effectiveness is the The distribution of recidivism
odds ratio (see Fleiss, 1994) and was effects across the 44 boot camp ver-
the index of effect (see below). The sus comparison group samples is
mean odds ratio and homogeneity of shown in Figure 1. Each row of this
effects across studies was computed forest plot represents a distinct sam-
using the inverse variance weight ple, identified by the label in the left
method. A random-effects model was column. The recidivism odds ratio
assumed, and the random-effects (effect size) is represented by the
variance component was estimated small diamonds, and the line spans
using the methods outlined by the 95 percent confidence interval
Dersimonian and Laird (1986) and around the odds ratio. The samples
Raudenbush (1994). The computa- are sorted with the largest positive
tions were performed using macros effect at the top and the smallest neg-
written by the second author that are ative effect (odds ratios between 1
available for use with SAS, SPSS, and 0) on the bottom. At the very bot-
and Stata (Lipsey and Wilson 2001). tom of the plot is the overall random-
A total of 155 recidivism effect effects mean odds ratio.
sizes were extracted from the stud- The effects across these studies
ies. Recidivism effects that reflected ranged from large reductions to large
technical violations only were increases in the risk of recidivating
excluded from the analyses reported for the boot camp participants rela-
below, reducing the set of effect sizes tive to the comparison groups. The
to 142. The recidivism effects were overall mean odds ratio was 1.02 (95
examined in two ways. First, multi- percent confidence interval of 0.90 to
ple recidivism effects from a single 1.17), indicating an almost equal
study and sample were averaged odds of recidivating between the boot
prior to analysis, producing a set of camp and comparison groups, on
44 recidivism effect sizes for the average. Thus there appears to be no
analysis. The second set of analyses relationship between program par-
used arrest as the measure of recidi- ticipation (boot camp or comparison)
vism if it was available; if not, and recidivism. The equivalent recid-
reconvictions were used as the mea- ivism rates for the average boot camp
sure, and if neither of these was and comparison group, given this
available, reinstitutionalizations overall odds ratio, would be 49.4 per-
were used. The results from the two cent for the boot camp and 50 percent
methods of measuring recidivism for the comparison condition. This is
were compared and did not yield any a small difference by most any stan-
substantive differences in the dard. Thus, overall, the evidence sug-
results. Therefore, results based on gests that boot camps do not reduce
the second method of measuring the risk of recidivism relative to
EFFECTS OF CORRECTIONAL BOOT CAMPS ON OFFENDING 131
FIGURE 1
A FOREST PLOT SHOWING THE RECIDIVISM ODDS RATIOS AND
95 PERCENT CONFIDENCE INTERVAL FOR EACH STUDY AND
SAMPLE AND THE OVERALL MEAN ODDS RATIO
other existing criminal justice sys- between the boot camp samples and
tem forms of punishment and reha- the comparisons.
bilitation. From the forest plot, it is The distribution of odds ratios was
also evident that 9 studies observed a highly heterogeneous, Q = 464.6, df =
statistically significant positive ben- 43, p < .0001, suggesting the pres-
efit of boot camps, whereas 8 studies ence of moderators of the effects,
observed a statistically significant either methodological or substantive,
positive benefit of the comparison such as the nature of the boot camp
condition. The remaining 27 studies program and comparison conditions
found no significant differences and the types of offenders served.
132 THE ANNALS OF THE AMERICAN ACADEMY
TABLE 1
CROSS-TABULATION OF QUALITATIVE METHODOLOGICAL
QUALITY SCORE AND OTHER METHOD DESCRIPTORS (N = 44)
ies had high attrition or excluded different from each other (for exam-
program dropouts from the recidi- ple, confidence interval includes 1).
vism analysis, creating a potential In contrast to studies that did not use
threat from selection bias. Thus statistical controls in the analysis of
there were no evaluations of the recidivism outcomes, studies that
effectiveness of boot camps that were used controls observed smaller
free from methodological blemishes. effects that were negative in direc-
That said, however, many of the stud- tion. Once again, neither category
ies (19 of 44, or 43 percent) were differed significantly from the null
judged to be methodologically solid hypothesis. All other methodological
(method score of 4). These studies variables were unrelated to the
were generally the higher-quality observed odds ratios.
quasi-experimental designs that
either carefully selected the compari- Offender characteristics
son group so as to maximize similar- across studies
ity with the boot camp group (for There was generally little infor-
example, selecting boot camp eligible mation regarding the characteristics
offenders and matching the groups of the offenders in the studies. For 11
on demographic characteristics) or of the 44 samples, the authors did not
used statistical controls in the analy- indicate the gender, although it is
sis of recidivism effects. Only 8 of the reasonable to assume that in these
44 evaluations (18 percent) were cases the samples were all male.
judged to be of poor methodological Only 3 of the 44 samples were all
quality. female, and the mean odds ratio for
To assess the robustness of the these samples was 1.06 and statisti-
general finding of no effect, a sepa- cally nonsignificant. This mean odds
rate mean odds ratio was computed ratio is roughly the same as that for
for each category of the different the overall sample. Four samples
methodological variables (see Table were mixed gender, although they
2). The mean effect size was slightly were predominantly male (equal to
lower for the studies judged to be of or greater than 80 percent). Thus
overall higher methodological qual- there are insufficient data to ade-
ity, although the trend was statisti- quately explore whether boot camps
cally nonsignificant. Studies that are differentially effective for males
used a prospective research design and females, as some theorists have
had observed larger positive effects hypothesized (Morash and Rucker
(although not significantly different 1990).
from a null odds ratio of 1) than did All samples were successfully
retrospective designs. That is, while classified as either juvenile or adult.
the mean odds ratio of prospective The adult samples were typically
and retrospective designs are signifi- young adults and in some cases
cantly different from each other, nei- included at least a small percentage
ther design produces an odds ratio of juveniles who were adjudicated as
that suggests that the experimental adults. As shown in Table 3, the mean
and control samples are significantly odds ratio for the studies evaluating
134 THE ANNALS OF THE AMERICAN ACADEMY
TABLE 2
MEAN ODDS RATIO AND 95 PERCENT CONFIDENCE
INTERVAL BY METHOD VARIABLES (N = 44)
95 Percent
Confidence Interval
Mean
Method Variable Odds Ratio Lower Upper ka
TABLE 3
MEAN ODDS RATIO AND 95 PERCENT CONFIDENCE
INTERVAL BY OFFENDER CHARACTERISTICS (N = 44)
95 Percent
Confidence Interval
Mean
Offender Characteristic Odds Ratio Lower Upper ka
TABLE 4
MEAN ODDS RATIO AND 95 PERCENT CONFIDENCE INTERVAL BY
PROGRAM CHARACTERISTICS (JUVENILES n = 16, ADULTS n = 28)
95 Percent
Confidence Interval
Program Characteristic Mean Odds Ratio Lower Upper ka
criminal activities of adults and juve- and (2) whether the participants
niles. Critics argue that the pro- were limited to those convicted
grams are poorly conceived as thera- or adjudicated for nonviolent/
peutic interventions, they will not nonperson crimes or mixed violent
reduce recidivism, and they may and nonviolent crimes. Again we
actually have the opposite effect by found no evidence that differences in
increasing criminal activities. Our these characteristics explained the
results do not support either side of differences in the results.
this argument because we found no We were able to code and analyze
differences in recidivism between the the possible impact of six program
44 boot camp samples and the com- characteristics, including whether
parisons. Correctional boot camps the boot camps had aftercare, aca-
are neither as good as the advocates demic education, vocational educa-
assert nor as bad as the critics tion, drug treatment, counseling, or
hypothesize. manual labor components. It is
An examination of the forest plot important to note that this informa-
of the individual studies (see Figure tion was limited to general informa-
1) and our analysis of the data dem- tion about the characteristics of the
onstrated large differences in the programs. We assume the quality
studies in terms of the effect of boot and intensity of the programs dif-
camps. Some studies found boot fered greatly. From our knowledge of
camp participants did better than the boot camps we know that some
the comparisons, and others found programs consider Narcotics Anony-
comparison samples did better. For mous or Alcoholics Anonymous meet-
this reason, we explored whether the ings drug treatment, whereas others
differences among studies could be provide a more intensive drug treat-
attributed to the methods or design ment experience using a Therapeutic
of the studies or to characteristics of Communitytype model. We did not
the programs or individual partici- have enough information to code
pants. In our examination of the such differences. Almost no informa-
methodological variables, we did not tion was given about what happened
find any evidence that differences in to the comparison samples. The
the results of studies could be potential impact of these differences
explained by the study methodology. on recidivism cannot be overlooked.
Our examination of the offender When we examined the impact of
characteristics was disappointing program characteristics, the only dif-
because very few studies reported ferences we found were for adult
sufficient information to enable us to studies and, after controlling for
code and analyze the possible impact methodological differences, the only
of these characteristics on study out- difference was for boot camps that
comes. Few studies even reported on included an aftercare component. In
the gender of the samples. The only other words, whereas the odds ratios
variables we could examine were (1) differed for boot camps with and
whether the studies focused on adult without aftercare, in neither case did
offenders or adjudicated juveniles, the boot camp samples differ
EFFECTS OF CORRECTIONAL BOOT CAMPS ON OFFENDING 139
*Boyles, Cecilia E., Eric Bokenkamp, and ment Boot Camp: A Follow-Up Study
William Madura. 1996. Evaluation of of the First Five Platoons. Tallahassee:
the Colorado Juvenile Regimented Florida Department of Juvenile Jus-
Training Program. Golden: Colorado tice, Bureau of Data and Research.
Department of Human Services, Divi- *. 1996b. Manatee County Sheriffs
sion of Youth Corrections. Boot Camp: A Follow-Up Study of the
*Burns, Jerald C. and Gennaro F. Vito. First Four Platoons. Tallahassee:
1995. An Impact Analysis of the Ala- Florida Department of Juvenile Jus-
bama Boot Camp Program. Federal tice, Bureau of Data and Research.
Probation 59:63-67. *. 1996c. Pinellas County Boot
*California Department of the Youth Au- Camp: A Follow-Up Study of the First
thority. 1997. LEAD: A Boot Camp and Five Platoons (research rep. no. 33).
Intensive Parole Program; the Final Tallahassee: Florida Department of
Impact Evaluation (Report to the Cali- Juvenile Justice, Bureau of Data and
fornia Legislature). Sacramento: Au- Research.
thor.
*. 1997a. Bay County Sheriffs Of-
*Camp, David A. and Harjit S. Sandhu.
fice Boot Camp: A Follow-Up Study of
1995. Evaluation of Female Offender
the First Seven Platoons (research rep.
Regimented Treatment Program
no. 44). Bay County: Florida Depart-
(FORT). Journal of the Oklahoma
ment of Juvenile Justice, Bureau of
Criminal Justice Research Consor-
Data and Research.
tium 2:50-57.
*. 1997b. Martin County Sheriffs
Clark, Cheryl L. and David W. Aziz. 1996.
Office Boot Camp: A Follow-Up of the
Shock Incarceration in New York
First Four Platoons (research rep. no.
State: Philosophy, Results, and Limi-
43). Martin County: Florida Depart-
tations. In Correctional Boot Camps:
ment of Juvenile Justice, Bureau of
A Tough Intermediate Sanction, ed.
Data and Research.
Doris Layton MacKenzie and Eugene
E. Hebert. Washington, DC: U.S. De- *. 1997c. Polk County Juvenile
partment of Justice, National Insti- Boot Camp: A Follow-Up Study of the
tute of Justice. First Four Platoons. Tallahassee:
Dersimonian, Rebecca and Nan Laird. Florida Department of Juvenile Jus-
1986. Meta-Analysis in Clinical tice, Bureau of Data and Research.
Trials. Controlled Clinical Trials *. 1997d. Polk County Juvenile
7:177-88. Boot CampFemale Program: A Fol-
*Farrington, David P., G. Hancock, M. low-Up Study of the First Seven Pla-
Livingston, Kate Painter, and G. Towl. toons. Polk County: Florida Depart-
2000. Evaluation of Intensive Regimes ment of Juvenile Justice, Bureau of
for Young Offenders (Home Office re- Data and Research.
search findings). London: Home Office *Flowers, Gerald T., Timothy S. Carr, and
Research, Development and Statistics R. Barry Ruback. 1991. Special Alter-
Directorate. native Incarceration Evaluation. At-
Fleiss, Joseph L. 1994. Measures of Effect lanta: Georgia Department of Correc-
Size for Categorical Data. In The tions.
Handbook of Research Synthesis, ed. Gendreau, Paul, Tracy Little, and Claire
Harris Cooper and Larry V. Hedges. E. Goggin. 1996. A Meta-Analysis of
New York: Russell Sage. the Predictors of Adult Offender Re-
*Florida Department of Juvenile Justice. cidivism: What Works! Criminology
1996a. Leon County Sheriffs Depart- 34:575-607.
142 THE ANNALS OF THE AMERICAN ACADEMY
Gendreau, Paul and Robert R. Ross. 1987. venting Crime: What Works, What
Revivication of Rehabilitation: Evi- Doesnt, Whats Promising, ed. Law-
dence from the 1980s. Justice Quar- r e n c e W. S h e r m a n , D e n i s e C.
terly 4:349-408. Gottfredson, Doris Layton MacKen-
Gowdy, Voncile B. 1996. Historical Per- zie, John Eck, Peter Reuter, and
spective. In Correctional Boot Camps: Shawn Bushway. Washington, DC:
A Tough Intermediate Sanction, ed. U.S. Department of Justice, National
Doris Layton MacKenzie and Eugene Institute of Justice.
E. Hebert. Washington, DC: U.S. De- MacKenzie, Doris Layton and Eugene E.
partment of Justice, National Insti- Hebert, eds. 1996. Correctional Boot
tute of Justice. Camps: A Tough Intermediate Sanc-
*Harer, Miles D. and Jody Klein-Saffran. tion. Washington, DC: U.S. Depart-
1996. An Evaluation of the Federal ment of Justice, National Institute of
Bureau of Prisons Lewisburg Inten- Justice.
sive Confinement Center. Unpub- MacKenzie, Doris Layton and Dale G.
lished manuscript, Federal Bureau of Parent. 1991. Shock Incarceration and
Prisons, Research and Evaluation, Prison Crowding in Louisiana. Jour-
Washington, DC. nal of Criminal Justice 19:225-37.
*Jones, Mark. 1996. Do Boot Camp Grad- MacKenzie, Doris L. and Dale G. Parent.
uates Make Better Probationers? 1992. Boot Camp Prisons for Young
Journal of Crime and Justice 19:1-14. Offenders. In Smart Sentencing: The
*. 1997. Is Less Better? Boot Camp, Emergence of Intermediate Sanctions,
Regular Probation and Rearrest in ed. James M. Byrne, Arthur J. Lurigio,
North Carolina. American Journal of and Joan Petersilia. Newbury Park,
Criminal Justice 21:147-61. CA: Sage.
*Jones, Robert J. 1998. Annual Report to MacKenzie, Doris Layton and Alex
the Governor and the General Assem- Piquero. 1994. The Impact of Shock In-
bly: Impact Incarceration Program. carceration Programs on Prison
Springfield: Illinois Department of Crowding. Crime & Delinquency
Corrections. 40:222-49.
*Kempinem, Cynthia A. and Megan C. MacKenzie, Doris Layton and James W.
Kurlychek. 2001. Pennsylvanias Moti- Shaw. 1990. Inmate Adjustment and
vational Boot Camp (2000 Report to Change During Shock Incarceration.
the Legislature). Quehanna: Pennsyl- Justice Quarterly 7:125-50.
vania Commission on Sentencing. *MacKenzie, Doris L. and Claire Souryal.
Lipsey, Mark. 1992. Juvenile Delin- 1994. Multi-Site Evaluation of Shock
quency Treatment: A Meta-Analytic Incarceration: Executive Summary.
Inquiry into the Variability of Effects. Washington, DC: U.S. Department of
In Meta-Analysis for Explanation: A Justice, National Institute of Justice.
Casebook, ed. Thomas Cook, Harris . 1995. Inmate Attitude Change
Cooper, David S. Cordray, Heidi During Incarceration: A Comparison
Hartmann, Larry V. Hedges, Richard of Boot Camp with Traditional Prison.
J. Light, Thomas A. Louis, and Freder- Justice Quarterly 12:325-54.
ick Mosteller. New York: Russell Sage. *MacKenzie, Doris L., Claire Souryal,
Lipsey, Mark W. and David B. Wilson. Miriam Sealock, and Mohammed Bin
2001. Practical Meta-Analysis. Thou- Kashem. 1997. Outcome Study of the
sand Oaks, CA: Sage. Sergeant Henry Johnson Youth Lead-
MacKenzie, Doris L. 1997. Criminal Jus- ership Academy (YLA). Washington,
tice and Crime Prevention. In Pre- DC: University of Maryland, U.S. De-
EFFECTS OF CORRECTIONAL BOOT CAMPS ON OFFENDING 143
Cognitive-Behavioral
Programs for Offenders
144
COGNITIVE-BEHAVIORAL PROGRAMS FOR OFFENDERS 145
category for Wilson, Allen, and Mac- convincingly tests this promising
Kenzie to closely examine this factor. program approach.
The only other meta-analysis of In this article, we present another
the effectiveness of CBT programs on meta-analysis of research on the
the reoffense rates of offenders of effects of cognitive-behavioral pro-
which we are aware is Pearson and grams on the reoffense rates of
his colleagues (no date). This meta- offenders. To provide a clear view of
analysis included 69 research stud- the best evidence available on the
ies but covered both behavioral (for effectiveness of this specific treat-
example, contingency contracting, ment modality, we have restricted
token economy) and cognitive-behav- the studies eligible for inclusion in
ioral programs. Pearson and his col- the analysis in four important ways.
leagues found that the cognitive- First, we have selected only studies
behavioral programs were more with experimental or strong quasi-
effective in reducing recidivism than experimental designs so that the
the behavioral ones were, with a most methodologically credible evi-
mean recidivism reduction for dence is represented. Second, to
treated groups of about 30 percent. better isolate the distinctive feature
Moreover studies of higher method- of CBT, we have adopted a relatively
ological quality showed the largest narrow definition of CBT that
effect sizes. The criteria for identify- requires the intervention to focus
centrally on cognitive change. Third,
ing cognitive-behavioral programs in
we have included only studies that
this meta-analysis were rather
applied CBT to samples of the gen-
broad, however. They included not
eral of f en der popu l at i on an d
only interventions directed specifi-
excluded those using samples of spe-
cally toward altering cognitions but
cialized offenders. While the issue of
also social skills training and prob-
the effectiveness of CBT for different
lem-solving programs for which cog-
types of offenders is important, the
nitive change was not the main focus.
broader issue is whether CBT can be
These two meta-analyses provide used with good results in routine cor-
strong indications of the effective- rectional practice with typical mixed
ness of CBT for reducing the recidi- samples. Last, we have focused the
vism of offenders. Both, however, analysis exclusively on reoffense
encompass considerable diversity recidivism as an outcome variable.
within their scope, including a range Though disciplinary infractions
of offender types, quality of study within a correctional institution,
design, and (especially in Pearson technical parole violations, and other
et al. no date) variations in what is such outcomes may be proxies
counted as a CBT. Against this back- for criminal behavior, the effects of
ground, it seems wise to confirm the CBT on subsequent criminal behav-
effectiveness of CBTs for offenders ior is best shown by results on
with an analysis of the set of avail- such direct measures as rearrest and
able studies that most directly and reconviction.
148 THE ANNALS OF THE AMERICAN ACADEMY
FIGURE 1
LOGGED ODDS RATIOS FOR RECIDIVISM EFFECTS
Kownacki 1995
Bottcher 1985
Shivrattan 1988
Finn 1998
Robinson 1994
Robinson 1995
Walters 1999
Grand Mean
ment, though about one-third were ivism among treated offenders rela-
measured more than a year after tive to the odds among offenders in
treatment was completed. the control group, where the odds are
defined as the number recidivating
Intervention effects di v i ded by t h e n u mber n ot
recidivating. Thus an odds ratio of
Figure 1 shows the logged odds .50 means that the ratio of recidivists
ratios and their confidence intervals to nonrecidivists among treated
for the treatment versus control offenders was half that of the control
group differences in reoffense recidi- group. The logged odds ratio is more
vism across the 14 studies. The odds convenient for statistical computa-
ratio represents the odds of recid- tions and takes negative values
152 THE ANNALS OF THE AMERICAN ACADEMY
when the treatment group recid- the control groups (that is, 26 percent
ivates less than the control does, pos- versus 38 percent recidivating in the
itive values when the control respective groups). Thus only about
recidivates less than the treatment two-thirds as many offenders receiv-
group does, and zero when recidivism ing CBT recidivated as did offenders
rates are the same for both groups. in the control groups.
Figure 1 shows that with one The grand mean odds ratio and
slight exception, the odds ratios from treatment versus control recidivism
these 14 studies all indicate lower rates, however, do not take into
recidivism for the offenders receiving account the variation in treatment
CBT than for those in the control effects across the 14 studies. A Q test
groups with whom they were com- of the homogeneity of the logged odds
pared. As the confidence intervals ratios (Lipsey and Wilson 2001;
reveal, however, only a few of these Shadish and Haddock 1994) showed
odds ratios are statistically signifi- that even with only 14 studies, the
cant (that is, the confidence interval variation in findings was statisti-
around the logged odds ratio does not cally significant. Adjusting the
1
include zero). The weighted mean weighting for the one large sample
odds ratio across all 14 studies, com- reduced the heterogeneity but still
puted as .66 (log = .42), however, left it marginally significant (p = .07).
takes advantage of the statistical Before drawing any conclusions
power of the full set of studies and is about the overall effectiveness of
statistically significant (p < .05). CBT from these studies, therefore,
Of the 14 studies, 1 (Robinson some investigation is warranted of
1995) had an exceptionally large the potential sources of this variabil-
sample size (more than 2000; next ity in findings across studies.
largest was less than 300) and thus One potentially important moder-
exercised a disproportionate influ- ator of effect sizes is the type of
ence on the weighted mean odds research design. Although only
ratio. With the weight on the odds the strongest quasi-experimental
ratio from that study scaled back to designs were eligible for inclusion in
match the next highest, the weighted the meta-analysis, they still might be
mean across studies was .55 (log = sufficiently biased to yield findings
.59), which gives a more generally systematically different from those
representative value for this collec- resulting from randomized experi-
tion of study findings. This value ments. This did not prove to be the
indicates that on average, the odds of case, however. The mean odds ratio
recidivating for offenders receiving from the six nonrandomized designs
CBT programs were almost half (.62) was not significantly different,
those for offenders in the control Q(1) = .75, p = .39, from the mean for
groups who did not receive CBT. the eight randomized designs (.50).
More specifically, the weighted mean Of the remaining coded variables
recidivism rate for the intervention dealing with important study char-
groups was .26 compared with .38 for acteristics, exploratory analysis
COGNITIVE-BEHAVIORAL PROGRAMS FOR OFFENDERS 153
TABLE 2
CHARACTERISTICS OF STUDIES WITH DIFFERENT
MAGNITUDES OF RECIDIVISM EFFECTS
Studies showing largest effects (n = 4); Mean odds ratio = .16 (log = 1.85);
Weighted mean recidivism rate; Treatment groups = .15; Control groups = .52
Demonstration 4 Juveniles 2 Posttest < 6 months 4 Published before 1995 3
Practical 0 Adults 2 Posttest 6 months 0 Published in or after 1995 1
Studies showing in-between effects (n = 4); Mean odds ratio = .41 (log = .89);
Weighted mean recidivism rate; Treatment groups = .37; Control groups = .53
Demonstration 3 Juveniles 4 Posttest < 6 months 1 Published before 1995 4
Practical 1 Adults 0 Posttest > 6 months 3 Published in or after 1995 0
Studies showing smallest effects (N = 6); Mean odds ratio = .76 (log = .28);
Weighted mean recidivism rate; Treatment groups = .26; Control groups = .31
Demonstration 0 Juveniles 1 Posttest < 6 months 1 Published before 1995 1
Practical 6 Adults 5 Posttest > 6 months 5 Published in or after 1995 5
THEEVIDENCE-BASED
AN ANNALS OF THE AMERICAN
APPROACHACADEMY
TO PREVENTING CRIME
Toward an Evidence-Based
Approach to Preventing Crime
158
AN EVIDENCE-BASED APPROACH TO PREVENTING CRIME 159
excluded) and the need to reconcile tioners, policy makers, and the gen-
differences in coding of study charac- eral public.
teristics and outcomes by multiple A meta-analysis involves the sta-
researchers (that is, interrater tistical or quantitative analysis of
reliability). the results of prior research studies.
As noted in the article by David Since it involves the statistical sum-
Farrington and Anthony Petrosino mary of data (for example, effect
(2001 [this issue]), in systematic sizes), it requires a reasonable num-
reviews, quantitative techniques ber of intervention studies that are
are used, when appropriate and pos- sufficiently similar to be grouped
sible, in analyzing results (37). The together. For example, there may be
use of quantitative techniques such little point in reporting a mean effect
as meta-analysis may not be suitable size based on a very small number of
due to a small number of studies, het- studies. Nevertheless, quantitative
erogeneity across studies, and differ- methods can be very important in
ent units of analysis of the studies helping the reviewer determine the
(that is, a mix of area- and individual- average effect of a particular
based studies). Another instance in intervention.
which a meta-analysis does not equal David Wilsons (2001) article
a systematic review is when the for- reviews meta-analysis in the context
mer fails to conform to the methodol- of criminological interventions. He
ogy of systematic review, for example, identifies a number of strengths and
using biased search strategies to limitations in the use of meta-analy-
locate relevant studies. (For other sit- sis. Among the strengths, he lists its
uations in which a meta-analysis transparent naturethe explication
of its methods and the studies
should not be performed, see David
involvedwhich makes it easily rep-
Wilsons [2001] article.)
licated by other researchers and its
High-quality systematic reviews ability to handle a very large number
are the core ingredient of the newly of studies that may be overwhelming
created Campbell Collaboration. for other review methods, and he
Named after the influential experi- notes that the statistical methods of
mental psychologist Donald T. Camp- meta-analysis help guard against
bell, the Campbell Collaboration was interpreting the dispersion in results
set up for the purpose of preparing, as meaningful when it can just as
maintaining, and disseminating evi- easily be explained as sampling
dence-based research on the effects error (84). Limitations of meta-anal-
of interventions in the three fields of ysis include, on a practical side, its
education, social welfare, and crime time-consuming nature and its in-
and justice. The Campbell Collabora- ability to synthesize complex pat-
tion Crime and Justice Group aims to terns of effects often found in individ-
prepare and maintain systematic ual studies (Wilson 2001, 84).
reviews of criminological interven- Overall, as David Wilson notes,
tions and to make them accessible meta-analysis provides a defensible
electronically to scholars, practi- strategy for summarizing crime
AN EVIDENCE-BASED APPROACH TO PREVENTING CRIME 163
Both medical and social science used for including (and excluding)
(criminal justice included) litera- studies is perhaps the most impor-
tures on the subject of research de- tant and controversial (42) in con-
sign affecting study outcomes are ducting systematic reviews. How
mixed, with some systematic reviews high to set the bar of methodological
reporting that nonrandomized stud- rigor as part of a review of the litera-
ies produce results with significantly ture, systematic review method or
larger effect sizes than randomized otherwise, is a question that all
studies while other reviews report researchers face. (For a brief discus-
the opposite. Using a sample of crimi- sion of this issue in the context of the
nal justice studies with outcome vote-count review method, see the
measures of crime (N = 308) from the note 1 and MacKenzie 2000.) We sup-
University of Maryland report port the inclusion of intervention
(Sherman et al. 1997), Weisburd, studies that use experimental or
Lum, and Petrosino (2001) find a quasi-experimental designs in sys-
moderate inverse relationship be- tematic reviews. Systematic reviews
tween the quality of the research de- of criminological interventions con-
sign, defined in terms of internal va- ducted under the name of the Camp-
lidity, and the outcomes reported in a bell Collaboration should, however,
study (64). A number of different set as a minimum for inclusion stud-
analyses (for example, highest-qual- ies with before and after measures in
ity nonrandomized versus random- comparable experimental and con-
ized experimental studies, excluding trol conditions, together with statis-
quasi-experimental studies) confirm tical control of extraneous variables.
the results. The main implication of Of course, for certain intervention
this finding, despite the authors not- modalities like multisystematic
ing that their work is preliminary therapy for chronic and violent
and hence that the findings should be young offenders and cognitive-
interpreted with caution, is that the behavioral skills training for offend-
findings of systematic reviews may ers, it will be possible for researchers
be bias ed by the inclus i on of to use only randomized experiments
nonrandomized studies. More re- because of the comparatively large
search is needed on this subject, ex- number of experimental evaluations
amining different intervention mo- in these areas.
dalities (for example, paren t M ore h i g h - qu al i t y res earch
training, hot spots policing) and set- designs of this sort and, of course,
tings in which crime prevention randomized experiments are needed
takes place (for example, families, in criminology. Experiments and
communities), as well as examining quasi-experiments should have large
if research design affects study out- samples, long follow-up periods, and
comes differently for individual- and follow-up interviews (Farrington
area-level studies. 1999). As noted above, sample size
In their article, Farrington and is particularly important for both
Petrosino (2001) note that the crite- individual- and area-based studies.
rion of methodological quality that is Long-term follow-ups are needed to
166 THE ANNALS OF THE AMERICAN ACADEMY
assess how long effects persist after here, of course, is fostering high-
the intervention ends. This informa- quality researchusing experimen-
tion may point to the need for booster tal and quasi-experimental evalua-
sessions. Long follow-ups are a rarity tion designson the effects of inter-
in criminological interventions and ventions. Another important process,
should be a top priority of funding likely (but disappointingly) to be
agencies. even more important than the
Research is needed to identify the research product, which we take up
active ingredients of successful (and in the next section, is the political
promising) crime prevention pro- and policy considerations that are
grams (Farrington 2000). Many pro- made about what evidence gets used
grams are multimodal, making it dif- and what does not.
ficult to isolate the independent Systematic reviews are the most
effects of different components. comprehensive method to assess the
Future experiments that attempt to effectiveness of crime prevention
disentangle the effects of different measures and, in an evidence-based
elements of the most successful pro- society, they would be the source that
grams are needed. Although not a governments would turn to for help
specific focus of this special issue, it is in the development of policy. This
also important that programs issue of The Annals reports on four
include, as part of the original systematic reviews of different crimi-
research design, provision for an eco- nological interventions. Each follows
nomic analysiseither a cost-benefit as closely as possible the methodol-
or cost-effectiveness analysisto ogy for conducting systematic
allow for an assessment of the eco- reviews as advocated by the Camp-
nomic efficiency of the program (see bell Collaboration, although it is
Welsh and Farrington 2000; Farring- important to note that none of these
ton, Petrosino, and Welsh 2001; reviews (or their research protocols)
Welsh, Farrington, and Sherman has yet been approved by the Camp-
2001). bell Collaboration. In this part we
briefly summarize the main findings
and discuss policy implications of
USING WHAT WORKS AND
STOPPING USING WHAT DOES these four systematic reviews.
NOT AND WHAT IS HARMFUL In the systematic review of parent
training and support before age 3 by
In an evidence-based society, gov- Odette Bernazzani, Catherine Ct,
ernment crime prevention policy and and Richard Tremblay (2001 [this
local practice would be based on issue]), seven studies were included,
interventions with demonstrated and effectiveness was assessed on
effectiveness in preventing crime the outcome measures of disruptive
and offending. Equally important, behavior (for example, opposition to
governments would put an end to adults, truancy, aggression) and
those interventions that do not work delinquency. The authors found that
and, more important, to those that three of the studies reported some
are harmful or iatrogenic. The key beneficial effects on disruptive
AN EVIDENCE-BASED APPROACH TO PREVENTING CRIME 167
THE ANNALS
BOOK DEPARTMENT
OF THE AMERICAN ACADEMY
174
BOOK DEPARTMENT 175
itary regimes, are probably the future Amendment would have provided for the
models. United States. This was in 1982; the rele-
Brookers book is a well-documented, vant sections of the Canadian Charter of
sobering analysis of the oldest and most Rights would have been significantly
common form of government: non- weaker without the intervention of
democracy. His book also has important women activists. A second attempt at
foreign policy implications for the United constitutional revision, the decentraliz-
States in its relationships with a plethora ing, Quebec-focused Meech Lake Accord,
of less than democratic systems around failed to achieve the necessary provincial
the world. This is an excellent book for ratification in 1987, partly because of op-
upper-division or graduate courses in po- position by womens groups. Then, in
litical science. 1992, a national referendum rejected a
PAUL C. SONDROL packet of constitutional amendments
University of Colorado that would, among other things, have
Colorado Springs subordinated womens rights to commu-
nal claims. Canadas largest national
womens organization, the National Ac-
tion Committee on the Status of Women,
was conspicuous among the many oppo-
DOBROWOLSKY, ALEXANDRA. 2000. nents of the Charlottetown Accord; other
Th e Po l i t i c s o f P r a g m a t i s m . women and womens groups were among
Women, Representation, and Constitu- the prominent supporters.
tionalism in Canada. Pp. xii, 320. Don The role of women activists in the first
Mills, Canada: Oxford University of these episodes has already been stud-
Press. Paperbound, Canada$24.95. ied at some length, but Dobrowolsky has
The Politics of Pragmatism recounts been able to add previously unpublished
the activism of national womens groups material from interviews and private ar-
in Canada as they attempted to influence chives. Womens political responses to the
the series of constitutional changes pro- Meech Lake Accord have had almost no
posed in the years 1980 to 1992. These attention, and their organized involve-
proposals, which first fascinated and ment with the campaign to defeat the
then appalled even the mavens of Cana- Charlottetown Accord has had little
dian politics, are hardly familiar to any- more, so most of the relevant, detailed
one else. Even fewer are aware of the role discussion in this book is new. Dobro-
played by women and womens groups in wolsky shows that as we might expect,
the process. As Alexandra Dobrowolsky there has been continuing attention to
is able to demonstrate, in Canada as else- the constitutional process from national
where, womens groups are now active, womens groups. She also interestingly
continuously, in all forms of conventional demonstrates how media and politicians
and unconventional politics. If anything, have lumbered womens activism with la-
the books focus on national groups (as bels that would delegitimize its political
well as the identity of those interviewed) role: the womens movement is only a spe-
understates the amount and variety of cial interest group (therefore elitist and
feminist activism across Canada. unrepresentative) or a social movement
Dobrowolskys topic includes one of (therefore with no political or policy
the icons of contemporary Canadian fem- standing).
inism, the successful enshrinement in As its title and subtitle suggest, The
the Canadian constitution of equality Politics of Pragmatism is inclined to draw
rights superior to what the Equal Rights rather sweeping theoretical conclusions
176 THE ANNALS OF THE AMERICAN ACADEMY
from its case study material. Certainly, a as global forces and internal challenges
more experienced scholar would not be so produce competing claims for entitle-
surprised to learn of excellent efforts by ments, recognition, and resources. Few
womens groupsand less excellent re- states are exempt from such tensions.
sults. Nor would she feel obliged to refute However, crises related to sovereignty
the many monocausal theories that ab- pose a special significance for nations
stract from the complex realities of politi- that are struggling to acknowledge and
cal action. For example, it is hardly redefine their relationships with indige-
necessary to prove that womens organi- nous populations. Augie Fleras and Paul
zations in Canada have, in spite of mis- Spoonley, in an insightful and provoca-
givings, worked with and through tive recent book, highlight critical dimen-
political parties. In addition, the knowl- sions of these challenges for New Zealand
edgeable reader will look at The Politics (Aotearoa) and its relations with Maori.
of Pragmatism and remember what the Fleras and Spoonley present a con-
protagonists are like as individuals. Tact vincing case for a reconstituted vision of
and self-protection together probably dic- New Zealand through multiculturalism
tated the absence of such a personal di- with a bi-national framework, which ac-
mension, but it is important for the study knowledges in substantive ways the need
of politics in a large country with a small to create effective spheres of Maori self-
population of activists. government within a nation constituted
Feminism is not extinct, and the by increasing diversity. While their posi-
womens movement has not been ren- tion is rooted in an understanding of de-
dered obsolete; these facts make a differ-
tails specific to the history, composition,
ence for mainstream politics. This book is
and dynamics of New Zealand society,
therefore to be recommended to those in-
their analysis carries implications that
terested in constitutional politics as well
extend even beyond the most obvious ref-
as that more obvious audience whose con-
erents to comparable societies like the
cern is the involvement of women with
United States, Canada, and Australia in
politics.
which recent changes in immigration
NAOMI BLACK
patterns are layered on a white settler co-
York University lonialism imposed over long-standing
Toronto forms of social organization among indig-
Canada enous peoples.
Mount Saint Vincent University Fleras and Spoonley begin with a re-
Halifax thinking of New Zealand history from a
Canada Maori perspective to highlight contend-
ing visions that accompany efforts to re-
alize a meaningful postcolonial state.
They do not define the issue by whether
there is any legitimacy to claims to
FLERAS, AUGIE and PAUL SPOON-
indigenous self-determination, since
LEY. 1999. Recalling Aotearoa: Indig-
indigenous rights are recognized and
enous Politics and Ethnic Relations in
guaranteed in a legal and historical
New Zealand. Pp. xiv, 288. Auckland,
framework; they define an issue of dis-
NZ: Oxford University Press. $29.95.
agreement over the nature, scope, and
The meaning and nature of nation- specific manifestations of Maori sover-
states, national identity, and social cohe- eignty in relation to the sovereignty of
sion are becoming increasingly contested the nation-state. The British crowns rec-
BOOK DEPARTMENT 177
The first thing that should be noted Bukovina, the Donbass, and much stud-
about this book is that it is primarily his- ied Crimea and Galicia). There is slightly
torical and descriptive. Despite the in- more than one page on Chechnya and six
triguing conceptualization that is pages on Tatarstan. Even a specialist on
contained in the title, University of Oslo the Baltic region would appreciate
professor Pl Kolst is primarily con- Kolsts study of the Latgale province of
cerned with the background conditions Latvia.
that gave former Soviet republics the his- Another refreshing feature of this
toric opportunity to become nation build- book is its Scandinavian perspective.
ers. His survey is more far reaching than Whether it is invoking the pioneering
comparable studies published in recent work of Norwegian Stein Rokkan, citing
years. The countries examined include the research results of not-well-known
Russia (the Russian diasporas found in Danish and Swedish academics, or draw-
the other new independent states are ing comparisons to the unique status of
treated in a separate chapter), Latvia, Finlands land islands, the North Amer-
Ukr a i ne, Bel a r us, Mol d o v a , a n d ican reader will be enriched by exposure
Kazakhstan. A decade after Soviet disin- to new ideas. Indeed Kolsts work is just
tegration, few authors are brave enough one example of the high-quality scholar-
to draw comparisons between so many ship produced in recent years by Scandi-
and such diverse former Soviet republics. navian academics studying former Soviet
To be sure, Kolst is writing for under- republics. The anecdotal information
graduate students and a general reader- that Kolst relates on national airlines,
ship. As an introductory text, this book postage stamps, soccer matches, and
has much to recommend it. Instead of Olympic athletes in the new states
simply a sequence of case studies of select should not be treated as trivia. Rather
new states, it begins with useful over- they confirm his extensive knowledge
views of such topics as nation building, about the societies he writes about. More-
construction of historiography, and reli- over he demonstrates the degree to which
gious pathways. Certainly a more robust nation building has been micromanaged
chapter on theorizing about nation build- in these new countries.
ing and social integration would be desir- The last chapter of Political Construc-
able. As it is, the attention given to a tion Sites attempts to test certain theo-
handful of experts (Walker Connor, Karl r i e s o f n a t i o n b u i l d i n g ( s u ch a s
Deutsch, Ernest Gellner, and Arend consociationalism and ethnic democracy)
Lijphart are the featured scholars) does in light of Kolsts empirical findings
not do credit to this exceptionally fertile about the new states. Kolst concludes,
field. rightly, that not all of them will be suc-
Kolsts major achievement is to alert cessful in nation building. But he infers,
the reader to the diversity between and confusingly, that there is good reason to
within the new states. His accounts of mi- expect that at least some of them will join
norities in the new countries are particu- the international community of nation-
larly informative and also idiosyncratic. states. Even if they may qualify only as
He piques the interest of the reader with quasi-states today, that does not mean
references to the ethnic complexities of they are doomed to remain in this cate-
Moldova (for example, the recently estab- gory forever. The contemporary interna-
lished Gagauz Yeri territory) and of tional system already comprises quasi
Ukraine (Transcarpathia, Northern states from several continents. This unof-
BOOK DEPARTMENT 181
ficial status does not affect their legal sta- tional developments, and ultimate suc-
tus. Nor do quasi states consist only of cesses and failures are conditioned. To
failed nation builders. describe these networks, and to trace
RAY TARAS their ongoing impacts, they draw on the
abundant secondary literature on Chil-
Tulane University
New Orleans ean parties and politics and on open-
Louisiana ended interviews of leaders, militants,
and members of both parties that they
conducted during the 1980s and early
1990s. On these bases, they describe in of-
ten rich detail the origins, early develop-
LOMNITZ, LARISSA ADLER and ANA ment, and coming of age of these parties
MELNICK. 2000. Chiles Political Cul- and nicely capture the flavor of their re-
ture and Parties, An Anthropological spective projects and their contributions
Explanation. Pp. ix, 161. Notre Dame, to pre-1973 Chilean politics, although in
IN, University of Notre Dame Press. each instance they largely reinforce the
$26.00. Paperbound, $14.00. images or portraits that the existing lit-
erature provides of these parties.
This is an English-language version of
It is not clear to me, however, that
a study published in Chile in 1998 by the
Lomnitz and Melnick really test their ap-
Fondo de Cultura Economica. The au-
proach or show its superior explanatory
thors, one a naturalized Mexican citizen
power (vis--vis alternative perspectives)
who lived in Chile for many years, and
in dealing with either Chilean parties
the other a Chilean, both teach at
generally or their evolution beyond 1973.
Universidad Nacional Autnoma de
They argue that contrasting Radical and
Mxico but frequently work on Chilean
Christian Democratic identities have en-
topics. In this essay/study, they analyze
dured over the years thanks to the con-
two political parties, the avowedly secu-
lar and entirely middle-class Radical tinuing vitality of underlying social
party, and the multiclass and largely networks, the ongoing appeal of values
Catholic Christian Democratic party, and symbols associated with their re-
from the vantage points of their respec- spective subcultures, and their success-
tive political cultures. In their view, polit- ful and not so successful electoral and
ical parties cannot be understood political experiences. They further con-
independent of the subjective attitudes of tend that horizontally linked networks
their leaders, militants, and rank-and- continue to provide the basis for internal
file members and supporters. But they do tendencies and factions within the par-
not believe that such attitudes are ade- ties and that they enable party members
quately captured by survey data alone to restrain or hold their party leaders
and opt for studying the formative expe- more accountable in Chile than in coun-
riences and relationships that in their tries like Mexico. But I do not believe that
view give them meaning and legitimacy. Lomnitz and Melnick look carefully
Lomnitz and Melnick view party cul- enough at either the distant or the not-so-
tures as emerging from formative social distant past to establish connections
networks (which are clubs, regions, among social networks, party tendencies,
schools, and family settings) to which and party life. Although claiming to ex-
leaders and activists are initially at- plain key party choices (of candidates, al-
tracted or exposed, and by which their liances, or platforms) at crucial junc-
subsequent initiatives, decisions, fac- tures, they do not look closely enough at
182 THE ANNALS OF THE AMERICAN ACADEMY
any of these events, at the changing con- perous, socially progressive, and Euro-
texts in which they take place, or at the peanized republics in Latin America.
other factors that may or may not have During the 1970s, these characteristics
played important roles. And it is there- were supplemented by an unflattering
fore difficult to compare their assertions one: all three were governed by authori-
or conclusions with those reached by tarian regimes that exercised repression
other analysts. as a preferred tool of political control and
In fact, aside from asserting that fac- social intimidation.
tions or tendencies do exist and are based This book recounts the events leading
on underlying social networks, Lomnitz to the installation of dictatorial rule and
and Melnick tend to treat these parties as the subsequent use of repression against
monoliths. In the case of the PDC, for ex- political activists and dissenters. When
ample, Lomnitz and Melnick emphasize authoritarian rule ended (1983 in Argen-
its strong identification with the progres- tina, 1985 in Uruguay, and 1989 in Chile)
sive social Christian tradition but fail to these nations entered a period of democ-
note that both the meaning and content ratization during which societies and po-
of this social Christian tradition were litical elites had to deal with the
vigorously debated from the late 1940s on restoration of civil government and si-
(when a second generation of activists en- multaneously engage in a critical exami-
tered the Falange), and rival internal ten- nation of what occurred in the past. To
dencies have attempted to fuse social different degrees, in each of these coun-
Christianity with things as dissimilar as tries the politics of reconciliation had to
neocapitalist and development (Frei balance the search for truth and justice
Montalva), purist independence from with the urgent need to establish viable
both Left and Right (Castillo), and Chris- democratic regimes. Public agendas had
tian socialism in alliance with the tradi- to incorporate a necessary commitment
tional left (Silva, Chonchol, and others) toward human rights protection, but for
as early as the PDCs 1959 founding con- the sake of keeping the state functioning,
gress. a compromise had to be established be-
MICHAEL FLEET tween memory and oblivion. Obviously,
Marquette University this threw into question the primacy that
Milwaukee the human rights discourse held in the
Wisconsin public affairs and private lives of individ-
uals. Not dealing with the crimes of the
past and allowing for impunity disturbed
the sensitivities of many who held that
collective morality was an integral part of
RO NI G ER, L UI S a nd M A R I O
the national soul. On the other hand, the
SZNAJDER. 1999. The Legacy of Hu-
majority sentiment was that indulging in
man Rights Violations in the Southern
recriminations or fueling revenges did
Cone, Argentina, Chile and Uruguay.
little to appease the national social con-
Oxford Studies in Democratization.
science.
Pp. xvi, 367. Oxford, UK: Oxford Uni-
Although all three countries share
versity Press. $82.00.
many common sociocultural traits, their
Argentina, Chile, and Uruguay, the historical and political paths diverge con-
three temperate countries of South siderably, so it is erroneous to assume
America, were considered the most pros- that repression and expiation ran similar
BOOK DEPARTMENT 183
courses in each of them. Argentina was tation of the sad events that occurred in
shaken most by both processes, as lack of the recent history of Argentina, Chile,
accountability and an attitude of disre- and Uruguay. The book is structured in a
spect for the powerless nurtured a defi- way that permits a clear understanding
ance and cynicism that made recognition of the issues involved, and the abundant
of past atrocities more difficult and con- background information substantiates
trition less frequent. Consequently, the every aspect of the redemocratization
healing process has taken longer and has and reconciliation processes. Moreover,
been more bitter than in its sister repub- in the discussion of the transgressions
lics. Uruguays use of violent repression committed, Roniger and Sznajder invoke
during the dictatorship years (1972- enlightening interpretations that come
1985) was less encompassing, and the from current theories explaining authori-
subsequent reconciliation was carried tarianism, political abuse, and violations
out in a spirit of compromise. The concil- of civil rights. Such frameworks allow for
iatory and pragmatic character of Uru- a theory-informed view of the issues at
guayans permitted the resumption of play in the Southern Cone during these
political dialogue, and redemocratization troubled years. In this light, the abuses
proceeded without the antagonistic tones do not appear as unique South American
found in Argentina. Developments in configurations but figure as recurrent
Chile followed a different course from themes in contemporary history from
those of the Platine countries. The dicta- which, it is hoped, general lessons can be
torship of General Augusto Pinochet was drawn.
accompanied by an economic develop- CESAR N. CAVIEDES
ment not mirrored in the other two coun- University of Florida
tries, which contributed to a greater Gainesville
acceptance of military rule by the rich
and middle-class segments. The poor and
the Left suffered so deeply that their for-
giveness of authoritarian abuses has
been more difficult to achieve. The legal-
SWAINE, MICHAEL D. and ASHLEY J.
istic character of Chileans and their
TELLIS. 2000. Interpreting Chinas
belief that institutions and public actors
Grand Strategy: Past, Present and Fu-
are accountable for their actions ex-
ture. Pp. xx, 283. Santa Monica, CA:
plains the ongoing search for justice and
RAND. $35.00. Paperbound, $20.00.
punishment.
One can dispute the propriety of using Of the many large unknowns facing
the term legacy of human rights viola- the United States in the new century,
tions, referenced in the title, when refer- China is perhaps the most tangible,
ring to the postauthoritarian political though not the most predictable. Inter-
events in Argentina, Chile, and Uruguay. preting Chinas Grand Strategy is a com-
Even Roniger and Sznajder find difficul- prehensive attempt by two senior RAND
ties in specifying what is meant by this Corporation analysts to fit Chinapast,
expression that certainly represents an present, and futureinto a general pat-
unpleasant bequest to the countries in tern of predictability. The result is not
question. Despite the improper use of the convincing, but it is the best-argued pes-
term legacy, this work is a well-docu- simistic estimate of Chinas security tra-
mented account and a credible interpre- jectory.
184 THE ANNALS OF THE AMERICAN ACADEMY
Swaine and Tellis take structural re- through catastrophe or internal dissen-
alism as their analytic approach, arguing sion is considered and dismissed. The
that China, like other nations, wants to possibility that the global context of a fu-
maximize its interests and power, and ture powerful China will constrain it to
therefore its strategic posture is set by its be cooperative is considered and rebutted
situation of relative power. Since power is in some detail. We are left with an asser-
ultimately a function of a states capabil- tive China, one whose interests are more
ity to coerce other states, the argument concentrated on its periphery but whose
is relevant to current American policy to- memories of imperial glory and confi-
ward China, and since Chinas capacity is dence of new strength lead it to view the
likely to continue to grow in both absolute existing global order as an old one.
and relative terms, it will be even more Any book of this scope requires gener-
critical for the United States after 2020. alizations and judgments that can be dis-
Despite their expectation that China will puted, but two deeper questions should
become more assertive as it approaches be raised. First, the book is not an inter-
great power capacity, Swaine and Tellis pretation of Chinas grand strategy, if
recommend a long-term policy of engage- strategy involves intent, and by interpre-
ment rather than one of preemptive con- tation we mean the effort to understand
tainment. anothers intent. In contrast to such clas-
The book begins with a lengthy discus- sics by RAND alumni as Allen Whitings
sion of Chinas history, arguing that Chinas Calculus of Deterrence or Mel
China has not been reluctant to use vio- Gurtovs China and Southeast Asia,
lence in the past. Chinas wars were usu- there is little careful analysis of Chinese
ally on its periphery, and they were often diplomacy here. Indeed, the five princi-
limited and defensive, but there were ples of peaceful coexistence, Chinas ma-
many of them. The second major part of jor diplomatic mantra since 1954, is not
the book concerns Chinas strategic pos- mentioned. The book argues that China
ture since Deng Xiaopings reform era be- is not an exception from the predictable
gan in 1979. Swaine and Tellis describe a behavior of power-sensitive states, and
foreign policy that is based on economic therefore its grand strategy is a question
progress, requires peaceful international of capacity, not intent.
relations, and does not unduly concen- A second problem is that while the
trate resources on the military. This book is well researched and presents
peaceful posture is explained as a enough data to make its argument plau-
calculative strategy, one based on the sible, China contains sufficient ambigu-
fact that China is currently weak as a ous and contradictory data that a
modern military power and therefore its contrary argument could also be made
current purposes are best served by non- plausible. In such a situation an analyst
military means. Moreover, economic must be careful not to ride a tangent too
growth is a necessary prerequisite of far. For example, Swaine and Tellis say
global military prominence, so Chinas that if all the territories claimed, occu-
shortest path to great power status is pied, or directly controlled by China since
through peaceful expansion of its econ- its unification in the 3rd century BC were
omy for the next two decades. matched against its current physical
The final third of the book discusses holdings, the presently disputed territo-
possible future trajectories for Chinas ries would fade into insignificance. True,
security strategy. The possibility that but not true enough. After all, cozily
China will eliminate itself as contender within the same time frame the same
BOOK DEPARTMENT 185
termination camps of Poland in the their hurt and humiliation, the public
1940s, Hitler told them to kill men, turned to old myths from German cul-
women, and children without pitytheir ture. Such myths helped them to explain,
acts would be forgotten. Who remembers for example, their indigestible defeat in
now the massacres of the Armenians? he World War I and their having been
sneered. There is a chilling similarity be- stabbed in the back by evil doers, that is,
tween Hitlers words and Stalins com- Jews. The German people reactivated a
ment while signing death warrants group fantasy of rebuilding a Holy Ro-
during the Great Purge of the 1930s: man Empire of the German nation (a con-
Whos going to remember all this riff-raff cept that requires a closer look).
in ten or twenty years time? No one. Gonen rightly tells us that Hitler was
BRIAN VANDEMARK a borrower of ideas. He states that Hitler
United States Naval Academy borrowed ideas on the psychology of
Annapolis crowds from the French Gustave Le Bon
Maryland who was a racist himself and was ad-
mired by the fascist leadership in Italy.
While there is no proof that Hitler him-
self read Le Bon, the Nazi literature sug-
gests that Joseph Goebbels was familiar
GONEN, JAY Y. 2000. The Roots of Nazi with Le Bons writings. Interestingly,
Psychology: Hitlers Utopian Barba- Gonen does not focus on cronies of Hitler
rism. Pp. 224. Lexington: University of such as Goebbels. Perhaps future studies
Kentucky Press. $25.00. will explore the importance of the psy-
After the collapse of Yugoslavia, the chology between a leader and his or her
way in which Serbian leaders Slobodan entourage as a factor in the success of
Milosevic and Radovan Karadzic revived propaganda campaigns. Nevertheless,
historical traumas and enflamed malig- Gonen should be complimented on his il-
nant nationalism has increased contem- lustration of how a leaders internal
porary scholars interest in the relation- needs dovetail with the primitive desires
ship between such leaders and their and fears of the masses. Leaders, even ab-
millions of followers. How is political pro- solute dictators like Hitler, do not exist in
paganda related to the psychology of the a vacuum. The psychology of both sides,
masses? Nazi Germany still remains an leaders and followers, needs to be taken
arena for further exploration of this ques- into account for a comprehensive view of
tion. In The Roots of Nazi Psychology: Hit- history.
lers Utopian Barbarism, Jay Gonen At the end of his book, Gonen looks at
focuses on Nazi ideology to illustrate the present-day Germany and notes, for ex-
contact points between the Fhrer and ample, that the reunification of East and
the German people. Adolf Hitler often West Germany did not bring back old
used the repetition of short slogans to dreams of malignant nationalism. This
feed his followers Nazi ideology, includ- makes him optimistic that Germans ac-
ing ideas on race, the Jewish threat, liv- knowledgment of responsibility for the
ing space, world domination, the folkish past will secure a future that is neither
state, the femininity of masses, and the utopian nor barbaric. This reviewer, as a
omnipotence of the leader. Hitler was clinician, would point out that it has
able to sell this ideology in part because, taken Jewish people several decades to
after being traumatized during World fully recognize the transgenerational
War I, Germans were looking for an om- transmission of trauma and its influence
nipotent savior. As a reaction against on the second and third generations,
188 THE ANNALS OF THE AMERICAN ACADEMY
The strengths of the volume are its missing in purely political narratives.
comprehensiveness, which makes it use- And his quip from Casablanca, the
ful for teachers and students, and its at- Claude Rains line, of the French to
tention to detail. Its broad scope fills a round up the usual suspects, adds a mar-
need to have an up-to-date narrative of velous light touch to the proceedings.
the French empire; while imperial stud- G. WESLEY JOHNSON
ies are out of fashion in some quarters,
Brigham Young University
they are essential if we are to understand
Provo
the modern postidependence world. Utah
Quinn is at his best when he brings per-
sonal insights to bear on his material,
such as his arriving in Vietnam to be
given a library of Victor Hugos works
hidden since colonial days. UNITED STATES
One problem for Quinn is that of
sources. Generally, it is a fully docu- ALTSCHULER, GLENN and STUART
mented work. He was well advised to pe- BLUMIN. 2000. Rude Republic: Amer-
ruse the most relevant French sources, icans and Their Politics in the Nine-
which he has done, but in the process, his teenth Century. Pp. xii, 316. Princeton,
notes show a lack of familiarity with NJ: Princeton University Press.
works by British and American scholars $35.00.
of the French empire. For instance, the In the 1960s political historians devel-
African materials do not reflect the rele- oped what has become an enduring para-
vant works of Crowder, Wallerstein, digm of nineteenth-century American
Morgenthau, Austen, Foltz, and others. history. In its elemental form this inter-
Part of the problem is that his French ex- pretation holds that politics after 1840
perts themselves often do not cite the was pervasive, interesting, and accessi-
English-speaking sources, which Quinn ble to large numbers of Americans who
needed to pick up. displayed their commitment to the pro-
One of the empires most important cess by voting in the frequent elections
legacies of empire in our century was the that mark this period, by joining the po-
formation of political parties, which from litical parties that became institutional-
the 1920s onward played a capital role in ized, and by participating in the
the decolonization process. This subject campaign rituals that grew to large-scale
area, as well as the evolution of elites celebrations such as parades and pole
(who eventually took over the empire), raisings. Politics, in the clich that now
needed fuller coverage. We learn about summarizes this period, was in the air
Bourguiba and Ho Chi Minh but not everywhere, as the American electorate
much about Blaise Diagne or Ferhat created a golden age of politics.
Abbas. Glenn Altschuler and Stuart Blumin,
Last, more copyediting was needed in their remarkable study The Rude Re-
misspellings of Henri Brunschwig, spahi, public: Americans and Their Politics in
Lamine Guye, Ousmane Sembne, and the Nineteenth Century, challenge this
so forth mar an otherwise literate and ruling wisdom. Theirs is not so much a
readable text. On the positive side, heads-on confrontation as it is a flanking
Quinns sections on the literary aspect of operation that seeks to demonstrate the
colonial life add a sensitive touch often variability and different levels of political
190 THE ANNALS OF THE AMERICAN ACADEMY
events of the presidential election year eral presentation (p. 300, Table 10), rank-
rest on diligent and remarkably wide re- ing the states in which Blaine fared worst
search in manuscript collections, news- and indicating his Gain/Loss over 1880,
papers, and magazines of opinion. somehow generally confuses 1876 to
Summerss familiarity with relevant 1884 percentage point shifts with those of
scholarship, unpublished as well as pub- 1880 to 1884. Certainly James A. Gar-
lished, further strengthens his study. field carried neither Louisiana nor South
In the end, Summers concludes that Carolina. (A finaland idiosyncratic
Clevelands election marked a signifi- concern: I wonder whether the cotton
cant moment in Gilded Age politics in fields of Iowa were not fields of dreams).
several waysreal issues were at stake My caveats notwithstanding, I strongly
in 1884; politics remained the main- endorse Rum, Romanism, & Rebellion:
stream politicians game; the demon- The Making of a President, 1884.
strated need of the parties for campaign SAMUEL T. MCSEVENEY
funds would influence their subsequent Vanderbilt University
behavior. Likewise, the Democrats Nashville
tightening grip on the South and recap- Tennessee
ture of the presidency meant that the Re-
publicans would seek to strengthen
further their position in the North by
sharpening their protective tariff rheto-
ric and by seeking admission of Western WEBBER, MICHAEL J. 2000. New Deal
states. Summerss arguments merit close Fat Cats: Campaign Finances and the
attention by political historians, particu- Democratic Party in 1936. Pp. xiii, 208.
larly those of us who have stressed the New York: Fordham University Press.
significance of developments and elec- $39.95. Paperbound, $19.95.
tions during the 1890s and perhaps failed The rise and fall of Franklin D. Roose-
to appreciate fully the importance of velts New Deal coalition has captured
those of the previous decade. the imagination of generations of histori-
Rum, Romanism, & Rebellion stands ans, political scientists, and sociologists.
as a fine work of historical analysis, but it For most, the story begins with FDRs im-
is not without limitations and weak- pressive victory in the 1932 general elec-
nesses. Summers writes well, very well, tion and continues with the Democratic
but occasionally he strikes me as trying presidents skill in solidifying his broad-
too hard for effect. He is a master of de- based coalition in the 1936, 1940, and
tail, but some of that detail could have 1944 electoral contests.
been deleted without serious loss, indeed In New Deal Fat Cats, Michael Webber,
perhaps with benefit. The generally high chair of the Department of Sociology at
standard of writing to the contrary not- the University of San Francisco, offers a
withstanding, a handful of sentences con- powerful and convincing social scientific
tain gaffes that should have been caught. analysis of one of the centurys most im-
Summers appears more comfortable ana- portant elections. Noting an abundance
lyzing the behavior of politicians and of literature on class conflict and the rise
other public figures than analyzing that of labor unions, Webber focuses his study
of the electorate. Quantitative election of the 1936 election on the interactions
analyses are generally adequate, no between business, organized labor, and
more; for the most part they focus on se- political parties. The guiding thread of
lected districts rather than illuminate this study, Webber argues, is the as-
broader patterns in the period. One gen- sumption that campaign finance contri-
BOOK DEPARTMENT 197
butions can be a reliable empirical the activist role taken by the federal gov-
indicator of the political preferences of ernment since the inception of the New
people following their real material inter- Deal.
ests. Campaign donations, he explains, New Deal Fat Cats offers a powerful
act as an important tracer element and convincing mix of social scientific
marking the tracks of business and labor data and historical narrative. Webbers
influence in the political system. ability to translate the theoretical to the
In an eight-chapter book, Webber of- real makes this an excellent book for both
fers a historiographic overview of New the specialist and the interested ob-
Deal writings in both the social sciences server. His opening chapter surveying ex-
and humanities fields, reviews the role of isting New Deal scholarship was
business between 1932 and 1936, dis- especially useful and informative.
cusses the mass-consumption sector and Webbers use of secondary sources from
Democratic party finances, evaluates the across disciplinary lines also demon-
investment theory of politics, offers a fas- strated great attention to detail and re-
cinating and novel investigation of the spect for competing methodological
1936 Democratic convention book soft approaches.
money scandal, points out important dif- MARC DOLLINGER
ferences among business leaders in New Princeton University
York City and the South, and closes with New Jersey
a study of FDR and the unions.
For Webber, the story of business and
the 1936 election reaffirms the impor-
tance of FDRs far-reaching political co-
alition. If there is a lesson to be learned SOCIOLOGY
from the New Deal, Webber explains, it
is that the Democratic party successfully BAYOR, RONALD H. [1996] 2000. Race
mobilized groups outside the main- and the Shaping of Twentieth Century
Atlanta. Pp. xvi, 334. Chapel Hill: Uni-
stream of the American power structure
versity of North Carolina Press.
to bring about far-reaching changes in
Paperbound, $19.95.
how government responded to the needs
of the American people. The publication of a paperback edition
Webber, in one of his most important of Ronald H. Bayors book on race in
observations, discovered strong pro-FDR twentieth-century Atlanta is a welcome
sentiment among business leaders from event. Bayor is a careful scholar who
New York City, the South, and the alcohol writes with clarity and insight, and he de-
industry. Avoiding a common tendency to serves a wide reading.
blur all business leaders together, Historians, journalists, and social sci-
Webber showed how Jews in New York entists have perhaps studied and written
City, southern Democrats, and opponents about Atlanta more than any other city in
of prohibition dissented from the old-line the South. Site of Booker T. Washingtons
Protestant corporate elite. In this sense, famous speech at the Cotton States and
marginalized business leaders mirrored International Exposition, and just a few
the political attitudes of the larger elec- years later scene of a particularly violent
torate. The presidential election of riot precipitated by inflammatory anti-
1936, as Webber notes in his conclusion, black newspaper rhetoric, birthplace of
was, in many respects, a referendum on Martin Luther King, Jr., a one-time cen-
198 THE ANNALS OF THE AMERICAN ACADEMY
ter of Klan activity, headquarters for of land-use policy and a continuing drive
many progressive organizations seeking by whites to distance their commercial
to disestablish the Jim Crow system and and residential lives from African Ameri-
diminish its legacy, and the only South- cans remained a substantial force
ern city whose mayor testified in favor of throughout the second half of the twenti-
the public accommodations section of the eth century.
1964 Civil Rights Act, Atlanta has a ra- In mid-century, when the Civil Rights
cial history with twists and turns un- movement challenged the racial order of
matched by any other city in the nation. the South, apologists for the status quo
In the second half of the twentieth cen- often spoke of that order as resting on lo-
tury, Atlanta, with its widely celebrated cal custom. Historian Ronald Bayor sets
biracial coalition, was known as the city the record straight. The status quo of the
too busy too hate. Yet, as Bayor shows, midtwentieth century rested on a policy
such a simple slogan fails to capture the past constructed by a multitude of gov-
complexity of Atlanta in that period, and ernment actions. Once constructed, as
it overlooks the weight of the citys racial subsequent events demonstrated, that
past. Bayors book provides needed depth past could not be undone by a simple shift
to the story of how racially motivated pol- of political power and a wave of policy de-
icy actions taken at an earlier time re- cisions based in that shift. Contemporary
sound through the growth and change of conservatives who complain that African
a city that has often prided itself on rising Americans look too much to past griev-
to new heights out of the ashes of its past. ances reveal a superficial grasp of history.
Georgia is a deep South state, and for the They would do well to read Ronald
first half of the twentieth century there Bayors book and reflect on its lessons.
were few checks on white disregard for There is much that a deeper understand-
and antagonism toward the interests of ing of history can teach.
Atlantas black citizens. Black Atlanta CLARENCE N. STONE
proved adept at making use of limited re- University of Maryland
sources, but in harsh circumstances College Park
adeptness can only carry a group so far.
Focusing on public policy, Bayor shows
how large a part the drive to establish
and maintain a Jim Crow system played
in the shaping of the city, not just in the BROWDER, LAURA. 2000. Slippery
first half of the twentieth century, but Characters: Ethnic Impersonators and
American Identities. Pp. xii, 312. Cha-
through that early legacy in the second
pel Hill: University of North Carolina
half as well. The biracial coalition was
Press. $49.95. Paperbound, $18.95.
not so much a thorough triumph of en-
lightenment as a new arena in which At- An elderly WASP man poses as a La-
lantas African American community tino youth, a colored Southern janitor
could ply its efforts to end old practices goes native as a Cherokee Indian, and the
and cope with their consequences. To be daughter of a Welsh Baptist mother cre-
sure, electoral enfranchisement and ates an identity as an orthodox Jew.
other federal civil rights actions came to These are only three of the many ethnic
play a major part in giving black impersonators whose lives are discussed
Atlantans a larger body of resources with by Laura Browder in this intriguing
which to pursue their own attempt to rise study.
from the ashes, in their case the ashes of Browders guiding premise is that the
oppression. Nevertheless, early patterns autobiographical narrative is a key genre
BOOK DEPARTMENT 199
in American culture, which has been used Lance, and carved out a successful career
for a variety of personal and political pur- as an authentic Indian, starring in both
poses. She argues, convincingly, that pop- fictional and ethnographic films. In ear-
ular autobiographies have come to claim lier chapters, Browder carefully de-
a moral authority, as the power of indi- scribes how Indian representations in
vidual testimony can help reshape public theatre, museums, and Wild West shows
thinking and public discourse, in this laid the groundwork for the mlange of
case about race and ethnicity. Browder imagery that came to define Indian iden-
traces these autobiographies from white- tity and on which both Carter and Long
authored slave narratives to contempo- traded.
rary accounts of individuals passing from Turning to black identity, Browder
one ethnic category to another. traces the use of fictional slave narra-
An irony that unites them all is that tives by abolitionists to make the case for
autobiographies by impersonators were the essential humanity of African Ameri-
often greeted with greater public enthu- cans. Yet abolitionists also perpetuated
siasm than those by real members of the the stereotypes: unwilling to trust the
groups concerned. A case in point is the slaves to tell their own stories, these writ-
now-notorious 1976 memoir, The Educa- ers created slave personae who were dis-
tion of Little Tree, written under a pseud- gusted by black bodies . . . and unable to
onym by Forrest Carter, a former Ku see the truth about slavery without the
Klux Klansman and speechwriter for Al- intercession of the abolitionists. Later,
abama Governor George Wallace. In the she discusses the development of black-
persona of the Cherokee Little Tree, face entertainers, and describes how
Carter wrote of being raised by saintly, groups like the Symbionese Liberation
spiritual grandparents before enduring Army took on blacker-than-thou perso-
torment in a brutal orphanage and fi- nae in an attempt to demonstrate their
nally being rescued. The book fit perfectly radical authenticity in the 1970s.
with the new age mood of the time and be- Throughout, Browder makes a good
came a huge best seller. As Browder ex- case for the primacy of the personal nar-
plains, Little Tree was a fantasy about rative in American cultural history and
Indian spirituality, and it is also a fan- explores the complicated and ever-chang-
tasy perfectly attuned to an American ing notions of race and ethnicity. She
strains her argument when she includes
public well versed in the rhetoric of self-
a chapter on hoboes, asserting that pov-
actualization and the recovery move-
erty behaves as an ethnicity, and I would
ment. In the wake of the success of
have liked to see a discussion of a major
Carlos Castanedas equally spurious
autobiographical genre, the Indian cap-
Yacqui medicine man Don Juan (whom
tivity narrative, but these are quite mi-
Browder does not mention), Little Tree
nor flaws in an innovative and engaging
epitomized the noble savage image, and
book.
the cruel irony is that Carter succeeded
S. ELIZABETH BIRD
in his impersonation by trading on his
deep knowledge of racial and ethnic ste- University of South Florida
reotypes. Tampa
While Carter used his impersonation
to escape his racist past, Sylvester Long
became an Indian to escape his racial
identity as a colored man in the 1930s
South. In a series of calculated steps, COSTA, DORA L. 1998. The Evolution of
Long became Chief Buffalo Child Long Retirement: An American Economic
200 THE ANNALS OF THE AMERICAN ACADEMY
History, 1880-1990. Pp. xiii, 234. Chi- has decreased over time, or as economists
cago: University of Chicago Press. would put it, the elasticity of demand for
$40.00. Paperbound, $19.00. retirement with respect to income has de-
creased. Costa concludes that rising in-
Dora Costas book on retirement is come was important perhaps up until
very much in the cliometric tradition, 1940, after which other factors became
containing much evidence and statistical more important in the labor force with-
analysis. The big issue she tackles is ex- drawal decision. This result is important
plaining the well-documented with- for academic researchers, but it is also
drawal from the labor force of men 65 and very important for policy makers who
older in the United States, a trend that will be increasingly faced with a growing
started in the late nineteenth century. number of retirees as the baby boom gen-
The key factors she examines are rising eration ages. An example of such a policy
real income, improvements in health, and issue is, How much of the retired popula-
changes in the structure of occupations.
tion should social security retirement
It is based in part on several of her previ-
benefits cover? Has income, and indirect
ously published papers combined with
saving, become high enough that some re-
some new material that covers the politi-
strictions on coverage or even means
cal history of retirement and speculates
testing become sensible? Likewise, what
about future patterns of retirement.
would the impact of partial privatization
While it is filled with charts, graphs, and
of social security be on labor force activ-
econometrics, her story is still quite ac-
ity?
cessible to the careful, noneconomist
Given the scope of Costas subject,
reader.
The core of her evidence comes from a some questions arise. Equating declining
large sample of the military pension re- labor force participation rates with in-
cords of Union Civil War veterans. She creasing retirement rates is troublesome,
supplements these samples with samples as retirement has strong voluntary con-
drawn from the 1900 and 1910 Public notations to it, while being out of the la-
Use Microdata Samples of the Decennial bor force can include withdrawal owing to
Censuses. The Union Army Veteran disability, injury, or unemployment. I also
Sample allows the observation of how in- suspect that labor force activity of older
creases in real income, a pure income ef- men may start increasing as the rela-
fect, affected the labor force behavior of tively smaller supply of labor coming
veterans. As the sample of veterans is from the cohort following the baby boom
representative of the general population, generation puts pressure on the labor
Costa is able to show that the income market, increasing real wages. Perhaps
from the pension was large enough to in- the substitution effect, as economists put
duce many older men to withdraw from it, will induce older men to stay in the la-
the labor force and that until recently, ris- bor force longer now that health issues
ing real income was central in reducing are less of an issue than they were earlier
the labor force participation of men 65 in the twentieth century. This book, nev-
and older. ertheless, provides the basis from which
Rising real income has been key in most economic analysis of retirement
spurring the withdrawal from the labor should proceed.
force, particularly before social security JON R. MOEN
and other government pensions were put
in place. The sensitivity to changes in in- University of Mississippi
come on the part of older men, however, University
BOOK DEPARTMENT 201
teacher salary block grants to schools. Soviet states are going through the pro-
This would have changed the practice of cess of devolving education responsibili-
assigning and paying teachers directly ties to local and school levels. Some
from the national Ministry of Education. variant of the New Zealand model may
The idea was to allow schools to use these well allow these states to continue fund-
and operating grants in a flexible fash- ing schools on a national level while de-
ion. Unions strongly resisted, and after volving authority and responsibility to
10 years and continuous sweetening of localities and schools. Policy makers in
the pot, there still were a number of those and other countries would be well
schools not participating in this funding advised to read this wonderful book.
scheme. The model and implementation JOHN F. WITTE
problems are important because many
University of Wisconsin
education systems in former Eastern and
Madison
THE ANNALS OF THE AMERICAN ACADEMY
INDEX
INDEX
203
204 THE ANNALS OF THE AMERICAN ACADEMY
WELSH, BRANDON C., and DAVID P. WILSON, DAVID B., see MACKENZIE, DORIS
FARRINGTON, Toward an Evidence- LAYTON, coauthor
Based Approach to Preventing Crime, WILSON, DAVID B., Meta-Analytic Methods
158-173 for Criminology, 71-89