Title: Responding to Gangs: Evaluation and Research
Series: NIJ Research Report
Author: Winifred L. Reed and Scott H. Decker, editors
Published: July 2002
Subject: Gangs and juvenile delinquency prevention
252 pages
630,000 bytes

------------------------------

Figures, charts, forms, and tables are not included in this ASCII plain-text file.
To view this document in its entirety, download the Adobe Acrobat graphic file
available from this Web site or order a print copy from NCJRS at
800-851-3420 (877-712-9279 for TTY users).

------------------------------

U.S. Department of Justice
Office of Justice Programs
National Institute of Justice

Responding to Gangs: Evaluation and Research

------------------------------

U.S. Department of Justice
Office of Justice Programs
810 Seventh Street N.W.
Washington, DC 20531

John Ashcroft
Attorney General

Deborah J. Daniels
Assistant Attorney General

Sarah V. Hart
Director, National Institute of Justice

Office of Justice Programs
World Wide Web Site
http://www.ojp.usdoj.gov

National Institute of Justice
World Wide Web Site
http://www.ojp.usdoj.gov/nij

------------------------------

Responding to Gangs: Evaluation and Research

Editors

Winifred L. Reed
Social Science Analyst
National Institute of Justice

Scott H. Decker, Ph.D.
Department of Criminology and Criminal Justice
University of Missouri-St. Louis

NCJ 190351
July 2002

------------------------------
National Institute of Justice
Sarah V. Hart Director

Findings and conclusions of the research reported here are those of the authors
and do not necessarily reflect the official position or policies of the U.S.
Department of Justice.

The National Institute of Justice is a component of the Office of Justice
Programs, which also includes the Bureau of Justice Assistance, the Bureau of
Justice Statistics, the Office of Juvenile Justice and Delinquency Prevention, and
the Office for Victims of Crime.

------------------------------

Preface

This collection of papers presents a representative selection of NIJ's portfolio
of gang-related research. The genesis was the upsurge in gang crime beginning
in the mid-1980s, which prompted NIJ to expand research in this area.

NIJ's major focus is to generate research-based knowledge that can inform
policy and be useful for practitioners. For that reason, the emphasis in this
volume is on evaluations. An evaluation may indicate a program works or
doesn't, but in either case the information can be applied by those seeking
solutions to similar problems. The evaluation of Boston's initiative to halt youth
gang violence offers hope that focused law enforcement deterrence works. The
assessment of the G.R.E.A.T. program demonstrates that school-based
prevention can have favorable results.

Police chiefs and local policymakers can use evaluation findings to help develop
strategy and deploy resources or move in new directions. Not only were the
studies in this volume launched with an eye to practice, but some are themselves
the product of researcher-practitioner collaboration. Again, the Boston project
is an example, with its partnership of local law enforcement and academics
tackling youth homicide. Practitioners and researchers working together in
Orange County, California, developed a system for tracking gang crime.

In NIJ research, collaboration plays a major role. Through the National Youth
Gang Consortium, a coalition of Federal agencies, NIJ works to reduce youth
involvement in gang crime. The Consortium is administered by the Office of
Juvenile Justice and Delinquency Prevention (OJJDP), an agency with which
NIJ has had a long and productive relationship in addressing juvenile offending.
In addition to OJJDP, several other Consortium members supported studies in
this volume, sponsored the programs evaluated, or otherwise contributed. This
pooled expertise in different fields adds to the strength of the volume.

We are fortunate in having as editors Winifred L. Reed and Scott H. Decker.
Dr. Decker is with the Department of Criminology and Criminal Justice at the
University of Missouri and has been conducting research on gangs and related
issues for more than two decades. His extensive experience is no doubt the
source of his perceptive and forthright introductory comments. Winifred Reed,
a Social Science Analyst with NIJ, has similar depth of experience, having
managed the agency's portfolio of gang research for the past 10 years.

What's presented between these covers is not the final word on gang research.
If we have learned a great deal, a great deal is still unknown. Working with
practitioners and researchers, NIJ continues to identify knowledge gaps. The
risk factors for gang membership are known, but are there protective factors?
Are prison gangs related to street gangs? What can be done to encourage gang
members to leave the gang? What is the role of the media and popular culture in
the genesis and spread of gangs? These are only a few questions in a long list,
and as Scott Decker makes clear in his elaboration on research gaps, the list in
no way exhausts the possibilities.

This volume helps fill the knowledge gap with information that can be applied to
operations. It suggests that through research we may achieve a degree of
understanding that enables those in the field to maximize prevention,
intervention, and suppression.

Sarah V. Hart 
Director 
National Institute of Justice

------------------------------

Contents

Preface

Chapter 1. A Decade of Gang Research: Findings of the National Institute of
Justice Gang Portfolio

Chapter 2. The Evolution of Street Gangs: An Examination of Form and
Variation

Chapter 3. Young Women in Street Gangs: Risk Factors, Delinquency, and
Victimization Risk

Chapter 4. Youth Gang Homicides in the United States in the 1990s

Chapter 5. National Evaluation of the Gang Resistance Education and Training
(G.R.E.A.T.) Program

Chapter 6. Evaluating Nevada's Antigang Legislation and Gang Prosecution
Units

Chapter 7. Evaluation of a Task Force Approach to Gangs

Chapter 8. Gang Prevention Programs for Female Adolescents: An Evaluation

Chapter 9. Reducing Gang Violence in Boston

Chapter 10. Developing a GIS-Based Regional Gang Incident Tracking System

------------------------------

Chapter 1

A Decade of Gang Research: Findings of the National Institute of Justice Gang
Portfolio

Scott H. Decker, Ph.D.
Department of Criminology and Criminal Justice
University of Missouri-St. Louis

In 1983, sociologists Hedy Bookin-Weiner and Ruth Horowitz asked whether
the end of the youth gang was "fad or fact." Although their concern was largely
with the influence of politics and ideology on both the funding and the nature of
research about gangs, the fundamental question regarding the "end" of youth
gangs seemed a reasonable one. After all, the major gang "epidemic" of the
1960s had appeared to subside in most cities except, most notably, in Chicago
and Los Angeles. Gangs had certainly faded from the research agendas of most
criminologists, as shown by the lack of empirical and theoretical work using
gangs as a focal point. But like many cyclical behaviors, including crime (Klein
1995b), gangs returned. And with their return came increased attention from
the research community.

This volume represents one segment of the increased attention that gangs
received during the past decade. That decade saw a dramatic increase in the
level of funding for gang research. Federal agencies, led by the U.S.
Department of Justice's National Institute of Justice (NIJ) and Office of Juvenile
Justice and Delinquency Prevention, the U.S. Department of Health and Human
Services' (HHS's) Administration for Children and Families and National
Institute on Drug Abuse, and the U.S. Department of Education, have made
substantial contributions to the expansion of knowledge about gangs, gang
interventions, and the characteristics of individual gang members. The selections
in this volume were funded primarily by NIJ. Although these are not the only
federally funded projects in this topic area, they represent some of the more
prominent and visible ones.

Bookin-Weiner and Horowitz suggested (1983: 599) that gang research,
particularly work funded by Federal agencies, will be influenced profoundly by
the dominant ideology of the time regarding crime, and that suppression has
been the dominant ideology. It is important in this context to distinguish between
the ideology underlying gang research and the ideology underlying gang
intervention. Clearly, their suggestion was that suppression and deterrence
strategies will be reflected in research and practice funded by the Federal
Government.

One of the most influential federally funded research projects (Spergel and
Curry 1993) documented that although suppression was the dominant response
to gangs, it was perceived to be the least effective. Even among law
enforcement respondents in that study, it was perceived as less effective than
providing social opportunities through job training and education. In addition, as
Malcolm Klein (1995a) noted, little serious evaluation research has focused on
antigang suppression efforts, which suggests that the relationship among Federal
funding, political ideology, and research findings is much more complicated than
might be expected. Indeed, much of the gang research presented in this volume
suggests that simple predictions from researchers or law enforcement are likely
to be wrong. Research is not motivated by political ideology or a commitment
to a particular outcome, but is conducted because there is a pressing need for
information to guide decisionmaking.

The works collected in this volume reflect a diverse set of methodologies and
substantive interests and range from field studies to surveys of classroom
students to analyses of official records. Basic research issues as well as applied
policy issues are examined. The volume includes field research, survey
research, program evaluation, and records research. In addition, these chapters
reflect an interest in gender, in minorities, and in improving criminal justice
system interventions. Several researchers whose work is presented here have
also used the platform of Federal funding to examine issues well beyond the
scope of their initial grant, enhancing our knowledge of gangs as well as
leveraging scarce Federal research dollars in important ways. In all, this is a
broad-based collection of studies that will be useful to policymakers,
practitioners, and researchers. The impressive basic and applied work that
appears here adds to what we know about gangs and may affect what we do
about gangs. This research should help lay the foundation for the research and
intervention agenda for the next decade.

This introduction summarizes the chapters that follow and offers a substantive,
methodological, and policy response to each. The summaries attempt to place
the research in the broader context of research on gangs, delinquency, and
juvenile justice. The introduction then turns to an examination of areas of
research that need more attention and concludes by identifying potential future
directions and offering some methodological suggestions for gang research.

Summaries of the Research Projects

The best of the research projects in this volume combine a strong conceptual
focus with a sound methodology. Because NIJ funded most of these projects, it
follows that they have a singular focus on policy or program evaluation. What
sets the best of these projects apart, however, is their breadth of focus, strong
commitment to sound methodologies, and ability to identify both basic and
applied research questions.

The dominant paradigm in research is that a basic or scientific research question
must be identified. This leads to the choice of appropriate methodologies that
attempt to rule out rival hypotheses, controlling for threats to internal and
external validity. Data are collected, controls applied, and analyses conducted.
This is the prototypical scientific model, in which theoretical questions guide
research about basic scientific questions, which, in turn, produces findings that
lead to policies, programs, or interventions. In the pure form of this model, a
researcher would begin with a question about gang behavior derived from the
literature, formulate hypotheses about how gangs and their members behave,
and then test those hypotheses in various ways. Following several successful
tests of the hypotheses, a second researcher or research team may decide to
determine whether the basic research findings can change the behavior of
individual gang members or gangs. This process, which may last several years
or decades, may result in suggestions for program or policy intervention.

Unfortunately, however, such an approach is not feasible in practice.
Policymakers simply do not have the luxury of waiting 10 or 20 years for an
answer to the "gang problem." In many American cities, gang violence has so
disrupted the social fabric of some neighborhoods that socialization,
employment, and education can no longer work successfully. Without
overdramatizing the extent to which this is true, one need only look at such
cities as Chicago, Los Angeles, and St. Louis, where gangs are responsible for
more than 25 percent of all homicides and assaults. The reality of gang violence
has placed a significant amount of pressure on those who fund Federal
intervention and research programs to "do something." The question, of course,
is what to do. This volume documents many of the responses to gangs as well
as some of the research about the impact of such responses (see exhibit 1).

Reducing Gang Violence in Boston

In a review of his tenure at the National Institute of Justice, a former director
said that funding the research on Boston's Operation Ceasefire was one of the
most substantive investments that the Institute had made during that time (Travis
and Blumstein 2000). Anthony A. Braga and David M. Kennedy's "Reducing
Gang Violence in Boston" (chapter 9) provides ample evidence to support that
claim. As noted above, the strongest research combines a basic and applied
focus with a strong methodology. The body of research on Operation Ceasefire
has all of these attributes. Its strongest feature is its commitment to the
problem-solving process. Braga and Kennedy engaged a broad array of local,
State, and Federal justice officials, as well as community and neighborhood
leaders, in a data-driven effort to craft a more effective and efficient response to
youth violence. This intense effort ultimately led to a focus on gang violence.
Braga and Kennedy's dogged commitment to focus on data and analysis may
be the longest lasting contribution of this research.

The early stages of this research had a stronger qualitative and descriptive
focus, as would be appropriate in the early stages of a problem-solving model.
As rates of gun homicide in Boston decreased, however, a new research
question emerged: how to explain the precipitous and unprecedented decline in
gun homicides. The research process addressed this key issue later in the
project, providing time-series evidence that the intervention and the decline
were indeed linked statistically. This illustrates another strength of the research:
that it was flexible and farsighted enough to adapt to the project's emerging
needs. The growth of the project over time, from its initial problem-solving
focus on guns and youth, and its integration of divergent constituent groups are
examples of how research can influence policy.

Braga and Kennedy's youth gang and youth violence research has also
contributed to our conceptual understanding of deterrence. The search for the
appropriate "levers" to pull to ensure compliance with criminal justice mandates
remains a long-lasting addition to our knowledge about this process. Braga and
Kennedy stress that the threat of criminal sanctions will not be an effective
deterrent unless it reaches those individuals least likely to be reached by public
messages. It is ironic that a project initially funded to understand and respond to
youth violence and lead to a focus on gang violence has as yet provided less
direct knowledge about gangs than about the process of addressing the broader
issue of youth violence.

Gang Programs for Young Women

Among the many gaps in our knowledge of gangs, perhaps none looms larger
than the paucity of research on young women and gangs. Two of the chapters
in this volume address this crucial issue directly. Although we know too little
about young women and gangs, we know even less about programming for
gang girls. Katherine Williams, G. David Curry, and Marcia I. Cohen's "Gang
Prevention Programs for Female Adolescents: An Evaluation" (chapter 8)
attempts to address the second of these issues. The research presented here
summarizes an evaluation of gang programming targeted to young women in
Boston, Pueblo (Colorado), and Seattle and supported by HHS's Family and
Youth Services Bureau. The choice of these diverse cities served to maximize
the variation in ethnicity among the young female gang members studied. For
this reason, the results of this study should have been able to shed light on both
process and impact issues.

The description of the research process is a textbook example of what can go
wrong in program evaluation. The proposed randomization could not be fully
implemented, the control groups of matched individuals could not be
successfully constructed, the dropout rates in the control and program groups
were large and unequal, the number of subjects in the program group was often
too small for meaningful analysis, and the nature of the intervention changed
from the initial plan. None of these shortcomings was directly the fault of the
research team; indeed, the team had negotiated access to the programs and had
fostered agreement among the agencies on a process of assigning individuals to
control and treatment groups. In addition, the agencies were receiving Federal
funding to support their programs, often a predicate for participation in an
evaluation.

The research reported here is the result of both process and outcome
evaluations. In the Boston and Seattle sites, few goals of the initial research
design could be implemented because of low participation rates, high dropout
rates among those who did participate, and the changing nature of the program.
Despite these obstacles, the research team produced a process evaluation of
those two sites that should be required reading both for those involved in
program evaluation and for those who would fund such evaluations. It identifies
the myriad factors in program design and delivery that can frustrate an outcome
evaluation. The research produced an outcome evaluation of the program in
Pueblo, and the results of that intervention were mixed with regard to the
program's success.

This chapter leaves many unanswered questions about the state of programming
for young women in gangs. Among the key questions that cannot yet be
assessed are the following:

o What is the extent of programming for young women in gangs?

o Do female gang members need programming different from that for their male
counterparts?

o Do female gang members need programming different from that for young
women who are not gang members but are involved in other forms of serious
misconduct and delinquency?

o Can programming for female gang members operate independently of other
interventions for gang members?

o Can programming for female gang members operate independently of other
interventions for young women who engage in other forms of serious
misconduct and delinquency?

o Which types of programming are most successful in reducing the involvement
of female gang members in serious misconduct and delinquency?

The chapter begins to fill in some of the gaps that must be addressed in
developing, implementing, and evaluating programs targeted to young women in
gangs. Unfortunately, such programs often have mirrored programming for
youth in general, thereby preventing the research team from answering these
questions and from providing firmer guidance to others interested in the
answers.

Trends in Youth Gang Homicides

Few topics in the area of gang research have attracted as much attention as
gang homicide. Past research has focused on the correlates of gang homicides,
the definitions of gang homicide, trends in gang homicides, and the differences
between gang and nongang homicide. Cheryl L. Maxson, G. David Curry, and
James C. Howell have all made important contributions to the past literature on
this topic, so their current work (chapter 4 in this volume) is a welcome addition
to an already large body of literature. Given the already impressive body of
knowledge on gang homicide--much of it produced by these three
coauthors--can anything new be said on the subject? The answer is a
resounding
yes.

"Youth Gang Homicides in the United States in the 1990s" carefully examines
three issues of critical importance to the understanding of gang homicide:
measurement, trends, and correlates. The measurement issue, as Maxson and
her colleagues demonstrate, is central to understanding gang homicide and gang
crime in general. Depending on whether one applies a "gang membership"
definition (as the Los Angeles Police Department does) or a "gang motivated"
definition (as the Chicago Police Department does), one ends up with estimates
of the volume of gang crime that are "twice as great" or "half as great" (Maxson
and Klein 1990; Maxson, Gordon, and Klein 1985). Since respondents to the
National Youth Gang Center's (NYGC's) annual survey (which the authors
used as a data source) are given guidance on which crimes are to be classified
as gang crimes, there is reason to believe that the data have at least some
validity.

Yet this chapter also stresses that even such guidance does not always result in
valid estimates of the number of gang homicides. The authors compare NYGC
estimates with estimates found by other researchers and find modest levels of
convergence for Chicago and Los Angeles data. Although this convergence is
encouraging, it probably represents the best-case scenario, as Chicago and Los
Angeles have been using a formal definition to classify gang-related crimes for a
longer time than most cities have. The findings suggest that for smaller
jurisdictions or cities where the gang problem emerged in the past decade, the
validity of even homicide data may leave much to be desired.

An important finding of this research is the decline in gang homicides observed
at the end of the 1990s. The decline in homicide has been chronicled in various
sources (Blumstein and Wallman 2000). It is encouraging to see documentation
that gang homicides followed this decline, albeit later in the decade and at a
somewhat slower pace. If the trend continues, it will have major implications for
policymakers and practitioners. A decline in gang homicides may signal a
decline in the level of gang activity, gang membership, gang violence, or all
three. Researchers are advised to track such trends in their jurisdictions and use
the data to help map policy and problem-solving responses.

The section on correlates of gang homicide is the most intriguing, but least
developed, part of the paper. Owing to the limited number of years of data
available, it is not possible to examine a time-series analysis of the correlates of
the decline in gang homicides. Yet Maxson and her colleagues report that the
size of a jurisdiction and the magnitude of the reported gang problem (as
measured by the number of gangs and gang members) are likely correlates of
gang homicide. Although this finding is based on a small number of years, it is
likely to persist through time-series analyses. This likelihood suggests a pattern
of gang homicides that will invite closer scrutiny from researchers over the next
decade, as more data become available. In addition, gang homicide may be a
sentinel for other problems in a community and bears watching for other
reasons. The utility of NYGC data has been validated by this paper.

Evaluating G.R.E.A.T.

The American response to youth problems has seldom been based on a clear
conceptual model that has strong empirical support. As Gottfredson (1997)
notes, a host of "feel good" programs attempt to prevent problems before they
emerge or worsen. Much of what passes as school-based prevention falls into
this category of well-meaning but low-impact programs. Since 1998, the U.S.
Department of Education has worked to promote the use of effective programs.
"National Evaluation of the Gang Resistance Education and Training
(G.R.E.A.T.) Program," by Finn-Aage Esbensen and his colleagues (presented
in chapter 5), reflects a large-scale research effort to evaluate the success of the
Gang Resistance Education and Training (G.R.E.A.T.) program. Oversight for
G.R.E.A.T. is currently provided by the Bureau of Alcohol, Tobacco and
Firearms (ATF), the Federal Law Enforcement Training Center, and
representatives from five local law enforcement agencies (Phoenix; Portland,
Oregon; Philadelphia; La Crosse, Wisconsin; and Orange County, Florida). It
is offered in more than 2,100 schools to more than 340,000 students
nationwide. G.R.E.A.T. is a classroom-based program of instruction that
consists of eight lessons delivered in nine sessions designed to teach (primarily)
middle school students life skills that will enable them to resist the pressures of
gangs, drugs, and delinquency. It is essentially a cognitive program that depends
on teaching students facts and offering suggestions about how to respond to
situations they may encounter. As such, it resembles the Drug Abuse Resistance
Education (D.A.R.E. (registered trademark)) program.

The evaluation team chose the opportunity to evaluate G.R.E.A.T. to enhance
knowledge of gangs, delinquency, and the risk and resiliency factors related to
involvement (or noninvolvement) in gangs or delinquency. The team proposed a
theory-driven evaluation, that is, a research process in which key theoretical
questions about the nature of gangs and gang behavior are assessed. In addition
to providing information on the impact of the G.R.E.A.T. program, this study
generated an extensive and impressive body of research during the
evaluation.[1] The careful attention to data collection, concern for measurement
issues, and large and diverse sample should furnish information about gangs and
delinquency for the next decade. The theory-driven nature of the evaluation
suggests that the research team was also attentive to the conceptual issues
involved in the intervention, a strength also noted in Boston's Operation
Ceasefire research. Gang programs are evaluated too infrequently, and
evaluated well even less frequently. When evaluations of large-scale programs
are conducted as well as this one was, they will offer hope for the future of
evaluation research. The broad perspective on evaluation research held by
Esbensen and his colleagues argues that such a view should be required by
funding agencies in the future.

The results of the evaluation are interesting in themselves. The initial
cross-sectional results suggested that G.R.E.A.T. had an impact on increasing
students' knowledge of how to stay out of trouble and on reducing
self-reported delinquency and adolescent misbehavior. These findings exceeded
appropriate levels of statistical significance and caused a stir among the
research team and many gang researchers. How could a nine-session program
delivered by a noneducator pay such dividends? The answer came shortly
thereafter, when more powerful and appropriate longitudinal results were made
available. As expected, few differences were found between the control and
program groups over time, and program effects diminished. The longitudinal
evaluation did produce favorable results, however. In this phase of the study,
the researchers used a quasi-experimental design and a sophisticated
multivariate statistical analysis strategy--hierarchical linear modeling. The result
was a number of small but important differences between students who
participated in the G.R.E.A.T. program and those who did not. All these
favorable outcomes, which could not have occurred by chance, emerged 3 or 4
years after program completion. They constitute important evidence of the utility
of such programs.

When the initial cross-sectional results were reported, the program did not
appear to have favorable effects. Then a surprising thing happened. Evaluation
researchers do not look forward to telling program officials that the program
simply did not produce the desired results despite the commitment of program
staff to the welfare of participants. When ATF received the results of the
evaluation, it contracted with the researchers to convene a process that would
bring gang researchers, G.R.E.A.T. trainers, curriculum experts, and educators
together to redesign the curriculum. That process is currently underway, and
ATF remains committed to conducting a full-scale process and impact
evaluation of the newly designed curriculum.

Mapping Gang Activity

James W. Meeker and Bryan J. Vila are one of the most productive gang
research teams.[2] The significance of their work in Orange County, California,
which was initially funded by NIJ in 1996, is highlighted by the demographic
characteristics of that county, a diverse and rapidly changing environment for
youth gangs. As in many communities in the early 1990s, it was unclear whether
the public hysteria in Orange County over youth gangs reflected the reality of
the gang problem. Such circumstances are rife with the potential to create what
Huff (1990) identifies as overreaction and misidentification. Under such
circumstances, the public pressures law enforcement to "do something" about
gangs. Law enforcement rarely requires much motivation to identify and
suppress gangs; it casts a net so wide that many believe it overcriminalizes
youths (particularly minority youths) and defines the problem so broadly that it
cannot deal with it successfully. This research was undertaken in partnership
with law enforcement in an attempt to circumvent these problems.

Like the best of the research in this volume, the study by Meeker, Vila, and
Katie J.B. Parsons combines a focus on applied policy issues with a larger
concern with basic research issues about gangs. "Developing a GIS-Based
Regional Gang Incident Tracking System" (chapter 10 in this volume) describes
the evolution of a collaborative process between the University of
California-Irvine research team and local law enforcement. The paper is
virtually a textbook description of the dilemmas encountered when working
with multiple law enforcement agencies to arrive at common definitions of
gangs, gang members, and gang incidents. The process by which these
definitions were negotiated, implemented, and tested is described in
considerable detail and serves as a useful blueprint for other
researcher-practitioner teams planning to engage in a similar process.

At first reading, it would appear that the research team's primary contribution
was to develop a management information system that combined sophisticated
mapping and analytic capabilities. Such a reading, however, would miss the key
ingredients of the collaborative process that led researchers to help law
enforcement frame both strategic and tactical questions. This was possible only
through the use of multiple methods of data collection and analysis. Like the
process employed in Operation Ceasefire in Boston, data analysis here was
continually fed back to the law enforcement partners in Orange County, whose
input into measurement, data collection, and analysis was a key component of
the project. Because of the essential nature of the collaboration, it is important
to track the longevity of this working group. Whether such collaboration lasts is
a key question raised about many such groups. The extent to which the
researcher-practitioner partnership outlives the grant is an important measure of
institutionalization.

Evaluating Nevada's Antigang Legislation

A common refrain in this review of NIJ-funded gang research is that a particular
topic has received too little scrutiny. The same can be said of the legislative and
prosecutorial responses to gangs. Both the Orange County study and the work
of Terance D. Miethe and Richard C. McCorkle in Nevada were motivated by
interest in, and concern with, overreaction to, and misidentification of gangs.
"Evaluating Nevada's Antigang Legislation and Gang Prosecution Units" (found
in chapter 6) examines the evolution of Nevada's specialized gang legislation
and how it is being implemented in that State's two largest counties (Clark and
Washoe, which contain Las Vegas and Reno, respectively). Miethe and
McCorkle employed multiple methods, including content analyses of public
hearings and records, statistical analyses of court cases, and field observations
and interviews with criminal justice officials. Their research gets to the heart of
the issue of responding to gangs and gang crime, as legislative responses can
help frame the overall criminal justice response.

When a legislature responds in an overly punitive manner, extending the reach
of the criminal justice system beyond what a "rational" analysis of the gang
problem would dictate, the system will adapt appropriately. Miethe and
McCorkle document this process quite effectively. Ironically, Nevada's gang
legislation was used by prosecutors most often for minor firearms offenses
(such as "aiming" a firearm) and less commonly against more serious and
common instances of gang crime such as driveby shootings. Equally important,
they find that after controlling for case and offender characteristics, conviction
and incarceration rates of specialized units are remarkably similar to those of
nonspecialized units. The researchers do note, however, that the threat of such
prosecutions may have played a role in defendants' deciding to plead guilty to
other charges.

One key question that law enforcement must face is whether special laws, law
enforcement units, prosecution units, or sentencing enhancements are a
necessary response to gang crime. Miethe and McCorkle conclude that
specialized units may not be necessary, as they produce little more than the
"typical" approaches to crime with regard to sentence length and probability of
incarceration. When combined with work by Klein (1995a), which suggests
that specialized gang enforcement units add little to enforcement beyond
traditional approaches, the strong implication is that specialized approaches
have little to recommend them. Although this implication raises obvious and
important questions for the criminal justice system, it must also cause
researchers to assess whether gang crime is different from other crime. If it is
not, it may be appropriate to reassess whether using "gang" as a category is a
meaningful approach.

Young Women in Gangs

Despite the veritable explosion of gang research in the past decade, insufficient
attention has been paid to female gangs and gang members. Jody Miller's
paper, "Young Women in Street Gangs: Risk Factors, Delinquency, and
Victimization Risk" (chapter 3), helps to correct that deficiency. Using field
interviewing techniques, Miller interviewed 94 young women from Columbus,
Ohio, and St. Louis, Missouri. The approach called for interviews of both gang
members and nonmembers to assess the reasons for joining gangs, the meaning
and nature of gang life, and the impact of gender on these issues. Miller cites the
role of neighborhoods, serious family problems, and gang-involved family
members as three critical factors that influenced the decisions of young women
she studied to join gangs. Indeed, these three factors were critical in
differentiating gang members from nonmembers on a variety of measures.
Interestingly, Miller reports considerable variation among female gang members
with regard to such factors as participation in delinquency and violence and
victimization experiences.

It is one thing to conduct research on young women and quite another to
consider the role of gender in their lives. Miller does a masterful job of
explaining the role of gender in gang life, particularly as regards victimization
experiences.[3] The findings presented here clearly support the argument that
gang membership increases young women's risk of victimization. This was less
true for risk of sexual assault victimization, which gang members and
nonmembers experienced at roughly similar rates, but important differences
were found between gang members and nonmembers for other forms of
victimization, particularly assaults. Although it increased their exposure to
violence and victimization, the gang was also viewed by its members as a
source of strength, owing to the protection it provided for members. Ironically,
they saw the association with male gang members as providing such protection,
particularly against sexual violence, because such exploitation occurred
primarily in the content of the family.

Although this research addresses more basic research issues, the implications
for policy are clear and significant. Gang membership shows as much variation
for young women as it does for men, suggesting that a monolithic intervention is
unlikely to be successful. In addition, gang membership creates substantial
crime victimization risks for young women, which, in turn, solidify their ties to
the gang. Designing interventions that address these factors is a major challenge
for the next decade of gang research.

A Task Force Approach

The use of task forces to respond to drug, gang, and gun problems is not new
in law enforcement. In "Evaluation of a Task Force Approach to Gangs"
(chapter 7 of this volume), Susan Pennell and Roni Melton document such an
approach, initiated in San Diego County, California, in 1988. This task force,
known as JUDGE (Jurisdictions Unified for Drug Gang Enforcement), included
representatives from law enforcement, probation, parole, the State's narcotics
enforcement agency, the district attorney, and the local department of social
services, in an effort coordinated through the State Office of Criminal Justice
Planning. Funded by the Edward Byrne State and Local Law Enforcement
Assistance Program through the Bureau of Justice Assistance, U.S. Department
of Justice, JUDGE targeted offenders under supervision by State probation or
parole. The research presented in this chapter tracks the evolution of this group
as it dealt with the ebb and flow of drug, gang, and violent offending.

As was the case in some of the other research discussed above, the research
plan could not be fully implemented despite almost ideal circumstances. Pennell
is a veteran of applied research who has strong ties to the agencies
implementing JUDGE. Despite this, the design fell through for reasons beyond
the control of the researchers. Pennell and Melton suggest that different
definitions of gangs and gang members lie at the heart of the research team's
inability to find a suitable comparison group. The initial proposal was for a
pre-post quasi-experimental design that would compare probation violations
and offense rates for juvenile probationers targeted by the JUDGE intervention
with a comparable group of juveniles who were on probation before the
program was implemented. Although this design could not adequately control
for such design threats as history, selection bias, and maturation, it still had the
potential to control several threats to the external validity of an evaluation. The
inability to conduct the evaluation as planned was a considerable loss, as we
know too little about the impact of task force approaches.

Despite the fact that the task force approach is widely accepted as an effective
means of combating drug sales and gang problems, there is scant evidence to
support this contention. As a consequence, this research represents a lost
opportunity to learn more about several important topics. First, too little is
known about the effectiveness of task forces, from either a process or an
outcome standpoint. Consequently, the full impact of Byrne-funded task forces
on drug crime, gangs, or other offenses is not yet understood. Second, the
impact of drug-focused enforcement on gangs remains unknown. One
remaining dilemma in gang research is the extent to which gangs and drugs
overlap. Assessing the extent to which drug task forces "capture" gang
members in their enforcement efforts could lead to a better understanding of this
question. It is still unclear whether specific gang enforcement yields better--or
different--results than do traditional forms of law enforcement. Finally, a major
responsibility of researchers is to document specifically where their proposed
research design went wrong and how to avoid these problems in future
research.

The Evolution of Street Gangs

The research reported by Deborah Lamm Weisel in "The Evolution of Street
Gangs: An Examination of Form and Variation" (chapter 2) attempts to assess
the degree to which street gangs are evolving into organized crime groups. This
is an important issue that bears on both basic and applied questions about the
nature of gangs. If gangs are indeed becoming organized crime groups, the
implications for law enforcement and social services are quite clear. Such
groups will require more intensive intelligence, infiltration, and targeted
prosecution--tactics that have proven effective in addressing organized crime
groups. If street gangs are not evolving toward organized crime, however, such
suppression techniques are likely to exacerbate the gang problem by
strengthening gang leadership structures and making gangs more attractive
(Klein 1995a). In addition, understanding the nature of change in gangs and the
evolution of social groups would be an important addition to the basic
understanding of gangs.

Weisel adopts a two-pronged research design to address this issue. The first
part of the research used a nationwide mail survey of police agencies to identify
law enforcement views of the nature and degree of gang organization in their
jurisdictions. The response rate (74 percent) was high, although more detail on
selecting the respondents in individual police departments would be desirable.
The second part of the research used field study techniques to interview gang
members in San Diego and Chicago. These cities were chosen because they
represent both a "weak" and a "strong" test of the hypothesis that gangs are
evolving into organized crime groups. San Diego was a strong test because it is
not a chronic gang city, and Chicago, a chronic gang city, represented a weak
test because its gangs had been in operation longer and were more established.

Not surprisingly, the two approaches yielded different results. Although Weisel
asserts that delinquent gangs were the most typical group identified by the
police, little in the definitions offered by law enforcement differentiates
delinquent groups from groups of juveniles on the one hand or friendship groups
of offending juveniles on the other. The use of accepted typologies of gangs,
such as the empirical typology offered by Klein and Maxson (1996), would
make this research more useful for practitioners and researchers. The effort to
reconcile police survey results with gang member interviews is an important step
in this process. Because of the sample size of gang members interviewed and
the lack of compatibility between the instruments used for the two groups, it is
difficult to know what the differences between the two groups mean. Weisel
faced a difficult task trying to reconcile these differences; indeed, law
enforcement expressed concern about some of her conclusions.[4] These
disparities point to a critical issue: As Weisel correctly notes, the disparity has
important consequences for the ability of law enforcement to deal effectively
with gangs. After all, if law enforcement defines gangs in ways that are
considerably different from the reality in the community, they can hardly be
successful in responding to such groups.

Gaps in Our Knowledge

Each research project, each hypothesis, spawns a dozen more. The agenda of
gang research is large and NIJ's role in addressing this research agenda is an
important one. Many important topics demand the attention of the research and
practitioner communities.[5]

Too little is understood about ethnicity and gangs, particularly for new
immigrant groups. This focus has been a historic component of gang research,
from Thrasher (1927) through Whyte (1943) to the present day and is
especially important in light of increasing levels and complexities of immigration.
Research also needs to explain more precisely the role of the media and
popular culture in the genesis and spread of gangs. Too much information about
these topics is anecdotal or generated by the media. Ongoing study of the
diffusion of gangs, gang forms, and gang symbols is necessary to better
understand this topic.

A major gap in our knowledge of gangs is the link (if any exists) between prison
gangs and street gangs. It is known that a large number of gang members go to
prison and then return to the streets, but the links that gang members maintain
between prisons and the street; the nature of relationships between imprisoned
gang members and their counterparts on the streets; and the nature, structure,
and activities of prison gangs remain largely unknown.

Despite the growing emphasis on understanding the role of women in gangs, this
area, too, needs to be expanded. And as the chapter by Miller in this volume
underscores, the role of gender remains an important topic in gang research,
whether the subjects include males, females, or both.

Many aspects of American culture, both positive and negative, are exported
around the world. This is true of gangs as well. A better understanding of the
importation and exportation of gang symbols, structure, culture, and behavior
should be an important part of future gang research.

The use of the gang as the unit of analysis is also an important topic often
ignored in gang research. After all, what makes gangs different from individuals
or institutions is the group context and group impact on behavior. This leads
naturally to the development and testing of typologies of gang structure and
behavior. Too little contemporary gang research focuses on the group context
of gang behavior.

Finally, and perhaps most important, we need to know "what works." The lack
of even basic knowledge about the impact of interventions on gangs should be a
clarion call to researchers, practitioners, and policymakers. This conclusion
applies to a variety of agencies and a variety of forms of interventions. In the
area of suppression, for example, too little is understood about the role of
specialized police units that focus on gangs, the impact of community policing
on gangs, and the long-term consequences of policies that emphasize
suppression over other aspects of intervention.[6] Very little is known about
factors that inhibit or enhance gang membership among young children below
the age at which they typically join and belong to a gang. Programs that might
affect such decisions must be built on solid research, and it is safe to conclude
that current programs have little on which to base their intervention strategies.
Too little is known about the relative merits of comprehensive, broad-based
interventions involving several agencies and targeted, single-agency
interventions. In short, there has not been enough high-quality evaluation
research on gang programs. The need for such evaluations is critical. Funding
agencies have a key role in achieving the goal of high-quality evaluation designs.
These agencies must insure that when programs are funded, the assignment of
cases to control and experimental status is a condition of funding. This must be
a high priority for the next decade of federally funded gang research.

Future Directions

Where does the future of gang research lie? The contributions to this report
suggest seven important recommendations. Klein (1995a: 138) notes that there
is a "paucity of respectable evaluations of gang intervention programs."
Unfortunately, this observation remains true and continues to serve as the
catalyst for future gang research. Thus, the first recommendation is that
meaningful gang interventions must be evaluated. The level of funding for gang
intervention programs demands evaluation and is--in Klein's words (1995a:
138)--"an inexcusable exercise in public irresponsibility." Both process and
outcome must be evaluated, as our understanding of how programs are
implemented is probably as important as the impact of such programs.

The second recommendation to emerge from these collected chapters is the
importance of documenting failure. This refers to failures both of research
design and in program implementation and impact. These, of course, are often
linked. The research community has much to learn from the results of
evaluations, but it cannot do so without full disclosure of the problems and
prospects of such evaluations. This suggests that researchers will make the
study design--as proposed and as implemented--available in full detail.

The third recommendation is that all gang research should be driven by a
conceptual understanding of the problem. The research in this report by
Esbensen, Freng, Taylor, Peterson, and Osgood is perhaps the best and
strongest argument for this position. The research team began with a
commitment to a theory-driven evaluation, which made possible an assessment
of a broader range of questions that resulted in greater utility of the findings.
This highlights the importance of addressing basic and applied questions, as
they are inexorably linked. The use of logic models in this context is appropriate
and necessary. Logic models display the underlying relationships and
assumptions behind a program or policy by linking goals and objectives to
outcomes.

The fourth, and related, recommendation is that future gang research should
incorporate some of the insights of research literature outside the gang field.
Too often, gang research is described in terms that seem to identify it as new
and unique, when a larger body of related research already exists. One
suggestion to emerge from this recommendation is that funded research must
have a commitment beyond submitting a final report and should be expected to
contribute to the broader scientific body of knowledge regarding gangs.

The fifth recommendation is that funded research should be required to adopt
sound methodologies. This means that projects should submit letters of
commitment regarding the evaluation in conjunction with the research proposal.
This procedure would allow funding agencies to ensure that programs provide
access to data and follow assignment protocols and that official data are
available to evaluators.

The sixth recommendation is that future gang research should look to
occasionally fund collaborative efforts; that is, groups of researchers who have
identified substantively or methodologically linked issues. By employing such a
tactic, it is possible to expand our knowledge of gang processes and gang
programs, particularly in multicity contexts. The use of collaborative efforts
does not stop with the research team. Strong peer review to select high-quality
products and the involvement of research and practitioner peers to provide
technical assistance, review progress of the research, and assist in overcoming
hurdles is a responsibility that NIJ should shoulder.

The final recommendation is that government-funded gang research projects
should consider the use of a dynamic problem-solving approach. The work
reported in this volume by Braga and Kennedy and by Meeker, Vila, and
Parsons reflects the importance of using research to help define problems,
structure responses, and determine the success of responses. Program
intervention seldom remains focused on the initial target. As a consequence, it is
important to document the manner in which a program differs from its blueprint.
In doing so, we can better understand how programs interact with their
environments and adapt to the needs of those environments and the individuals
such programs are designed to serve. A collaborative problem-solving
approach can best accomplish these important goals.

This introduction began by noting the questions raised by Bookin-Weiner and
Horowitz in 1983 about the end of gangs and the influence of politics in framing
gang research agendas. As for the former, there is considerable evidence that
gangs will be around for a while. The spread of gangs from urban to suburban
and rural areas has been well documented, as has their impact on popular
culture. Unlike the earlier emergence of gangs in this country, gangs in their
current form seem to be gaining both cultural and institutional foundations that
may signal their presence for some time to come.

The more interesting question raised by Bookin-Weiner and Horowitz concerns
the role of politics and research, particularly in the context of
government-funded research. It is often contended that government-funded
research necessarily supports the status quo, is repressive in nature, and
reinforces dominant political ideologies. The chapters in this report have a clear
focus on the criminal and juvenile justice systems and, especially, on the impact
of programs and policies. Yet it would be shortsighted to conclude that this is
their sole focus. Clearly the work of Miller, Esbensen et al., and Braga and
Kennedy has a strong theoretical focus that extends well beyond whether a
policy or program simply works. These chapters also have a critical tone,
expressed in Miller's conclusions about the role of gender, Esbensen's
conclusions about the impact of G.R.E.A.T., and Weisel's concern about law
enforcement views of the level of organization within gangs. But Miethe and
McCorkle capture the critical edge best, noting the failure of gang legislation
and prosecution to add much value beyond traditional efforts in these areas.
Their work illustrates that it is possible to be critical of criminal justice policies
while linking that criticism to a theoretical perspective. Indeed, researchers and
practitioners alike probably learn more from such efforts than from more
narrowly focused evaluations. In addition, these chapters illustrate the value of a
diverse set of methods and the combination of multiple methods within a
specific study.

Notes

1. The data from the evaluation are being submitted to NIJ and will be available
to researchers through NIJ's Data Resources Program (See
www.ojp.usdoj.gov/nij.)

2. It is ironic that an issue like gangs seems to lend itself to a team approach.

3. These issues are discussed in greater detail in Miller (2001).

4. Of course, gang members did not have the opportunity to express their
reservations about law enforcement's definitions and conclusions.

5. These comments are more a reflection on the state of gang research than a
direct reflection on the chapters in this volume.

6. There is, however, emerging research on these issues. See Katz (2001).

References

Blumstein, Alfred, and Joel Wallman. 2000. The Crime Drop in America. New
York: Cambridge University Press.

Bookin-Weiner, Hedy, and Ruth Horowitz. 1983. "The End of the Youth
Gang: Fad or Fact?" Criminology 21(4): 585-602.

Gottfredson, Denise. 1997. "School-Based: Crime Prevention" in Preventing
Crime: What Works, What Doesn't, What's Promising: A Report to the United
States Congress. Washington, D.C.: Office of Justice Programs. Prepared for
the National Institute of Justice by Laurence W. Sherman, Denise Gottfredson,
Doris MacKenzie, John Eck, Peter Reuter, and Shawn Bushway. Chapter 5.
NCJ 165366.

Huff, C. Ronald. 1990. "Denial, Overreaction, and Misidentification: A
Postscript on Public Policy," in Gangs in America, ed. C. Ronald Huff.
Newbury Park, CA: Sage: 310-317.

Katz, Charles M. 2001. "The Establishment of a Police Gang Unit: An
Examination of Organizational and Environmental Factors." Criminology 39(1):
37-73.

Klein, Malcolm. 1995a. The American Street Gang: Its Nature, Prevalence,
and Control. New York: Oxford University Press.

-------. 1995b. "Gang Cycles." In Crime, ed. James Q. Wilson and Joan
Petersilia. San Francisco: ICS Press: 217-236.

Maxson, Cheryl, and Malcolm Klein. 1990. "Street Gang Violence: Twice as
Great, or Half as Great." In Gangs in America, ed. C. Ronald Huff. Newbury
Park, CA.: Sage Publications: 71-100.

Maxson, Cheryl, M.A. Gordon, and Malcolm Klein. 1985. "Differences
between Gang and Nongang Homicides." Criminology 23(2): 209-222.

Miller, Jody. 2001. One of the Guys: Girls, Gangs, and Gender. New York:
Oxford University Press.

Spergel, Irving, and G. David Curry. 1993. "The National Youth Gang Survey:
A Research and Development Process." In The Gang Intervention Handbook,
ed. C. Ronald Huff and Arnold Goldstein. Champaign, IL: Research Press:
359-400.

Thrasher, Frederic. 1927. The Gang: A Study of 1,313 Gangs in Chicago.
Chicago: University of Chicago Press.

Travis, Jeremy, and Alfred Blumstein. 2000. "Departing Thoughts from an NIJ
Director." NIJ Journal 243(April): 22-29.

Whyte, William F. 1943. Street Corner Society: The Social Structure of an
Italian Slum. Chicago: University of Chicago Press.

------------------------------

Chapter 2

The Evolution of Street Gangs: An Examination of Form and Variation

Deborah Lamm Weisel, Ph.D.
Department of Political Science and Public Administration
North Carolina State University

Acknowledgments

The research presented in this chapter was supported under award
92-IJ-CX-K036 from the National Institute of Justice, U.S. Department of
Justice.

------------------------------

Abstract

If street gangs are metamorphosing into organized criminal enterprises and
large, highly organized structures, that would suggest they are becoming an even
more serious threat. To determine whether this is happening, researchers
examined the prevalence of different types of gangs and identified changes over
time in their forms and functions. They obtained information from surveys
conducted among almost 300 large police agencies and from interviews with
members of four Chicago and San Diego gangs. The police and gang members'
perceptions were surprisingly similar. Little specialization of roles and
organization in the gangs was identified; instead there was a wide range of gang
types with various structures and a great variety of criminal activities. Even
though some gangs were large and had existed for many years, and even though
some gangs exhibited some features of highly structured organizations, overall
there was little evidence of evolution into formal organizations resembling
traditional organized crime. Instead, the gangs appeared to represent an
adaptive or organic form of organization, featuring diffuse leadership and
continuity despite the absence of hierarchy. Gangs in both cities experienced
considerable organizational change over time--consolidating, merging, acquiring
smaller gangs, reorganizing, and splintering. Their "generalist" orientation may
have contributed to their ability to adapt to these changes and survive in a
volatile environment. The criminal versatility of gangs suggests that law
enforcement directed at particular criminal behavior will work primarily for
gangs that are specialized, but most are not; and enforcement and prosecution
directed at targeting gang leadership may be suitable only to the few gangs that
have distinctive leadership patterns.

------------------------------

Street gangs continue to be a pervasive problem in America's cities. They
contribute to high rates of violent crime, instill fear in citizens, and engage in a
range of troublesome behavior, from vandalism and graffiti to drug dealing and
property crime. Problems related to gangs--especially those that are more
organized, engage in serious criminal activity, or are violent--are a major
concern. Although gangs have been around since at least the beginning of the
20th century, it is plausible that at least some of these organizations are
changing, developing into criminal enterprises that may be similar in structure
and criminal activity to traditional organized crime.

Much of the evidence that gangs may be metamorphosing into
organized-crime-like enterprises is anecdotal, suggested by high-profile
prosecutions, media coverage, or the actions of law enforcement agencies.
Cases such as the 1987 conviction of members of Chicago's El Rukns on
terrorism charges, which linked that gang with Libya's Moammar Gadhafi,
raised concerns about the seriousness and possible transformation of
contemporary gangs. Other high-profile cases, in which Federal RICO
(Racketeer Influenced and Corrupt Organizations Act) charges were used to
convict gang members, reinforced these concerns. Indeed, the Federal
response to gangs bolstered the perception that gangs were becoming highly
organized criminal enterprises. In 1991, the FBI created Operation Safe
Streets, a program in which the bureau's 52 field offices participate in a series
of multiagency task forces targeting gangs and violent crime (Freeh 1999), in
which 300 agents were reassigned from counterintelligence to violent crime
investigations. The media coverage of and political reaction to gang violence
contributed to perceptions that the gang problem was becoming increasingly
serious (Jackson and Rudman 1993; Zatz 1987; McCorkle and Miethe 1998).

As a result of widespread evidence of rising gang violence and the incursion of
gangs into middle America, numerous questions have been raised about gangs.
In the 1990s, many gangs were widely described as rather disorganized groups
(Klein 1995a; Spergel 1995). As Thrasher (1963) pointed out, however,
"under favorable conditions," gangs can undergo a "natural evolution" from a
loosely organized group into a mature form.[1] How does this occur? As gangs
become more prevalent, do they become more highly organized, taking on the
features of formal organizations? Do gangs naturally become larger and develop
greater labor and criminal specialization? While a large body of literature on
organizational evolution suggests that successful organizations become larger
and more formally organized over time (see, for example, Simmel 1902-3;
Starbuck 1965; Greiner 1972; Kimberly, Miles, and Associates 1980; Staw
and Cummings 1990), scant attention has been paid to the ways in which gangs
change over time. These and other questions shaped this inquiry into the form
and evolution of contemporary gangs.

This study was conducted to generate information about the different types of
gangs and to document the changes occurring in them over time. The first part
of the study was designed to identify and describe the different types of gangs
through police sources, focusing on distinctions between the typical gang and
the more serious gangs in a jurisdiction--violent gangs, drug-dealing gangs, and
entrepreneurial or money-making gangs. The second part of this study was
designed to examine highly organized gangs to determine how they are
organized and shed light on their evolution over time.

Approach to Research

Two methods of research were used--a nationwide mail survey of police
agencies, conducted in 1995, and structured, in-person interviews with gang
members in four gangs in two cities, conducted in 1996 and 1997.

The aim of the survey, administered to 385 large municipal police agencies, was
to identify the various types of gangs and highlight distinctions among them by
organizational characteristics, demographic composition, criminal activities, and
other factors. The survey drew from Fagan's (1989) typology of gangs: violent,
drug-dealing, entrepreneurial, delinquent, or social--a typology based
predominantly on the behavior of gang members.[2] Also examined were the
evolution of gangs over time, including changing patterns of leadership,
organizational characteristics, and gang duration.

In the field portion of the study, four highly organized criminal gangs in Chicago
and San Diego were examined. "Organized" gangs are the exception, not the
rule, among the universe of gangs in the United States. In the fieldwork, the
organizational structure of these gangs was investigated and documented and
changes in these gangs over time were identified, including their transformation
or transition into groups or organizations resembling traditional organized crime
groups. Differences and similarities between gangs within and across the two
cities were highlighted.

By focusing on the most serious or organized gangs within jurisdictions, this
study, in effect, ignores the most common or typical gangs. These more typical
or prevalent gangs are no less important than the more organized gangs, and
may be more troublesome for communities on a day-to-day basis. Yet this
research sought to understand how serious gangs operated and the extent to
which typical gangs develop or evolve into more serious or mature gangs over
time. Concentrating on a particular type of gang made it possible to gather
specific information about individual gangs rather than being limited to collecting
general information about a broader range of gangs. By focusing on the "most
organized" gangs, the research team was able to home in on characteristics of
specific gangs and examine how they have evolved in recent years. As Blau and
Scott (1962: 224) note, large organizations did "not spring into existence
full-blown but develop[ed] out of simpler ones." An examination of these
mature gangs may therefore provide unique insight into the effect of
organizational processes on an important subset of contemporary gangs. Thus,
although an examination of four gangs limits the generalizability of findings to all
gangs, the greater depth of the investigation makes it possible to arrive at
reasoned judgments about the extent of organization likely among other, less
organized gangs. In addition, by examining the dynamics of highly organized
gangs, the study lays the groundwork for a reexamination of how law
enforcement monitors and responds to criminal gangs and other criminal
groups.

The mail survey was distributed to all law enforcement agencies serving
populations of 100,000 or more and to a randomly selected third of all agencies
serving populations between 50,000 and 100,000.[3] Responses were
obtained from 82 percent of agencies serving large populations and 57 percent
of agencies serving small populations--a total of 286 agencies.[4] This number
represented 74 percent of all agencies surveyed.

Two cities--Chicago and San Diego--were selected for the field portion of the
study. On their face, these two cities and their gang problems were quite
different. Chicago, with a population of nearly 3 million, reported an estimated
130 gangs and 60,000 gang members in 1997.[5] Gangs have existed in
Chicago since at least the beginning of the last century. In stark contrast, street
gangs are a much more recent phenomenon in San Diego, although the roots of
the city's gangs in its Latino car clubs date to the 1950s (see, for example,
Pennell et al. 1994). San Diego had an estimated 65 gangs and nearly 5,000
gang members in 1997. The two cities also vary in demographics, economic
conditions, urban geography, and in other important ways. In Klein's (1995a)
terms, Chicago can be characterized as a "chronic" gang city, while San Diego
is considered an "emerging" gang city. Gang crime has been estimated as linked
to more than 50 percent of crime in Chicago, while reported gang involvement
in crime is much lower in San Diego (National Drug Intelligence Center 1998).
Indeed, the two cities were selected precisely because of these differences as
well as the presumed differences in the nature and characteristics of their gangs.

Two gangs in each city were selected for study--a Hispanic and a black gang in
each. The Black Gangster Disciples (BGDs) and the Latin Kings were selected
in Chicago, and the Logan Calle Treinta/Red Steps and Lincoln Park
Piru/Syndo Mob gangs in San Diego.[6] Like their home cities, the gangs also
varied. The BGDs are one of the largest and most well-established gangs in the
country. With an estimated membership of 10,000 to 30,000, this gang has
been heavily involved in drug trafficking. It was established about 1974,
although its roots are in the 1960s. The BGDs have been remarkably tenacious
despite the conviction and death of key leaders. In the 1990s, they formed a
prosocial group called Growth and Development, which shares the same initials
as the gang, to further educational and economic objectives. The Latin Kings,
with 3,000 to 15,000 members, are also an extremely large gang. This gang
was established in the 1960s or 1970s, but its roots go back to the 1940s.

In San Diego, the gang known as Syndo Mob was a set of Lincoln Park Piru, a
predominantly black gang formed in the early 1980s with an initial membership
of 12 people. The organization was heavily involved in drug trafficking, but
some 26 members were indicted on Federal charges, and by the late 1980s its
ranks had been decimated. In the mid-1990s, the gang had approximately 165
members. In contrast to Syndo, the Logan gang factions of Calle Treinta and
Red Steps are Hispanic. The roots of the Logan gang go back to the 1940s,
while Calle Treinta and Red Steps were both established in the mid-1970s.
Each set has approximately 200 members.

The gang members interviewed for this study were identified through probation
and prison sources in each city. Researchers sought to identify gang members
who were more intensively involved with the gang and hence presumably more
knowledgeable about its organization and activities. Gang members were asked
to participate in a semistructured interview, assured of confidentiality, and paid
$20 for participating. A total of 85 gang members were interviewed--26 Black
Gangster Disciples, 18 Latin Kings, 20 from the Logan factions, and 21 from
the Syndo Mob. Of the gang members interviewed, 61 percent (52) were
contacted through the local probation department and the remainder were
identified through prison records. The sample was opportunistic and is in no
way random or representative.

Among the research issues this study examined were the nature of the
organizational structure and criminal activity of the specific gangs and gang
types. The study sought to determine the ways in which criminal gangs may be
organizationally similar to traditional crime groups. The types of criminal activity
in which these gangs engage and some of the organizational characteristics of
these gangs were also examined.

Examining gangs both from the police perspective and from within the gang
itself made it possible to compare and contrast these two (presumably quite
divergent) points of view. Because police are concerned primarily with criminal
activity, in many jurisdictions they tend to concentrate their attention on the
most serious gangs and the most serious offenders in those gangs. Nonetheless,
because local police deal with a wide range of behavior, from disorder,
vandalism, and loitering to driveby shootings and drug dealing, they have a
broad perspective on gang behavior.[7] In contrast, gang members tend to
view their gang from a different perspective, focusing more on their gang's
friendship networks than on its criminal activity.

The Police Perspective

Police agencies surveyed in this study were able to identify the various types of
gangs that coexist in their communities and how these gangs differed in some
important ways. When asked to categorize the most typical gang in their
jurisdiction, police responded that the delinquent gang is most common.
Forty-six percent of police respondents said that the typical gang in their
jurisdiction is a delinquent gang consisting primarily of juveniles who engage in
vandalism and other delinquent behavior, or a more socially oriented or "party"
gang. Twenty-six percent of respondents reported drug-dealing gangs (or other
entrepreneurial gangs) as the most typical in their jurisdiction, whereas 28
percent reported violent gangs as most typical.[8]

Overall, more police respondents described the typical gang in their jurisdiction
as a loose-knit organization (45 percent) with no formal structure (47 percent),
territorial (50 percent), and primarily oriented toward criminal purposes (60
percent). Respondents were divided in their view of the leadership structure of
the typical gang: 30 percent said that their typical gang had no formal
leadership, whereas 37 percent reported formal leadership as a component of
the typical gang.

Police see important regional differences in gang structure and activities. As
expected, the larger cities typically have more gangs, larger gangs, and gangs
that have been in existence for longer periods of time. Consistent with that
observation, these cities also tend to have gangs that are more involved in
serious criminal activity, are more highly organized, and have a more identifiable
leadership structure. Delinquent gangs, more common in smaller cities, tend to
be more loosely organized, with ephemeral leadership; these gangs are newer
and lack the historic roots of gangs established generations ago. Delinquent
gangs were reported as most typical among Southeast and Midwest
respondents (by 56 and 54 percent, respectively). In the Western States,
violent gangs were most common, whereas income-generating gangs (including
drug-dealing gangs) were reported most commonly in the Northeast (see
exhibit 1). Despite the predominance of gang type by region, large numbers of
respondents reported other types of gangs in the region. For example, 35
percent of respondents in the Northeast and 42 percent of respondents in the
West reported delinquent gangs as most typical.

The distribution of number of gangs within jurisdictions was consistent with the
findings for types of typical gang by size of jurisdiction; violent gangs were
identified less frequently in small cities, and delinquent gangs were identified
most frequently. In cities with populations of 100,000 or less, 13 percent of
respondents classified their typical gang as violent while 66 percent of
respondents in these cities classified their typical gang as delinquent. In large
cities (with populations of 200,000 or more), 44 percent of respondents
classified their typical gang as violent. Yet even a large proportion of
respondents from large cities--nearly one-third (31 percent)--reported
delinquent
gangs as their most typical gang.

Police were asked a series of questions about the structure of more serious
gangs in each jurisdiction--the violent, drug-dealing, and entrepreneurial gangs.
Police respondents most often described serious gangs as lacking a clear or
hierarchical organizational structure: More than half of respondents reported
that their violent gangs and drug-dealing gangs (51 and 56 percent,
respectively) had no clear organizational structure. A similar number reported
that violent and drug-dealing gangs had no clear leadership. Entrepreneurial
gangs were the type of gang the police viewed as most likely to feature a
hierarchy: 36 percent said that entrepreneurial gangs in their jurisdictions had an
organizational structure, and 41 percent said that these gangs had a clearly
identifiable leadership.

Criminal Activity of Gangs

Klein's observations (1995a) about street gangs and the Youth Gang Survey's
findings about youth gangs (Office of Juvenile Justice and Delinquency
Prevention 1999) were confirmed by police, who reported a great deal of
criminal versatility among their serious gangs.[9] Assaults, crack cocaine sales,
graffiti, intimidation, vandalism, violence as a means of discipline, and violence
as a means of retaliation were the most common criminal activities of gangs as
reported by police respondents.[10] Each gang type tended to favor certain
sorts of crimes. For example, entrepreneurial gangs were reported to have the
highest involvement in motor vehicle theft and theft in general, whereas violent
gangs had the highest involvement in assault, intimidation, graffiti, and vandalism.
As expected, drug-dealing gangs were the most involved in selling crack,
powder cocaine, marijuana, and other drugs, according to the police.

Yet police reported that most gangs, regardless of type, participated in many
different types of crime (see exhibit 2). For example, entrepreneurial gangs
frequently also sell crack cocaine, with 39 percent of police respondents
reporting that these gangs often or very often engage in such activity. Similarly,
violent gangs frequently commit burglary, with 36 percent of police respondents
reporting high levels of participation. Police reported that all gang types mark
their territories with graffiti, although the highest level of participation was
associated with violent gangs (67 percent of respondents). Although
drug-dealing is featured prominently in police estimates of gang activity, only
one type of gang--drug-dealing gangs--engages predominantly in this activity.

The breadth of criminal activity identified within the various gang types suggests
that while police may characterize gangs as concentrating on a specific type of
crime (such as drug dealing), gangs as criminal organizations (and their
members) have great criminal versatility, participating in a range of crimes rather
than specializing in a few crime types. Police responses in this study indicated a
clear recognition of this criminal versatility.

From the police perspective, serious gangs are changing over time in ways that
create more problems for police and the community: 78 percent of police
respondents said serious gangs had grown larger in the past 3 years, while 72
percent said serious gangs had become more violent during that period. (The
term "serious gang" combines gangs identified by respondents as violent,
drug-dealing, and entrepreneurial gangs.) Police respondents said that serious
gangs had grown larger in the past 3 years, both in size and in geographic
coverage. Some of the growth had occurred through retention of older
members who failed to leave the gang, effectively increasing the gang's size.
Fifty-four percent of police respondents reported that the average age of
members of serious gangs had increased.

Gangs had also expanded geographically over the past 3 years: 53 percent of
police respondents said serious gangs had migrated into their community, and
43 percent reported that gangs in their city had expanded to other jurisdictions,
including suburban communities. Serious gangs are also causing more problems
for police, evolving into organizations that have some features of traditional
organized crime. Seventy-two percent of police respondents reported that in
the past 3 years serious gangs were using more sophisticated weapons, and 46
percent reported that gangs had developed links with other crime groups. A
total of 19 percent of police respondents in the past 3 years said that gangs
were committing more sophisticated crime, 17 percent reported that they were
using more sophisticated technology, and 16 percent said they had acquired
legitimate businesses.

Summary of Survey Findings

Police respondents portrayed a picture of gangs that reflected recognition of a
wide array of sizes, organizational structures, and activities. Rather than
characterizing all gangs in much the same way, police drew clear distinctions
among different gangs and among different types of gangs in their jurisdictions.
Notably, most police respondents did not appear to stereotype serious gangs
as highly organized or highly specialized.

For this study, information from the police provided a context for examining the
organizational evolution of serious gangs from a different perspective--through
the eyes of gang members. Because of their institutional objectives, police tend
to look at gangs from the perspective of their criminal behavior; examining gang
structure, leadership, and noncriminal gang activities are of secondary
importance. Yet understanding how gangs emerge, grow, and evolve over time
has substantial implications for police in developing effective strategies in
response to the gang problem.

The Perspective of Gang Members

In the second part of this study, the nature of four gangs and their changes over
time were examined from the perspective of gang members. The gangs were
the Syndo Mob/Lincoln Park Piru and Calle Treinta/Red Steps in San Diego
and the Black Gangster Disciples and Latin Kings in Chicago. Members of two
of the gangs described their gang and its operations as disorganized. By
measures of formal organization, two of the four gangs were found to be highly
organized, but only one could be characterized as exhibiting features of
traditional organized crime.

The gangs were examined with a view to determining whether they exhibited
the characteristics of formal organizations or bureaucratic structures, including
leadership, role differentiation, participation in formal meetings, compliance with
formal rules and discipline, specialization, and goal orientation.[11] The two
Chicago gangs were higher on every measure of organization than those in San
Diego. Both Chicago gangs featured more formal and distinctive roles of
leadership, more explicit (and even written) rules and clearer consequences for
breaking them, more routinely held and purposeful meetings, and the collection
of dues. Only in the Black Gangster Disciples were there high levels of
relationships with neighborhood businesses (including ownership and control of
these businesses), relationships with gangs in other cities across the country,
formal contacts with prison gangs, and involvement in political activities. These
features are also characteristic of organized crime.

Although some law enforcement agencies, such as the FBI, would characterize
San Diego's Latin Kings as highly organized (see the sidebar on the FBI's
Enterprise Theory of Investigation), the gang members interviewed did not
uniformly support that view. Some Latin Kings described some features of
formal organizations, but most did not. This inconsistency between the study
findings and the perspective of Federal law enforcement agencies is not
unexpected. The FBI's focus is on economic enterprise where there is gang
leadership; this view may exaggerate the role of older or more crime-involved
gang members. Although leaders, adult members, and heavy crime involvement
may be present within the gang, these features are unlikely to characterize the
gang as a whole.

Unlike the Chicago gangs, the gangs examined in San Diego featured little
formal leadership, and were described by the gang members interviewed as
primarily friendship and kinship networks rather than criminal enterprises.
Gangs in Chicago also included these traits of friendship and brotherhood, but
they were subordinate to the objectives of economic opportunity and
protection.

Many of the organizational differences between Chicago and San Diego gangs
may be attributable to the vastly different size of the gangs and the respective
gang population. Although Chicago is about twice as large as San Diego in
population, it has nearly 12 times more gang members. Of the gangs studied,
the largest Chicago gang is approximately 25 times larger than either San Diego
gang. Even by the most conservative estimates of the number of BGD
members, this exponential difference in size likely contributes to much of the
organizational characteristics.

Criminal Activity of Gangs

Like the police respondents, gang members in both cities reported that their
gang is extensively involved in a wide range of criminal activity. Indeed, gang
members reported much greater participation of their gang in specific criminal
activities than police attributed to specific serious gangs. In the interviews, gang
members reported about two to three times as much criminal activity as did
police.

The wide variety of criminal activity reported by gang members indicates little
specialization of the gang as a criminal enterprise (see exhibit 3). Again, this
finding is consistent with Klein's (1995a) depiction of gangs as criminally
versatile. Assaults and drug sales were the activity most often reported, but
most gang members said that their gangs were involved in almost every criminal
activity. In fact, the only exception was the Latin Kings' involvement in
shootings: Only for this crime did less than a majority indicate their gang was
involved.

Although the San Diego gangs were less formally organized, the members
interviewed reported levels of criminal behavior as high as those reported by
the Chicago gang members. In fact, for all but three crimes, gang members from
San Diego reported even higher gang participation than Chicago gangs. Gang
members from the two cities reported similar gang participation rates in
assaults, driveby shootings, and crack sales.

The high levels of participation in a wide range of criminal behavior by all four
gangs suggests that these gangs can be considered criminal generalists. Indeed,
according to organizational theory, organizations operating in highly volatile
environments are much more likely to be generalists than specialists (Katz and
Kahn 1966; Meyer 1978). Generalist organizations can adapt more quickly to
changing conditions and are more likely to survive than specialists because the
latter must learn a new set of complex skills to create a new specialty (or
"niche") under changing conditions. While organizational specialization may be
sustained under conditions of short-term volatility, it will not be useful in the
continuously volatile gang environment (Katz and Kahn 1966), in which there is
ongoing competition with other gangs and a presumed high level of attention
from police and other criminal justice organizations.[12] Such environmental
volatility would likely discourage specialization both within and among gangs.

Organizations can also be classified by the extent of specialization of their
individual members. Like the gangs they belonged to, the members of all four
gangs studied reported participating in a wide range of criminal activities, from
vandalism to drug sales (see exhibit 4). When group criminal activities were
compared with individual criminal activities (exhibits 3 and 4), it was found that
gang members participated substantially less in some of the gang's criminal
activities than did the gang itself. This finding could reflect a respondent's
reluctance to identify all of his individual criminal behavior,[13] but it also
suggests that there are criminal activities in which some gang members
participate but others do not.

All gang members reported participating in a wide range of criminal activities. A
majority of all gang members reported that they had engaged in assaults (74
percent) and sold marijuana (68 percent). Nearly half had engaged in motor
vehicle theft, theft, burglary, and vandalism.

The differences between the two cities in gang member participation in criminal
activity were consistent with the differences in gang participation. Gang
members in San Diego reported greater individual participation in almost all 11
crime types than did Chicago gang members. A large majority (90 percent) of
San Diego gang members reported individual involvement in assault. San Diego
gang members also reported greater involvement in assault and marijuana sales
than gang members in Chicago. When members' responses from each city were
combined, it was found that gang members in Chicago were more involved in
cocaine sales and shootings than were gang members from San Diego. Gangs in
the two cities reported similar rates of involvement in motor vehicle theft.

The breadth of criminal activities reported by individuals helps confirm the
notion that most gang members are criminal generalists much like the gangs to
which they belong. Just as generalist organizations adapt more quickly to
changing environmental conditions, generalist gang members are inherently
easier to replace than specialists. Because they report involvement in fewer
types of crime, the Chicago gangs and their members appear to be slightly more
specialized than the San Diego gangs and gang members.

Goal Orientation

Centrality of goals is a key and defining feature of formal organization (Parsons
1987; Weber 1947; Blau and Scott 1962; Katz and Kahn 1966; Stinchcombe
1965; Lippitt 1982). Organizations are established to attain certain goals and
are structured to maximize their attainment. It is widely recognized, however,
that organizational goals may be vague, changing, numerous, contradictory, and
not always closely linked with the organization's day-to-day activities. In some
forms of organization, the goals of individuals and those of the organization are
consistent, and the former may be thoroughly integrated into the latter
(Popielarz and McPherson 1995; Scott 1993).

Among the features of formal organization in the gangs Thrasher (1963)
studied, he recognized that they were directed toward goals. According to
Klein (1995a), their members, on the contrary, typically have a "rather low
focus on group goals" because these goals may come into conflict with the
individual needs of gang members. "Gangs," says Klein, "are not committees,
ball teams, task forces, production teams, or research teams. The members are
drawn to one another to fulfill individual needs, many shared and some
conflicting: they do not gather to achieve a common, agreed-upon end" (1995a:
80). Yet group rewards, Klein contends, are an important individual motivation
for joining a gang. These include status, companionship, excitement, and
protection. Among individual motivations, gang members routinely join for a
sense of "belonging" or of "family." Material rewards associated with group
crime are also a factor in promoting gang membership.

Among gang members studied here there was strong evidence of organizational
goals and purposefulness focused on making money.[14] These economic
objectives were often described in quite varied terms, and organizational goals
were often embedded in broader descriptions of the purpose of the gang. Many
gang members showed evidence of integrating or blending the larger
organizational objectives with their own individual needs. Among San Diego
gang members, moneymaking appeared incidental to the friendship and social
networks. As San Diego gang members put it--

We kick it together, smoke marijuana, maybe jack something. If we think we
can make some money selling it, we will take it.

Throughout the whole time [I've been in the gang] I was a party cat. . . . The
one thing I did throughout the whole time was party. I always liked to party.
Liked to have fun. If you are going to have fun you got to make your money.
So you sell drugs or you do violent crimes to get money.

We get together and have little picnics and things with the community.
Sometimes we get together and set up moves to make on other gangs. When
you go to other neighborhoods and they selling certain things we go and take
theirs, we make moves. . . . We do more transactions, money transactions. We
mainly making money. That is a main part of the gang, making the money. The
riding on other neighborhoods and shooting and all that, that's part of the gang
too but the main thing is getting our money.

[We] smoke weed, sell dope, I would say that's about it that I know of. [We]
have parties and stuff. . . . [We p]robably hang out, that's about it, make
money.

In contrast, gang members from Chicago tended to describe their gangs as
primarily focused on making money:

Well, in my words, a gang ain't nothing but people come together to do crime
and make money and be a family to each other. That's the original idea.

A gang, nowadays, would be money making, make money.

[A gang is] a bunch of brothers hooked up, trying to make money.

[The gang is] really [about] making money, it's holding your own neighborhood
so nobody can come into your neighborhood and try to take the bread out of
your mouth.

But Chicago gang members also articulated the objectives of the gang in a
broader way that emphasized the social and familial rewards of gang
membership:

A gang is a group of individuals bound together for a common purpose. [They
are bound together] a lot of ways, socially, economically, emotionally
sometimes.

Basically [a gang is] a group of people with the same objectives, trying to reach
for the same goals.

Gang members in Chicago and San Diego viewed their gangs as an opportunity
for social interaction (including partying, "hanging out," getting women) and as a
family, brotherhood, or support system. Gang members in both cities described
the role of the gang as a means of protection from rivals, a means of survival,
and a source of respect. Many gang members appear to seamlessly combine
social interaction with making money: For San Diego gangs, social interaction,
friendship, and self-protection appeared to be the primary purpose, and
moneymaking was incidental or opportunistic. Although the concept of gang as
brotherhood or family was also extremely important to the Chicago gangs,
making money was their central organizational objective or defining feature.

Organizational Transformation: Consolidation and Splintering

Perhaps the strongest evidence of the formal organizational character of the
gangs studied was the organizational transformation each had experienced over
time.[15] Like other organizations, gangs are seldom static entities. Although
the temporary gangs described by Thrasher (1963) typically disintegrated, the
gangs in this study reflected patterns of consolidation (primarily through merger
with, or acquisition of, smaller gangs), reorganization, and the splintering of
larger gangs into spinoff gangs. Such findings are consistent with those of
Thrasher, who noted that groups form and re-form over time.

Gang members' richly detailed descriptions of the organizational transformations
that occurred in each of the gangs were evidence of both the tenacity and
enduring nature of the gang and offered insights into its organizational growth or
decline. One gang member described the merger that created the BGDs:

[Black Gangster Disciple Nation, Black Gangsters, Disciples, and High
Supreme Gangsters] actually was all brought together in 1981. Although you
had the same members that came together, all these different gangs all came
together under one name. It's like we'll take some of your doctrines, we'll take
some of your doctrines to satiate everybody's upbringing, what they originally
were, and bring them all together under one thing. . . . There was many different
smaller organizations just in the process of being brought together. And then a
couple of years later everybody was brought together under one law, under a
one-people concept, all of us being the same thing.

Two other gang members told a similar story of merger and acquisition:

Before I became a Black Disciple, I was a Rod; it was an extremely small
organization and we converted over to GD's, to Black Gangster Disciples. . . .
Disciples been around for ages. Like I said, in the beginning there was Devil
Disciples. But the GD organization has been in existence since 1971.

They were attempting to form a conglomerate. . . . [I]t was three organizations
that come together and formed one big organization.

Unlike the Chicago gangs, which showed evidence of mergers, gangs in San
Diego appeared to exhibit a pattern of splintering through the division of larger
gangs and the creation of new, spinoff gangs. Logan gang members described
the evolution of their gang this way:

Thirtieth Street was originally all of Logan. Then along came another 'hood
called Logan Trece--another little gang. So they had to get permission from
30th
to start their little gang. Give them a little part of Logan. So Red Steps came
along and asked them too. So Red Steps and Logan Trece to gain some
respect from other gangs, they had to start fighting and all that. So that's where
it all started.

There is two more [gangs] in Logan Heights. There is Red Steps and Trece
Logan Heights. But back in the old days, it used to just be Logan. But as time
went by, they started separating because of freeways getting built and
boundaries started separating them apart. But they are still united though.
Except there is always family disputes between the different family cliques.

Spergel (1990: 204) believes that the reason for splintering is "competition
between cliques [within the gang]." Such competition, he wrote, "may be a
central dynamic leading to the gang splitting into factions or separate gangs." It
seems reasonable that external pressures--law enforcement, neighborhood
dynamics, competition with other gangs--and other "pull" factors could also
splinter the gang. These same dynamics could also lead small gangs to join
forces to provide protection or to form economic alliances.

The notion of gang mergers is occasionally mentioned in the gang literature but
has not been fully discussed. During the 2-year period in which Huff (1989)
studied gangs in two Ohio cities, the number of gangs in Cleveland declined
from 50 separately named gangs to 15 or 20; in Columbus, the number of
gangs dropped from 20 to 15 during the course of the study. Mergers
accounted for the reduction in the number of gangs, according to Huff, although
some gangs dissolved and some groups originally identified as gangs may have
actually been splinter groups rather than gangs. While mergers may result in
larger gangs (that is, more gang members in a gang), such an increase in size
does not necessarily take place. Mergers may serve only to offset attrition of
gang members, resulting in no net increase in gang size.

As Monti (1993) noted, gang cliques and sets can combine and reassemble in
different ways over time; a portion of the gangs he studied in St. Louis were
"absorbed" into other gangs during a 2-year period. The growth of the
Blackstone Rangers in Chicago was described by Sale (1971) as occurring
through takeovers of existing gangs and "renovation" of cliques. The gang's
original street clique clashed with rival gangs and then later combined with
them. The result, after 10 years, was a much larger version of the Blackstone
Rangers. The observation that the merger is an organizational feature of
contemporary gangs and occurs over time is an accepted but poorly
understood dimension of the organizational growth of gangs.

Similarly, there has been little discussion in the field of gang research of
splintering, although there is recognition that cliques or subsets of gangs have a
life of their own. Spergel et al. (1991) reported that internal competition within
the gang may cause it to split into factions or form a separate gang. He also
suggested that gangs might splinter and dissolve if more criminal opportunities
become available to members through drug trafficking gangs or other criminal
groups. According to Goldstein and Huff (1993), there is serious intragang
rivalry between sets within the Bloods or the Crips in Los Angeles, especially
when the profits of drug dealing are at issue. They note that there can be as
much violence between different sets of the same gang as between rival gangs,
a fact that may contribute to further splintering of the gang. Decker (1996)
described how the rise of violence in larger gangs can result in the emergence of
splinter gangs. Monti (1993) suggested that when gangs reach a certain size,
they would split when friction occurred among members and remained
unsettled. This splintering occurs because, when the gang is small, the gang can
exercise cohesion and control through face-to-face personal interactions
(Kornhauser 1978), but large increases in the size of individual gangs appear to
lead to breaches within gangs, resulting in more gangs in a jurisdiction.

The concepts of gang merger and splintering can be framed within
organizational theory. Organizational theory holds that populations of
organizations of a similar form and function tend to reach an equilibrium. Over
time, some organizations die out and others form--some through schism, which
occurs when subgroups break away to create a new organization. The process
continues until the number of organizations is stable (Tucker et al. 1988;
Hannan and Freeman 1987; Hannan and Carroll 1992). Historically, this
phenomenon is driven by organizational creation and failure--two processes
that
are much more common than adaptation.

Summary of Gang Member Interviews

The gang members interviewed in this study provided richly textured
descriptions of the character of their gangs. Framed in an organizational
context, their narratives illuminate and clarify police observations about gang
size, gang activities, and changes in gangs over time. The evidence from these
interviews demonstrates that the four gangs studied have been in existence for
many years and have experienced major organizational changes. Changes in
gang name and size are the most observable indications of these organizational
shifts. The generalist orientation of the gangs and their members may have
contributed to their survival through periods of organizational upheaval and
environmental uncertainty.

Unexpected Agreement

It might be expected that the views of police and gang members would be quite
divergent on issues such as the nature of gang leadership and activities. This
study, however, did not identify any major differences. In large part, police
descriptions of gangs were not inconsistent with those of gang members.
Indeed, the interviews with gang members tended to elaborate on police
observations of the variation within and among gangs. For the most part, the
police did not tend to stereotype gangs; they did not, for example, describe
them as showing evidence of formal organization and hierarchy where this was
not well established. Nor did the police tend to see role and organizational
specialization within gangs. Instead, they recognized a wide range of gang
types, identified different structures within different gangs, and described a wide
range of behaviors of gangs and gang members. In the interviews, gang
members confirmed and elaborated on these observations.

It is worth noting that both police and gang members indicated that the large
and enduring gangs examined in this study exhibited some distinctive features of
formal organizations. There was little evidence, however, that their structure is
highly bureaucratic--an anomaly, according to organizational theory. Despite
their large size and organizational longevity, of the four gangs examined, in only
two was there evidence of the bureaucratic structure of large organizations
(leadership, membership levels, regular meetings, specialization, and written
rules). Only one gang exhibited the more elaborate features of traditional
organized crime groups--relationships with other gangs nationwide (including
prison gangs), incursion into legitimate businesses, and involvement in political
activities (Decker, Bynum, and Weisel 1998). In all four gangs, however, there
was some evidence of formal organization, such as organizational continuity and
an orientation toward goals.

The gangs studied here appear to represent a fundamentally different form of
organization, one that can be described as adaptive or organic rather than
bureaucratic. These forms of organization have also been called "federations,
networks, clusters, cross-functional teams, lattices, modules, matrices, almost
anything but pyramids" (Bennis 1993).[16]

In stark contrast to the myriad literature on bureaucratic organizations, there
have been few studies of organic organizations--and thus there is a dearth of
descriptions of this form even among organizations such as legitimate
businesses. An examination of the recognized features of the organic-adaptive
model, however, suggests that the gangs studied here feature the attributes
associated predominantly with this form of organization. These include an
emphasis on individual goals concurrent with organizational goals, diffuse
leadership, the active role of subgroups, a generalist orientation, persistence in a
volatile environment, and continuity despite the absence of hierarchy. Adaptive
or organic organizations thrive in a volatile or changing environment, and
organizations that survive under such conditions are more likely to maintain
multipurpose and flexible structures, with flexible leadership and little
differentiation among member roles (Meyer 1978; Burns and Stalker 1961).

Despite some evidence of bureaucratic features in the gangs studied, the results
of the interviews with gang members make it difficult to support the hypothesis
that these gangs have evolved into formal organizations mirroring traditional
organized crime.[17] Although the Chicago gangs have some explicit features of
formal organizations, on other dimensions, the evidence is less clear: Group and
role specialization, for example, appear minimal.

Their organizational continuity and their expansion and membership growth
suggest that these four groups can be considered successful organizations. They
have endured and thrived at times when environmental exigencies might have
logically selected against them. Their survival and growth have been punctuated
by organizational changes--mergers, splintering, consolidation, and other
organizational dynamics. These changes have been responsible for the gangs'
growth to their current, large size.

The substantial organizational change these gangs experienced is consistent with
organizational theories that see social organizations as not static but changing in
important ways, adapting to changing environmental conditions. According to
one widely held notion, organizations proceed through temporal and sequential
stages of development, a process commonly known as the organizational life
cycle. The gangs in this study--with their patterns of consolidation and
fragmentation--showed clear evidence of such a process. According to Klein
(1995a), the proliferation of gangs in the 1980s resulted from the establishment
of many small, autonomous organizations. The result was that gangs were large
in number but small in size. As Klein points out, however, "there are a lot of
acorns out there that could become stable, traditional oaks" (p. 104). In other
words, small, autonomous gangs can grow into stable, traditional, and much
larger gangs, the latter, of course, being of far greater concern to the public and
to police.

Life cycles of gangs suggest that smaller, socially oriented gangs can evolve into
more serious gangs, often merging or aligning with larger or more organized
gangs for protection. While transformation into large, networked gangs such as
the BGDs is clearly an exception, for that gang, the merger and acquisition
process was a major contribution to its growth. Of course, gangs are also
growing larger for a number of reasons, among them that the age range is
expanding, with members remaining in the gang longer and the gang retaining
members who move outside the neighborhood.

Size will continue to be a major factor in predicting the extent of organization in
a gang. The largest gang in this study showed the clearest evidence of
bureaucratic organizational features, suggesting the need for a systematic
process for counting gangs and their members and for close attention to
changes that affect the number and size of gangs in a specific area or
jurisdiction. Klein (1995b) describes the cyclical processes of the seasonal and
epochal variations in gang activity (crime) through which gangs proceed. Many
of these epochal cycles, Klein believes, are city-specific and may reflect upturns
and downturns in sub-areas and neighborhoods within cities. This view, of
course, suggests that cities or counties are the relevant population boundary for
monitoring gang crime and changes in the size and number of gangs. As Klein
(1995b) states, aggregate numbers of gangs, gang members, and gang crime
tend to mask important changes in gangs that occur at smaller geographic
levels.

Implications for Police Practice

For the police, monitoring the growth of individual gangs or of organizational
changes taking place among or within gangs in a jurisdiction provides insights
that can aid in developing effective responses to gang problems. Differences
among gang types and among specific gangs--especially organizational
differences such as patterns of leadership, membership age, size, duration,
criminal involvement, and so forth--have significant law enforcement
implications. Hierarchical or organized crime models that target gang leadership
by using vertical prosecution and applying Federal statutes[18] may be
appropriate for law enforcement agencies or prosecutors that focus on the few
gangs having particularly distinctive leadership patterns. This model is probably
not useful for addressing the vast majority of the country's street gangs. Spergel
(1990), for example, has warned against exaggerating the organized character
of gangs. Such exaggeration may be a byproduct of an organized crime model
that targets gang leadership, which tends to characterize most gangs by the
troublesome features of a few gangs. Indeed, the present study indicates that
street gangs do not necessarily progress into highly organized crime
organizations. Organizational processes, however, may contribute to marked
increases in gang size, influencing the structure, operations, and relationships of
some gangs.

The criminal versatility of the gangs and gang members observed in this study
suggests that law enforcement efforts that target particular criminal behavior will
work primarily for highly specialized gangs. Most gangs are not. Even the most
troublesome gangs in this study appeared to be highly adaptive generalists.
Most research suggests that rather than employing generic antigang strategies,
approaches to gang problems should be framed very narrowly to address
identified problems of concern to communities and police (e.g., Sherman et al.
1998; Bureau of Justice Assistance 1997).

Gangs and gang-related problems vary. Indeed, Spergel's model requires that
law enforcement agencies carefully assess local problems before implementing
any antigang strategy. Because there have been few useful evaluations of such
strategies, there is a need for rigorous evaluations, which could help move the
country to more quickly identify the most effective methods for solving the
various gang problems.

Police have tried a wide variety of measures to address the problem of gangs at
the local level. They have employed situational crime prevention, for example,
altering the flow of vehicular traffic to reduce gang-related violence[19];
enforcing antiloitering statutes to keep gangs from intimidating and menacing
community members (Regini 1998); using civil injunctions to keep gang
members out of areas where they cause trouble (Cameron and Skipper 1997;
Gibeaut 1998; Regini 1998); setting up traffic checkpoints (Crawford 1998);
carrying out aggressive curfew and truancy enforcement (Fritsch, Caeti, and
Taylor 1999); and cracking down on weapons violations, often using Federal
laws that impose stiffer penalties. Some jurisdictions have used a technique
known as "lever pulling," targeting specific chronic offenders with warrants,
close supervision of probation conditions, and other measures.[20]

Many of these law enforcement approaches have been integrated into
community or problem-oriented approaches to gangs. The approaches also
include mediation, situational crime prevention, working with families, and other
strategies (Sampson and Scott 2000; Bureau of Justice Assistance 1997).
Klein (1997) warns that policing that involves only enforcement will solidify
gangs by increasing cohesion among gang members. Policing strategies are
most effective when teamed with intervention programs such as providing
economic opportunities, job training, remedial education, and other services
and community involvement (see, for example, Spergel 1995; Spergel et al.
1991). These diverse strategies may be necessary to deal with highly versatile
and adaptive gangs.

At the turn of the 21st century, a wide variety of organizations fall under the
umbrella category of "gangs." One type or size of criminal organization may
differ from another only by a matter of degree. Howell and Decker (1999)
distinguish between drug gangs and street gangs, and between youth gangs and
adult criminal organizations. These distinctions (and others) will have the most
value when employed at the local level. As Starbuck, Howell, and Lindquist
(forthcoming) point out, their changing nature makes it increasingly difficult to
categorize and characterize them. These authors describe the new gang form as
a hybrid organization. Even 70 years ago, Thrasher (1963 [1927]) recognized
that the distinctions between gangs and other criminal organizations crime were
rather illusory. There is, according to Thrasher, "no hard and fast dividing line
between predatory gang boys and criminal groups of younger and older adults.
They merge into each other by imperceptible gradations." It is important to
identify and monitor the inherent distinctions and similarities among different
criminal organizations within jurisdictions, since these elements reflect local
conditions, criminal opportunity, and other explanatory variables. Such an
examination will aid in building an accurate perception of local problems and
will be useful in avoiding the stereotypes and overreaction that may lead to
well-intentioned but misguided policies and practices. Just such a continued
examination may aid in building a corpus of information that will offer greater
insight into the form and structure--and inherent variation--of contemporary
gangs.

Notes

1. Not all gangs "mature"; indeed, many disintegrate. Some gang researchers
have described this maturation process. Knox (1994) characterizes gangs by
their stage of development, from pre-gang to emergent gang, from crystallized
gang to formalized gangs, suggesting that gangs move from one category to
another as they grow larger and more like a formal organization. Thrasher
described the transformation of gangs to a solidified form as a function of
longevity, conflict, and the age of their members (1963: 47-62).

2. Fagan (1989) examined gang participation in criminal activity and drugs,
classifying gangs as party gangs, which engage in few nondrug criminal
behaviors except vandalism; social gangs, which engage in few delinquent
activities; delinquent gangs, which engage in violent and property crime but few
drug sales; and organized gangs, which are extensively involved in the sale and
use of drugs along with predatory crime. Other typologies of gangs based on
police information have taken different approaches. For example, Maxson and
Klein (1995) identified common structures for 59 gangs and looked at how
offending was related to these structures. As these authors note, however,
"police attend far more to gang crime than gang structure," leading the present
study to focus on gang behaviors as an organizing characteristic.

3. Surveys were addressed to police chiefs, who were asked to have it
completed by the person in the department most knowledgeable about gangs.
As a result, respondents varied from police chiefs to investigations commanders
to gang or youth unit supervisors. Since populations served by respondents
varied from 50,000 to more than 3 million, it seemed appropriate that
knowledgeable police respondents could be identified at different levels of
different organizations. The technique of surveying the police chief is also used
by the National Youth Gang Survey (see Office of Juvenile Justice and
Delinquency Prevention 1999).

4. The higher response rate from larger police departments likely reflects the
greater prevalence of gangs in large jurisdictions; smaller jurisdictions with few
or no gangs were probably less likely to respond to the survey.

5. Estimates of numbers of gangs and gang members vary by source. City-level
estimates were provided by the National Youth Gang Center based on annual
surveys conducted in 1996-1999.

6. Federal and local law enforcement sources were used to identify the most
organized gangs in each city. General gang history was provided by local police
sources.

7. In contrast to local law enforcement, which focuses on the criminal behavior
of individual gang members, Federal responses to gangs, such as the FBI's Safe
Streets Task Force, appear to focus on the structure of gangs, especially
leadership, using tools such as Racketeering Influenced and Corrupt
Organizations (RICO) statutes to dismantle gangs. Indeed, the stated policy of
the FBI is to address violent street gangs through long-term, proactive
investigations by concentrating on criminal enterprise and conspiracy. See
Freeh (1999) and the sidebar "The FBI's Enterprise Theory of Investigation"
for a more detailed description of the FBI's work on gangs and organized
crime.

8. Fagan (1989) notes that social gangs accounted for 28 percent of all gangs;
party gangs accounted for 7 percent; serious delinquents constituted 37
percent; and "organization"-type gangs represented 28 percent of all gangs.
These proportions varied, however, from one city to another. In Chicago,
gangs were predominantly serious delinquents and organized gangs; in Los
Angeles, gangs were social (38 percent) and serious delinquents (36 percent),
whereas San Diego gangs consisted of more serious delinquents (39 percent)
and organized gangs (31 percent).

9. Klein (1995) called this versatility "cafeteria style" crime, a type in which
gang members combine opportunistic crime with crime requiring more planning.
It should be noted that participation of the gang in crime is different from
participation of individual gang members in crime; the former term aggregates
activities of gang members, the latter reflects individual behavior. Both concepts
of criminal activity are examined in this study.

10. Although this study revealed higher levels of criminal versatility in serious
gangs, the findings of this study are parallel with findings about criminal activity
of youth gangs in the National Youth Gang Survey (Office of Juvenile Justice
and Delinquency Prevention 1999). Police participants in the youth gang survey
reported criminal versatility among gang members, with high involvement by
youth gang members in aggravated assault, larceny/theft, motor vehicle theft,
and burglary. Slightly more than one-fourth of all NYGS respondents reported
high involvement by gang members in those crimes. If youth gangs evolved into
the more serious criminal gangs examined in this study, one would anticipate
substantial increases in assaults, robbery, theft, and burglary and some increase
in motor vehicle theft.

11. An earlier report of the findings from this study described the extent to
which these organizational characteristics were present in these four gangs
(Decker, Bynum, and Weisel 1998).

12. One respondent described the volatile environment as follows: "Standing on
the street corner and talking and sitting there getting high all day, you make
plenty of money, I'm not gonna lie about that, but it gets tiresome always
looking for the police too or looking for somebody that is gonna try to kill you
for your money."

13. Taylor (1990a) and Joe (1993) reported that older gang respondents
minimized gang activity, and Goldstein (1991) noted that gang members may
exaggerate or hide information. Although the gang members interviewed for this
study appeared to have been mostly honest in their responses, it seems
reasonable to assume that many would be disinclined to own up to the range of
their own criminal activity.

14. Gang members were asked to define a gang, to describe the purpose of the
gang (including why they joined the gang), and to describe what is good about
being in a gang.

15. Gang members were asked to tell interviewers about the history of their
gang.

16. Similar terms include negotiated order, federation, loosely coupled system,
temporary system, organic-adaptive organization, coalition, external model,
post-bureaucracy, colleague model, interactive organization, network, and
blended or open organization.

17. As with contemporary gangs, there is a great deal of debate about the
definition and degree of organization of organized crime. See Maltz (1985) and
Kenney and Finckenauer (1995) for a discussion of this issue as it relates to
organized crime.

18. Limitations of these strategies are described by Johnson, Webster, and
Connors (1995) and Miethe and McCorkle (1997).

19. See Lasley (1998) for a discussion of Operation Cul de Sac, an example of
a situational crime prevention tactic.

20. "Lever pulling" and RICO statutes were employed in a widely publicized
reduction of violence program in Boston. See Kennedy (1997).

References

Bennis, Warren G. 1993. Beyond Bureaucracy: Essays on the Development
and Evolution of Human Organization. San Francisco: Jossey-Bass.

Blau, Peter M., and W. Richard Scott. 1962. Formal Organizations: A
Comparative Approach. San Francisco: Chandler Publishing Co.

Bureau of Justice Assistance. 1997. Addressing Community Gang Problems: A
Model for Problem Solving, Washington, DC: U.S. Department of Justice,
Bureau of Justice Assistance. NCJ 156059.

Burns, Tom, and G.M. Stalker. 1961. The Management of Innovation,
London: Tavistock.

Cameron, Jeffrey R., and John Skipper 1997. "The Civil Injunction: A
Preemptive Strike Against Gangs," FBI Law Enforcement Bulletin, 66(11):
12-15.

Crawford, Kimberly A. 1998. "Checkpoints: Fourth Amendment Implications
of Limiting Access to High Crime Areas," Law Enforcement Bulletin 67(3):
27-32.

Decker, Scott. 1996. "Collective and Normative Features of Gang Violence,"
Justice Quarterly, 13(2): 243-264.

Decker, Scott H., Timothy S. Bynum, and Deborah Lamm Weisel. 1998. "A
Tale of Two Cities: Gangs as Organized Crime Groups," Justice Quarterly
15(3): 395-425.

Fagan, Jeffrey. 1989. "The Social Organization of Drug Use and Drug Dealing
Among Urban Gangs," Criminology 27(4): 633-667.

Freeh, Louis J. 1999. "Ensuring Public Safety and National Security Under the
Rule of Law: A Report to the American People on the Work of the FBI
1993-1998," Washington, DC: U.S. Department of Justice, Federal Bureau of
Investigation.

Fritsch, E.J., T.J. Caeti, and R.W. Taylor. 1999. "Gang Suppression Through
Saturation Patrol, Aggressive Curfew, and Truancy Enforcement: A
Quasi-Experimental Test of the Dallas Anti-Gang Initiative," Crime and
Delinquency 45(1): 122-139.

Gibeaut, John. 1998. "Gang Busters," ABA Journal, 84(January): 64-68.

Goldstein, Arnold P. and C. Ronald Huff, eds. 1993. The Gang Intervention
Handbook. Champaign, IL: Research Press.

Greiner, Larry E. 1972. "Evolution and Revolution as Organizations Grow,"
Harvard Business Review, 50(July-August): 37-46.

Hannan, Michael T., and Glenn R. Carroll. 1992. Dynamics of Organizational
Populations: Density, Legitimation, and Competition. New York: Oxford
University Press.

Hannan, Michael T., and John Freeman. 1987. "The Ecology of Organizational
Founding: American Labor Unions, 1836-1985," American Journal of
Sociology 92(4): 910-43.

Howell, James C., and Scott H. Decker. 1999. "The Youth Gangs, Drugs and
Violence Connection," Juvenile Justice Bulletin. January. Washington, DC: U.S.
Department of Justice, Office of Juvenile Justice and Delinquency Prevention.

Huff, C. Ronald. 1989. "Youth Gangs and Public Policy," Crime and
Delinquency 35(4): 524-537.

Jackson, Patrick, with Cary Rudman. 1993. "Moral Panic and the Response to
Gangs in California," in Gangs: The Origins and Impact of Contemporary Youth
Gangs in the United States, ed. Scott Cummings and Daniel J. Monti. Albany:
State University of New York Press: 257-275.

Jankowski, Martin Sanchez. 1991. Islands in the Street: Gangs and American
Urban Society. Berkeley: University of California Press.

Joe, Karen. 1993. "Issues in Accessing and Studying Ethnic Youth Gangs,"
Gang Journal 1(2): 25-36.

Johnson, Claire, Barbara Webster, and Edward Connors. 1995. Prosecuting
Gangs: A National Assessment. Research in Brief. February. Washington, DC:
U.S. Department of Justice, National Institute of Justice. NCJ 151785.

Katz, Daniel, and Robert L. Kahn. 1966. The Social Psychology of
Organizations. New York: John Wiley and Sons.

Kennedy, David M. 1997. "Pulling Levers: Chronic Offenders, High-Crime
Settings, and a Theory of Prevention," Valparaiso University Law Review
31(2): 449-484.

Kenney, Dennis Jay, and James O. Finckenauer. 1995. Organized Crime in
America. Belmont, CA: Wadsworth Publishing Co.

Kimberly, John R., Robert H. Miles, and Associates, eds. 1980. The
Organizational Life Cycle: Issues in the Creation, Transformation, and Decline
of Organizations. San Francisco: Jossey-Bass.

Klein, Malcolm W. 1995a. The American Street Gang: Its Nature, Prevalence
and Control. New York: Oxford University Press.

-------. 1995b. "Street Gang Cycles," in Crime, ed. James Q. Wilson and Joan
Petersilia. San Francisco: Institute for Contemporary Studies Press.

-------. 1997. "The Problem of Street Gangs and Problem-Oriented Policing,"
in Problem-Oriented Policing: Crime-Specific Problems Critical Issues and
Making POP Work, ed. Tara O'Connor Shelley and Anne C. Grant.
Washington, DC: Police Executive Research Forum: 57-88.

Knox, George W. 1994. An Introduction to Gangs. Bristol, IN: Wyndham Hall
Press.

Kornhauser, Ruth Rosner. 1978. Social Sources of Delinquency: An Appraisal
of Analytic Models. Chicago: University of Chicago Press.

Lasley, James. 1998. "Designing Out" Gang Homicides and Street Assaults.
Research in Brief. November. Washington, DC: U.S. Department of Justice,
National Institute of Justice. NCJ 173398.

Lippitt, Gordon L. 1982. Organization Renewal: A Holistic Approach to
Organization Development, 2nd ed. New York: Appleton-Century-Crofts.

Maltz, Michael. 1985. "Toward Defining Organized Crime," in The Politics and
Economics of Organized Crime, ed. Herbert E. Alexander and Gerald E.
Caiden. Lexington, MA: D.C. Heath: 21-35.

Maxson, Cheryl, and Malcolm Klein. 1995. "Investigating Gang Structures,"
Journal of Gang Research 3(1): 33-40.

McCorkle, Richard C., and Terance D. Miethe. 1998. "Political and
Organizational Response to Gangs: An Examination of a 'Moral Panic' in
Nevada," Justice Quarterly 15(1): 41-64.

Meyer, John W. 1978. "Strategies for Further Research: Varieties of
Environmental Variation," in Environments and Organizations, ed. Marshall W.
Meyer and Associates. San Francisco: Jossey-Bass: 352-393.

Miethe, Terance D., and Richard C. McCorkle. 1997. "Evaluating Nevada's
Anti-Gang Legislation and Gang Prosecution Units," Washington, DC: U.S.
Department of Justice, National Institute of Justice. NCJ 171961.

Monti, Daniel J. 1993. "Origins and Problems of Gang Research in the United
States," in Gangs: The Origins and Impact of Contemporary Youth Gangs in the
United States, ed. Scott Cummings and Daniel J. Monti. Albany: State
University of New York Press: 3-25.

Moore, John P., and Ivan L. Cook. 1999. Highlights of the 1998 National
Youth Gang Survey. OJJDP Fact Sheet No. 123. December. Washington,
DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency
Prevention. FS-99123.

National Drug Intelligence Center. 1998. Street Gangs '98: National Street
Gang Survey Report-1998. Johnstown, PA: National Drug Intelligence Center.

Office of Juvenile Justice and Delinquency Prevention. 1999. 1997 National
Youth Gang Survey: Summary. Washington, DC: U.S. Department of Justice,
Office of Juvenile Justice and Delinquency Prevention. NCJ 178891.

Parsons, Talcott. 1987. "Suggestions for a Sociological Approach to the
Theory of Organizations," in Classics of Organization Theory, 2nd ed., ed. Jay
M. Shafritz and J. Steven Ott. Chicago: Dorsey Press: 132-146.

Pennell, S., E. Evans, R. Melton, and S. Hinson. 1994. Down for the Set:
Describing and Defining Gangs in San Diego. San Diego: San Diego
Association of Governments, Criminal Justice Research Unit.

Popielarz, Pamela A., and J. Miller McPherson. 1995. "On the Edge or In
Between: Niche Position, Niche Overlap, and the Duration of Voluntary
Association Memberships," American Journal of Sociology 101(3): 698-720.

Regini, Lisa A. 1998. "Combating Gangs: The Need for Innovation," FBI Law
Enforcement Bulletin 67(2): 25-32.

Sale, R.T. 1971. The Blackstone Rangers: A Reporter's Account of Time
Spent with the Street Gang on Chicago's South Side. New York: Random
House.

Sampson, Rana, and Michael S. Scott. 2000. Tackling Crime and Other
Public-Safety Problems: Case Studies in Problem Solving. Washington, DC:
U.S. Department of Justice, Office of Community Oriented Policing Services.
NCJ 181243.

Scott, W. Richard. 1993. Organizations: Rational, Natural, and Open Systems.
Englewood Cliffs, NJ: Prentice Hall.

Sherman, Lawrence W., Denise Gottfredson, Doris MacKenzie, John Eck,
Peter Reuter, and Shawn Bushway. 1998. Preventing Crime: What Works,
What Doesn't, What's Promising, A Report to the United States Congress.
Washington, DC: U.S. Department of Justice, National Institute of Justice. NCJ
165366.

Simmel, Georg. 1902-3. "The Number of Members as Determining the
Sociological Form of Groups, I and II," The American Journal of Sociology
8:1-46; 138-96.

Spergel, Irving A. 1990. "Youth Gangs: Continuity and Change," in Crime and
Justice: A Review of Research, ed. Michael Tonry and Norval Morris.
Chicago: University of Chicago Press.

-------. 1995. The Youth Gang Problem: A Community Approach. New York:
Oxford University Press.

Spergel, Irving A., with G. David Curry, Ron Chance, Candice Kane, Ruth E.
Ross, Alba Alexander, Pamela Rodriguez, Deeda Seed, Edwina Simmons, and
Sandra Oh. 1991. Youth Gangs: Problem and Response. Chicago: School of
Social Service Administration, University of Chicago.

Starbuck, David, James C. Howell, and Donna J. Lindquist (forthcoming).
Hybrids and Other Modern Gangs, OJJDP Bulletin, Washington, DC: U.S.
Department of Justice, Office of Juvenile Justice and Delinquency Prevention.

Starbuck, William H. 1965. "Organizational Growth and Development," in
Handbook of Organizations, ed. James G. March. Chicago: Rand McNally:
451-533.

Staw, Barry M., and L.L. Cummings, eds. 1990. The Evolution and
Adaptation of Organizations. Greenwich, CT: JAI Press.

Stinchcombe, Arthur L. 1965. "Social Structure and Organizations," in
Handbook of Organizations, ed. James G. March. Chicago: Rand McNally:
153-193.

Thrasher, Frederic M. 1963. The Gang: A Study of 1,313 Gangs in Chicago.
Chicago: University of Chicago Press (originally published in 1927).

Tucker, David J., Jitendra V. Singh, Agnes G. Meinhard, and Robert J. House.
1988. "Ecological and Institutional Sources of Change in Organizational
Populations," in Ecological Models of Organizations, ed. Glenn R. Carroll.
Cambridge, MA: Ballinger: 127-151.

Weber, Max. 1947. The Theory of Social and Economic Organization, trans.
A.M. Henderson and Talcott Parsons, ed. Talcott Parsons. New York: The
Free Press.

Zatz, Marjorie S. 1987. "Chicano Youth Gangs and Crime: The Creation of a
Moral Panic," Contemporary Crises 11: 129-158.

---------------------------

The FBI's Enterprise Theory of Investigation [sidebar]

Gangs vary greatly in size, criminal sophistication, modus operandi, and their
impact on the community. The vast majority of gangs in the United States are
community or neighborhood based and adversely affect small geographical
areas. Some gangs, however, grow rapidly in size and sophistication, becoming
multijurisdictional and even international in nature.

The goal of the FBI's national gang strategy, known as the Enterprise Theory of
Investigation (ETI), is to identify, disrupt, and ultimately dismantle violent gangs
whose activities constitute criminal enterprises. Stated succinctly, the strategy is
for the FBI, in conjunction with other Federal, State, and local law enforcement
agencies, to combat major domestic violent street gang/drug enterprises as
significant threats to American society through sustained, multidivisional,
coordinated investigations that support successful prosecution. This strategy
incorporates investigative and prosecutorial theories of enterprise investigations
that have proven successful in combating traditional organized crime. Although
ETI is not an appropriate way to deal with every jurisdiction's street gang
problem, it is an effective technique to use against large, multijurisdictional street
gangs. 

The FBI defines a "violent street gang/drug enterprise" as a criminal enterprise
that has an organizational structure and that functions as a continuing criminal
conspiracy, employing violence and any other criminal activity to sustain itself.
A criminal enterprise is any union or group of individuals engaged in a pattern of
criminal activity. For an enterprise to be "criminal," it is not necessary for it to
have written bylaws or written agreements between the individuals.

Street gangs are increasingly viewed as organized crime threats because they
are heavily involved in illegal drugs or guns, seek dominance in many new areas
of the country, and often use violence in pursuing their objectives. Street gang
drug and gun activities often produce criminal networks that span regional and
national boundaries and use modern weapons, communication technology, and
transportation in their operations. FBI analysis suggests that some gangs are
building coalitions to control and expand their operations more effectively.

As they grow, gangs involved in criminal operations must develop an
organizational structure if they are to function efficiently. Most gangs are loosely
knit coalitions of small, autonomous cliques. Apart from a general commitment
to their "hood" and the gang lifestyle, the only unifying force is combat with
outsiders. Gang leadership is usually decentralized, nonhierarchical, even
situational; it is more a function of individual prowess and reputation than a
formalized structure for making collective decisions. Leadership changes rapidly
and may vary by activity; for example, leaders in drug selling may differ from
leaders in "gangbanging." As with most groups, leadership is age-graded,
although some older members may represent powerful role models. The type of
gang leadership varies from gang to gang and by geographical area. Gangs in
Chicago, for example, tend to have a more defined leadership structure than
those in many western cities.

Two primary Federal criminal statutes are used to prosecute street gangs: the
Racketeer Influenced and Corrupt Organizations (RICO) Act and the
continuing criminal enterprise (CCE) statutes, part of the Comprehensive Drug
Abuse Prevention and Control Act of 1970. These statutes were enacted to
address the conspiratorial nature of the street gang and to enable prosecutors to
present evidence of multiple criminal acts committed by various gang members
that proves a pattern of criminal activity by the enterprise or gang. In RICO
prosecutions, evidence of acts of racketeering ("predicate acts")--such as drug
distribution, any violent acts, witness tampering, mail or wire fraud, or illegal
gambling--can be presented to the jury; thus, the full scope of the gang's
criminal conduct can be demonstrated. Prosecuting these crimes individually
would not convey a clear overall picture of the gang's criminal activity. The
CCE statute is an excellent prosecutorial tool to combat gangs that have some
form of defined leadership structure and are involved in drug distribution.

Prosecuting Street Gangs Under ETI

Normally, evidence developed in an investigation is presented to a prosecutor,
who structures the indictment and thus the method of prosecution. Under the
Enterprise Theory of Investigation (ETI), a strategy originally devised by the
FBI in 1981, this is not the case; the prosecution is structured from the
inception of the investigation.

The first step in investigating any violent street gang is a review of the available
intelligence base to estimate the group's structure, membership, and criminal
activities. Then, a background investigation is conducted through agency file
reviews, agency and public record checks, and by assigning informants in the
group to check on the individual members and their criminal activities. At this
stage, the investigator attempts to identify individual and group assets, as well as
proprietary interests of the gang.

In the next phase, investigators seek to identify meeting places, methods of
communication among gang members, and entities used to facilitate their
criminal activities. This is done through the use of various investigative
techniques. At this stage, the investigator evaluates and defines the criminal
group, the enterprises that may be the investigative focus, and the potential
predicate criminal acts.

Before the proactive investigative stage begins, consideration is given to
determining which sanctions are appropriate and can be obtained and which
type of legal relief will solve this particular problem. If every member of the
group is identified and potentially prosecutable, criminal RICO prosecution and
confinement will suffice. If the group has amassed assets, their identification for
possible forfeiture should be included in the strategy. If criminal prosecution will
not solve the problem, the civil provisions of the RICO statute should be
applied.

Next, the question of whether to pursue an overt or covert investigation must be
decided. In an overt investigation, the investigator interviews witnesses,
subpoenas records, locates expert witnesses, and convenes a grand jury. The
use of covert investigative techniques in the initial stages has proven to be more
successful. While these techniques are being used, direct and circumstantial
evidence must be recorded separately, based on the sanctions being pursued.
Separate administrative systems must be established as a repository for
evidence of the structure, membership, and purpose of each enterprise; assets
must be identified for each individual and enterprise; illegal profit generation and
criminal evidence of RICO predicate criminal acts must be identified; and
support must be compiled for the projected civil relief. The predicate criminal
acts, the defined enterprises, and the required civil sanctions must be continually
reevaluated during the course of the investigation. When the criminal RICO
indictment is structured, all assets subject to forfeiture must be identified. These
assets must be frozen at the time of indictment so that they are not transferred
and liquidated before they can be seized.

The Weisel Study

The conclusions of Dr. Weisel's report as they relate to the Logan Street,
Syndo Mob, and Black Gangster Disciples gangs are reasonably consistent
with investigative information developed by the FBI on these groups, but the
report's findings on the Latin Kings are inconsistent with the FBI's investigative
and criminal intelligence information. ETI would be an appropriate investigative
approach for all four gangs, however, because a criminal enterprise need not
have a "traditional" hierarchical organization and can be as simple as a group of
individuals "associated in fact" who engage in a pattern of illegal activity.

Source: FBI Violent Crimes and Major Offenders Section, Criminal
Investigation Division


------------------------------

Chapter 3

Young Women in Street Gangs: Risk Factors, Delinquency, and Victimization
Risk

Jody Miller, Ph.D.
Department of Criminology and Criminal Justice
University of Missouri-St. Louis

Acknowledgments

The research presented in this chapter was supported under award
92-IJ-CX-0005 from the National Institute of Justice, U.S. Department of
Justice.

------------------------------

Abstract

This study addresses the extent of young women's involvement in gangs, the
reasons for joining, and what their experiences have been. Two cities were
selected as the study sites: St. Louis, where gang emergence paralleled
deteriorating socioeconomic conditions, and Columbus, Ohio, where conditions
are conducive to the emergence of gangs despite a vibrant economy. Research
has shown that young women join gangs to solve myriad problems in their lives,
but gang involvement tends to exacerbate the situation, increasing the likelihood
they will engage in delinquency and exposing them to victimization by rival
gangs and fellow gang members. The current study confirmed this, revealing
that the neighborhood context, the existence of severe family problems (such as
drug addiction), and having gang-involved family members are risk factors for
membership. It was no surprise that female gang members were more prone to
delinquency than nongang members. Although there was no difference between
the two cities in rates of serious delinquency, female gang members in
Columbus tended to commit delinquent acts more frequently. Gang involvement
also increased the risk of being exposed to violence or victimized; nevertheless,
many young women described gang membership as providing a sense of
empowerment and a measure of protection. Prevention would optimally focus
on girls exposed to the risk factors identified here. And because young women
begin associating with gang members when relatively young, prevention would
best begin early in life. Gender-specific interventions may sometimes be useful,
but some programming for young men may be applicable to young women.

------------------------------

Scholarly concern with young women's gang involvement has grown in recent
years, not only because of the enormous growth in gangs and gang research
over the past decades, but also because of the expansion of feminist
criminology. Recent estimates suggest that more than 1,000 cities and towns
nationwide have reported gangs in their communities--more than five times the
number reported in 1980 (Klein 1995; Maxson et al. 1995). As more
researchers have attempted to understand this phenomenon, feminist scholars
have insisted that gender be considered as part of the research equation
(Campbell 1990; Joe and Chesney-Lind 1995; Messerschmidt 1995; Moore
1991). As a result, more is known about girls' experiences within gangs today
than at any point in history, including the heyday of gang studies in the 1960s
(see, e.g., Klein 1971; Short and Strodtbeck 1965).

Recent estimates also suggest that female participation in gangs is fairly
widespread. Although official sources continue to underestimate the extent of
female gang membership,[1] survey research indicates that it is relatively
extensive, particularly in early adolescence, with the percentage of young
women in gangs only slightly below that of young men. For instance, the
Rochester Youth Development Study, based on a stratified sample of youths in
high-risk, high-crime neighborhoods, found that 22 percent of females,
compared with 18 percent of males, claimed to be gang members (Bjerregaard
and Smith 1993). Later evidence from this longitudinal study suggests that girls
tend to remain gang members for a shorter time than boys do, with peak
involvement in the eighth and ninth grades (Thornberry 1999). Based on a
sample of eighth graders in 11 cities, Esbensen and Deschenes (1998) report
that 14 percent of males and 8 percent of females in their study were members
of gangs. Young women are estimated to comprise between 20 and 46 percent
of all gang members (Esbensen and Huizinga 1993; Esbensen and Winfree
1998; Moore 1991; Winfree et al. 1992).

Drawn from a larger study of female gang involvement in two Midwestern cities
(see Miller 2001), this chapter reports findings on the following issues:

--Why young women join gangs.

--The extent of young women's involvement in gang-related delinquency.

--Victimization risks associated with young women's participation in gangs.

These findings are discussed in the context of recent research on female gang
involvement and will highlight the salience of gender in shaping young women's
gang experiences.

Methods

This project incorporates several research strategies, including the use of
multiple methodologies and a comparative study design. The work is primarily
based on survey and indepth interviews with female gang members in
Columbus, Ohio, and St. Louis, Missouri, but it also includes survey interviews
with girls from the same communities who are not gang members.

In all, 94 young women were interviewed for the project: 48 who were gang
members (21 from Columbus and 27 from St. Louis) and 46 (25 from
Columbus and 21 from St. Louis) who were not. Interviewing girls in Columbus
and St. Louis made it possible to explore variations that emerged in different
geographical settings, while interviews with at-risk young women not in gangs
provided useful comparative information on differences in life experiences that
help explain why some girls join gangs while others in the same communities do
not. The goal of this project was to improve understanding of girls' gang
involvement--why they join gangs, the nature of their gangs and gang life, and
the
meaning of gang involvement--focusing particularly on how gender shapes these
issues.

Respondents ranged in age from 12 to 20, with 80 percent between the ages of
14 and 17. The mean age of gang members was 15.3, compared with 15.2 for
nonmembers. Most of the young women participating in the study were
African-American (81 percent of gang members and 72 percent of
nonmembers). Four gang members and 6 nonmembers described themselves as
biracial or multiracial, and 12 girls (5 gang members, or 10 percent, and 7
nonmembers, or 15 percent, all from Columbus) were white. Interviews in
Columbus began in early 1995 and were finished in early 1996; interviews in
St. Louis began in the spring of 1997 and were completed that fall.

Young women were recruited to participate in the project with the cooperation
of several organizations working with at-risk youths. These included the local
juvenile detention facility in both cities, a shelter care facility for girls in
Columbus, and schools and local community agencies in both cities; in addition,
one young woman was referred by a friend. The interviews were voluntary,
with respondents promised strict confidentiality, and were conducted primarily
in private offices, empty classrooms, interview rooms, or secluded spots in
visiting rooms, with a handful conducted in respondents' homes. In Columbus,
girls were paid $10 per interview; in St. Louis, they were paid $10 per survey
and $10 or $20 for each indepth interview, depending on the site.

Because sampling was purposive (i.e., not random), the representativeness of
the sample is unknown (see Glassner and Carpenter 1985). Cooperation from
agency personnel generally is successful for accessing gang members (see
Bowker et al. 1980; Fagan 1989; Short and Strodtbeck 1965), although these
referrals pose the problem of targeting only youths officially labeled as gang
members. The comparative design of the original research mitigated this
problem, however. Though the sample came from agencies working with young
women, agencies working with gang members were not targeted, nor was there
any attempt to generate a pool of known gang members. Instead, agency
personnel were asked for references to girls they believed to be involved in
gangs and to girls living in neighborhoods where they might have gang contact.
Each respondent participated in an extensive survey, during which she was
asked if she was involved in a gang. Girls were identified as gang members
through self-nomination--in short, they were classified as gang members if they
said that they were.[2]

Young women who identified themselves as gang members were asked a series
of questions about their gang during the survey, including what the gang was
like, how it was structured, what members did, and how and why they became
involved. They were then asked to participate in a followup interview, which
was completed just after the survey or within a few days. The indepth
interviews were semistructured, with open-ended questions; all but one were
audiotaped. These interviews have made it possible to gain a greater
understanding of gang life from the perspective of female gang members (see
Glassner and Loughlin 1987; Miller and Glassner 1997).

The indepth interviews were structured around several groups of questions and
allowed considerable probing. First discussed were when and how respondents
became involved in gangs and what was going on in their lives at the time.
Discussion then turned to the gang's structure--its history, size, leadership, and
organization--and their place within it. The next series of questions addressed
gender within the gang: for example, how females got involved, what activities
they engaged in and whether these resembled the males' activities, and what
kind of males and females had the most influence in the gang and why. The next
series of questions explored gang involvement more generally: what being in the
gang meant, what gang members did together, and so on. Participants were
then asked how safe or dangerous they felt gang membership was, and how
they dealt with risk. At the end of the interview, girls were asked to speculate
about why people their age joined gangs, what they liked or disliked about
being in the gang and have learned from that experience, and what they liked
best about themselves. This format was followed for each interview, although
additional topics that arose during the interview were pursued as well.

Study Settings

Since the 1980s, gangs have emerged in many cities where they had not been
observed before. Sites chosen for this study were based on the author's
concern with female gang involvement in these "emergent" gang cities and a
desire to study gangs in cities that offered an interesting comparison because of
their differences: A city that fit the literature's dominant view of gang
proliferation based on socioeconomic shifts (see, e.g., Hagedorn 1998; Klein
1995) was compared with one that seemed to challenge this view. Maxson and
colleagues' (1995) national gang migration survey provided an initial pool of
Midwestern cities with emergent gang problems; social and economic indicators
from the 1990 U.S. Census were then used in choosing the final sites.

Many scholars have focused on compelling evidence that much of the growth in
gangs in recent decades has been spurred by deteriorating economic conditions
caused by structural changes brought about by deindustrialization (see
Hagedorn 1991, 1998; Huff 1989; Klein 1995). These changes have resulted
in the growth of what scholars refer to as an urban
"underclass"--disproportionately poor African-Americans and Latinos living in
conditions of entrenched poverty in inner-city communities. These communities
are experiencing intense racial and economic segregation and isolation, the
outmigration of middle-class families, and a precipitous decline in social
services for those left behind (see Moore and Pinderhughes 1993; Sampson
and Wilson 1995). The resulting lack of alternatives and sense of hopelessness
are believed to have contributed to the growth of gangs in many cities; scholars
see the gang as a means for inner-city youths to adapt to the oppressive
conditions imposed by their environments.

St. Louis fits this explanation very well. Its experiences are typical of the urban
distress facing many Midwestern cities affected by considerable
deindustrialization and population loss in the past several decades (Decker and
Van Winkle 1996; see Wilson 1987, 1996 for an overview). These problems
are exacerbated by racial inequality. In 1990, African-Americans made up 47.5
percent of the population of St. Louis (Community Development Agency
1993), one of the 10 most racially segregated cities in the United States (Rusk
1995). Moreover, there are considerable gaps in income and poverty and
unemployment rates between the city's African-Americans and whites.

In contrast, based on city-level indicators, Columbus appears an atypical site
for gangs. It is one of the few Midwestern cities that has continued to thrive
during periods of overall decline and has experienced continuous economic and
population growth since the 1950s (Columbus Metropolitan Human Services
Commission 1995; Rusk 1995). Columbus is also much less racially diverse
than St. Louis, with African-Americans constituting 22.6 percent of the city's
population.[3] Despite its overall economic vibrancy, however, the city has
conditions conducive to the emergence of gangs (Huff 1990, 1998). There
remain substantial pockets of impoverished neighborhoods with high
concentrations of African-Americans. While the median household income for
African-Americans is nearly $6,000 higher in Columbus than in St. Louis, the
racial disparity in each city is similar--African-Americans' median income is
about 60 percent that of whites in St. Louis, compared to 69 percent in
Columbus (U.S. Census Bureau 1990).

These characteristics made St. Louis and Columbus optimal choices for this
study, allowing the author to explore variations in girls' experiences in gangs that
may emerge in different settings. Indeed, there is evidence that the city context
has had some impact on the extent and nature of gangs in each of these cities.
Maxson and colleagues (1995) place the onset of gangs in both cities at around
1985, although gangs have waxed and waned in St. Louis for the last century
(Decker and Van Winkle 1996). According to police estimates at the time of
Maxson and colleagues' (1995) research, Columbus had about 200 gang
members, whereas St. Louis was home to roughly 900 gang members.[4] In
both cities, gangs appear to be "homegrown" problems (Decker and Van
Winkle 1996; Huff 1989), although evidence suggests that St. Louis gangs are
larger and more criminally involved than are gangs in Columbus (see Decker
and Van Winkle 1996; Mayhood and LaLonde 1995). Because law
enforcement data tend to underestimate female gang involvement (Curry et al.
1994), it is difficult to assess the variation in female gang participation rates
between the two sites.[5] As discussed below, evidence suggests that females
constitute a slightly larger percentage of gang membership in St. Louis.

Joining Gangs

Research shows that youths typically associate with gang members for some
time--often as much as a year--before joining (Decker and Van Winkle 1996).
In
the current study, young women typically began hanging out with gang members
when they were quite young--around age 12 on average--and joined at an
average age of 13. In fact, 69 percent of the girls in the sample joined their
gangs before turning 14.

In exploring why young women become involved in gangs, scholars have
considered how gender inequalities intersect with class and racial oppression to
shape young women's gang experiences. As Campbell (1990: 173)
summarizes, "the gang represents for its [female] members an idealized
collective solution" for addressing many problems in their lives, including limited
educational and occupational opportunities, subordination to men, and childcare
responsibilities, as well as the powerlessness of underclass membership they
share with males in their communities. This section explores girls' pathways into
gangs by presenting the broader contexts and precipitating events that lead
young women to spend time with gang members and join gangs.

The study revealed three themes about the factors that contribute to girls' gang
involvement. What is notable is that these themes emerged independently in the
surveys, as factors distinguishing gang members from nonmembers, and in the
indepth interviews, as reasons young women gave for joining gangs. The first
theme was girls' neighborhood contexts and their exposure to gangs through
neighborhood peer networks. A second theme was the existence of severe
family problems, such as violence and drug abuse, which decreased parental
supervision, led young women to avoid home, and to seek to meet their social
and emotional needs elsewhere. Finally, many young women described the
strong influence of gang-involved family members, particularly older siblings in
Columbus and siblings and cousins in St. Louis, on their decisions to join gangs.

Exhibit 1 compares gang members and nonmembers' reports on these issues.
As noted, there is a significant connection between contemporary youth gangs,
urban poverty, and racial segregation. In both cities, the vast majority of girls in
the study lived in neighborhoods that were economically worse off and more
racially segregated than the city as a whole. These neighborhoods had
substantially lower median incomes and higher rates of poverty and
unemployment than citywide averages[6] (see Miller 2001 for a more detailed
description). Although this was the case for gang members and nonmembers
alike, important differences emerged in girls' descriptions of their
neighborhoods, particularly about their exposure to gangs.

Neighborhood Exposure

In both Columbus and St. Louis, most of the young women described some
exposure to gangs in their neighborhoods. Yet, as exhibit 1 highlights, gang
members and nonmembers differed in their descriptions of the extent and
proximity of gang activity in their neighborhoods. Gang members were
significantly more likely than nonmembers to report "a lot" of gang activity in
their neighborhoods and to note that gang members lived on their street.[7] In
fact, 90 percent of the gang sample responded "yes" to one or both questions.
It appears, then, that coupled with other risk factors, living in neighborhoods
with gangs in close proximity increases the likelihood that young women will
decide to join a gang.[8]

Family Problems

Although neighborhood exposure helps explain why girls join gangs, other
precipitating factors also must be considered. The family situation has long been
considered crucial for understanding delinquency and gang behavior. Weak
supervision and low parental involvement (Esbensen and Deschenes 1998; Joe
and Chesney-Lind 1995; Thornberry 1997) and family violence and substance
abuse by family members (Fleisher 1998; Joe and Chesney-Lind 1995; Moore
1991) all have been found to contribute to the likelihood that girls will join
gangs.[9] Thus, not surprisingly, many of the gang girls in this study described
problems in their family lives that led them to spend time away from home, out
on the streets, and with gang members.

As exhibit 1 illustrates, gang members were significantly more likely to come
from homes with numerous problems than were the young women who were
not in gangs. Gang members were significantly more likely to have witnessed
physical violence between adults in their homes and to have been abused by
adult family members. In addition, gang members were much more likely to
report regular drug use in their homes. Most important, gang members were
significantly more likely to describe experiencing multiple family problems--with
60 percent describing three or more of the five problems listed in exhibit 1 and
44 percent reporting that four or more of these problems existed in their
families. In fact, only three gang members said none of these problems occurred
in their families, compared with nine (20 percent) of the nonmembers.

In the indepth interviews, the most common family themes described by young
women as contributing to their gang involvement were drug addiction among
primary caregivers and being physically or sexually abused by family
members.[10] Family problems facilitated girls' gang involvement in various
ways, but there was a common thread--young women began spending time
away from home as a result of difficulties or dangers there, seeking to get away
and to meet their social and emotional needs elsewhere. Several researchers
have suggested that "the gang can serve as a surrogate extended family for
adolescents who do not see their own families as meeting their needs for
belonging, nurturance, and acceptance" (Huff 1993: 6. See also Campbell
1990; Joe and Chesney-Lind 1995; for a dissenting view, see Decker and Van
Winkle 1996). Regardless of whether gangs actually fill these roles in young
women's lives, many young women believe that the gang will do so when they
become involved.

Influence of Gang-Involved Family Members

Some girls who lack close relationships with their primary caregivers can turn to
siblings or extended family members to maintain a sense of belonging and
attachment. But if these family members are involved in gangs, it is likely that
girls will join gangs as well. Moreover, even when relationships with parents or
other adults are strong, having adolescent or young adult gang members in the
family often heightens the appeal of gangs (see also Joe and Chesney-Lind
1995; Moore 1991). As exhibit 1 illustrates, the gang members in this study
were significantly more likely than nonmembers to report having family
members in gangs. Most important, gang members were much more likely to
have siblings in gangs and to have two or more family members involved in
gangs.

These relationships differed in the two study sites, with the relationship between
a girl's gang membership and that of her family most marked in St. Louis. In
Columbus, gang members were not significantly more likely than nongang girls
to have a family member in a gang--57 percent of gang members had family
members in gangs, versus 48 percent of nonmembers. By comparison, all but
one of the gang members surveyed in St. Louis (96 percent) reported having at
least one family member involved in gangs. In fact, a greater percentage of
nongang girls in St. Louis (62 percent) reported having a family member in a
gang than did gang members in Columbus (57 percent). Gang members in both
cities, however, were significantly more likely to report having a sibling involved
in gangs than were nonmembers. In all, 52 percent of St. Louis gang members
and 48 percent of Columbus gang members had siblings in gangs, compared to
19 and 16 percent of the nonmembers in these cities, respectively. Moreover,
in indepth interviews, many girls described the significant influence of older
siblings and relatives on their decisions to join their gangs. Often, girls who
joined gangs to be with or like their older siblings also experienced the family
problems noted earlier (see Miller 2001).

In fact, the themes just reviewed--neighborhood exposure to gangs, family
problems, and gang-involved family members--overlapped in most gang
members' accounts, further distinguishing them from the nonmembers. Taken
individually, a majority of girls fit within each category: 96 percent of gang
members described living in neighborhoods with gangs (versus 59 percent of
nonmembers). Of these, 69 percent cited their neighborhood and peer
networks as factors in their decisions to join a gang. Likewise, 71 percent
recognized family problems as contributing factors (26 percent of nonmembers
reported similar problems), and 71 percent had siblings or multiple family
members in gangs or were influenced in their decision to join by gang-involved
family members (compared to one-third of the nonmembers who had gang
members in their immediate family or multiple gang members in their extended
family). In all, 90 percent of the gang members reported two or more
dimensions of these risk factors, and fully 44 percent fit within all three
categories. In contrast, only one-third of the nonmembers experienced multiple
risk factors for gangs, and only four nonmembers (9 percent, versus 44 percent
of gang girls) reported all three dimensions.

Gang Characteristics

Gender

In Columbus and St. Louis, mixed-gender gangs appear to predominate. Of the
48 gang members in the study, 42 (88 percent) described their gangs as
including both male and female members.[11] Most girls in mixed-gender gangs
were in groups that were primarily male, though St. Louis gangs tended to have
more female members than Columbus gangs. In Columbus, 17 girls (85 percent
of those in mixed-gender gangs) described their gangs as having more males
than females; in St. Louis, 14 of 22 girls (64 percent) did so. In fact, about
one-third of the girls in both cities were in gangs with 20 percent or fewer
female members. Yet the girls in Columbus were more consistently a minority in
their gangs than were the young women in St. Louis. Half of the girls in St.
Louis were in gangs in which more than one-third of the membership was
female, and more than a third (37 percent) were in gangs in which females
comprised at least half of the members. In contrast, only 25 percent of
Columbus girls were in gangs in which more than one-third of the members
were female, and only three described girls as comprising at least half of the
members. These differences are noteworthy because recent evidence suggests
that the gender composition of gangs affects the level of gang-member
delinquency (Peterson, Miller, and Esbensen 2001).

Size and Racial Composition

Columbus gangs tended to be small: All of the girls described their gangs as
having 50 or fewer members, and 85 percent were in gangs with 30 or fewer
members. St. Louis gangs were more varied and somewhat larger, as 14 girls
(52 percent) described their gangs as having 30 or fewer members, 9 girls (33
percent) reported being in gangs with 35 to 50 members, and 4 girls (15
percent) said that their gangs had more than 50 members. Girls' gangs in
Columbus were principally African-American, though some had a small number
of white members. Although nearly 25 percent of the gang members
interviewed in Columbus were white, all of the girls in St. Louis were
African-American (89 percent) or multiracial (11 percent). Their gangs were
almost exclusively African-American as well. Only four young women (15
percent) said that there were non-African-Americans in their gangs. The greater
racial segregation in St. Louis, combined with the greater strength of
neighborhood ties in defining the boundaries of St. Louis gangs (see Decker
and Van Winkle 1996; Miller 2001), probably helps explain the less racially
diverse character of these groups.

Age

Columbus gangs also were primarily adolescent groups. Of the 21 Columbus
gang members surveyed, 10 reported that all their members were 20 or
younger, while the rest said their gangs included members who were 21 or
older. Yet almost without exception, these groups were primarily teenagers with
either one adult who was considered the O.G. ("Original Gangster," or leader)
or just a handful of young adults. In contrast, girls in St. Louis were more likely
to report having adult members in their gangs. In all, 19 St. Louis girls (70
percent) reported having adults in their gangs, and 6 girls (22 percent) reported
members over age 30 in their gangs. Significantly, they characterized these
groups as being made up of adolescents and young adults, rather than as
primarily adolescent groups.

Regional Differences

These differences probably can be explained by the divergent socioeconomic
circumstances of the two cities. Several scholars in gang research have noted
that in communities like St. Louis, gang members have a more difficult time
leaving the gang because they have limited options in the legitimate economy
(Decker and Van Winkle 1996; Hagedorn 1998; Klein 1995; Moore 1991).
Thus, a likely explanation for the greater number of older members in St. Louis
girls' gangs is that a larger proportion of youths who joined in their early to
mid-teens stayed in the gang into young adulthood for economic reasons. In
addition, the much stronger neighborhood character of St. Louis gangs was
probably a contributing factor: Unless young people left their neighborhood,
they likely did not leave the gang, even if their ties to the group weakened with
age.

Perhaps because of the greater age range within their gangs, 70 percent of the
girls in St. Louis described these groups as having age-based subgroups,
though they did not describe role specialization distinguishing between
subgroups. Gangs in both cities were territorial, though the nature of their
territoriality differed--diffuse in Columbus, but closely tied to specific
neighborhoods and blocks in St. Louis (see Miller 2001). In addition, girls'
gangs were best characterized in each city as involved in versatile, rather than
specialized, delinquency (see below). Based on the dimensions already
discussed--size, age range, duration, subgrouping, territoriality, and crime
patterns--Klein and Maxson (1996) documented five gang types in cities
around
the country: traditional, neotraditional, compressed, collective, and specialty
gangs.[12] According to this typology, it appears that the gangs in St. Louis are
more diverse structurally than the gangs in Columbus.

All of the girls interviewed in Columbus were in gangs that can be characterized
as compressed: they were small, without subgroups, and with a narrow age
range, a duration of 10 years or less, some territoriality, and criminal versatility.
Seven (26 percent) of the interviewees in St. Louis also described gangs that fit
this category, and one young woman described her gang in terms that fit Klein
and Maxson's designation for a collective gang (no subgroups, may or may not
be territorial, versatile crime patterns, 50 or more members, an age range often
greater than 10 years, and longer duration than compressed gangs). Most of the
gang members interviewed in St. Louis (70 percent), however, described
groups that can best be described as neotraditional gangs. These groups
resemble potential traditional gangs in the making: Their duration is probably no
more than 10 years, (Maxson and Klein 1995: 39) but they are of medium size,
with around 50 to 100 members, have age-based cliques, and are
territorial.[13] A final difference that emerged in comparing the Columbus and
St. Louis gangs was the impact of cultural diffusion on gang characteristics.
Girls in Columbus described gangs heavily influenced by Chicago gang style;
young women in St. Louis did not. Thus, though Columbus gangs did not
appear more organized or criminally sophisticated than those in St. Louis, their
leadership structure was more rigid, and they were more likely to have
initiations, meetings, and rules (see Miller 2001 for further discussion).

Gangs, Delinquency, and Violence

Gang Membership and Delinquency

Recognition of the strength of the relationship between gang membership and
crime participation has generated much gang research in the last decade.
Several important studies have shown that gang youths account for a
disproportionate amount of delinquency, especially of serious and violent acts
(Esbensen et al. 1993; Esbensen and Huizinga 1993; Fagan 1989, 1990; Huff
1998; Thornberry et al. 1993; Thornberry 1997). Moreover, studies show that
participation in delinquency increases dramatically when youths join gangs and
declines significantly once they leave their gangs (Huff 1998; Thornberry 1997).
These patterns also hold true for female gang members: Young women in gangs
have higher rates of delinquency than nonmembers, male or female
(Bjerregaard and Smith 1993; Deschenes and Esbensen 1999; Esbensen and
Winfree 1998; Thornberry et al. 1993). Recent evidence also suggests that the
gender composition of gangs affects female and male gang members' levels of
participation in gang delinquency, with female and male members of
majority-male gangs reporting the highest levels of delinquency (Peterson et al.
2001).

Despite gang girls' involvement in delinquency, however, young men are much
more extensively involved in the most serious forms of gang crime. For instance,
Fagan (1990) reports a bimodal distribution of delinquency for gang girls but
not gang boys: Approximately 40 percent of the gang girls in his study were
involved in only petty delinquency while a third were involved in multiple Index
offending, compared to 15 percent and 56 percent, respectively, for young
men. Moreover, evidence suggests that gun use and homicide are much more
prevalent among male than female gang members (Decker, Pennell, and
Caldwell 1996; Miller and Decker 2001).

In view of this research, it is not surprising that female gang members surveyed
in Columbus and St. Louis reported greater involvement in delinquency than did
nonmembers. Exhibit 2 compares prevalence rates for gang members and
nonmembers for several moderate and serious offenses, as well as frequencies
for minor, moderate, and serious offenses, and drug sales.[14] Gang members
were significantly more likely than nonmembers to have engaged in nearly every
offense measured, with the exceptions of stealing over $100 and hitting
someone with the intent of causing injury. Moreover, gang members committed
all types of offenses--minor, moderate, and serious, as well as drug sales--more
frequently than did girls who were not in gangs. These differences were most
striking for serious delinquency and drug sales. On average, gang members had
committed 42.21 serious offenses in the past 6 months (averaging about 7
serious crimes per month), compared with only 2.7 for nongang girls (less than
1 every two months).[15] Comparison of gang members and nonmembers'
frequency of participation in drug sales reveals even more disparity. On
average, gang members sold drugs 53.72 times in the past 6 months, or about
twice a week, compared with 4.17 for nonmembers.[16]

Few variations emerged in comparing delinquency prevalence rates among
female gang members in both cities, and no variation was evident in comparing
rates of serious delinquency or drug sales, but there were striking differences
between the two sites in the frequency of their involvement in delinquency and
drug sales (see exhibit 3). On average, female gang members in Columbus had
committed 3.5 times more minor delinquency, twice the amount of moderate
delinquency, and just more than 3.5 times more serious delinquency in the 6
months before being interviewed. In contrast, girls in St. Louis were more
frequently involved in drug sales. On average, female gang members in St.
Louis had sold drugs 11 times per month in the past 6 months, as compared to
slightly more than 6 times for girls in Columbus. Recent evidence from a
multisite survey indicates that young women in majority-male gangs have higher
delinquency rates than their counterparts in gender-balanced groups (Peterson
et al. 2001). Thus, the differences in delinquency between Columbus and St.
Louis may result from Columbus girls' greater likelihood of membership in
gangs with a majority of male members. On the other hand, differences in drug
sales between the two sites is likely indicative of the greater economic focus of
St. Louis gangs (see Decker and Van Winkle 1996; Miller 2001; Rosenfeld,
Bray, and Egley 1999).

The patterns evident in exhibits 2 and 3 do not show another important
dimension of gang girls' delinquency--the wide variation in their frequency of
offending. As noted above, research on female gang members' delinquency has
revealed that some girls are very delinquent, whereas others are involved only
in sporadic or minor offending. This pattern was clear among the girls
interviewed for this study. Nearly 25 percent of the gang members interviewed
had committed no serious delinquent acts in the last 6 months; the same number
said they had done so fewer than six times during this period. Yet slightly more
than a quarter of them reported more than weekly involvement in serious
delinquency--with six girls (13 percent) reporting everyday involvement in these
activities. More than half of the gang members interviewed had not sold drugs in
the past 6 months, yet 27 percent sold drugs more than once a week, including
seven (17 percent) who sold drugs daily. In short, some female gang members
are quite delinquent, whereas others seem to temper their routine involvement in
criminal offending, particularly serious offending (see also Fagan 1990).

Gender and Gang Crime

Turning to the context of girls' offending within gangs, several key points bear
highlighting. First, except for regular drug-selling, most of the gang delinquency
young women and their gangs were involved in was not organized or planned.
Sometimes youths went out "looking for trouble" or to "start something," but
even then, crime was not planned. It was merely something to do, and
particular crimes were committed as opportunities presented themselves. As
Erica[17], a 17-year-old from Columbus, surmised, "If somebody's bored and
they have nothin' to do, then they'll start a fight." At other times, groups of
youths came looking to make trouble with other youths, and fights or
confrontations ensued. Yet very little crime was premeditated in either
Columbus or St. Louis. The second point is that youths' involvement in gang
crime is shaped--but not determined--by gender. The kind of activities girls
were
routinely involved in, the level of their involvement, and the kinds of activities
they did not engage in were influenced by gang members' perceptions of
appropriate "male" or "female" interactions and behavior. Each of these points is
more fully explored below in the context of two facets of gang delinquency:
girls' descriptions of conflicts with rival groups and their involvement in the drug
trade.

An important element of gang life involves challenging and fighting with rival
gangs; it is often at the level of these antagonisms that youths stake out the
identity of their gang in opposition to its enemies (see Decker and Van Winkle
1996; Sanders 1993). Opposition to rival gangs is a central theme in the
cultural imagery and symbolism that gang youths adopt, and young women
often described confrontations with rivals as a consequence of these symbolic
displays. Yet these confrontations rarely resulted in serious violence among
girls. In fact, gender differences most clearly emerged in girls' discussions of
serious violence. Most serious affronts to the gang--and retaliations--were
committed by and against young men. When girls were involved, they reported
that male gang members left retaliation against rival girls to young women in the
gang, except in unusual circumstances (see Miller 1998).

Female-on-female violence within gangs generally involved fists or such
weapons as knives, rather than guns. In fact, only one young woman reported
having shot a rival, and only three said they had participated in a driveby
shooting. Instead, girls said young men were most likely to use guns in their
confrontations with rival gang members. Crystal, a 13-year-old from St. Louis,
noted, "Girls don't be up there shooting unless they really have to." Likewise,
17-year-old Tonya, also from St. Louis, said, "I ain't never carried no strap
[gun] before. I was too nervous. I ain't used no guns." Explaining why girls used
fists and knives rather than guns, Pam, an 18-year-old from St. Louis,
suggested it was because "We ladies, we not dudes for real." Likewise,
Veronica, a 15-year-old from Columbus, said she preferred to fight with her
fists, commenting, "I think it's dumb that they have to use weapons and
everything." Not surprisingly, a recent analysis of St. Louis homicide data
revealed only one gang homicide involving a female perpetrator between 1990
and 1996, compared with 229 gang homicides involving male perpetrators[18]
(Miller and Decker 2001). Despite some female gang members' regular
involvement in serious gang crime, young women are rarely involved in the most
serious and lethal forms of gang violence.

As for drug sales, it was noted above that 52 percent of the female gang
members in Columbus described having sold marijuana and 48 percent, crack,
whereas around 63 percent of the St. Louis gang members had sold both. In
all, 14 girls (6 from Columbus and 8 from St. Louis) were involved in drug sales
on at least a weekly basis. As exhibit 3 shows, however, girls in St. Louis had a
considerably higher mean frequency of drug sales than girls in Columbus. In
fact, five young women in St. Louis (19 percent) sold crack on a daily or
near-daily basis, whereas only one young woman in Columbus did. Moreover,
young women in St. Louis were more likely to describe status within the gang
as deriving from success in the drug trade (see Miller 2001).

Young women in St. Louis who described selling drugs routinely emphasized
that it was a key element of their gang activities. Tonya said, "Mostly dudes was
. . . selling guns and jacking cars and stuff like that. But everybody selling
drugs." In St. Louis, the older girls emphasized the economic benefits of drug
sales within their gangs.[19] The two cities' different socioeconomic
circumstances and their likely relationship to St. Louis gangs' older and adult
membership, could help explain why St. Louis gang members appeared more
deeply involved in drug sales than those in Columbus (see also Decker and Van
Winkle 1996; Rosenfeld et al. 1999). Most young women in Columbus who
talked about girls' involvement in drug sales described it--especially street-level
crack selling--as primarily a male endeavor. For example, 17-year-old
LaShawna noted, "We don't really let the females do that unless they really
wanna and they know how to do it and not to get caught and everything."
Asked why girls would be less likely to know how, she explained, "Like they
don't get taught . . . 'cause some females in mine don't know how, don't even
know how to make dope or whatever."

Gender and Victimization in Gangs

All this focus on the link between gang membership and delinquency has
overlooked the equally important relationship between gang membership and
crime victimization risk. Though few studies have addressed this issue, it is
important. For instance, there is strong evidence that participation in
delinquency increases youths' risk of victimization (Lauritsen, Sampson, and
Laub 1991). Given ample evidence linking youths' participation in gangs with
increases in delinquency, as well as research showing that the primary targets of
gang violence are other gang members (Decker 1996; Klein and Maxson
1989; Sanders 1993), it seems self-evident that gang members are more likely
than nonmembers to become victims of crime. Moreover, evidence of gender
inequality within gangs (Fleisher 1998; Miller 1998, 2001; Moore 1991) and of
young men's greater participation in serious gang crime suggests that
victimization risk within gangs is likely to be shaped by gender.

There is strong evidence that "adolescent involvement in delinquent lifestyles
strongly increases the risk of both personal and property victimization"
(Lauritsen et al. 1991: 265). In fact, Lauritsen and her colleagues found that
gender decreases in significance as a predictor of victimization risk if one
controls for participation in delinquent lifestyles. That is, much of young men's
greater victimization risk (with the exception of sexual assault) can be
accounted for by their greater involvement in offending behaviors. Given gang
members' greater involvement in delinquency, it is not surprising that the gang
members interviewed were more likely than nonmembers to report having been
victims or witnesses of violent acts.

Gang Membership and Victimization

Exhibit 4 compares gang members' and nonmembers' exposure to violent acts
and their victimization experiences. Gang members were significantly more
likely than nonmembers to report having seen physical violence involving
weapons and to have been sexually assaulted, threatened with a weapon, or
stabbed. The figures are startling: Two-thirds of the young women in gangs had
witnessed at least one homicide in their lives, and 79 percent had seen someone
shot. More than half of the gang members had been sexually assaulted and
threatened with a weapon, and a third of them had been stabbed. No significant
differences emerged between gang members in Columbus and St. Louis,
although a larger percentage of gang members in St. Louis reported witnessing
shootings (89 percent, versus 67 percent in Columbus) and killings (74 percent
and 52 percent, respectively), and more gang members in St. Louis (41
percent, versus 24 percent in Columbus) had been stabbed.

Gang involvement appears to raise the risk of becoming the victim of the kinds
of violence indicative of girls' gang activities.[20] As described above, the vast
majority of young women described confrontations with rival gang members
that involved fists or knives, but not guns. Thus, it is not surprising that only a
handful of girls had been shot, whereas a sizeable minority had been stabbed. In
fact, of the four young women who reported having been shot, two of these
were accidental shootings that occurred when they were children. These
findings are further evidence of the gender-based nature of weapons use and
weapons assaults within gangs.

Twenty-five gang members (52 percent) reported having been sexually
assaulted, for a total of 35 assaults. Most of the sexual victimization young
women reported occurred in the context of the family. In all, 23 of the 35
incidents described (66 percent) were committed by family members, or by
men whom young women were exposed to through their families. Gang
membership did not seem to increase girls' risk of being sexually assaulted, but
it increased their risk of becoming the victim of other forms of violent crime.
That is, they were at tremendous risk for sexual assault in other arenas of their
lives and perhaps saw the gang as a refuge from the sexual aggressors around
them. This is not to suggest, however, that sexual assaults within gangs did not
occur or that other forms of sexual exploitation were not present (see Fleisher
1998; Miller 1998, 2001; Miller and Brunson 2000). In fact, as will be
discussed, the less girls were involved in gang crime, the more vulnerable they
appeared to be to sexual exploitation and other mistreatment within the gang.

Because intergang rivalries and delinquency are important elements of gang
activities, some level of victimization risk is an expected part of gang life.
Members recognized that they may be the targets of rival gang members and
were expected to "be down" for their gang at those times, even when it meant
being physically hurt. Young women's descriptions of the qualities they valued in
gang members revealed the extent to which the ability to handle violence was
deemed an important part of successful gang involvement. Members were
supposed to be tough and able to fight and to engage in criminal activities, while
remaining loyal to the gang and willing to put themselves at risk for it.

A Sense of Empowerment

Despite the violence that was a structured and expected part of gang
membership, many young women described their gang involvement as providing
a sense of empowerment. Young women in both cities focused on the
camaraderie of having a group of friends, who, as Lisa, a 13-year-old from
Columbus, said, "got my back," and also the sense of self that developed from
earning a reputation for being tough. Erica explained, "It's like you put that
intimidation in somebody . . . they don't bother you." Perhaps because of their
greater involvement in the drug trade, some St. Louis gang members also
mentioned the empowerment that came from economic self-sufficiency and
independence. An important element of the empowerment that gangs provided
was that of having a group of people who could offer protection, backup, and
retaliation--necessities for success in the drug trade and survival on the streets.

Many young women in both cities felt a gender-based sense of protection from
being a member of a predominantly male group. Gangs operate within larger
social milieus characterized by gender inequality and sexual exploitation, as
evidenced by the widespread sexual abuse young women had experienced
outside of their gangs. In part, they saw the gang as a refuge from these
dangers. Being in a gang with young men offered at least the semblance of
protection from, and retaliation against, predatory men in the larger social
environment. Heather, a 15-year-old from Columbus, noted, "You feel more
secure when, you know, a guy's around protectin' you." She explained that as a
gang member, "You get protected by guys . . . not as many people mess with
you." Likewise, Rhonda, a 15-year-old from St. Louis, said that people didn't
often "mess with" her, "Cause they know if they try to take advantage of the
girls [in the gang], then it gonna be a problem."

Gender as a Shield From Risk

Moreover, in gangs, young women could use their gender to shield and control
their exposure to gang violence, at least to some extent. Because status
hierarchies in most of their gangs were male-dominated, young women actually
seemed to have greater flexibility in their gang activities than did young men (see
Miller 1998). Female gang members were not expected by their male and
female peers to be heavily involved in such criminal activities as gun use, drug
sales, and other serious crimes. As discussed, most of the gang members
surveyed in both cities were not routinely involved in serious crime and drug
sales, and few of them used guns. Compared with the experience of more
seriously delinquent gang youths, these girls' limited involvement in such
activities decreased their risk of becoming a victim of gang-related crime,
because they could avoid activities likely to place them in danger. Moreover,
girls' routine confrontations were more likely to be female-on-female, rather
than male-on-female, which further reduced their risk of serious victimization.
As Vashelle, a 14-year-old from St. Louis, explained, "Most girls, they ain't
gonna do nothing for real but try to stab you, cut you, or something like that. As
far as coming by shooting and stuff like a dude would do, no." Getting stabbed,
while serious, is much less likely to be lethal than getting shot.

Yet despite the protections offered by the gang and their attempts to avoid
risky behaviors, the young women surveyed admitted that gang membership
posed personal dangers. They were concerned not only about the risk of
running into rival gang members but also about the ever-present threat of
violence in their dangerous neighborhoods. As Rhonda explained, "I worry
about [violence] all the time and it's like, violence is gonna be here until the day
everybody kill one another, until God comes down from heaven and takes
over." Leslie, a 16-year-old from Columbus, said that even though she tried to
avoid danger by not "doin' all that they was doin'," she recognized that as a
gang member, "You can get shot when you're sittin' on the porch. I mean, you
can get hurt anywhere you are, you can still get hurt. It doesn't matter where
you are or what you're doin', you can get hurt." Crystal said, "[When] you got
one gang member that got a problem with another person, they'll shoot at the
whole set or whatever. . . . A whole bunch of them will start shooting at people
that got nothing to do with it."

As these descriptions suggest, young women did not fear being the target of
rival gang members' gunshots; rather, they recognized the danger of being shot
as a result of being in a group context where someone opened fire. These
descriptions echo homicide reports from St. Louis: Young women made up
only 8 percent of gang homicide victims from 1990 to 1996, and the vast
majority of them were not the intended target, but were shot when rival gangs
opened fire into a group (see Miller and Decker 2001).

Victimization Within Gangs

Much of the violence discussed thus far has been tied to intergang rivalries and
conflict. It is important to note, however, that some young women in gangs face
mistreatment and abuse at the hands of fellow gang members. Many girls did
not participate in serious gang crime, not just because they chose to avoid it but
also because they were excluded by young men (see also Bowker et al. 1980).
Although girls' exclusion from some gang crime was framed as protective (and
thus likely reduced their victimization by rival gangs), it also served to
perpetuate the devaluation of female members as less significant to the gang--as
weak and not as tough, or not as ready to protect the gang as are male
members. Describing the masculine basis of status hierarchies within her gang,
Sheila, a 17-year-old from St. Louis, noted, "The dudes think they run it all."

In terms of gender-specific victimization risk within the gang, the devaluation of
young women meant several things. It could lead to mistreatment and
victimization of some girls by other gang members, especially when they did not
have specific male protection in the gang (e.g., a brother or boyfriend) or were
not able to stand up for themselves to male members. This was exacerbated by
activities that led young women to be viewed as sexually available. The most
obvious example was the existence of "sexing in"--having sexual relations with
multiple male members of the gang--as an initiation ritual. Other gang members,
both male and female, viewed young women initiated in this way as
promiscuous and sexually available and were more likely to mistreat them (see
Miller 1998, 2001). This stigma could extend to female members in general,
creating a sexual devaluation that all the girls had to contend with.

Nonetheless, girls' experiences within gangs varied greatly. Some young women
were able to carve out a niche for themselves that put them, if not equal in
standing to young men, at least on par in terms of their treatment. Other young
women were severely mistreated, and there was a range of experiences in
between. Ironically, the young women most respected within the gang were
more likely to face gang-related victimization by rivals; those defined as "weak,"
because they did not participate in gang-related fighting and delinquency, were
more likely to be abused by their gang peers.

Implications

This chapter has addressed several issues about young women's involvement in
gangs--the extent of that involvement, why they become involved, and what
their
experiences in gangs are like, including their participation in delinquency and
exposure to violence. As has been discussed, young women join gangs in
response to a myriad of problems in their lives. Yet gang involvement tends to
exacerbate, rather than improve, their problems. It increases the likelihood that
young women will engage in delinquency and exposes them to risks of
victimization, both by rival gangs and by fellow gang members. Though young
women are not as involved as young men in the most serious forms of gang
violence, and gangs are less likely to be life-threatening for females than for
males, gang involvement often narrows young women's life options even further
(see Moore 1991).

Although this research was not explicitly focused on policy, it may yield several
suggestions for policy and practice. With regard to efforts to prevent gang
involvement, the research highlights several notable issues. The girls in this study
began associating with gang members at age 12 on average, and 69 percent of
them joined gangs before turning 14. Thus, prevention must begin early in a
girl's life. In addition, a series of risk factors appear to increase the likelihood
that gangs will become an alluring option. These include living in neighborhoods
where gangs are close by; having multiple family problems, including violence
and drug abuse; and having siblings or multiple family members in gangs. Most
significant, it appears that the convergence of such risk factors heightens girls'
risks for gang involvement: Fully 90 percent of the gang members in this study
reported two of these three risk factors. Prevention should focus on girls
exposed to these problems and should begin early.

This project also has implications for gang intervention with young women.
Most notable is the need to recognize variations in young women's experiences
and activities within and across gangs. The research presented here suggests
that although gender-specific interventions may be useful in some instances
(particularly in addressing sexual assault and abuse), some features of gang
programming for young men are likely to be just as important for girls. For
example, although girls are rarely as involved as boys in serious violence, group
processes, conflicts, and rivalries provoke girls' participation in confrontations
with rival gang members just as they do for young men. Thus, interventions for
girls need to take such issues into account (see Klein 1995). Younger girls, in
particular, appear to respond to gang group processes, whereas older
girls--like
older male gang members--view the gang as providing support for such
income-generating activities as drug sales. This suggests that such intervention
strategies as providing job training and employment opportunities are just as
important for female gang members as for their male counterparts.

Intervention strategies should also be tailored to meet the diverse needs of
female gang members. Specific knowledge about girls' gangs--for instance, their
participation in economic versus symbolic crimes--can suggest effective
approaches. In addition, knowledge about the gender composition of girls'
gangs is likely to provide information about the level of young women's
involvement in delinquency. As noted, young women in gangs having a majority
of male members appear to be more deeply immersed in criminal endeavors
than are young women in gender-balanced or all-female groups (see Peterson
et al. 2001). In particular, the level and nature of young women's participation
in gang crime may be an indicator of the particular types of victimization risks
these young women face: Girls who are heavily involved in street crime are at
heightened risk for physical violence, such as assaults and stabbings; those who
are not are at greater risk for ongoing physical and sexual mistreatment by male
gang members (see also Miller and Brunson 2000). Most important is the need
to recognize that young women in gangs are not a monolithic group and should
not be treated as if they were.

Unfortunately, responses to gangs and gang members are often primarily
punitive and disregard the social, economic, and personal contexts that lead to
gang participation. This punitiveness means that gang members are not seen as
needing assistance and protection, which--coupled with the problems they face
in their daily lives--further harms these youths (see Fleisher 1998; Moore and
Hagedorn 1996). Moreover, few programs and policies have been targeted to
the needs of female gang members (see Curry, 1999b; Williams, Cohen, and
Curry, this volume). Given the findings detailed above, the best strategy for
addressing young women's gang involvement should include policies that take
into account the social, economic, and personal contexts that influence gang
participation, gang crime, and young women's victimization within these groups.
Initiatives that consider the best interest of youths are needed to respond
rationally to gangs and young women's involvement in them.

Notes

1. For instance, Curry, Ball, and Fox (1994) found that some law enforcement
policies officially exclude females from their counts of gang members.
Controlling for data from these cities, they still found that females were only 5.7
percent of the gang members known to law enforcement agencies. Part of the
reason for this underestimation may be the greater likelihood of male gang
members to be involved in serious gang crimes and the average age differences
between male and female gang members (see Curry 1999a).

2. An extensive body of evidence suggests that self-nomination is a robust
measure of gang membership (see Bjerregaard and Smith 1993; Esbensen et al.
1993; Esbensen and Huizinga 1993; Fagan 1990; Thornberry et al. 1993;
Winfree et al. 1992). It is not foolproof, however. One potential problem is that
"wannabes" may claim gang membership when they are not actually involved.
To contend with this potential problem, the veracity of youths' accounts and
their depth of knowledge was checked by comparing them to other youths'
descriptions of their gangs and other available information. No significant
inconsistencies were found. An additional benefit of using self-nomination was
that it provided a means of applying limited definitional criteria to capture what
may be a varied phenomenon (see Horowitz 1990).

3. In both cities, the population is primarily African-American and white--in
Columbus, the two groups together comprise 95.8 percent of the population; in
St. Louis, 98.4 percent (U.S. Census Bureau 1990).

4. In response to a survey by the National Youth Gang Center, the St. Louis
Police Department recently described the city as having 75 active gangs and
1,300 gang members. In contrast, the Columbus Police Department recently
adopted an official policy that the city has no recognized youth gangs (Curry
1998). (For a discussion of the politics of gang definitions and law enforcement
response to gangs, see Huff 1990.)

5. In fact, the St. Louis police department's recent response to the National
Youth Gang Center estimates that St. Louis gangs are 100 percent male (Curry
1998).

6. Findings from the Rochester Youth Development Study (RYDS) also
suggest that growing up in disorganized, violent neighborhoods is a risk factor
for gang involvement for young women (Thornberry 1997). Yet researchers
have found no differences between gang and nongang youths in these
communities in perceived limited opportunities (Esbensen et al. 1993), and
family poverty was not a significant predictor of gang membership for young
women in the RYDS study.

7. Significance levels are based on chi-square tests. Note that although statistics
are used to make comparisons throughout the chapter, the sample is purposive
and thus violates key assumptions about random or representative sampling.
Although, technically, statistical methods are inappropriate for this sample, the
methods are used here not in an attempt to generalize to a larger population but
to highlight the strength of the patterns uncovered.

8. No young woman in this study reported having been forced to join a gang.

9. Although qualitative studies are most likely to find family problems at the
heart of girls' gang involvement, several studies based on surveys of juvenile
populations note school-based problems (see Thornberry 1997 for an
overview). Bjerregaard and Smith (1993) found that low expectations for
completing school were a significant predictor of gang membership for young
women. Likewise, Bowker and Klein (1983) report that female gang members
are less likely than nonmembers to intend to finish high school or go to college.
More recently, Esbensen and Deschenes (1998) report that school commitment
and expectations are associated with gang involvement for girls. In the current
study, differences between gang members and nonmembers did not emerge in
school measures, nor did gang members describe school contexts as
contributing to their decisions to join. This may be a result of difference across
sites, or the result of sampling or methodological differences. (See Esbensen
and Winfree 1998 for a discussion of sampling and methodology.)

10. While the RYDS study did not find family violence a significant risk factor
for female gang involvement (Thornberry 1997), a growing body of literature
supports the link between childhood maltreatment and youths' subsequent
involvement in delinquency. (See Smith and Thornberry 1995; Widom 1989.)

11. One member of an all-girl gang in Columbus and two in St. Louis were
interviewed, in addition to three St. Louis girls who were members of a female
set affiliated with a neighborhood mixed-gender gang. Though there have been
recent critiques of Walter Miller's (1975) tripartite classification of female gang
involvement as occurring in mixed-gender gangs, independent female gangs,
and auxiliary subgroups of male gangs (see Hagedorn and Devitt 1999;
Peterson et al. 2001; Nurge 1998), for the purpose of this paper, this
classification is maintained.

12. Traditional gangs are found in cities with longer histories of gangs; they have
a history spanning 20 years or more, include more than 100 members, have a
wide age range (20 to 30 years), subgroupings, and are territorial with versatile
crime patterns. Neotraditional gangs also have subgroupings, are territorial, and
have versatile crime patterns, but they are somewhat smaller (50 to 100
members) and have a narrower age range and a shorter duration. Compressed
and collective gangs have no subgroups, may or may not be territorial, and have
versatile crime patterns. While compressed gangs are small (fewer than 50
members), with a narrow age range (typically less than 10 years) and short
duration, collective gangs are larger (50 or more members), have a wider age
range (typically more than 10 years), and their duration can be somewhat
longer. Finally, specialty gangs, which were not found in either city, specialize in
particular types of crimes (such as drug distribution) and are less social; they
typically do not have subgroups, are quite territorial, and are small, with a
narrow age range and short duration. (See Klein and Maxson 1996; Maxson
and Klein 1995.)

13. Yet there is evidence that gang involvement has declined in St. Louis over
the last few years, and police suggest a splintering of gangs has occurred
(Decker 1998). This is also reflected in the decline in gang homicide and the
decline in the percentage of gang homicides that are gang motivated (see
Rosenfeld, Bray, and Egley 1999). Consequently, the likelihood that these
groups will emerge as traditional gangs does not appear to be great.

14. Minor delinquency includes the following: running away, skipping class,
lying about one's age, being loud or rowdy in public, avoiding paying for things,
and stealing $5 or less. Six items were classified as moderate delinquency:
being drunk in public, damaging or destroying property, stealing items valued
between $5 and $50, using a car without permission, throwing objects like
bottles or rocks at people, and hitting someone with the intent to cause injury.
Finally, serious delinquency included carrying a concealed weapon, stealing
items valued at more than $50, stealing a car, attacking someone with a
weapon or with the intent to cause serious injury, participating in a gang fight,
and committing a robbery. (See Bjerregaard and Smith 1992.)

15. Yet it is important to note that female gang members' most frequent serious
offense was carrying concealed weapons. When this item was removed from
the count, gang girls committed an average of 13.78 serious crimes in the last 6
months--slightly more than 2 per month.

16. Seven gang members said that they sold drugs every day. Each day was
counted as an incident of drug sales, rather than each sale, making the mean
frequency reported an underestimation of the actual number of times they sold
drugs. Moreover, one nonmember reported selling marijuana every day. If
these sales (n = 180) were removed from the calculation of nonmembers' mean
frequency of participation in drug sales, all nonmembers reported selling drugs
12 times in the last 6 months (for a mean of .26).

17. All of the names of the gang members quoted in this chapter are fictitious.

18. The woman in question was not a gang member, but because her victim
was a gang member, the homicide was classified as gang related. (See Miller
and Decker 2001.)

19. All of the young women interviewed described drug sales for individual
profit, rather than for their gangs.

20. In an analysis not shown here, bivariate correlations were run between the
frequency of victimization and exposure to violence (measured as never, once,
a few times, many times) and the frequency of serious offending and crack sales
(measured as the number of times young women reported having committed
these offenses in the past 6 months). Frequent involvement in serious
delinquency was significantly correlated with several types of victimization and
exposure to violence, including being threatened with a weapon and stabbed
and witnessing guns shot, stabbings, shootings, driveby shootings, and
homicides. Frequent involvement in crack sales was correlated with having
been stabbed, seen guns fired, and witnessed driveby shootings. These findings
further suggest that girls' patterns of offending shape their exposure to violence.
(See Miller 2001.)

References

Bjerregaard, Beth, and Carolyn Smith. 1993. "Gender Differences in Gang
Participation, Delinquency, and Substance Use." Journal of Quantitative
Criminology 9(4): 329-355.

Bowker, Lee H., Helen Shimota Gross, and Malcolm W. Klein. 1980. "Female
Participation in Delinquent Gang Activities." Adolescence 15 (59): 509-519.

Bowker, Lee H., and Malcolm W. Klein. 1983. "The Etiology of Female
Juvenile Delinquency and Gang Membership: A Test of Psychological and
Social Structural Explanations." Adolescence 18(72): 739-751.

Campbell, Anne. 1990. "Female Participation in Gangs." In Gangs in America,
ed. C. Ronald Huff. Newbury Park, CA: Sage Publications: 163-182.

Columbus Metropolitan Human Services Commission. 1995. State of Human
Services Report--1995. Columbus, OH: Columbus Metropolitan Human
Services Commission.

Community Development Agency. 1993. Neighborhood Demographic Profiles:
1990. St. Louis: City of St. Louis.

Curry, G. David. 1998. Personal correspondence, August 13.

-------. 1999a. "Race, Ethnicity, and Gender Issues in Gangs: Reconciling
Police Data." In Problem-Oriented Policing: Crime-Specific Issues and Making
POP Work, vol. 2, ed. Corina Sole Brito and Tracy Allan, Washington, DC:
Police Executive Research Forum: 63-89.

-------. 1999b. "Responding to Female Gang Involvement." In Female Gangs
in America, ed. Meda Chesney-Lind and John Hagedorn. Chicago: Lake View
Press: 277-294.

Curry, G. David, Richard A. Ball, and Robert J. Fox. 1994. Gang Crime and
Law Enforcement Recordkeeping. Research in Brief. Washington, DC: U.S.
Department of Justice, National Institute of Justice, June, NCJ 148345.

Decker, Scott H. 1996. "Collective and Normative Features of Gang
Violence." Justice Quarterly 13(2): 243-264.

-------. 1998. Personal correspondence. September 6.

Decker, Scott H., Susan Pennell, and Amy Caldwell. 1996. Arrestees and
Guns: Monitoring the Illegal Firearms Market. Final Report submitted to the
National Institute of Justice. Washington, DC: National Institute of Justice, NCJ
163497.

Decker, Scott H., and Barrik Van Winkle. 1996. Life in the Gang. Cambridge:
Cambridge University Press.

Deschenes, Elizabeth Piper, and Finn-Aage Esbensen. 1999. "Violence Among
Girls: Does Gang Membership Make a Difference?" In Female Gangs in
America, ed. Meda Chesney-Lind and John Hagedorn. Chicago: Lake View
Press: 277-294.

Esbensen, Finn-Aage, and Elizabeth Piper Deschenes. 1998. "A Multi-Site
Examination of Gang Membership: Does Gender Matter?" Criminology 36:
799-828.

Esbensen, Finn-Aage, and David Huizinga. 1993. "Gangs, Drugs, and
Delinquency in a Survey of Urban Youth." Criminology 31(4): 565-589.

Esbensen, Finn-Aage, David Huizinga, and Anne W. Weiher. 1993. "Gang and
Non-Gang Youth: Differences in Explanatory Factors." Journal of
Contemporary Criminal Justice 9(2): 94-116.

Esbensen, Finn-Aage, and L. Thomas Winfree. 1998. "Race and Gender
Differences Between Gang and Non-Gang Youth: Results from a Multi-Site
Survey." Justice Quarterly 15: 505-525.

Fagan, Jeffrey. 1989. "The Social Organization of Drug Use and Drug Dealing
Among Urban Gangs." Criminology 27 (4): 633-667.

-------. 1990. "Social Processes of Delinquency and Drug Use Among Urban
Gangs." In Gangs in America, ed. C. Ronald Huff. Newbury Park, CA: Sage
Publications: 183-219.

Fleisher, Mark S. 1998. Dead End Kids: Gang Girls and the Boys They Know.
Madison: University of Wisconsin Press.

Glassner, Barry, and Cheryl Carpenter. 1985. The Feasibility of an
Ethnographic Study of Property Offenders. Final Report submitted to the
National Institute of Justice, Washington, DC: National Institute of Justice, NCJ
107044.

Glassner, Barry, and Julia Loughlin. 1987. Drugs in Adolescent Worlds:
Burnouts to Straights. New York: St. Martin's Press.

Hagedorn, John M. 1991. "Gangs, Neighborhoods, and Public Policy." Social
Problems 38: 529-542.

-------. 1998. People and Folks: Gangs, Crime and the Underclass in a
Rustbelt City, 2nd ed. Chicago: Lake View Press.

Hagedorn, John M., and Mary Devitt. 1999. "Fighting Female: The Social
Construction of Female Gangs." In Female Gangs in America: Essays on Girls,
Gangs and Gender, ed. Meda Chesney-Lind and John Hagedorn. Chicago:
Lake View Press: 256-276.

Horowitz, Ruth. 1990. "Sociological Perspectives on Gangs: Conflicting
Definitions and Concepts." In Gangs in America, ed. C. Ronald Huff. Newbury
Park, CA: Sage Publications: 37-54.

Huff, C. Ronald. 1989. "Youth Gangs and Public Policy." Crime and
Delinquency. 35(4): 524-537.

-------. 1990. "Denial, Overreaction, and Misidentification: A Postscript on
Public Policy." In Gangs in America, ed. C. Ronald Huff. Newbury Park, CA:
Sage Publications: 310-317.

-------. 1993. "Gangs in the United States." In The Gang Intervention
Handbook, ed. Arnold P. Goldstein and C. Ronald Huff. Champaign, IL:
Research Press: 3-20.

-------. 1998. Comparing the Criminal Behavior of Youth Gangs and At-Risk
Youths. Research in Brief. October. Washington, DC: U.S. Department of
Justice, National Institute of Justice, NCJ 172852.

Joe, Karen A., and Meda Chesney-Lind. 1995. "'Just Every Mother's Angel':
An Analysis of Gender and Ethnic Variations in Youth Gang Membership."
Gender and Society 9(4): 408-430.

Klein, Malcolm W. 1971. Street Gangs and Street Workers. Englewood Cliffs,
NJ: Prentice-Hall.

-------. 1995. The American Street Gang: Its Nature, Prevalence, and Control.
New York: Oxford University Press.

Klein, Malcolm W., and Cheryl L. Maxson. 1989. "Street Gang Violence." In
Violent Crime, Violent Criminals, ed. Neil Weiner and Marvin Wolfgang.
Newbury Park, CA: Sage Publications: 198-231.

-------. 1996. "Gang Structures, Crime Patterns, and Police Responses." Final
Report to the National Institute of Justice. Washington, DC: National Institute
of Justice, NCJ 188511.

Lauritsen, Janet L., Robert J. Sampson, and John H. Laub. 1991. "The Link
Between Offending and Victimization Among Adolescents." Criminology 29(2):
265-292.

Maxson, Cheryl L., and Malcolm W. Klein. 1995. "Investigating Gang
Structures." Journal of Gang Research 3(1): 33-40.

Maxson, Cheryl L., Kristi Woods, and Malcolm W. Klein. 1995. "Street Gang
Migration in the United States." Final Report to the National Institute of Justice.
Washington, DC: NCJ 172235.

Mayhood, Kevin, and Brent LaLonde. 1995. "A Show of Colors: A Local
Look." The Columbus Dispatch, September 3.

Messerschmidt, James W. 1995. "From Patriarchy to Gender: Feminist
Theory, Criminology and the Challenge of Diversity." In International Feminist
Perspectives in Criminology: Engendering a Discipline, ed. Nicole Hahn Rafter
and Frances Heidensohn. Philadelphia: Open University Press: 167-188.

Miller, Jody. 1998. "Gender and Victimization Risk Among Young Women in
Gangs." Journal of Research in Crime and Delinquency 35(4): 429-453.

-------. 2001. One of the Guys: Girls, Gangs, and Gender. New York: Oxford
University Press.

Miller, Jody, and Rod K. Brunson. 2000. "Gender Dynamics in Youth Gangs:
A Comparison of Male and Female Accounts." Justice Quarterly 17(3):
801-830.

Miller, Jody, and Scott H. Decker. 2001. "Young Women and Gang Violence:
An Examination of Gender, Street Offending and Violent Victimization in
Gangs." Justice Quarterly 18(1): 115-140.

Miller, Jody, and Barry Glassner. 1997. "The 'Inside' and the 'Outside': Finding
Realities in Interviews." In Qualitative Research ed. David Silverman. London:
Sage Publications: 99-112.

Miller, Walter. 1975. Violence by Youth Gangs and Youth Groups as a Crime
Problem in Major American Cities. Washington, DC: U.S. Department of
Justice, Office of Juvenile Justice and Delinquency Prevention, NCJ 137446.

Moore, Joan. 1991. Going Down to the Barrio: Homeboys and Homegirls in
Change. Philadelphia: Temple University Press.

Moore, Joan, and Raquel Pinderhughes. 1993. In the Barrios: Latinos and the
Underclass Debate. New York: Russell Sage Foundation.

Nurge, Dana. 1998. "Female Gangs and Cliques in Boston: What's the
Difference?" Paper presented at the annual meeting of the American Society of
Criminology, Washington, DC.

Peterson, Dana, Jody Miller, and Finn-Aage Esbensen. 2001. "The Impact of
Sex Composition on Gangs and Gang Member Delinquency." Criminology
39(2): 411-439.

Rosenfeld, Richard, Timothy Bray, and Arlen Egley. 1999. "Facilitating
Violence: A Comparison of Gang-Motivated, Gang-Affiliated, and Non-Gang
Youth Homicides." Journal of Quantitative Criminology 15(4): 495-516.

Rusk, David. 1995. Cities Without Suburbs. 2nd ed. Washington, DC:
Woodrow Wilson Center Press.

Sanders, William. 1993. Drive-Bys and Gang Bangs: Gangs and Grounded
Culture. Chicago: Aldine.

Sampson, Robert J., and William Julius Wilson. 1995. "Toward a Theory of
Race, Crime and Urban Inequality." In Crime and Inequality, ed. John Hagan
and Ruth D. Peterson. Stanford, CA: Stanford University Press: 37-54.

Short, James F., and Fred L. Strodtbeck. 1965. Group Process and Gang
Delinquency. Chicago: University of Chicago Press.

Smith, Carolyn, and Terence P. Thornberry. 1995. "The Relationship Between
Childhood Maltreatment and Adolescent Involvement in Delinquency."
Criminology 33(4): 451-479.

Thornberry, Terence P. 1997. "Membership in Youth Gangs and Involvement
in Serious and Violent Offending." In Serious and Violent Juvenile Offenders:
Risk Factors and Successful Interventions, ed. Rolf Loeber and David P.
Farrington. Thousand Oaks, CA: Sage Publications: 147-166.

-------. 1999. Personal correspondence, April 2.

Thornberry, Terence P., Marvin D. Krohn, Alan J. Lizotte, and Deborah
Chard-Wierschem. 1993. "The Role of Juvenile Gangs in Facilitating
Delinquent Behavior." Journal of Research in Crime and Delinquency 30(1):
75-85.

U.S. Census Bureau. 1990. 1990 Decennial Census of Population and Housing
Characteristics. Washington, DC: U.S. Census Bureau.

Widom, Cathy Spatz. 1989. "Child Abuse, Neglect, and Violent Criminal
Behavior." Criminology 27(2): 251-271.

Wilson, William Julius. 1987. The Truly Disadvantaged: The Inner City, the
Underclass, and Public Policy. Chicago: University of Chicago Press.

-------. 1996. When Work Disappears: The World of the New Urban Poor.
New York: Alfred A. Knopf.

Winfree, L. Thomas, Jr., Kathy Fuller, Teresa Vigil, and G. Larry Mays. 1992.
"The Definition and Measurement of 'Gang Status': Policy Implications for
Juvenile Justice." Juvenile and Family Court Journal 43(1): 29-37.

------------------------------

Chapter 4

Youth Gang Homicides in the United States in the 1990s

Cheryl L. Maxson, Ph.D.
Department of Criminology, Law, and Society
University of California-Irvine

G. David Curry, Ph.D.
Department of Criminology and Criminal Justice, Department of Sociology
University of Missouri-St. Louis

James C. Howell, Ph.D.
Adjunct Researcher, National Youth Gang Center

Acknowledgments

This research was supported by cooperative agreement 95-JD-MU-K001
between the Institute for Intergovernmental Research (IIR) and the Office of
Juvenile Justice and Delinquency Prevention, U.S. Department of Justice. The
datasets used for this report were acquired from IIR's National Youth Gang
Center (NYGC), funded by the Office of Juvenile Justice and Delinquency
Prevention, and from studies conducted by the first and second authors that
were supported by the National Institute of Justice. Initial work on the chapter
was performed under NYGC auspices. The authors are grateful for the
technical assistance of Arlen Egley, NYGC, and suggestions made on an earlier
version of the chapter by Scott H. Decker, University of Missouri-St. Louis,
and an anonymous reviewer. A summary of this study was published as an
OJJDP Fact Sheet (Curry, G. David, Cheryl L. Maxson, and James C.
Howell, Youth Gang Homicides in the 1990's, Washington, DC: U.S.
Department of Justice, Office of Juvenile Justice and Delinquency Prevention,
March 2001, FS 200103.)

------------------------------

Abstract

In this study of youth gang homicides in U.S. cities in the 1990s, the authors
found that although the numbers fell, they varied from city to city and in the
early and later parts of the decade. From 1991 to 1996, when 408 cities were
studied, the total number of gang homicides declined nearly 15 percent from
1,748 to 1,492. Of the 237 cities that reported having both a gang problem and
a gang homicide in all 3 years of the period 1996 to 1998, the total number of
gang homicides fell 18 percent from 1,293 to 1,061. Although these decreases
are heartening and relatively few cities reported large numbers of gang
homicides, they conceal the fact that one city--Los Angeles--accounted for 30
percent of the decrease in the early part of the decade. Along with Chicago,
Los Angeles stands out as having the highest gang homicide rate. Moreover, the
percentage of cities reporting an increase in that period (29 percent) was about
the same as the percentage reporting a decrease (32 percent); 39 percent
reported no change. The second part of the decade brought news that was
more encouraging, as more cities reported a decrease (49 percent) than an
increase (36 percent).

------------------------------

Gangs have become a focus for lawmakers, political candidates, and police
chiefs, as well as for researchers and Federal agencies, because gang members
commit a lot of crime, particularly as compared with other young people.
Findings from recent long-term studies of youths in several U.S. cities reveal
that those identifying themselves as gang members commit the vast majority of
serious violent offenses reported for their age group (Thornberry 1998). For at
least the past two decades, law enforcement has responded to gang violence
with a variety of suppression efforts, many of which are described in this
volume. These efforts are based primarily in individual jurisdictions, and thus are
most appropriately framed in the historical, political, and sociodemographic
context of those communities.

This chapter focuses on national patterns of gang homicide, specifically on
whether this most serious form of gang violence increased or decreased during
the 1990s. This analysis is limited to homicide because homicide is the only
form of gang crime for which there are national data. Moreover, homicide is the
gang crime most likely to raise community concern, elicit political reaction, and
trigger suppressive responses by law enforcement. Thus, in a volume dedicated
to "responding to gangs," it is useful to include a broad, empirically based
depiction of the changes in national gang homicide levels over the past decade
of gang research and program evaluation. This information would also be
important for policy purposes, such as the allocation of Federal resources. The
identification of factors associated with changes in gang homicide is essential to
understanding gang violence and to planning policies and strategies that address
it.

Available data on national gang homicide trends have several limitations; local
area analyses provide far richer prospects to guide policy and interventions. Yet
local and regional policymakers may benefit from placing gang violence trends
within the broader context of the national patterns reported here. Federal
policymakers also should find the conclusions of this study useful.

A secondary objective of this study was to integrate gang homicide data from
three studies to build a reliable national database that can be used to examine
future gang homicide trends. The limitations of the available data restrict the
scope of the analysis, but if expanded support for gang homicide reporting were
forthcoming, it could substantially increase the utility of such analyses.

This chapter reports the results of a study of youth gang[1] homicides in the
1990s. The study comprises three components: analysis of the trend in gang
homicides from the early to mid-1990s; analysis of the pattern of gang
homicides in the late 1990s; and examination of selected city-level variables
(population size, number of gangs, and number of gang members) associated
with gang homicides.

Three gang databases were used to examine gang homicides reported in
selected cities in 1991 and again in the mid-1990s. For the mid- to late 1990s
comparison, National Youth Gang Survey (NYGS) data on 99 percent of all
cities with a population of more than 25,000 were analyzed. The NYGS
database was also used to examine several correlates of gang homicides. The
study set out to determine whether gang homicides increased during the decade
and to examine factors related to different homicide trends.

Background

Two national databases are commonly used by researchers to examine overall
homicide trends: mortality data from Vital Statistics of the United States
(compiled by the National Center for Health Statistics) and the Supplementary
Homicide Reports (SHR) in the Uniform Crime Reports compiled by the
Federal Bureau of Investigation (FBI). Unfortunately, gang homicide data are
not reported in Vital Statistics of the United States, and the data contained in
the Supplementary Homicide Reports are of questionable utility for gang
research (see Maxson 1998a for a detailed discussion; see also Curry, Ball,
and Decker 1995, 1996a, 1996b; Curry, Ball, and Fox 1994; and Miller
1992). Thus, surveys of law enforcement agencies are used to obtain data on
gang homicides that are maintained in law enforcement records. The information
in these records is limited to what is known and documented by law
enforcement investigators. Researchers attempting to extract information from
police investigation files quickly confront these limitations. As one researcher
noted, "[I]nformation on some coded items may be missing, conflicting, or
otherwise confusing so as to challenge the skills of the most skilled data
collectors" (Maxson 1992: 199). Surveys of law enforcement that span multiple
cities, such as the studies included in this analysis, generally are based on
reports of the number of gang homicides during a specified period, obtained
from a jurisdiction's designated gang expert or data analyst. Such surveys do
not capture the incident-level descriptors (e.g., characteristics of the incident
circumstances, settings, participants) obtained by researchers that extract data
elements directly from homicide investigation files.

A growing body of empirically based literature describes the gang aspects of
homicide in a number of U.S. cities. Several studies of Los Angeles (by
Maxson, Klein, and colleagues), Chicago (by C. Block, R. Block, and
colleagues) and St. Louis (by Rosenfeld, Decker, and colleagues) have tracked
the characteristics of gang homicides over several decades (see Howell 1999
and Maxson 1998a for a review of these studies). Other gang homicide studies
have been conducted in Boston (Kennedy, Braga, and Piehl 1997; Kennedy,
Piehl, and Braga 1996), Chicago (Curry and Spergel 1988), Houston (Brewer,
Damphousse, and Adkinson 1998), Los Angeles (Hutson, Anglin, and Pratts
1994; Hutson et al. 1995, 1999), Minneapolis (Kennedy and Braga 1998),
and Pittsburgh (Cohen and Tita 1998). Comparative studies of gang homicide
have been conducted in St. Louis and Chicago (Cohen et al. 1998), and in
Philadelphia, Phoenix, and St. Louis (Zahn and Jamieson 1997).

Previous Multicity Gang Homicide Surveys

The first multicity gang homicide data were reported by Miller (1975, 1992) in
conjunction with his national survey of cities with gang problems. He tabulated
gang-related killings from 1967 to 1980 for 9 of the Nation's largest cities and
50 other cities for which statistics were available and made various
comparisons among the major cities with gang problems (Los Angeles,
Chicago, Philadelphia, San Francisco, Boston, Miami, Detroit, San Antonio,
and New York). Arrest data Miller compiled show that gang killings were
concentrated in large cities. Reported gang killings in the 59 cities increased
from 181 in 1967 to 633 in 1980, an increase of 250 percent. The total number
of gang homicides over the 14-year period was 3,509--an average of 250
killings per year. The leading gang homicide cities during the 14-year period of
Miller's study, in average number of killings per year, were Los Angeles (75),
Chicago (55), and Philadelphia (25). Homicide rates for youths aged 10 to 19
in these cities were 31.9 per 100,000 in Los Angeles, 18.5 in Chicago, and
14.7 in Philadelphia.

Two surveys elicited gang homicide data for 1991 from law enforcement
agencies. In the one conducted by Curry and his colleagues (Curry et al. 1992;
Curry, Ball, and Decker 1995) among the 72 largest U.S. cities, 41 agencies
reported a total of 964 gang homicides, 53 percent of which were committed in
Chicago and Los Angeles. In the Maxson-Klein survey (Klein 1995; Maxson
1998b; Maxson, Woods, and Klein 1995), police departments in 792 cities
were asked about gang homicides during 1991. A total of 299 cities (40
percent of those surveyed) reported 2,166 gang-related homicides in that year,
40 percent of which were committed in just 12 cities and 23 percent of which
(500 homicides) were committed in Los Angeles and Chicago alone (Klein
1995; Maxson 1998a).

Measuring Gang Homicides

Counts of gang homicides are vulnerable to challenges about reliability and
validity. Curry and his colleagues note that "the major sources of error in
conducting national assessments of gang problems are matters of policy, not
technology" (Curry, Ball, and Decker 1996b: 32). Local law enforcement
databases are designed to track individual gang members for possible
apprehension, not to compile aggregate statistical information. Compiling
national gang homicide data through surveys of law enforcement agencies
involves asking them "to provide us with a service they may not routinely
provide for local assessment and policy making" (p. 33). This must be kept in
mind when using law enforcement data in compiling aggregate statistics.

Major Measurement Issues

The most critical issue in measuring gang homicides concerns the definition used
in attributing homicides to gangs. Some law enforcement agencies count only
gang-motivated homicides--incidents that grow out of a gang function, such as
territorial disputes, affiliation challenges, or ongoing rivalries between gangs.
Other agencies use a more expansive definition, counting all events in which a
gang member is involved in any capacity (gang-related or gang-involved). This
factor can make a big difference in the resulting homicide count. Maxson and
Klein (1990, 1996) found that the motive-based definitional approach law
enforcement agencies use in Chicago produced homicide estimates only half as
large as in Los Angeles, where agencies use the member-based recording
method, although the depictions of the nature of Los Angeles gang homicides
were remarkably similar, regardless of the definition used.

Fifty-eight percent of those who responded to the 1998 National Youth Gang
Center (NYGC) survey said they used the member-based definition, 32
percent used the motive-based definition, and 10 percent used another
definition (National Youth Gang Center 2000; see also Spergel and Curry
1993). Some agencies have positioned themselves to report counts using either
definition (e.g., Los Angeles and Bloomington, IL). Nevertheless, the difficulty
that police have in distinguishing gang-motivated from gang-involved homicides
(Weisel and Painter 1997) produces considerable variation in the recording of
gang homicides.

There are significant differences among law enforcement agencies and State
statutes in how they define "gang" (Ball and Curry 1995; Curry et al. 1992;
Klein 1995; Spergel 1995; Spergel and Curry 1993). The terms "youth gang"
and "street gang" often are used interchangeably. "Criminal street gangs," a term
formerly used exclusively to refer to adult criminal organizations, is now applied
to youth gangs. Some jurisdictions call youth gangs by other names, such as
"crews" or "posses." This ambiguity in gang definitions and offense
classifications casts doubt on the validity of official gang homicide records and,
consequently, on national surveys based on these records and police estimates.

These are not the only measurement problems that limit the quality of gang
homicide data (Maxson 1998a). Another is underreporting and
underestimation. The proportion of unsolved homicides has increased
significantly since the mid-1980s (Snyder, Sickmund, and Poe-Yamaguchi
1996), partly as a result of witness intimidation (Maxson 1992). Other factors
that lead to inaccurate reporting of gang homicides include the difficulties police
encounter in investigating them, the lack of valid rosters of gang members, and
the chaotic situations in which many homicides are committed (Maxson 1998a).
Finally, acknowledging the existence of a gang problem is a prerequisite for
counting gang homicides, at least for the analyses reported here.[2] The
NYGC, which conducts the annual National Youth Gang Survey (NYGS) used
in the analyses reported here, readily acknowledges that law enforcement
reporting on gangs is subject to local political considerations that may lead to
exaggeration or denial of gang problems (NYGC 1999a). Approximately half
of the respondents reported gang problems in each of the 1996, 1997, and
1998 NYGC surveys.

Underestimation of gang homicide is thus a concern for researchers who use
law enforcement survey estimates. Yet underestimation is even more apparent
in the Supplemental Homicide Report (SHR) data (Bailey and Unnithan 1994;
Riedel forthcoming). The SHR reports only two categories of gang homicides:
"gangland killings" (organized crime) and "juvenile gang killings." By definition,
the latter category excludes young adult gang homicides. At least half of all gang
homicide perpetrators and victims are at least 18 years old (Block et al. 1996;
Klein and Maxson 1989).

How law enforcement officials view gang crime as a public safety concern
could affect homicide reporting. For various reasons, officials in some cities
deny the existence of gang problems altogether. In others, gang problems are
downplayed as a matter of policy. For example, Curry and colleagues (Curry,
Ball, and Decker 1996b) report that when their 1991 gang survey was
conducted, the New York City Police Department included only crimes
committed by Asian gangs. (At that time, Asian gangs were the gang-crime
unit's only concern; African-American and Latin "drug organizations" came
under the jurisdiction of the narcotics unit.) By 1994, New York City officials
had developed a new system for identifying gang-related crimes and incidents
that included African-American and Latin gang crime, and what was formerly
identified as Asian gang crime was reclassified as part of "organized crime."
Conversely, local law enforcement officials might overestimate gang homicides
in response to "moral panics" (Jackson 1993; Zatz 1987), local political
pressure, or unreliable witness accounts.

Reporting Trends

Two trends can be seen in gang data reporting today. States are increasingly
expanding the domain of illegal gang activities for adjudication and sentencing
purposes (Bureau of Justice Assistance 1997). Examples include enhanced
penalties for shooting from an automobile, prohibitions against congregating in
public areas, nuisance abatement laws, and severe sentences for gang
offenders. Furthermore, gangs are increasingly defined more broadly for law
enforcement intelligence and recordkeeping purposes. The new National
Incident Based Reporting System (NIBRS) data collection uses the National
Crime Information Center's definition of "gang": "An ongoing organization,
association, or group of three or more persons [that] have a common interest
and/or activity characterized by the commission of or involvement in a pattern
of criminal or delinquent conduct" (FBI 1997a: 1). Similar definitions,
incorporated into statutory codes in several States (for example, in California's
Street Terrorism, Enforcement and Prevention Act), have likely influenced local
jurisdictions' practices of defining gangs.

The Uniform Crime Reporting (UCR) program is in the process of changing the
way it collects information about gangs. In the late 1970s, the law enforcement
community called for a thorough evaluative study of UCR to recommend ways
the program could be expanded and enhanced to meet law enforcement needs
into the 21st century. NIBRS was implemented in 1987 to provide more
detailed information on crime incidents. As part of the Violent Crime Control
and Law Enforcement Act of 1994 (also known as "The Crime Act"),
Congress mandated that the Director of the FBI acquire and collect information
on incidents of gang violence for inclusion in its annual Uniform Crime Report.
The FBI began assembling information on incidents of gang violence for 11
violent offenses in September 1997. The Bureau has made substantial progress
in establishing NIBRS as a viable, national gang-related crime statistics
collection system, but it will be several years before the scope of coverage is
sufficient to provide representative national data.

In the meantime, surveys of law enforcement agencies are the only source of
national-level data on gang homicide and other gang crimes. A national
database of gang homicides was constructed for this study using previous gang
surveys of major cities and the NYGC national gang surveys conducted from
1995 to 1998.

Study Methodology

Cities were selected for inclusion in the study if they reported the number of
gang-related homicides in their jurisdiction on surveys conducted by three
teams of researchers covering the period from 1991 to 1998. The questions
eliciting gang homicide data were similar in all three studies: the 1991 Curry
survey (Curry et al. 1992), the 1991 Maxson-Klein survey (Maxson 1998a),
and the 1995-1998 National Youth Gang Surveys (Egley 2000; NYGC 1997,
1999a, 1999b, 2000).

Gang homicide data from these three surveys were examined in two ways.
First, the trend from 1991 to the mid-1990s was identified for selected cities,
surveyed by either Maxson and Klein or Curry in 1991, that also provided
homicide data in the NYGC survey in 1996. The city matching for 1991 and
1996 was the only explicit comparison that could be made, because the three
surveys used different samples. If no 1996 data were available for one of the
cities, data from the 1995 NYGS were used for the mid-1990s comparison.
This city-by-city comparison produced a view of gang homicide trends in
selected cities in the first half of the decade. Second, gang homicide trends from
1996 to 1998 were examined using a single source, the NYGC survey.
Together, these two sets of analyses produce a composite view of gang
homicide trends from 1991 to 1998.

Curry's 1991 Survey

In the National Assessment of Law Enforcement Anti-Gang Information
Resources study (Curry, Ball, and Fox 1994), G. David Curry surveyed police
departments in all 79 U.S. cities with populations of at least 200,000 by 1990
census estimates. He also included 43 smaller cities and selected counties that
had been surveyed in 1988 by the National Youth Gang Suppression and
Intervention Program (Spergel and Curry 1990, 1993). In Curry's survey,
a group had to meet the following criteria to be identified as a youth gang:

o Be identified as a "gang" by law enforcement.

o Involve some level of participation by juveniles.

o Engage in criminal activity.

More than 90 percent of the police departments contacted responded to the
survey, and the 72 largest cities all reported at least 1 gang-related homicide.

The Maxson-Klein 1991 Survey

The Maxson-Klein National Gang Migration study (Klein 1995; Maxson
1998a; Maxson, Woods, and Klein 1995, 1996) surveyed police departments
in 1,105 cities. All 190 U.S. cities with populations of more than 100,000 were
included. Smaller cities and towns were added to the sample, based on other
information about cities that either had local street gangs or had experienced
gang migration. For purposes of the survey, gangs were defined as "groups of
adolescents or young adults who see themselves as a group (as do others) and
have been involved in enough crime to be of considerable concern to law
enforcement and the community" (Maxson, Woods, and Klein 1995: 10).
Survey recipients were instructed to include drug gangs in their response. They
were asked to provide the number of gang-related homicides in the reporting
jurisdiction for 1991. More than 90 percent of the police departments
contacted responded, and 299 cities reported at least 1 gang-related homicide.

The National Youth Gang Center Surveys

The 1996-98 NYGC surveys were administered to a representative sample of
U.S. city and county jurisdictions. The same jurisdictions were contacted in
each survey. The sample consisted of--

o All 1,216 police departments serving cities with populations of 25,000 or
more.

o All 662 suburban-county police and sheriffs' departments (FBI 1995).

o A randomly selected representative sample of 400 police departments serving
cities with populations between 2,500 and 25,000. --A randomly selected
representative sample of 742 rural-county police and sheriffs' departments (FBI
1995).

The youth gang definition in the NYGC surveys was broader than those used in
the two 1991 surveys. For the purposes of the 1996-98 NYGC surveys, a
youth gang was defined as "a group of youths in your jurisdiction that you or
other responsible persons in your agency or community are willing to identify or
classify as a 'gang'" (NYGC 1999a: 56). Respondents were asked to exclude
motorcycle gangs, hate or ideological groups, prison gangs, and adult gangs.
The 1996 survey asked respondents: "How many homicides involving gang
members occurred in your jurisdiction in 1996?" A followup question clarified
whether this figure was limited to gang-motivated homicides. Subsequent
surveys requested separate counts of member- and motive-based homicides
(NYGC 1999b: 48; NYGC 2000: 54).

In 1996, 2,629 of the 3,018 agencies contacted responded to the survey (an
87 percent response rate) (NYGC 1999a). In 1997, the response rate was 92
percent (NYGC 1999b), and in 1998, it was 88 percent (NYGC 2000). (For
additional information on the methods used in the three surveys, see NYGC
1999a, 1999b, 2000.) Of the cities with populations greater than 25,000, 73.7
percent responded to all 3 surveys and 99.1 percent responded to at least 1 of
them. Reports of gang problems in these cities were rather stable from 1996 to
1998. More than 8 in 10 cities (82.6 percent) with populations greater than
25,000 reported the same pattern over the 3-year period (either consistently
reporting gang problems or consistently reporting no gang problems).
Approximately 6 percent reported an emergent gang problem during the 3-year
period, 8 percent reported that their gang problems had desisted, and only 3
percent perceived an intermittent gang problem.

The estimated number of gang homicides (based on responses to NYGC's
annual survey and extrapolated for the Nation as a whole) decreased somewhat
from 1996 to 1998.[3] In 1996, gang members were responsible for an
estimated 2,925 homicides in large cities and suburban counties in the United
States (NYGC 1999a: 32).[4] This estimate dropped to 2,676 in 1997
(NYGC 1999b: 15) and to 2,642 in 1998 (NYGC 2000), but 5 large cities
reported erratic numbers of gang homicides, making the national figures
somewhat suspect. These cities are excluded from the analysis reported here.

With this exception, all cities that indicated gang problems in response to at
least one of the two 1991 surveys and all cities that reported gang problems in
the 1996-98 NYGC surveys were included in the analyses. Cities were chosen
as the unit of analysis for this study because of the authors' interest in examining
city trends, the possibility of some overlap in survey responses between
counties and cities, and the number of gang-related homicides in small towns
and rural areas that produce less reliable statistical comparisons. Where the
same cities were surveyed in both 1991 surveys, responses to the
Maxson-Klein survey were used for consistency because many more cities
were contacted in that survey.

Census data for 1990 were used to determine city size for the cities in the 1991
dataset, and 1994 Census Bureau population estimates were used for the cities
in the 1996-98 dataset. All three NYGC surveys asked respondents to
estimate the number of gangs and gang members.

Findings

City-by-City Homicide Trends: Early to Mid-1990s

Gang homicide trends from the early to mid-1990s were examined by
comparing homicide data from the 408 cities from which homicide reports were
received in 1991 by Maxson and Klein or Curry with figures from the same
cities reported in the 1996 NYGC survey.[5] In 1991, the 408 cities reported
1,748 gang homicides. By 1996, that figure had dropped to 1,492, a total
decrease of 14.6 percent. This substantial net decrease in gang homicides
among mid- to large-sized cities is encouraging. Yet these total counts mask
conflicting patterns among individual cities. A decrease in gang homicides was
observed in 31.9 percent of the 408 cities, whereas 28.7 percent showed an
increase. The remainder (39.4 percent) stayed the same from 1991 to 1996.

Approximately one-third of the cities whose homicide counts remained the
same reported no homicides at either data point. In most cases, the observed
patterns of change (i.e., increasing or decreasing) involved small numbers of
homicides. Most cities reported fewer than 10 gang homicides in any given
year. The magnitude of the change in homicide counts within cities is not
addressed systematically here, but can be assumed to be minor in most, but not
all, cities. The overall net decrease in gang homicide numbers suggests that the
magnitude of the change was greater in cities with a decreasing pattern. This
issue will be revisited later in the chapter.

National City Trend: 1996-98

Gang homicide data from 1996 to 1998 were taken from the annual NYGC
survey data for 1996,[6] 1997,[7] and 1998[8] and analyzed to determine
whether the net decrease in the first half of the decade continued into its second
half. The bulk of reported gang homicides occur in the largest cities, but until
this study, gang homicide figures for smaller cities had not been examined. Yet
the distribution of gang homicides nationwide by city size is a question that can
be resolved empirically. The 1997 NYGC survey (1999b: 15, Table by
Jurisdiction Type) shows that two-thirds to three-fourths of all gang homicides
occur in jurisdictions with populations greater than 25,000. Sparsely populated
areas are even less likely to have persistent gang homicide problems. Of the
347 municipalities with populations of 25,000 or less, only 15 reported 1 or
more homicides for all 3 years in the NYGC survey. Therefore, this analysis
focuses on gang homicide trends in 3 years of NYGC surveys in the 1,216
municipalities with populations greater than 25,000.[9]

The analysis begins with a composite description of gang homicides in these
cities over the 3-year period. In the next section, year-to-year homicide trends
are examined. The unit of analysis for this section is cities that reported gang
problems and gang homicides over the 3-year period from 1996 to 1998.
There was a great deal of stability in the reporting of gang problems among
these jurisdictions. Almost two-thirds of the sample (64 percent) reported gang
problems on all three surveys. Only 231 (19 percent) of the 1,216 cities with
populations greater than 25,000 reported "no gang problem" on all 3 surveys.
The remaining 985 (81 percent) reported a gang problem in at least 1 of the 3
years.

Of these 985 cities, 968 provided gang homicide statistics to one of the NYGC
surveys.[10] Exhibit 1 displays the highest annual count of gang homicides
reported from 1996 to 1998 in these 968 cities and groups them by population
size. In an era in which there is great concern about lethal gang violence, it is
worth noting that most cities that reported having a gang problem (54 percent)
experienced no gang homicides during this period.

Clearly, the number of gang homicides is associated with city population size.
Ninety-nine percent of the cities that reported no gang homicides have a
population of less than 200,000 (see exhibit 1). Only 5 cities with populations
greater than 200,000 reporting gang homicides at least once on the 1996,
1997, or 1998 surveys reported no gang homicides on either of the other two
surveys. Further research might examine the gang situations in these five cities
to determine likely explanations for this unexpected finding. The absence of
gang homicides in these cities may reflect departmental recording policies.

The overwhelming majority (88 percent) of the 436 cities reporting at least 1
gang homicide over the 3-year period reported between 1 and 10 homicides as
the highest number for any year within that period. Ten percent (45 cities)
reported between 11 and 50 homicides, and just 2 percent (8 cities) reported
more than 50 homicides in any 1 year. Cities with larger populations reported
more gang homicides. Fifty percent of cities with gang problems and
populations greater than 200,000 experienced more than 10 gang homicides in
at least 1 of the 3 years. Comparable levels of gang homicides were reported in
just 10 percent of the cities with populations between 100,000 and 200,000
and in 1 percent or less of cities in the two smaller population categories.

The 3 years of NYGC survey data also permit analysis of trends over a brief
period. Among the 985 cities, 237 reported a gang problem in all 3 years and
provided a gang homicide statistic in 1996, 1997, and 1998. There were 1,293
gang homicides in these cities in 1996, 1,260 in 1997, and 1,061 in 1998. For
these cities, the peak homicide year was about evenly distributed: 32.5 percent
each reported their peak number of gang homicides in 1996 and 1997, and 35
percent reported theirs in 1998. Among these 237 cities, 49 percent reported a
decrease in gang homicides, 15 percent reported no change, and 36 percent
reported an increase. A similar pattern was observed in a comparison of the
first and last homicide reports among the 376 cities that provided a homicide
statistic for at least two of the three surveys. Nearly half (48 percent) of these
cities reported a decrease in gang homicides, 14 percent reported no change,
and 38 percent reported an increase.

Yet this analysis of national trends masks some interesting changes within a few
large cities. Five large cities--Denver, Knoxville, Phoenix, Rochester, and
Seattle--showed steady increases in gang homicides over the 3-year period
(exhibit 2); three additional cities--Kansas City, KS (from 6 to 34), Detroit
(from 5 to 75), and New York (from 17 to 59)--showed dramatic increases
from 1997 to 1998; and five cities--Long Beach, CA, Las Vegas, Santa Ana,
CA, St. Paul, and Toledo--showed steady decreases in gang homicides over
the
3-year period (exhibit 3).

Of course, none of these cities is comparable to Los Angeles and Chicago with
respect to the annual number of gang homicides. These two cities stand out
among the 8 cities with the most gang homicides and also show marked
changes in homicides over this period (exhibit 4). In 1998, Los Angeles
reported 173 gang homicides and Chicago reported 180.[11] Both cities
reported substantial decreases in gang homicides from 1996 to 1998, with Los
Angeles reporting a far greater decrease than did Chicago. (Los Angeles
reported a drop from 295 homicides in 1996 to 173 in 1998, a 41-percent
decrease, whereas gang homicides in Chicago dropped from 223 in 1996 to
180 in 1998, a 19-percent decrease.) Together, gang homicides in these two
cities dropped by 165. Thus, Los Angeles and Chicago contributed significantly
to the drop in gang homicides from 1996 to 1998, accounting for 71 percent of
the decrease in homicide numbers over this period.

This finding emphasizes the importance of considering magnitude of change in
homicide levels. Although 49 percent of cities reported decreasing trends in
homicide numbers, the vast majority of the decrease in lethal gang violence
came in the two cities with the largest gang populations. Chicago and Los
Angeles have been the focus of gang research for several decades (see Maxson
and Klein, forthcoming, for a detailed review), and this analysis of gang
homicide trends both underscores the importance of these two cities in any
discussion of gang violence in America and reinforces how the gang patterns in
these two cities differ from the gang situation in other parts of the country.

Correlates of Youth Gang Homicides

The original intent of this study was to determine the characteristics of cities
associated with different national trends in gang homicide levels during the
1990s, using population change characteristics, such as changes in economic
and sociodemographic patterns, to explain increases or decreases in gang
homicides. Such analyses could be quite useful for policy and theoretical
purposes, especially as a basis for an examination of law enforcement
programs, policies, and practices in cities experiencing changes in homicide
levels. The value of such an exercise, however, was diminished by the results of
the analyses reported above. In a nutshell, gang homicide levels changed very
little during this period. Most cities with gang problems report no gang
homicides, and most cities reporting gang homicides report relatively few
incidents. Although many cities (85 percent of the cities included in the 1996-98
analysis) experienced minor fluctuations in small incident numbers, just a few
cities account for most of the decrease in the total number of homicides.
Researchers working with the far more detailed, incident-based datasets in
Chicago and Los Angeles can more ably address the dynamics in those two
cities.

Given the context of these findings, only a quite limited analysis of the correlates
of gang homicide during the 1996-98 NYGC survey period is possible. Los
Angeles and Chicago were excluded from the analysis as outliers because their
large numbers of gangs, gang members, and gang homicides would confound
the results. They are the only cities reporting more than 100 gang homicides in
any year from 1996 to 1998 (see exhibit 1). Five cities with large, erratic
numbers of reported homicides were also excluded as outliers.

In the first analysis, population size (using the 1994 Census Bureau estimates)
and number of gangs (measured as the highest number reported in the 1996 or
1997 surveys) were tested to determine whether they distinguished cities
reporting gang problems but no gang homicides from cities reporting at least
one gang homicide. As expected, both differences were found to be statistically
significant. Cities with gang homicides were more likely than others to have
larger populations (a mean of 128,237 versus 54,406 persons) and a larger
number of gangs (a mean of 30 versus 8 gangs).

Next, the correlation between population, largest number of gangs, largest
number of gang members, and largest number of gang homicides over the
3-year period (1996-98) was examined. For this analysis, cities reporting 1 to
49 gang homicides during the 3-year period were included. The relationships
between population, number of gangs, number of members, and number of
gang homicides are positive and significant. The number of gangs and gang
members are strongly correlated with population size (.668 and .701,
respectively) and with the number of gang homicides (.436 and .667,
respectively). As expected, gang homicides are highly correlated with
population size (.592).

Neither police practitioners nor researchers should be surprised by these
results. Gang homicides are more frequent in larger cities having more gangs
and more gang members. Subsequent analyses could have produced rates of
gang homicides per gang member or population size, but a previous effort in this
direction by one of the authors revealed several analytical and reporting
problems (see Maxson 1998a). The limited examination of correlates of gang
homicide reported in this section is not particularly helpful for guiding
intervention policy; these findings serve merely to reinforce the importance of
efforts to prevent adolescents from joining gangs.

Summary and Discussion of Findings

The examination of gang homicide trends from 1991 to the mid-1990s used
paired cities from three gang surveys (Curry, Ball, and Fox 1994; Maxson,
Woods, and Klein 1995; NYGC 1999a). This is the only explicit comparison
that could be made, given that the three surveys used different sampling frames.
This analysis showed that from the early to mid-1990s, reported gang
homicides in more than 400 cities decreased by 15 percent, slightly less than
the 17-percent decrease in the total number of homicides among 14- to
24-year-olds from 1991 to 199612 (Fox and Zawitz 2000).

The 15-percent drop provides little comfort when two other findings from this
analysis are considered. First, much of this decrease (30 percent) comes from a
drop in gang homicides in just one city--Los Angeles. Second, this analysis also
indicates that the percentage of cities with a decrease in gang homicides (32
percent) is similar to the percentage of cities with an increase (29 percent).
Most cities' (39 percent) reported gang homicide levels remained exactly the
same, and most cities that showed some change did so in small increments.

The examination of gang homicide trends from 1996 to 1998 used the annual
NYGC survey of all cities with populations of more than 25,000. This analysis
showed that nearly half of the more than 400 cities that reported gang
homicides from 1996 to 1998 reported a decrease in gang homicides, whereas
14 to 15 percent reported no change and 36 to 38 percent reported an
increase. Of the more than 200 cities that reported a gang problem and a gang
homicide statistic in all 3 years, 49 percent reported a decrease in gang
homicides, 15 percent reported no change, and 36 percent reported an
increase. Compared with the first part of the decade, a larger proportion of
cities in the late 1990s reported decreases as well as increases, and a smaller
proportion stayed the same. Although these observations suggest that the gang
homicide picture was more dynamic in the latter part of the decade than in the
early to mid-1990s, the magnitude of change within cities for both periods was
minor. There have been substantial reductions in gang homicides in certain large
cities (e.g., Los Angeles and St. Louis) and some stabilization (as in Chicago),
but there are still some cities, such as Denver, Knoxville, Phoenix, Rochester,
and Seattle, where gang homicides have continued to increase.

Four clusters of gang homicide cities are found among municipalities with
populations greater than 25,000. The first cluster contains Chicago and Los
Angeles, which stand alone, far above other cities, in the number of reported
gang homicides on the 1996-98 NYGC surveys. In 1998, they reported a total
of 353 gang homicides. The second cluster, 6 cities, reported more than 50
homicides in at least 1 of the 3 years (Baltimore, Compton, Detroit, Gary,
Indianapolis, and New York). The third cluster contains 45 cities that reported
between 11 and 50 gang homicides in at least 1 year in the 3-year period.
Nearly 400 (383) other cities with populations greater than 25,000 reported
from 1 to 10 gang homicides on at least 1 of the 3 surveys.

Finally, the analysis of the correlates of gang homicides showed that cities with
gang problems that report gang homicides are more likely than others to have
large populations, large numbers of gangs, and large numbers of gang members.
While these findings are not surprising, the extent of the concentration of gang
homicides in very large cities is revealing. Of the 532 cities with populations
greater than 25,000 that reported no gang homicides in the 1996-98 NYGC
surveys, only 1 percent had populations of at least 200,000. Despite the
proliferation of gangs in cities of all sizes in the 1990s (NYGC 1999a), gang
homicides remain far more of a problem in the Nation's largest cities.

Policy Recommendations

The finding that 436 of our Nation's cities experienced gang homicide in at least
1 of the 3 years (1996 to 1998) indicates that gang violence is a prevalent and
persistent problem in the United States. The encouraging finding of a net
decrease in the total number of gang homicides is offset by the more sobering
finding that the level of gang homicide did not change substantially in the vast
majority of cities. Despite annual declines in national youth homicide figures
during the 1990s, just a few cities produced most of the reduction in gang
homicides reflected in the trend analysis. Four policy recommendations are
suggested by this study's findings.

First, it would be useful if large cities with serious gang problems continue to be
the subject of intense scrutiny regarding the causes of and the solutions to
violent gang activity. Examples of such efforts can be found in other chapters of
this volume. Analysis of the detailed, incident-based homicide datasets already
produced in some large cities should continue to be supported. Information on
gang prevention, intervention, and suppression programs implemented in these
cities could be linked to these datasets to provide a better understanding of
changes in gang violence levels.

Second, the hundreds of cities that report gang homicides should continue to be
monitored via national surveys like the NYGS. If substantial changes in incident
numbers become evident, analysis of these changes should include population
measures derived from the census and other national databases (see Jackson
1991 for examples of demographic and economic transition variables used to
predict the presence of gangs). The issue of magnitude of change should be
considered carefully; big changes in small gang homicide numbers may be
related to factors different from those that explain small changes in big numbers.

Third, Federal policymakers might lead an effort to improve data sources on
gang homicides and other forms of gang activity. Any national assessment is
severely constrained by the data limitations described earlier in this chapter.
Research on the reliability and validity of gang reporting practices is also
essential. Variations in local practices for defining and recording gangs and gang
membership undoubtedly have an impact on the designation and reporting of
gang crime. There is not sufficient consensus among law enforcement
practitioners regarding optimal definitions to get them to agree to uniform
policies, but efforts to build such a consensus should be encouraged. Building
knowledge about the national patterns of gang homicide will be very difficult
until reporting practices are standardized.

As the effort to develop uniform definitions takes place, it would be useful if the
structures for gang homicide reporting could be expanded to incorporate
incident-level descriptions. The NIBRS system may eventually fill this need, but
pending its adoption in populous States, immediate changes could be made in
the SHR reporting system. Recording officially established gang membership for
any participant in a homicide and gang-related motives or circumstances in the
SHR system would permit the kind of detailed analysis of the characteristics of
different forms of gang homicide that is currently available in only a handful of
cities. Examination of patterns of weapon use, incident locations, relationships
between victims and offenders, and participant demographic characteristics
such as age, race or ethnicity, and gender would considerably improve our
understanding of the dynamics of gang homicide.

Finally, gang homicide trends and patterns of demographic and situational
characteristics should be compared with trends for other forms of homicide and
other forms of violence. In some cities, such as Chicago and Los Angeles, gang
homicides represent a substantial proportion of all homicides. Changes in gang
homicides can be fully understood only when placed in the context of the
patterns of other forms of homicide. The same observation applies to less lethal
forms of violence. Analysts and policymakers should be in the position to know
whether gang assaults increase during periods of homicide decreases.
Furthermore, since assaults are far more common than homicides, changes in
the levels and characteristics of gang assaults might be the more useful
barometer of the effectiveness of gang intervention programs.

These recommendations for further research and policies require improved data
sources and data collection strategies. This study reports the results of an
exploratory assessment of a national dataset built from previous studies. In a
Nation with more than 400 cities reporting gang homicides, it is critical to
establish a national database for the examination of trends in gang violence. This
study offers a starting point for analyzing patterns of gang homicide and strongly
encourages the further development of data sources to further understand the
dynamics of gang violence and responses to gangs.

Notes

1. Throughout this chapter, the term "gang" is used to refer to "youth gangs,"
"street gangs," or "criminal street gangs." Each term may have different
connotations for different audiences, but the authors make no distinction among
them here.

2. It is possible that a gang homicide may have occurred in a city that reported
it had no gang problem. The NYGS format terminated the survey following a
negative response to the gang problem item, so homicides in these cities would
not have been captured by the survey.

3. The number of youth gang homicides reported in the 1995 NYGC survey
cannot be compared with the reported numbers for 1996-1998, because the
sample of jurisdictions surveyed in 1995 was not nationally representative.

4. Too few respondents in less populated jurisdictions reported youth gang
homicides to extrapolate numbers nationwide.

5. No 1996 data were available for 8.3 percent of the cities. In these cities,
data from the 1995 NYGS were used for the second mid-1990s comparison
point.

6. In 1996, 785 of 1,037 cities with populations greater than 25,000 (that
could provide an answer) reported a gang problem (76 percent) and 693 (88
percent) of these provided a statistic on gang homicides (of which 62 percent
reported none). In 1996, 267 cities reported at least 1 gang homicide. These
267 cities reported 1,821 homicides in 1996.

7. In 1997, 786 of 1,071 cities with populations greater than 25,000 (that
could provide an answer) reported a gang problem (73 percent) and 728 (92
percent) of these provided a statistic on gang homicides (of which 64 percent
reported none). In 1997, 265 cities reported at least 1 gang homicide. These
265 cities reported 1,734 member-based and 890 motive-based homicides in
1997.

8. In 1998, 736 of 1,049 cities with populations greater than 25,000 (that
could provide an answer) reported a gang problem (70 percent), and 691 (94
percent) of these provided a statistic on gang homicides (of which 66 percent
reported none). In 1998, 234 cities reported at least 1 gang homicide. These
234 cities reported 1,783 member-based and 897 motive-based homicides in
1997.

9. Only one suburban county, Los Angeles County, reported more than 100
gang-involved homicides on any of the three NYGC surveys: 124 in 1996, 124
in 1997, and 70 in 1998.

10. Seventeen cities (1.7 percent) reported that the number of gang homicides
in the jurisdiction was not known for all years when a gang problem was
reported. The survey results made it possible to distinguish between
member-based and motive-based gang-involved homicides, and almost all
jurisdictions used one or the other of the definitions. In this analysis, the larger
of the two numbers (always the member definition) was chosen for the few
jurisdictions where both types of definitions are used.

11. Los Angeles reports gang-related homicides and Chicago reports only
gang-motivated homicides.

12. The comparison with national homicide data is not a precise one, because
youth gang membership is not restricted to 14- to 24-year-olds. Adults older
than age 24 could be responsible for a significant proportion of gang homicides.

References

Bailey, G., and N.P. Unnithan. 1994. "Gang Homicides in California: A
Discriminant Analysis." Journal of Criminal Justice 22(3): 267-275.

Ball, R.A., and G.D. Curry. 1995. "The Logic of Definition in Criminology:
Purposes and Methods for Defining 'Gangs.'" Criminology 33(2): 225-245.

Block, C.R., A. Christakos, A. Jacob, and R. Przybylski. 1996. Street Gangs
and Crime: Patterns and Trends in Chicago. Chicago: Illinois Criminal Justice
Information Authority.

Brewer, V.E., K.R. Damphousse, and C.D. Adkinson. 1998. "The Role of
Juveniles in Urban Homicide: The Case of Houston, 1990-1994." Homicide
Studies 2(3): 321-339.

Bureau of Justice Assistance. 1997. Urban Street Gang Enforcement.
Washington, DC: U.S. Department of Justice, Bureau of Justice Assistance,
NCJ 161845.

Cohen, J., D. Cork, J. Engberg, and G. Tita. 1998. "The Role of Drug Markets
and Gangs in Local Homicide Rates." Homicide Studies 2(3): 241-262.

Cohen, J. and G. Tita. (1998). "The Drugs-Gangs-Guns Nexus of Homicide in
Pittsburgh, 1987-1995." Unpublished manuscript. Carnegie Mellon University,
School of Public Policy and Management.

Curry, G.D., R.A. Ball, and S.H. Decker. 1995. "An Update on Gang Crime
and Law Enforcement Recordkeeping." Unpublished manuscript. St. Louis,
University of Missouri, Department of Criminology and Criminal Justice.

-------. 1996a. Estimating the National Scope of Gang Crime from Law
Enforcement Data. Research in Brief. Washington, DC: U.S. Department of
Justice, National Institute of Justice, NCJ 161477.

-------. 1996b. "Estimating the National Scope of Gang Crime from Law
Enforcement Data." In Gangs in America, (2nd ed.) Ed. C.R. Huff. Thousand
Oaks, CA: Sage: 21-36.

Curry, G.D., R.A. Ball, and R.J. Fox. 1994. Gang Crime and Law
Enforcement Recordkeeping. Research in Brief. Washington, DC: U.S.
Department of Justice, National Institute of Justice, NCJ 148345.

Curry, G.D., R.J. Fox, R.A. Ball, and D. Stone. 1992. National Assessment of
Law Enforcement Anti-Gang Information Resources. Washington, DC: U.S.
Department of Justice, National Institute of Justice, NCJ 147399.

Curry, G.D., C.L. Maxson, and J.C. Howell. 2001. Youth Gang Homicides in
the 1990's. Washington, DC: U.S. Department of Justice, Office of Juvenile
Justice and Delinquency Prevention, FS 200103.

Curry, G.D., and I.A. Spergel. 1988. "Gang Homicide, Delinquency, and
Community." Criminology 26(3): 381-405.

Egley, A., Jr. 2000. Highlights of the 1999 National Youth Gang Survey.
Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and
Delinquency Prevention, FS 200020.

Federal Bureau of Investigation. 1995. Crime in the United States, 1994:
Uniform Crime Reports. Washington, DC: U.S. Department of Justice, Federal
Bureau of Investigation.

-------. 1997. Uniform Crime Reporting. State Program Bulletin, June 30.
Washington, DC: U.S. Department of Justice, Federal Bureau of Investigation.

Fox, J.A., and Zawitz, M.W. 2000. Homicide Trends in the United States:
1988 Update. Washington, DC: U.S. Department of Justice, Bureau of Justice
Statistics, NCJ 179767.

Howell, J.C. 1999. "Youth Gang Homicides: A Literature Review." Crime and
Delinquency 45(2): 208-241.

Hutson, H.R., D. Anglin, and M.J. Pratts. 1994. "Adolescents and Children
Injured or Killed in Drive-By Shootings in Los Angeles." New England Journal
of Medicine 330: 324-327.

Hutson, H.R., D. Anglin, D.N. Kyriacou, J. Hart, and K. Spears. 1995. "The
Epidemic of Gang-related Homicides in Los Angeles County from 1979
through 1994." Journal of the American Medical Association 274 (13):
1031-1036.

Hutson, H.R., D. Anglin, C. Reek-Asa, D.N. Kyriacou, and J.F. Kraus. 1999.
"The Relationship Between Socioeconomic Factors and Gang Violence in the
City of Los Angeles. Journal of Trauma--Injury, Infection, and Critical Care
46(2): 334-339.

Jackson, P.I. 1991. "Crime, Youth Gangs, and Urban Transition: The Social
Dislocations of Postindustrial Economic Development."Justice Quarterly 8(3):
379-397.

Jackson, P. 1993. "Moral Panic and the Response to Gangs in California." In
Gangs: The Origins and Impact of Contemporary Youth Gangs in the United
States, ed. S. Cummings and D.J. Monti. Albany: State University of New
York Press: 257-275.

Kennedy, D.M., and A.A. Braga. 1998. "Homicide in Minneapolis: Research
for Problem Solving." Homicide Studies 2(3): 263-290.

Kennedy, D.M., A.A. Braga, and A. Piehl. 1997. "The (Un)known Universe:
Mapping Gangs and Gang Violence in Boston." In Crime Mapping and Crime
Prevention, ed. K.D. Weisburd and J.T. McEwen. New York: Criminal Justice
Press: 219-262.

Kennedy, D.M., A.M. Piehl, and A.A. Braga. 1996. "Youth Violence in
Boston: Gun Markets, Serious Youth Offenders, and a Use-Reduction
Strategy. Law and Contemporary Problems 59(1): 147-196.

Klein, M.W. 1995. The American Street Gang: Its Nature, Prevalence, and
Control. New York: Oxford University Press.

Klein, M.W., and C.L. Maxson. 1989. "Street Gang Violence." In Violent
Crime, Violent Criminals, ed. M.E. Wolfgang and N.A. Weiner. Newbury
Park, CA: Sage: 198-234.

Maxson, C.L. 1992. "Collecting Data from Investigation Files: Descriptions of
Three Los Angeles Gang Homicide Projects." In Questions and Answers
About Lethal and Non-Lethal Violence, ed. C.R. Block and R. Block.
Washington, DC: U.S. Department of Justice, National Institute of Justice, NCJ
142058: 87-95.

-------. "Gang Homicide." In Studying and Preventing Homicide, ed. M.D.
Smith and M.A. Zahn. Thousand Oaks, CA: Sage: 197-219.

-------. 1998b. Gang Members on the Move. Juvenile Justice Bulletin: Youth
Gang Series. Washington, DC: U.S. Department of Justice, Office of Juvenile
Justice and Delinquency Prevention. NCJ 171153.

Maxson, C.L., and M.W. Klein. 1990. "Street Gang Violence: Twice as Great,
or Half as Great?" In Gangs in America, ed. C.R. Huff. Newbury Park, CA:
Sage: 71-100.

-------. 1996. "Defining Gang Homicide: An Updated Look at Member and
Motive Approaches." In Gangs in America (2nd ed.), ed. C.R. Huff. Thousand
Oaks, CA: Sage: 3-20.

-------. (Forthcoming). "'Play Groups' No Longer: Urban Street Gangs in the
Los Angeles Region. In From Chicago to L.A.: Making Sense of Urban
Theory, ed. M.J. Dear. Newbury Park, CA: Sage Publications.

Maxson, C.L., K. Woods, and M.W. Klein. 1995. "Street Gang Migration in
the United States." Unpublished manuscript. Center for the Study of Crime and
Social Control. Social Science Research Institute, University of Southern
California.

-------. 1996. "Street Gang Migration: How Big a Threat?" National Institute of
Justice Journal 230, February 26-31.

Miller, W.B. 1975. Violence by Youth Gangs and Youth Groups in Major
American Cities. Washington, DC: U.S. Department of Justice, National
Institute of Justice. NCJ 137446.

-------. 1992. Crime by Youth Gangs and Groups in the United States.
Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and
Delinquency Prevention. NCJ 156221.

National Youth Gang Center. 1997. 1995 National Youth Gang Survey:
Program Summary. Washington, DC: U.S. Department of Justice, Office of
Juvenile Justice and Delinquency Prevention. NCJ 164728.

-------. 1999a. 1996 National Youth Gang Survey: Summary. Washington,
DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency
Prevention. NCJ 173964.

-------. 1999b. 1997 National Youth Gang Survey. Washington, DC: U.S.
Department of Justice, Office of Juvenile Justice and Delinquency Prevention.
NCJ 178891.

-------. 2000. 1998 National Youth Gang Survey. Washington, DC: U.S.
Department of Justice, Office of Juvenile Justice and Delinquency Prevention.
NCJ 183109.

Riedel, M. Forthcoming. "Homicide in Los Angeles County: A Study of Latino
Victimization." In Interpersonal Violence: The Ethnicity, Race, and Class
Nexus, ed. D.F. Hawkins. New York: Cambridge University Press.

Snyder, H.N., M. Sickmund, and E. Poe-Yamagata. 1996. Juvenile Offenders
and Victims: 1996 Update on Violence. Washington, DC: U.S. Department of 
Justice, Office of Juvenile Justice and Delinquency Prevention. NCJ 159107.

Spergel, I.A. 1995. The Youth Gang Problem. New York: Oxford University
Press.

Spergel, I.A., and G.D. Curry. 1990. "Strategies and Perceived Agency
Effectiveness in Dealing with the Youth Gang Problem." In Gangs in America,
ed. C.R. Huff. Newbury Park, CA: Sage: 288-309.

-------. 1993. The National Youth Gang Survey: A Research and
Development Process. In The Gang Intervention Handbook, ed. A. Goldstein
and C.R. Huff. Champaign, IL: Research Press: 359-400.

Thornberry, T.P. 1998. "Membership in Youth Gangs and Involvement in
Serious and Violent Offending." In Serious and Violent Juvenile Offenders: Risk
Factors and Successful Interventions, ed. R. Loeber and D.P. Farrington.
Thousand Oaks, CA: Sage: 147-166.

Weisel, D.L., and E. Painter. 1997. The Police Response to Gangs: Case
Studies of Five Cities. Washington, DC: Police Executive Research Forum.

Zahn, M.A., and K.M. Jamieson. 1997. "Changing Patterns of Homicide and
Social Policy." Homicide Studies 1(2): 190-196.

Zatz, M.S. 1987. "Chicano Youth Gangs and Crime: The Creation of a Moral
Panic." Contemporary Crises, 11(2): 129-158.

---------------------------

Chapter 5

National Evaluation of the Gang Resistance Education and Training
(G.R.E.A.T.) Program

Finn-Aage Esbensen, Ph.D.
Department of Criminology and Criminal Justice
University of Missouri-St. Louis

Adrienne Freng, Ph.D.
Department of Criminal Justice
University of Wyoming

Terrance J. Taylor
Dana Peterson
Department of Criminal Justice
University of Nebraska-Omaha

D. Wayne Osgood, Ph.D.
Department of Sociology
Pennsylvania State University

Acknowledgments

Funding for this research was provided by the Bureau of Alcohol, Tobacco and
Firearms, U.S. Department of the Treasury, with additional support (award
94-IJ-CX-0058) from the National Institute of Justice, Office of Justice
Programs, U.S. Department of Justice. The authors wish to thank their
colleagues who assisted with various aspects of this evaluation project: Drs.
Tom Winfree, Chris Sellers, Libby Deschenes, Fran Bernat, and Ron Taylor,
as well as numerous research assistants who assisted with data collection.

---------------------------

Abstract

Gang Resistance Education and Training (G.R.E.A.T.), a school-based
program for middle school students, is designed to help them avoid peer
pressure to join gangs through cultivation of such life skills as social
competence, problem solving, and responsibility. Funded by the Federal
Bureau of Alcohol, Tobacco and Firearms, the program is a prevention strategy
consisting of a series of classroom lessons taught by specially trained law
enforcement personnel. A national evaluation of the program was conducted to
determine whether it was succeeding. The cross-sectional evaluation was based
on an analysis of the program in 11 cities. The longitudinal study, based on an
analysis of G.R.E.A.T. in six cities, compared the experiences of participants
and nonparticipants at yearly intervals up to 4 years after program completion.
The results of the initial cross-sectional evaluation were favorable. The 2-year
results of the longitudinal evaluation were that there were no statistically
significant differences in social attitudes or behavior between the two groups.
Then, after 4 years, there appeared to be a lagged effect, with G.R.E.A.T.
students exhibiting more positive social attitudes than non-G.R.E.A.T. students.
The contradiction between the cross-sectional and 2-year longitudinal findings
prompted a review and revision of the G.R.E.A.T. curriculum. It suggested
greater involvement of the regular classroom teacher and more focus on active
learning than lecturing. The recommendations were accepted and the new 
curriculum was tested in the spring of 2001.

---------------------------

This chapter describes the Gang Resistance Education and Training
(G.R.E.A.T.) program, a school-based gang prevention program targeting
middle school students. Also described are the research design and results of a
longitudinal national evaluation of this program, funded in 1994 by the National
Institute of Justice (NIJ) in partnership with the Bureau of Alcohol, Tobacco
and Firearms (ATF). For that evaluation, a quasi-experimental research design
was implemented in six cities to assess the program's impact. The results of that
evaluation are reported here and reflect short-term (posttests administered
within 2 weeks of program completion) and long-term effects (2 and 4 years
after program completion). This chapter also details how these results have
been used to revise the G.R.E.A.T. curriculum.

Analyses based on the 2-year followup data failed to detect statistically
significant differences between the G.R.E.A.T. and non-G.R.E.A.T. students.
The 4-year followup analyses, however, resulted in significant differences
between the two groups: G.R.E.A.T. students reported more positive social
attitudes 4 years after program completion than did the non-G.R.E.A.T.
students.

Partly in response to the null findings from the 2-year followup data, the
G.R.E.A.T. administration sought help in enhancing the program. A working
group was formed, which included G.R.E.A.T. officers and administrators, staff
members from the national evaluation, and experts in gangs and school-based
gang prevention programs. The group reviewed the G.R.E.A.T. program and
recommended ways to improve it. Officers were retrained in the new
curriculum, and pilot programs were taught in middle schools during the spring
of 2001.

What Is G.R.E.A.T.?

In 1991, Phoenix law enforcement formed a partnership with local educators
and community leaders to develop an innovative, comprehensive anti-gang
program. The result was G.R.E.A.T., Gang Resistance Education and Training,
supported by funding from the U.S. Bureau of Alcohol, Tobacco and Firearms.

The preceding statement appears on most publications associated with the
G.R.E.A.T. gang prevention program, which consists of eight lessons provided
in nine 1-hour sessions taught by specially trained law enforcement personnel.
The intent of the lessons is to provide students with "the necessary skills and
information to say no to gangs and become responsible members of society"
(G.R.E.A.T. Brochure n.d., 3). According to G.R.E.A.T. publications, the
program's mission is "To provide a wide range of structured activities and
classroom instruction for school-aged children that result in a sense of
competency, usefulness and personal empowerment needed to avoid
involvement in youth violence" (G.R.E.A.T. News 1994, 1). To see how these
goals were accomplished, the authors reviewed the G.R.E.A.T. curriculum and
program delivery, which are described below.

G.R.E.A.T. is a classroom-based, officer-instructed program generally taught in
the seventh grade. The point of providing the program to students at this age
and grade level is to communicate the message that "gangs have nothing to
offer" before gang recruitment begins in earnest. As initially created, the lessons
included the following:

o Introduction. Students become acquainted with the program and the officer.

o Crimes/Victims and Your Rights. Officers demonstrate the impact crime can
have on victims and neighborhoods.

o Cultural Sensitivity/Prejudice. Students learn cultural differences and their
impact on the community.

o Conflict Resolution. Officers create an atmosphere of understanding to enable
all parties to better address problems and work on solutions together (two
sessions).

o Meeting Basic Needs. Students are taught how to become better equipped to
meet their basic needs.

o Drugs/Neighborhoods. Officers teach students the effects drugs can have on
a neighborhood.

o Responsibility. Students learn the diverse responsibilities of individuals in a
community.

o Goal Setting. Officers teach students how to set long-range goals.

A review of these lessons and the detailed lesson plans and workbook
exercises that accompany them shows that the content of the program is nothing
startling or entirely new for students in this age group. By the seventh grade,
they should have been exposed to most, if not all, of the ideas contained in the
lessons. A key element of the G.R.E.A.T. program is that the "teachers" are
police officers, sheriff's deputies, town marshals, military police officers, and (in
a few cases) agents from the Bureau of Alcohol, Tobacco and Firearms. In
addition, the eight lessons, spaced over nine classroom sessions, synthesize the
content of many other classes that students have been exposed to during their
school years. Thus, the officers in this program seek to link what seventh grade
students may view as unrelated pieces of information.

In their report on the early history of the G.R.E.A.T. program, Winfree,
Lynskey, and Maupin (1999) relied on interviews with "key players" and
documents provided by the Phoenix Police Department and the Bureau of
Alcohol, Tobacco and Firearms to recreate the program's historical
development. It

is evident from this report that the current G.R.E.A.T. curriculum was the
product of an intense effort by several Phoenix Police Department officers to
produce a product in a short time. Because of the officers' considerable
experience as Drug Abuse Resistance Education (D.A.R.E. (registered
trademark))[1] officers and mentors, the G.R.E.A.T. program, as developed,
resembled the D.A.R.E. (registered trademark) program. Little attention, it
appears, was given to pedagogical and developmental issues, let alone to the
prevention literature. Despite these shortcomings, the officers produced a
curriculum and a training model that were generally well received by educators,
parents, and other law enforcement representatives.

As the curriculum indicates, the G.R.E.A.T. program is intended to provide life
skills that empower adolescents to resist peer pressure to join gangs. The
strategy is a cognitive approach (similar to the D.A.R.E. (registered trademark)
and to the Law Related Education program, which is based on the notion that
juvenile delinquency is caused by a lack of awareness of the laws and seeks to
teach students about law and the legal process) designed to produce attitudinal
and behavioral change through instruction, discussion, and role playing.

Another notable feature of the G.R.E.A.T. program is its target population.
Unlike suppression and intervention programs, which are directed at youths
who already are gang members, G.R.E.A.T. is intended for all students. This is
an example of the classic, broad-based prevention strategy found in medical
immunization programs: broad intervention with a simple and relatively
unintrusive program well before any problem is detectable and without
attempting to predict who is most likely to be affected by the problem.

To date, two published evaluations have reported that the program has had
small, but positive, effects on students' attitudes and behavior (Esbensen and
Osgood 1999; Palumbo and Ferguson 1995). Esbensen and Osgood (1999)
reported the findings from the cross-sectional component of the National
Evaluation of the G.R.E.A.T. program. Based on surveys completed by 5,935
eighth-grade students (1 year after program completion) in 11 cities across the
continental United States, the study found that G.R.E.A.T. graduates reported
committing fewer delinquent acts and expressed more positive social attitudes,
including more favorable attitudes toward the police, higher levels of
self-esteem and attachment to parents, and greater commitment to school.
Using a multisite, pretest/posttest research design, Palumbo and Ferguson
(1995) found that G.R.E.A.T. graduates had a "slightly increased ability" to
resist pressures to join gangs. The authors acknowledged, however, that "the
lack of a control group prevents assessments of the internal validity. Therefore,
it cannot be concluded that the results . . . were due to G.R.E.A.T. as opposed
to other factors" (Palumbo and Ferguson 1995: 600).

The Longitudinal National Evaluation

In 1994, the National Institute of Justice sponsored a national evaluation of the
G.R.E.A.T. program. The research design involved a process evaluation that
examined program delivery (Sellers, Taylor, and Esbensen 1998), a preliminary
impact evaluation (Esbensen and Osgood 1999), and a longitudinal,
quasi-experimental outcome study.

Although development of the G.R.E.A.T. curriculum was not theory driven, the
design of the national evaluation was. The theories determined to be most
relevant in evaluating the program were social learning theory (Akers 1985)
and self-control theory (Gottfredson and Hirschi 1990). Identification of
relevant theoretical constructs is critical in evaluating prevention programs
because prevention, by definition, takes place well before the outcome that the
program is designed to prevent (i.e., gang membership) is likely to occur. In
brief, a positive effect on attitudes should produce a subsequent reduction in
delinquent behavior. This evaluation therefore emphasized theoretical constructs
that relate logically to the G.R.E.A.T. curriculum and are theoretically and
empirically linked to gang membership and delinquency. A more detailed
account of the theoretical foundations of the national evaluation is given in an
earlier report (Winfree, Esbensen, and Osgood 1996).

The student questionnaires used in the national evaluation consisted of attitudinal
and behavioral questions. Measures of perceptions of the appropriateness of
certain behaviors and measures of peer group conduct were of primary
importance. Given the significant role of peer pressure in gangs and
delinquency, several scales were used to measure the commitment the youths
felt to their peer group. The questionnaire also measured student involvement in
school and community activities.

One of the more important objectives of the G.R.E.A.T. program is to reduce
adolescent involvement in criminal behavior and gangs; this involvement was
measured through self-reports of illegal activity. This technique has been used
widely since the early 1970s and provides a good measure of actual behavior,
rather than a reactive measure of police response to behavior (e.g., Hindelang,
Hirschi, and Weis 1981; Huizinga 1991; Huizinga and Elliott 1986). Questions
measuring self-reported drug use and victimization were also included in this
section of the questionnaire.

Another important focus of the questionnaire was measurement of gang
membership and gang activity. Two filter questions were used to determine
gang membership: "Have you ever been a gang member?" and "Are you now in
a gang?" Students answering yes to either of these questions were then asked
about gang structure, gang activity, and attitudes about the gang. Also included
were queries to determine the following: identification of good and bad aspects
of gang membership, approval of gang membership, measures of gang
attachment, and reasons for joining the gang.

Longitudinal Research Design

The two previously published evaluations of the G.R.E.A.T. program discussed
above contained methodological limitations. The Palumbo and Ferguson (1995)
study did not include a comparison group; the Esbensen and Osgood (1999)
evaluation used a cross-sectional design, an approach that lacks a pretest
measure and requires the ex post facto creation of a comparison group.
Although statistical procedures can strengthen the validity of this method (e.g.,
Heinsman and Shadish 1996), it is generally considered a weak design (e.g.,
Sherman et al. 1997). The longitudinal research strategy implemented in the
current evaluation, with a strong quasi-experimental[2] research design and
assignment of classrooms to treatment and control groups, served two very
important functions. First, the evaluation team believed, this assignment process
should create groups of G.R.E.A.T. and non-G.R.E.A.T. students who are at
equal risk for future delinquency and gang involvement. Second, the longitudinal
research design would greatly increase statistical power for detecting program
effects by controlling for previous individual differences and examining change
over time.

Site Selection

Six cities were selected for the longitudinal phase of the National Evaluation.
The first criterion for selection was the existence of a viable G.R.E.A.T.
program in 1995; a second criterion was geographical location. A third criterion
was the cooperation of the school districts and the police departments in each
site. These criteria led to selection of an east coast city (Philadelphia), a west
coast location (Portland, Oregon), the site of the program's inception (Phoenix),
a Midwest city (Omaha), a small city with little or no gang presence (Lincoln,
Nebraska), and a small border town with a chronic gang problem (Las Cruces,
New Mexico).

Research Design

The longitudinal study included relatively equal-sized groups of treatment
(G.R.E.A.T.) and control (non-G.R.E.A.T.) students in the seventh grade at
five of the sites and sixth grade students in the sixth.[3] Because G.R.E.A.T. is
a classroom-based program, assignment was implemented for classrooms
rather than for individual students; that is, classrooms in each school were
assigned either to receive G.R.E.A.T. or to serve as a control classroom and
not receive G.R.E.A.T. The sample included a total of 22 schools, 153
classrooms, and more than 3,000 students. During the fall of 1995, students in
all the selected classrooms completed pretests before G.R.E.A.T. instruction
began in the treatment classrooms. Within 2 weeks of completing the 9-week
program, all students were surveyed again. The pretests made it possible to
compare the two groups on all measures before program intervention. These
analyses revealed no preexisting systematic differences between students in the
G.R.E.A.T. classrooms and those assigned to the control group. The posttests
allowed for examination of immediate, short-term programmatic effects. To
assess whether the G.R.E.A.T. program had any sustained effect, surveys were
administered during each of the subsequent 4 years (in the fall of 1996, 1997,
1998, and 1999) to all students for whom active parental consent had been
obtained (see Active Consent Procedures). Although some students were lost
because of the active consent process or mobility in subsequent years,
questionnaire completion rates exceeded industry standards (see Questionnaire
Completion Rates).

Results of the Longitudinal Analysis of Program Impact

In this section, results are reported from the analysis of outcome effects 2 and 4
years after program completion. Due in part to congressional inquiries about
program effectiveness, the G.R.E.A.T. management team had requested the
2-year outcome results. These interim results provided outcome measures
based on student responses to the pre- and posttest surveys collected in 1995
and the 1- and 2-year followup surveys collected in 1996 and 1997,
respectively. More complete outcome analyses, incorporating the third and
fourth years of followup surveys, were reported on completion of the last
scheduled surveys.

To assess program effectiveness, four levels of analysis were considered: the
individual, change across time, classroom, and school. Examining individual
change over time (i.e., identifying program effects on each individual who
completed the G.R.E.A.T. course compared with those who did not) was no
easy analytical task. Because students received the program within a classroom
context, researchers had to control for classroom-level information. Likewise,
classrooms were part of a larger school environment. Recently developed
statistical programs (Bryk and Raudenbush's Hierarchical Linear Models
[1992] and Goldstein's Multilevel Statistical Models [1995]) allow researchers
to examine individual change across time, while controlling for group change
across time and other "nested" conditions (individuals within classrooms and
within schools), allowing an examination of the unique effect of G.R.E.A.T. on
individual students. (For a detailed discussion of the design, analysis strategy,
and longitudinal results, see Esbensen et al. 2001.)

Two-Year Followup Results

The most direct indication of program impact was a comparison of preprogram
with postprogram survey data from the 1- and 2-year followup questionnaires.
This comparison identified the degree to which change for the treatment group
differed from that for the control group. Although the large sample afforded
ample statistical power, only one of the 31 pre-post change comparisons was
statistically significant at the standard .05 probability level (victimization, p =
.017). With this large number of significance tests, 1.55 nominally significant
findings could be expected by chance alone. Furthermore, most of the
differences in change were quite small, and almost half of them indicated that
the program had had an unfavorable impact. Based on these comparisons, it
did not appear that the G.R.E.A.T. program had reached its goals.

Program Impact Under "Optimal" Circumstances

Why did this analysis fail to find that the G.R.E.A.T. program had a significant
impact on student attitudes and behavior? Before accepting these results as
indicating that the program had no benefits, other alternatives needed to be
explored. For example, the program may have been implemented better in
some sites than in others. If so, positive results in more optimal circumstances
could have been masked by less favorable outcomes in others. To determine
whether this was the case, the evaluation team repeated the above analysis,
using only the three sites where process analysis indicated that program staff
had been most successful in delivering the program as designed. Analysis was
further restricted to classrooms in which at least 55 percent of students
participated in the study. This analysis included 1,074 students from 55
classrooms in 11 schools.

The results for this more selective analysis closely matched those for the entire
sample. Only 2 of 76 significance tests for program impact reached the .05
level of significance (fewer than would be expected by chance), and none
reached the .01 level. Again, the results were as likely to favor the control
group as the treatment group. In sum, the attempt to identify the best examples
of the G.R.E.A.T. program in the sample yielded no evidence of program
benefits. The lack of program effects in the overall analysis did not appear to be
a matter of weaker programs masking the impact of stronger ones.

Variation in Program Effectiveness by Prior Risk

The evaluation team also explored the possibility that the impact of the
G.R.E.A.T. program might depend on students' level of risk for delinquency
and gang membership. Earlier cross-sectional analyses of program impact had
examined the consistency of program effects across demographic groupings
(Esbensen and Osgood 1999) and found that G.R.E.A.T. was more effective
with groups at higher risk for delinquency, specifically males and minorities. In
the present longitudinal analysis, the data from the pretest measure allowed the
team to measure risk of future delinquency and gang membership directly,
rather than inferring from demographic proxies. As with the preceding analyses
of the 2-year followup data, however, no evidence was found that participating
in the G.R.E.A.T. program produced favorable outcomes for students at high
or low risk for antisocial outcomes.

Four-Year Followup Results

On completion of the fourth year of data collection, outcome analyses were
conducted incorporating all six waves of student responses: pretest, posttest,
year 1, year 2, year 3, and year 4 surveys. The same analytical strategies
reported above were used for the 2-year followup analyses. Contrary to the
interim results, however, the evaluation team found a 4-year lagged effect of the
G.R.E.A.T. program. Four years after program completion, students who had
been assigned to the G.R.E.A.T. program reported more positive social
attitudes and behaviors on 25 of the 29 outcome measures; five of these
differences were statistically significant at the .05 level. Students in the
G.R.E.A.T. program reported lower levels of risk-seeking and victimization,
more positive attitudes toward the police, more negative attitudes toward
gangs, and more friends involved in positive social activities than students in the
control group.

The results based on the full longitudinal dataset support a modest program
effect (effect sizes of approximately 0.10). Several questions, however, must be
addressed: Why was there no measurable program effect 2 years after program
delivery? Why did the cross-sectional study, which surveyed students 1 year
after program delivery, produce favorable programmatic effects? What factors
can explain a 4-year lagged effect? From a policy perspective, should interim
results, such as the 2-year followup, be reported? As discussed in the next
section, the null findings from the 2-year analyses contributed to a rigorous
review of the G.R.E.A.T. curriculum.

The G.R.E.A.T. Review

Because of the contradictory findings from the cross-sectional and the 2-year
longitudinal study results, the National Policy Board (NPB) of the G.R.E.A.T.
program wanted a board of experts to review the G.R.E.A.T. curriculum. In
response, NIJ funded a review of the G.R.E.A.T. program, and in 1999 the
G.R.E.A.T. Review Workgroup was convened to conduct a critical assessment
of G.R.E.A.T. The NPB deserves recognition for its willingness to seek
recommendations from researchers. The G.R.E.A.T. review process was
extraordinary in that G.R.E.A.T. program administrators took the evaluation's
findings seriously and sought to improve the content and implementation of their
program based on these findings. Their willingness to subject the program to
critical review, which could result in recommendations for substantial program
revision, is uncommon and demonstrates their commitment to the prevention of
gangs and violence.

The G.R.E.A.T. Review Workgroup was made up of G.R.E.A.T. officers and
administrators, staff members from the national evaluation, and experts in gangs
or school-based prevention programs. This group met three times during early
1999 and addressed four tasks. First, the workgroup reviewed the research on
American youth gangs on the following questions: What constitutes a gang, and
what risk factors are associated with gang membership? Next, since
G.R.E.A.T. is a school-based prevention program, the workgroup reviewed
the research evaluating the effectiveness of such programs to see what kinds of
strategies had been implemented and which elements had been found to reduce
delinquency or violence. The workgroup then turned its attention to
G.R.E.A.T.'s core curriculum, critically examining the extent to which it
contained elements found to be effective and ineffective in delinquency and
violence prevention. Finally, the workgroup outlined the structure of an
enhanced curriculum that incorporates elements known to be effective in
preventing delinquency and violence. Although some components of the
G.R.E.A.T. curriculum were retained, the proposed revised curriculum contains
many new elements.

An important first step in developing a gang prevention program is identifying
the risk factors associated with joining gangs. Thus, the workgroup reviewed
this literature, giving special consideration to individual, peer, family, and school
factors found to be predictive of gang membership. As this literature has been
summarized elsewhere (e.g., Curry and Decker 1998; Hill, Howell, Hawkins,
and Battin-Pearson 1999; Howell 1995, 1998), only a cursory overview is
provided below.

Youth gangs are found throughout the United States, and there has been an
apparent increase in the number of youth gangs and gang members since the
mid-1980s. Many theories have emerged to explain the formation of gangs and
why youths join them. Hagedorn (1988), Jackson (1991), and Klein (1995)
are among those who argue that gang formation is a product of postindustrial
development. The early work of Thrasher (1927) and other Chicago-based
gang researchers emphasized the importance of structural and community-level
factors and indicated that delinquency in general, and youth gangs in particular,
were a product of the social environment. Given that most youths residing in
areas where gangs are present choose not to join gangs, however, there must
be some other reason why some youths decide to join them. As Malcolm Klein
(1995, 75-76), in his summary of the literature of the demographic
characteristics of gang members, aptly observes:

In regard to who joins street gangs, then, first, it is not sufficient to say that gang
members come from lower-income areas, from minority populations, or from
homes more often characterized by absent parents or reconstituted families. . . .
[M]ost youths from such areas, such groups, and such families do not join
gangs (emphasis added).

It must also be added that youth gang membership is not an exclusively male
phenomenon (see, for instance, Bjerregaard and Smith 1993; Esbensen and
Huizinga 1993; Esbensen and Winfree 1998).

Several consistent attitudinal differences have been found between gang
members and nongang youths. Representative of the differences reported are
the findings from a Seattle study, in which Hill and colleagues (1999) found that
gang youths held more antisocial beliefs than youths who were not gang
members. Maxson and her colleagues (1998) found that gang members had a
more delinquent self-concept, were more likely to resolve conflict by threats,
and had experienced more critical stressful events than nongang youths. More
generally, both studies found significant differences between gang and nongang
youths within multiple contexts (i.e., individual, school, peer, family, and
community characteristics).

As with delinquency research, one consistent finding from the gang research is
the overarching influence of peers on adolescent behavior (e.g., Hill et al. 1999;
Menard and Elliott 1994; Warr and Stafford 1991). In their comparison of
stable and transient gang youths, Hill and colleagues reported that the strongest
predictors of sustained gang affiliation were a high level of interaction with
antisocial peers and a low level of interaction with prosocial peers. Researchers
have examined the influence of peers through various measures, including
exposure to delinquent peers, attachment to delinquent peers, and commitment
to delinquent peers. However this peer affiliation is measured, the results are the
same: Association with delinquent peers is one of the strongest predictors (i.e.,
risk factors) of gang membership.

Although less commonly examined by gang researchers, school factors also
have been consistently associated with the risk of joining gangs. Research
indicates that gang youths experience lower levels of commitment to school than
do nongang youths (Bjerregaard and Smith 1993; Esbensen and Deschenes
1998; Hill et al. 1999; Maxson, Whitlock, and Klein 1998). Some gender
differences have been reported, however; in the Rochester study (Bjerregaard
and Smith 1993), for example, school expectations were not predictive of male
gang membership. Ethnographic reports also attest to the role of school factors
in explaining gang membership (e.g., Campbell 1991; Fleisher 1998; Hagedorn
1988).

A second step in developing a school-based prevention program is to assess
the success of previous programs and current practices. (For a thorough review
of these school-based programs, see Gottfredson 1997.) A brief overview of
this literature is provided below.

Some evidence suggests that in schools that use a participatory management
style, where administrators and teachers actively communicate and collaborate,
teacher morale is higher and there is less disorder. Schools with clear rules and
reward structures also experience less disorder. And considerable evidence
suggests that "smaller is better"--larger schools experience higher levels of
violent crime than do smaller schools (Gottfredson and Gottfredson 1985;
Kaufman et al. 1998). Some attribute this to more effective informal social
control, whereas others are likely to attribute this to demographic variables.
Regardless, the consensus is that smaller schools experience less violence and
other disruptive behaviors.

Teaching styles and classroom organization also have been examined as
possible violence prevention strategies. One effective approach involves
cooperative learning strategies, whereby teachers provide initial instruction to
students, after which students are divided into groups of four to five students of
mixed skill levels; students help each other learn, but generally take tests
individually. This approach has been associated with higher academic
achievement, more positive attitudes toward school, better race relations, and
acceptance of special education students who have been mainstreamed. The
consensus is that these improvements in educational performance may also be
associated with reductions in violence (Brewer et al. 1995).

Consistent with research on parental discipline and parenting strategies, the
most important aspect of classroom and school management approaches is
consistency and fairness in applying the rules. Teachers and administrators have
learned that having clear rules, enforcing those rules, and positive feedback are
key elements of school safety (Gottfredson 1997).

Several programs target the individual, seeking to change attitudes and, thus,
behavior. These programs tend to focus on increasing knowledge and skills
while changing beliefs. Most of the programs that focus directly on crime and
violence prevention (including G.R.E.A.T.) can be classified as individual
change strategies. To date, evaluations of these strategies have provided mixed
results (Brewer et al. 1995; Drug Strategies 1998). In isolation, however, these
programs do not appear to have the desired effect.

One such program that has received considerable attention is the Life Skills
Training (LST) program developed by Gil Botvin (Botvin 1998). Developed as
a drug prevention program, it may also provide beneficial information for gang
prevention. LST is a 3-year intervention (15 lessons in the first year, with 10
booster sessions in the second year, and 5 in the third) designed to be
implemented in school classrooms. Students are taught a set of general
self-management and social skills and provided with information and skills that
are directly related to drug abuse. At face value, this program is not that
different from D.A.R.E. (registered trademark), G.R.E.A.T., and other
individual change strategies. It differs, however, in its instructional approach,
which emphasizes skills development rather than knowledge assimilation.
Problem-solving exercises and a combination of instructional strategies are key
aspects of the program. Evaluations of LST have reported reductions in drug
use and positive effects on mediating variables, such as interpersonal and
communication skills (Botvin 1998).

After reviewing the literature described above, the G.R.E.A.T. Review
Workgroup carefully examined the overall objectives of the G.R.E.A.T.
program and the content of its curriculum. The workgroup recognized that
unlike suppression and intervention programs, which target youths already in
gangs, G.R.E.A.T. targets all youths, seeking to provide life skills to empower
adolescents to resist peer pressure to join gangs. This strategy is meant to be a
cognitive approach that produces attitudinal and behavioral change through
instruction, discussion, and role playing. The workgroup found, however, that
the curriculum lacked many of the elements necessary to prevent delinquency.

In examining each curriculum lesson, the workgroup identified which, if any,
components were consistent with the elements of effective prevention
programs. The workgroup also searched for learning strategies within the
curriculum, such as cooperative learning and active student and teacher
participation, that were consistent with effective prevention efforts. For
example, classroom observations and reports from officers indicated that the
classroom teacher was not integrated into the actual G.R.E.A.T. lessons
(Sellers, Taylor, and Esbensen 1998). At G.R.E.A.T. officer training, the
officers are encouraged to engage teachers in the lessons and give them
supplemental activities. This rarely occurred, however; teachers tended to treat
the lesson time as a free planning period or a coffee break.

A primary concern was whether the curriculum focused sufficiently on providing
students with social competency skills. The curriculum contained lessons on
goal-setting, responsibility, and problem solving, and portions of lessons at least
touched on having empathy for victims, developing positive social affiliations,
and altering perceptions about gangs. Most of the lessons, however, were
found to be didactic, relying primarily on lecture and information dissemination.
Moreover, the skills being taught were presented in isolation from one another,
with little effort to revisit earlier skills and build upon them. Although each
lesson contained a group activity, the emphasis on information dissemination
precluded students from practicing the skills they were taught (Sellers, Taylor,
and Esbensen 1998). Finally, the curriculum notably lacked such social
competency skills as stress management, emotional control, and
communication.

Because it was charged with conducting a critical review of the G.R.E.A.T.
program and recommending improvements, the G.R.E.A.T. Review
Workgroup outlined a "revised" G.R.E.A.T. program and submitted it to the
NPB. As revised, this program would continue to be taught in the first year of
middle school, with recommended booster sessions in each subsequent year of
middle school and, if possible, in high school. The workgroup also
recommended that supplemental programs, such as the current summer and
parent components, be retained and possibly expanded.

The workgroup identified the following goals and objectives of the G.R.E.A.T.
program for the NPB's approval and verification:

o Goals--

-- To reduce gang membership.

-- To prevent violence and criminal activity.

-- To develop positive relationships with law enforcement.

o Objectives--

-- To improve social competency skills (emotional control; stress management;
conflict resolution; communication and listening; and decisionmaking,
problem-solving, and goal-setting skills).

-- To foster empathy for victims.

-- To encourage positive social affiliations.

-- To alter perceptions about gangs.

-- To increase social responsibility.

In developing a curriculum outline, the workgroup intended to structure the
curriculum around social competency skills, which would be reinforced in future
lessons. Each subsequent lesson added new material to address other program
objectives (i.e., having empathy for victims, altering perceptions about gangs,
developing positive social affiliations, and demonstrating social responsibility).

Given these goals and objectives, the workgroup produced the following outline
for the revised G.R.E.A.T. program:

o Introduction

-- Lesson 1: Introduction

o Unit I: Interpersonal Skills

-- Lesson 2: Empathy

-- Lesson 3: Communication/Listening

o Unit II: Decision-Making

-- Lesson 4: Evaluating Input

-- Lesson 5: Problem-Solving

-- Lesson 6: Goal-Setting

-- Lesson 7: Social Responsibility

o Unit III: Conflict Resolution

-- Lesson 8: Emotional Control and Stress Management

-- Lesson 9: Conflict Resolution

Wrap-up and Culmination Activity

Research shows that middle school is a crucial transition point for adolescents.
Thus, the workgroup recommended that G.R.E.A.T. be taught at middle school
entry, whether it be sixth or seventh grade. Existing prevention literature also
reinforced several issues that the Workgroup believed were important in any
reconceptualization of the G.R.E.A.T. program. First, the literature
recommended greater emphasis on incorporating the teacher into the lesson
plan to enhance the reinforcement of lessons and skills learned. Second, the
curriculum should focus more on active learning strategies, rather than relying
on the didactic style used by most officers (Brewer et al. 1995; Catalano,
Loeber, and McKinney 1999; Gottfredson 1997). Third, research documented
the desirability of booster sessions to reinforce skills learned in prior years
(Botvin 1998; Botvin et al. 1990).

The NPB accepted the workgroup's recommendations and by August 2000, a
team of G.R.E.A.T. officers, curriculum writers, gang researchers, and experts
in school-based prevention programs had written an "enhanced" curriculum.
During the fall of 2000, officers were retrained in the new curriculum, and pilot
programs were taught in the spring of 2001.

Conclusions and Policy Implications

The G.R.E.A.T. program seeks to reduce adolescent involvement in crime and
gangs. Before implementing the longitudinal, quasi-experimental study described
in this report, the authors conducted a preliminary cross-sectional survey of
students to assess program effectiveness. Findings from that research supported
continuing the G.R.E.A.T. program (Esbensen and Osgood 1999). Initial
results from the longitudinal quasi-experimental research design described in this
chapter failed to replicate those favorable results. In response, a rigorous
program review resulted in development of a revised G.R.E.A.T. curriculum.
The 4-year results, however, were consistent with the cross-sectional results.
At this juncture, it is necessary to assess these contradictory research findings
and their consequences.

It is not uncommon for evaluations conducted with different samples and at
different times to produce mixed results. The cross-sectional evaluation of the
G.R.E.A.T. program was completed in 1995 in 11 cities using anonymous
questionnaires completed by students under passive parental consent
procedures. The longitudinal evaluation was conducted in six cities (four of
which were included in the cross-sectional study) from 1995 to 1999 using
confidential questionnaires restricted to those students for whom active parental
consent had been obtained. After analyzing the year 2 followup data, the
evaluation team considered several factors that could account for the null
findings. Methodological differences (i.e., anonymous versus confidential
questionnaires, different consent processes, and different samples) could have
contributed to the contradictory results. With the subsequent findings of a
4-year lagged programmatic effect, however, attention turns to two questions:
(1) What could account for the lagged program effects? and (2) Was the
program review implemented after the interim results worthwhile?

The finding that G.R.E.A.T. students were more prosocial at the 4-year
followup than were control students is curious and unexpected, especially since
no such differences were observed at the 2-year followup. Yet other
evaluations (the Perry Preschool Project and the Seattle Social Development
Intervention) have reported similar lagged effects (Berrueta-Clement et al.
1984; Hawkins et al. 2000). The reason for this lag is unclear, although several
possible interrelated explanations come to mind. First, young adolescence is a
stressful and anxiety-filled stage, during which most adolescents are confused
about appropriate attitudes and behaviors. Second, the organizational structure
of American schools may contribute to this stress. At ages 11 or 12, children
move from the comfort of relatively small, stable elementary schools to larger,
more diverse middle or junior high schools; then, at ages 14 or 15, young
adolescents are forced to make another transition to even larger, more diverse
high schools. As they reach the ages of 16 and 17 (10th or 11th grade), some
of the angst of adolescence is resolved and they have adapted to the high
school setting. Thus, prior prevention or intervention experiences may begin to
manifest themselves at this time. Before this maturation occurs, other factors
may have obfuscated the effects of the prevention experiences.

In considering the utility of the program review, this exercise serves as an
excellent example of cooperative collaboration between practitioners and
researchers. The initial curriculum had been developed relatively quickly, with
little input from education and prevention specialists. Six years after its
development, the program had unexpectedly expanded nationwide. What had
been developed as a local program for Phoenix was experiencing "growing
pains," as some educators and G.R.E.A.T. officers called for a review of the
curriculum. Thus, during a 3-day meeting in August 1997, a panel of officers
and educators reviewed the G.R.E.A.T. lessons. This meeting failed to produce
a consensus on proposed modifications and no changes were implemented. The
presentation of the null findings from the 2-year followup evaluation to the NPB
in October 1998 again prompted a program review, including examination of
curriculum content and educational practices. The NPB should be
acknowledged for its willingness to respond to evaluation results that did not
support the program and for its desire to enhance the program's potential.
Unlike the earlier review, this review produced a consensus and, as described
earlier, led to the enhanced G.R.E.A.T. program that was tested in the spring of
2001.

Where does this leave us with regard to gang prevention policy? Can police
officers be effective instructors? Given the lack of consistent findings on the
effectiveness of the G.R.E.A.T. program, this is an important question. From a
school safety and community policing perspective, however, it may be
reasonable to continue this strategy. Some evidence indicates that the officers
may have a small positive effect on student attitudes and behavior. Additionally,
surveys completed with teachers and parents as part of the national evaluation
revealed that most respondents favored school-based prevention programs and
having officers instruct students and generally supported the G.R.E.A.T.
program. A lingering question remains, however: How effective can
individual-based prevention programs be in reducing gang involvement? As a
review of risk factors reveals, a significant reduction in gang activity may be too
much to expect from any program if the more fundamental causes and
attractions of gangs (i.e., social, structural, community, and family conditions)
are not simultaneously addressed.

To conclude, there may be no one "silver bullet" program or "best practice" for
preventing gang affiliation and gang-associated violence. The youth gang
problem may be best addressed through a comprehensive strategy (such as the
Office of Juvenile Justice and Delinquency Prevention's Comprehensive Gang
Model) with a multifaceted approach that targets individual youths, as well as
peer groups, family, school, and the community. G.R.E.A.T., in tandem with
other programs, may prove to be one piece of a much larger solution.

---------------------------

Notes

1. D.A.R.E. (registered trademark), Drug Abuse Resistance Education, is a
school-based drug prevention program taught by uniformed law enforcement
officers. This 16-lesson program targeting elementary school students has been
widely accepted and implemented by police departments across the United
States.

2. We had initially hoped to implement a true experimental design, but
real-world conditions precluded true random assignment in two of the sites.
Thus, the assignment of classrooms to G.R.E.A.T. and non-G.R.E.A.T.
participation was achieved through negotiations with school personnel at each
site.

3. Portland educators requested that the G.R.E.A.T. program be delivered at
the entry year to middle schools (i.e., sixth grade). The G.R.E.A.T.
management agreed to this arrangement and subsequently approved a policy of
preferably implementing the program during the entry year to either middle
school or junior high school.

References

Akers, Ronald, L. 1985. Deviant Behavior: A Social Learning Approach. 3rd
ed. Belmont, CA: Wadsworth.

Berrueta-Clement, J.R., L.J. Schweinhart, W.S. Barnett, A.S. Epstein, and
D.P. Weikart. 1984. Changed Lives: The Effects of the Perry Preschool
Program on Youths through Age 19. Ypsilanti, MI: High/Scope Press.

Bjerregaard, Beth, and Carolyn Smith. 1993. "Gender Differences in Gang
Participation, Delinquency, and Substance Use." Journal of Quantitative
Criminology 4: 329-55.

Botvin, Gil J. 1998. Life Skills Training. Book Five of the Blueprints for
Violence Prevention. Boulder: University of Colorado Press.

Botvin, Gil J., E. Baker, A.D. Filazzola, and E.M. Botvin. 1990. "A
Cognitive-Behavioral Approach to Substance Abuse Prevention: One-Year
Follow-up." Addictive Behaviors 15: 47-63.

Brewer, Devon, J. David Hawkins, Richard F. Catalano, and Holly J.
Neckerman. 1995. "Preventing Serious, Violent, and Chronic Juvenile
Offending: A Review of Evaluations of Selected Strategies in Childhood,
Adolescence, and the Community." In A Sourcebook: Serious, Violent, and
Chronic Juvenile Offenders, ed. James C. Howell, Barry Krisberg, J. David
Hawkins, and John J. Wilson. Thousand Oaks, CA: Sage Publications:
61-141.

Bryk, Anthony S., and Stephen W. Raudenbush. 1992. Hierarchical Linear
Models: Applications and Data Analysis Methods. Newbury Park, CA: Sage.

Campbell, Anne. 1991. The Girls in the Gang. 2nd. ed. Cambridge, MA: Basil
Blackwell.

Catalano, Richard F., Rolf Loeber, and Kay C. McKinney. 1999. School and
Community Interventions to Prevent Serious and Violent Offending. Juvenile
Justice Bulletin. Washington, DC: U.S. Department of Justice, Office of
Juvenile Justice and Delinquency Prevention, October, NCJ 177624.

Curry, G. David and Scott H. Decker. 1998. Confronting Gangs: Crime and
Community. Los Angeles: Roxbury.

Dillman, Don A. 1978. Mail and Telephone Surveys: The Total Design
Method. New York: John Wiley and Sons.

Drug Strategies. 1998. Safe Schools, Safe Students: A Guide to Violence
Prevention Strategies. Washington, DC: Drug Strategies.

Esbensen, Finn-Aage, and David Huizinga. 1993. "Gangs, Drugs, and
Delinquency in a Survey of Urban Youth." Criminology 31: 565-89.

Esbensen, Finn-Aage, and Elizabeth P. Deschenes. 1998. "A Multisite
Examination of Youth Gang Membership: Does Gender Matter?" Criminology
36: 799-828.

Esbensen, Finn-Aage, and L. Thomas Winfree, Jr. 1998. "Race and Gender
Differences Between Gang and Non-Gang Youth: Results from a Multi-Site
Survey." Justice Quarterly 15: 505-26.

Esbensen, Finn-Aage, and D. Wayne Osgood. 1999. "Gang Resistance
Education and Training (G.R.E.A.T.): Results from the National Evaluation."
Journal of Research in Crime and Delinquency 36: 194-225.

Esbensen, Finn-Aage, Michelle H. Miller, Terrance J. Taylor, Ni He, and
Adrienne Freng. 1999. "Differential Attrition Rates and Active Parental
Consent." Evaluation Review 23: 316-35.

Esbensen, Finn-Aage, D. Wayne Osgood, Terrance J. Taylor, Dana Peterson,
and Adrienne Freng. 2001. "How Great is G.R.E.A.T.?: Results from a
Longitudinal, Quasi-Experimental Design." Criminology and Public Policy 1:
87-118.

Fleisher, Mark. 1998. Dead End Kids. Madison: University of Wisconsin
Press.

Goldstein, Harvey. 1995. Multilevel Statistical Models. London: Arnold.

Gottfredson, Denise. 1997. "School-Based Crime Prevention." In Preventing
Crime: What Works, What Doesn't, What's Promising: A Report to the United
States Congress, ed. Lawrence W. Sherman, Denise Gottfredson, Doris
MacKenzie, John Eck, Peter Reuter, and Shawn Bushway. Washington, DC:
U.S. Department of Justice, National Institute of Justice, NCJ 165366.

Gottfredson, Gary D., and Denise C. Gottfredson. 1985. Victimization in
Schools. New York: Plenum.

Gottfredson, Michael R., and Travis Hirschi. 1990. A General Theory of
Crime. Stanford, CA: Stanford University Press.

G.R.E.A.T. Brochure. n.d. "Ask the Phoenix Police Chief How the Fight
Against Youth Gangs Is Going and He Will Give You A One Word Answer:
GREAT." Washington, DC: U.S. Department of Treasury.

G.R.E.A.T. News. 1994. G.R.E.A.T. Mission Statement. May 3: 1.

Hagedorn, John M. 1988. People and Folks: Gangs, Crime, and the
Underclass in a Rustbelt City. Chicago: Lakeview Press.

Hansen, William B., Linda M. Collins, C. Kevin Malotte, C. Anderson
Johnson, and Jonathan E. Fielding. 1985. "Attrition in Prevention Research."
Journal of Behavioral Medicine 8: 261-75.

Hawkins, J. David, Jie Guo, Karl Hill, Sara Battin-Pearson, and Robert
Abbott. 2000. "Long Term Effects of the Seattle Social Development
Intervention on School Bonding Trajectories." Seattle: University of
Washington, Social Development Research Group.

Heinsman, Donna T. and William R. Shadish. 1996. "Assignment Methods in
Experimentation: When Do Nonrandomized Experiments Approximate
Answers from Randomized Experiments?" Psychological Methods 1: 154-69.

Hill, Karl G., James C. Howell, J. David Hawkins, and Sara R. Battin-Pearson.
1999. "Childhood Risk Factors for Adolescent Gang Membership: Results
from the Seattle Social Development Project." Journal of Research in Crime
and Delinquency 36: 300-22.

Hindelang, Michael J., Travis Hirschi, and Joseph G. Weis. 1981. Measuring
Delinquency. Beverly Hills, CA: Sage Publications.

Howell, James C. 1995. "Gangs and Youth Violence: Recent Research." In A
Sourcebook: Serious, Violent, and Chronic Juvenile Offenders, ed. James C.
Howell, Barry Krisberg, J. David Hawkins, and John J. Wilson. Thousand
Oaks, CA: Sage Publications.

-------. 1998. "Promising Programs for Youth Gang Violence Prevention and
Intervention." In Serious and Violent Juvenile Offenders: Risk Factors and
Successful Interventions, ed. Rolf Loeber and David P. Farrington. Thousand
Oaks, CA: Sage Publications.

Huizinga, David. 1991. "Assessing Violent Behavior with Self-Reports." In
Neuropsychology of Aggression, ed. Joel Milner. Boston: Kluwer: chapter 3.

Huizinga, David, and Delbert S. Elliott. 1986. "Reassessing the Reliability and
Validity of Self-Report Delinquency Measures." Journal of Quantitative
Criminology 2: 293-327.

Jackson, Pamela I. 1991. "Crime, Youth Gangs, and Urban Transition: The
Social Dislocations of Post-industrial Economic Development." Justice
Quarterly 8: 379-97.

Kaufman, Phillip, Xianglei Chen, Susan P. Choy, Kathryn A. Chandler,
Christopher D. Chapman, Michael R. Rand, and Cheryl Ringel. 1998.
Indicators of School Crime and Safety, 1998. Washington, DC: U.S.
Department of Education and U.S. Department of Justice.

Klein, Malcolm W. 1995. The American Street Gang. New York: Oxford
University Press.

Maxson, Cheryl L., Monica L. Whitlock, and Malcolm W. Klein. 1998.
"Vulnerability to Street Gang Membership: Implications for Practice." Social
Service Review 72: 70-91.

Menard, Scott, and Delbert S. Elliott. 1994. "Delinquent Binding,

Moral Beliefs, and Illegal Behavior: A Three-Wave Panel." Justice Quarterly
11: 173-88.

Palumbo, Dennis J., and Jennifer L. Ferguson. 1995. "Evaluating Gang
Resistance Education and Training (G.R.E.A.T.): Is the Impact the Same as
That of Drug Abuse Resistance Education (D.A.R.E.)?" Evaluation Review 19:
591-619.

Sellers, Christine S., Terrance J. Taylor, and Finn-Aage Esbensen. 1998.
"Reality Check: Evaluating a School-Based Gang Prevention Model."
Evaluation Review 22: 590-608.

Sherman, Lawrence W., Denise Gottfredson, Doris MacKenzie, John Eck,
Peter Reuter, and Shawn Bushway. 1997. Preventing Crime: What Works,
What Doesn't, What's Promising: A Report to the United States Congress.
Washington, DC: U.S. Department of Justice, National Institute of Justice, NCJ
165366.

Tebes, Jacob K., Davis L. Snow, and Michael W. Arthur. 1992. "Panel
Attrition and External Validity in the Short-Term Follow-Up Study of
Adolescent Substance Use." Evaluation Review 16: 151-70.

Thrasher, Frederick M. 1927. The Gang: A Study of One Thousand Three
Hundred Thirteen Gangs in Chicago. Chicago: University of Chicago Press.

Warr, Mark, and Mark Stafford. 1991. "The Influence of Delinquent Peers:
What They Think or What They Do?" Criminology 29: 851-65.

Winfree, L. Thomas, Jr., Dana Peterson Lynskey, and James R. Maupin.
1999. "Developing a Local Police and Federal Law Enforcement Partnerships:
G.R.E.A.T. as a Case Study of Policy Implementation." Criminal Justice
Review 24: 145-168.

Winfree, L. Thomas, Jr., Finn-Aage Esbensen, and D. Wayne Osgood. 1996.
"Evaluating a School-Based Gang Prevention Program: A Theoretical
Perspective." Evaluation Review 20: 181-203.

---------------------------

Active Consent Procedures [sidebar]

The University of Nebraska Institutional Review Board approved a research
design that allowed passive parental consent during the pre- and posttest data
collection (i.e., students were surveyed unless a parent refused their
participation). These surveys were conducted 2 weeks before and after
delivery of the G.R.E.A.T. program. Active parental consent was planned for
the subsequent annual surveys (only those students for whom signed permission
was obtained from a parent were surveyed). These procedures were also
approved by each of the participating school districts.

A modified Dillman (1978) total design method was used to obtain the active
consent forms, although the specific procedures varied slightly for timing and
sequencing across the six sites. The following account serves as an "ideal type"
of the procedures followed.

During the spring and summer of 1996, three direct mailings were made to
parents of survey participants. Included in the mailings were a cover letter, two
copies of the parent consent form for student participation, and a business reply
envelope. As Phoenix and Las Cruces had substantial Spanish-speaking
populations, mailings to parents in these cities included Spanish versions of the
cover letter and consent form. All parents not responding after the second
mailing were contacted by telephone. School personnel also cooperated by
distributing consent forms and cover letters at school. Teachers in all
classrooms involved in the evaluation assisted with this process, rewarding
students with a new pencil upon return of the forms. Some teachers agreed to
allow us to offer incentives, such as pizza parties, to classrooms in which a
minimum of 70 percent of students returned a completed consent form. Other
teachers offered their own incentives, including earlier lunch passes and extra-
credit points. This process resulted in an overall response rate of 67 percent
(57 percent providing affirmative consent and 10 percent withholding consent),
while 33 percent of parents failed to return the consent forms. (For a more
detailed discussion of the active consent process and examination of the effects
of active consent procedures on the representativeness of the sample, see
Esbensen et al. 1999.)

Questionnaire Completion Rates [sidebar]

Completion rates for the student survey were excellent. Of the 2,045 students
for whom active parental consent had been obtained at the six sites, 1,761
surveys (86 percent) were completed at the 1-year followup and 1,550 (76
percent) at the 2-year followup. Given the multisite, multischool sample and that
respondents at five of the six sites made the transition from middle school to
high school between the year 1 and year 2 surveys, this completion rate is
commendable. Hansen and colleagues (1985) examined attrition in a meta-
analysis of 85 longitudinal studies and reported an average completion rate of
72 percent for the 19 studies with a 2-year followup period; few of these 19
studies included multisite samples. Tebes, Snow, and Arthur (1992) reported
on the attrition rates from middle school to high school. In their study examining
differential attrition for different age groups, they reported losing 41.3 percent
of their sample between eighth and ninth grade.

In the present study, considerable difficulty was encountered in retaining the
student sample for the year 2 followup. As the cohort moved from middle
school to high school and normal mobility patterns occurred, students in
Omaha, Phoenix, and Philadelphia were enrolled in more than 10 different high
schools in each city. To ensure these students' continued participation in the
study, the evaluation team contacted officials at these schools. In some
instances, these new schools were in different districts, which required approval
from the necessary authorities to survey their students. Despite these logistical
concerns, completed questionnaires were obtained from 76 percent of the
sample at the year 2 followup. In the third (1998) and fourth (1999) year
surveys, this high standard of questionnaire completion was again achieved,
with response rates of 69 and 67 percent, respectively, attained in 1998 and
1999.

---------------------------

Chapter 6

Evaluating Nevada's Antigang Legislation and Gang Prosecution Units

Terance D. Miethe, Ph.D.
Richard C. McCorkle, Ph.D.
Department of Criminal Justice
University of Nevada-Las Vegas

Acknowledgments

This project was supported by grant 94-IJ-CX-0053 awarded by the National
Institute of Justice. The authors appreciate the helpful comments of Winifred L.
Reed and Scott H. Decker on an earlier version of this paper.

---------------------------

Abstract

Like many other States, Nevada enacted antigang legislation to increase the
severity and certainty of punishment for gang members and established gang
prosecution units to improve the chances of successfully prosecuting gang
crime. Little research has focused on these types of laws and gang units,
however. The current study examines how often Nevada's antigang legislation
has been used and the extent to which the gang units in the State's two largest
counties, Clark (whose major city is Las Vegas) and Washoe (whose major
city is Reno) have succeeded. The study revealed that the antigang statutes are
widely used in some cases, when firearms are involved, but that convictions are
the exception rather than the rule. When gang cases were compared before and
after enactment of the legislation, no significant difference in conviction rates
were found. Such charges as being an accessory, aiding and abetting, and
racketeering are rare in both counties and have neither increased nor decreased
as a result of the legislation. There were no differences between the gang units
and the track (general) prosecution units in likelihood of conviction and
imprisonment. The antigang statutes may provide additional leverage in
prosecuting gang crime, because the threat of conviction under them may be an
enticement to plead guilty on other charges. Special prosecution units can be an
important arena for processing and adjudicating gang cases, although the
decision to establish them will depend on the gravity of the gang problem and
the expertise of the district attorneys.

---------------------------

Gang activity has become a major social problem in nearly all large U.S. cities.
Teen violence (much of which is gang related) increased dramatically from the
mid-1980s to the mid-1990s (Office of Juvenile Justice and Delinquency
Prevention 1993). Driveby shootings, drug-related homicides, and other types
of criminal activities performed by gang members have been major contributors
to this problem. Gang activity in the two largest cities in Nevada is no different.
Police estimate that up to 6,000 active gang members and associates reside in
Clark County (Las Vegas) and 1,000 in Washoe County (Reno). Las Vegas
had 17 gang-related homicides and 280 driveby shootings in 1993, and more
than 100 felonies involving gang members were reported to the Reno police in
1994. Rising gang violence in both cities, and a high-profile armed robbery of a
casino by Los Angeles gang members, helped fuel a major perceived threat to
the social and economic vitality of these metropolitan areas.

An increasingly popular response to gang activity has been the formation of
specialized gang units within police departments and the district attorney's
office. These collaborative efforts are designed to increase the gathering of
intelligence on gang activity and gang membership, develop key community
informants, and promote greater cooperation between crime witnesses and law
enforcement in prosecuting gang members. Rigorous, successful prosecution of
these offenders is especially important in overcoming the strong feelings of
invulnerability and immunity to prosecution many gang members hold. One way
to change these attitudes, and to ensure that the threat of legal sanction has any
deterrent value, is to use gang units and criminal statutes to increase the severity
and the certainty of punishment. This is the objective that underlies Nevada's
antigang legislation and the development of gang prosecution units in Las Vegas
and Reno. 

Using various methodologies (e.g., content analysis, statistical analysis of case
disposition, interviews of court personnel, and field observations), this study
examines the frequency of application of Nevada's antigang legislation in
criminal court practices and the relative effectiveness of the gang prosecution
units used in Nevada's two largest counties.

Review of the Literature

Although systematic empirical data on gang prosecution practices are limited
(see ILJ 1994; Conly 1993; Reiner 1992; Spergel 1990; McKinney 1988;
Zatz 1985; Dahmann 1983), previous studies have examined the structure of
gang units and special problems that arise in prosecuting gang cases. This
section explores the structure and function of gang prosecution units, antigang
legislation, the historical context of Nevada's antigang legislation, and
characteristics of gang prosecution units in Clark (Las Vegas) and Washoe
(Reno) Counties.

The Structure and Function of Gang Prosecution Units

It is widely recognized that gang offenses create severe problems for law
enforcement and prosecution (see, for review, ILJ 1994; Curry et al. 1994;
Reiner 1992; Spergel 1990; McKinney 1988; Klein et al. 1987; Maxson et al.
1987). Police often lack adequate intelligence on gang membership and
structure to prove gang participation; the level of victim and witness
cooperation is low, because of fear or intimidation or a preference for "street
payback"; and many victims and witnesses lack credibility because they are
also gang members. Gang offenses present major problems for prosecutors
because these cases often--

o Span both juvenile and adult systems.

o Require detailed attention (with most district attorneys unable to take on gang
cases because of their heavy caseloads).

o Demand specialized knowledge of how to use gang experts in courts, prove
gang membership, and execute search warrants.

These problems are less severe in gang cases involving drug trafficking in which
informants and police serve as witnesses. However, for cases involving violent
gang activity, victim and witness problems are so acute that many prosecutors
are more willing to use a current arrest to revoke a previous probation, which
has less stringent evidentiary requirements, than to initiate prosecution on a new
charge.

In response to these challenges, many jurisdictions have developed specialized
gang units within police departments and district attorneys' offices (see, for
review, ILJ 1994; Curry et al. 1994; Spergel 1991; Maxson et al. 1987). The
structure and function of these gang units vary widely, however. For example,
50 attorneys in Los Angeles County's Hardcore Gang Division deal almost
exclusively with gang-related homicides (see Genelin 1994; Reiner 1992),
whereas only 3 deputy district attorneys are assigned to the gang unit to
prosecute all gang offenses in Las Vegas. A national survey reveals that 32
percent of jurisdictions with populations of more than 250,000 have gang
prosecution units, compared with only 5 percent in smaller jurisdictions (see ILJ
1994). Instead of having gang units, small counties may treat gang cases
through an existing specialized unit (like drug, organized crime, or career
criminal units) or simply tag gang cases, as they are received, for specialized
attention or priority handling. Much of what is known about gang prosecution is
derived from prosecution studies in the largest U.S. cities (especially Los
Angeles and Chicago), which may provide little insight for law enforcers in
smaller metropolitan areas that have less serious gang problems and fewer
resources to devote to specialized units. Nonetheless, gang prosecution units
across geographical areas seem to share several major structural features.

For example, a key feature of many gang units is the adoption of vertical
prosecution. Under this approach, the same district attorney (and case
investigator, depending on the jurisdiction) handles the case from preliminary
hearing through sentencing. Vertical prosecution is widely advocated for gang
cases and other complicated cases because it provides continuity over the life
of the criminal complaint and takes advantage of specialized expertise. This
continuity is especially important for developing and maintaining rapport with
key witnesses. Prosecutors under a vertical system can more fully develop
expertise on such matters as how to use gang experts in court, how to prove
gang membership by issuing search warrants for multiple dwellings, why and
how to use gang graffiti in court, how to use tattoos to prove gang membership,
how to coordinate both juvenile and adult prosecutions, and how to provide
evidence to support complex charges of conspiracy or aiding and abetting (see
Genelin 1994). Although comprehensive studies of prosecutorial practices
before and after implementation of vertical prosecution are rare (see Spergel
1991; Daley 1985; Dahmann 1983), most experts support this approach for
gang prosecution (see Genelin 1994; ILJ 1994; McKinney 1988).

A second common feature of gang units across jurisdictions is the development
of close working relationships with agencies in the criminal justice community. A
cooperative exchange between the police, judiciary, and prosecution is a
fundamental necessity for any district attorney, but good working relationships
take on added importance in gang cases, in which evidentiary requirements for
determining that a case is "gang related" and victim or witness testimony are far
more problematic. Specialized gang units within police departments provide the
necessary intelligence on gang membership, initiation practices, and offense
specialization to establish elements required for successful prosecution (see
Genelin 1994; Reiner 1992; Klein et al. 1987; Maxson et al. 1987). As one
strategy for promoting cooperation, Los Angeles has developed an interagency
gang task force to coordinate all agencies working on gang issues and created
the Gang Reporting Evaluation and Tracking (GREAT) database to help law
enforcement (see Reiner 1992; McKinney 1988). Reno, Nevada, in contrast,
has made a full-scale commitment to community and problem-oriented policing
as another strategy to coordinate gang abatement efforts (see Conly 1993).

Antigang Legislation

Most States have found traditional criminal law to be adequate in prosecuting
crimes committed by gang members (ILJ 1994). In addition to existing criminal
codes for violent, property, and narcotic offenses, prosecutors employ
conspiracy, aiding and abetting, and drug kingpin statutes to reach those
indirectly associated with the offense. RICO (Racketeer Influenced Corrupt
Organization) statutes also have been proposed to target gang activity, but
these statutes are often limited because they are designed for more
sophisticated and sustained criminal activity than is typically found among street
gangs (see ILJ 1994). Some gang-related offenses, like driveby shootings, also
are not adequately covered under existing criminal laws. Over the past decade,
new statutes that target criminal street gang activity have been enacted in 14
States to augment the existing criminal codes.

Although there are no comprehensive evaluations of prosecution, adjudication,
and sentencing practices under antigang statutes, these legislative efforts may
not be fully applied in actual criminal court practices for several reasons. First,
under many State provisions, it is difficult to prove that an offense was gang
related and consequently that the statute applies to a particular case. Second,
the statute may be so narrow in its scope, or the penalty so mild, that
prosecutors believe it is not worth the extra effort. Third, some gang members
(especially in some Asian gangs) are more secretive of their gang involvement,
exhibit fewer external signs of gang membership, and are less likely than others
to "roll over" on codefendants, making successful prosecution more difficult.
Finally, criminal justice officials may oppose some types of antigang legislation
whose use affects charging and sentencing decisions. Under these conditions,
one would expect little change in prosecution, adjudication, and sentencing
practices after the implementation of particular antigang statutes.

Previous research on sentencing reform (see Miethe and Moore 1988, 1987;
Blumstein et al. 1983; Martin 1983) suggests several hypotheses about the
nature and consequences of opposition to antigang legislation. First, the greater
the level of support from prosecution and judges for particular types of antigang
legislation, the more frequently it will be used. However, the complexity of
proving that an offense was gang related or involved gang participation under
some State codes may be sufficient to discourage many prosecutors from filing
these charges even if they generally support this legislation. Similarly, judges
may resist efforts to use particular antigang statutes because such legislation
may have mandatory provisions thought to usurp their discretion. Second, use
of antigang statutes is likely to change as officials adapt to these external
changes. For example, as prosecutors become more accustomed to applying
these statutes, they may be more likely to use them in plea bargaining and
sentencing decisions. Alternatively, it is possible that antigang statutes will be
less frequently used over time because public defenders and judges who
oppose these efforts have developed strategies to nullify elements necessary for
successful prosecution.

The Historical Context of Nevada's Antigang Legislation

Throughout the late 1980s, Nevada experienced an alarming growth in gangs,
gang members, and gang-related crime. By 1990, 6,000 active gang members
in some 60 street gangs were estimated for Clark County alone (Lucherini
1991). That same year, criminal activity included 95 reported driveby
shootings, 13 gang-related murders, and more than 1,100 felony arrests by the
Las Vegas Metropolitan Police Department's Gang Unit. During this period,
Reno was first experiencing gang violence. In both jurisdictions, skepticism was
growing about the adequacy of the Nevada criminal codes to deal with
gang-related crime. Although local law enforcement had made arresting local
gang members under the State's RICO statute a priority, early prosecutorial
efforts to obtain convictions in court were a miserable and much-publicized
failure (Las Vegas Review-Journal 1991).

In response to rising public fears about the gang problem and the perceived
ineffectiveness of Nevada's statutes, State lawmakers enacted seven statutes
between 1989 and 1991 in the war against gangs. These statutes addressed the
following:

o Possessing a dangerous weapon on school property or in a vehicle at school.

o Discharging a firearm out of a motor vehicle.

o Aiming a firearm at a human being or discharging a weapon where a person
might be endangered.

o Imposing an additional penalty for the procurement or solicitation of a minor
to commit certain violations as an agent.

o Establishing additional penalties for committing certain violations at or near
schools, school bus stops, or recreational facilities for minors.

o Creating penalties for felonies committed on a school bus.

o Doubling the penalty for felonies committed to promote activities of a criminal
gang (referred to as the "gang sentencing enhancement statute," or SB230).

In often heated Senate Judiciary Committee hearings on SB230, considered the
most far-reaching of Nevada's antigang bills, elected officials, the ACLU, and
private citizens voiced their concerns. These individuals questioned the
definition of a "gang," methods of identifying gang members, the mandatory
nature of the additional penalties, rights, and the "inevitable" discriminatory
application of the law if passed. Nonetheless, after a series of revisions (that
many argued diluted the bill), SB230 passed through the committee, reached a
majority vote in the Senate, and was signed into law by Governor Bob Miller
on June 20, 1991. The other six antigang statutes in Nevada were passed with
less political opposition.

The gang sentencing enhancement statute, Nevada Revised Statute 193.168, is
in many ways similar to antigang measures enacted in other States. Like other
statutes, it provides a definition of a criminal gang (derived from Nevada's own
RICO statute),[1] stipulates how membership of a criminal defendant in a street
gang must be established ("expert testimony"), and prescribes the penalty for
conviction for a gang-related crime ("equal to and in addition to the term of
imprisonment prescribed by the statute for the primary crime").

When compared to other States' gang legislation, however, the Nevada gang
sentencing enhancement statute differs in at least three ways. First, unlike
statutes in many other jurisdictions (see ILJ 1994), the Nevada statute covers
all felonies. Thus, it is more inclusive than antigang laws enacted in Georgia,
Louisiana, Iowa, and several other States. California's Street Terrorism
Enforcement and Prevention Act, for example, targets only seven violent and
drug-related offenses for gang prosecution. Second, the additional penalties
provided by this Nevada statute are mandatory, with a limited exception for
rendering substantial assistance to law enforcement for apprehending
codefendants. Judges in Nevada retain far less discretion in imposing sentence
enhancements than judges in other States (e.g., Georgia and Louisiana). Third,
Nevada's statute is distinguished from other States' statutes by its severity. The
gang sentencing enhancement statute doubles the term of imprisonment to be
served by a gang member, and that additional term must run consecutively. By
contrast, in California, which admits to having the "worst gang problem in the
country" (Reiner 1992), the additional penalty for gang-related offenses is a
maximum of 3 years. In Louisiana, the enhanced penalty is at most half of the
maximum term for the primary offense (ILJ 1994). These three differences
make Nevada's statute particularly attractive for prosecutors in charging and
sentencing practices.

Gang Prosecution Units in Clark and Washoe Counties

The gang prosecution units in Clark (Las Vegas) and Washoe (Reno) Counties
vary greatly in structure and scope. These differences in the organizational
structure and community context provide an ideal basis for comparing the
effectiveness of different types of gang prosecution units.

The Clark County District Attorney's Office established a gang prosecution unit
in the fall of 1991 in response to escalating gang violence and gang activity.
Funded by the county at an annual budget of $340,000, the unit consists of
three full-time deputy district attorneys who deal with all gang-related cases in
Las Vegas. The unit screened more than 1,000 felony complaints involving gang
members in 1994 and has prosecuted more than 600 gang-related felony cases.
Because of strong collaboration with Federal law enforcement authorities on
drug cases (through the Southern Nevada Gang Task Force), the unit
prosecutes fewer drug-related crimes involving gang members than do most
other jurisdictions (see ILJ 1994). It instead deals mostly with gang violence
cases (e.g., attempted murders, robbery, and weapon possession) and
relatively fewer property crimes and drug offenses.

As is true of special units in other jurisdictions, the ultimate goals of the Clark
County Gang Prosecution Unit are to reduce the level of gang violence in the
community and to enhance communication among law enforcement agencies
(Federal, State, county, and school police), prosecutors' offices,
community-based organizations, probation departments, schools, community
leaders, and family members of gangs and potential gang members (see
Lucherini 1991). This prosecution unit attempts to achieve these goals by--

o Developing a computerized gang offender-based tracking system to monitor
gang activity.

o Improving operational effectiveness of gang prosecution through vertical
prosecution.

o Using a "team approach" to improve interagency coordination of gang
intelligence and facilitate multijurisdictional investigations.

o Monitoring parole and probation for gang members to facilitate revocations
for offenders who continue to participate in gang-related activities.

o Developing a gang hotline through the Las Vegas Metropolitan Police
Department to increase gang intelligence and contact with victims and witnesses
of gang activity.

o Creating a victim/witness protection program that offers physical security
against potential threats and violence by gang members.

Before the Clark County Gang Prosecution Unit was created, gang cases were
processed by regular prosecutors in different units within the District Attorney's
Office (e.g., "major crime" unit, juvenile, repeat offender, and economic
violators).

The gang prosecution unit uses both a proactive and reactive approach to
increase gang intelligence and investigation. It works closely with the Special
Enforcement Detail of the Las Vegas Police Department to compile information
on currently active gang members and associates in the community. The gang
prosecution staff meet regularly with this police unit to discuss use of particular
informants, constitutional issues, and victim/witness support resources.

Washoe County had no separate gang prosecution unit until the Dangerous
Youth Offender (DYO) specialized prosecution team was created and became
fully operational in July 1993. The DYO unit is not exclusively a gang unit, as it
deals with all youth (gang and nongang) who are considered a threat to the
community because of the seriousness of their crimes, their history of repeat
offenses, or their use of a weapon to commit their crimes. However, the unit
rapidly became a specialized gang unit after the number of gang-related
offenses in Reno rose from 77 to 193 in the last 6 months of 1994. Deputy
district attorneys assigned to the unit also work on nongang and adult cases.

The dual goals of the DYO unit are to rigorously prosecute the minority of gang
members who commit serious crimes and provide community alternatives for
at-risk youths who are just marginally involved in gang activity. It has endorsed
a variety of rigorous prosecutorial practices, including withholding plea
bargaining to reduced charges, opposing pretrial release for DYOs, seeking
maximum sentences, and publicizing convictions and sentences of gang
members in schools and neighborhoods where other gang members congregate.
Its commitment to providing alternatives to violence is evident in the strong
working relationship it has created with the Northern Nevada Youth Gang Task
Force, the Reno police force Community Action Team (CAT), and the
Washoe County School District to change the social conditions that have
promoted gang activity in the first place.

Differences and Similarities

The DYO unit resembles and differs from the Clark County Gang Prosecution
Unit in several ways. Both units use vertical prosecution, apply a "team
approach" that highlights multiagency collaboration, are relatively small, and
strongly endorse rigorous prosecution of violent and habitual youth offenders.
Key differences, however, include the following:

o Gang problems are far more pervasive in Clark County than in Washoe
County.

o The caseload for gang prosecutors is greater in Clark County.

o Clark County prosecutors have greater experience with gang cases and
specialize in gang cases, whereas DYO staff work with all types of youth
offenses.

o Gang cases processed in Clark County are generally more serious (involving
a higher concentration of violent crimes) than in Washoe County.

o Washoe County's DYO unit uses a more community-oriented response to
gangs (based on Reno's longstanding tradition of innovative programs for
community and problem-oriented policing), in contrast to Clark County's more
legalistic, law-and-order approach.

By comparing initial charging and sentencing practices before and after the gang
prosecution unit of each county was created, it was possible to assess the
impact of changes in organizational and community conditions on the successful
prosecution of gang cases.

Comparison of these two counties with other jurisdictions in the United States
(see ILJ 1994) makes it evident that the nature of gang problems and the
characteristics of the gang prosecution units in Clark and Washoe Counties
make them ideal case studies for national extrapolations. Clark County in the
early 1990s resembled the average U.S. county with a population greater than
250,000 in the following ways:

o The number of gang-related homicides and other violent crimes prosecuted.

o Use of vertical prosecution (the approach used by 30 percent of large
counties).

o Establishment of a separate gang prosecution unit (as with 30 percent of large
counties).

o Issuance of juvenile waivers to adult courts in some gang cases (as done by
the majority of jurisdictions).

o Use of alternative statutes (such as conspiracy and racketeering laws) in
prosecuting gang cases.

The DYO unit in Washoe County also employs many of these prosecutorial
practices, but Reno is more representative of cities of its size (60,000 to
100,000) in its emerging gang problem. Thus, the analysis of initial charging,
adjudication, and sentencing decisions in gang-related cases in Clark and
Washoe Counties has direct implications for improving understanding of how
best to prosecute these cases in other jurisdictions.

Evaluation Results

Although gang prosecution units and antigang legislation have emerged
nationwide in response to growing gang activity, the effectiveness of these
responses in reducing rates of case dismissal, promoting victim/witness
cooperation in gang cases, increasing conviction rates, and enhancing
imprisonment rates has not been systematically evaluated (see ILJ 1994;
Dahmann 1983). Much of what is currently known about gang prosecution and
gang activity is based on studies in the largest U.S. cities, Los Angeles and
Chicago. Unfortunately, such studies may have little relevance to increasing
understanding of gang prosecution in smaller jurisdictions with less serious gang
problems and fewer resources to combat these problems. This study was
designed to evaluate the effectiveness of antigang legislation and gang
prosecution units in two medium-sized cities in Nevada.

Three types of data were collected and analyzed: (1) arrest reports,
prosecutorial case files, and court records for gang and nongang cases; (2) field
observations of working relationships between police and prosecution gang
units, pretrial conferences, and criminal trials involving gang members; and (3)
interviews with police officers, prosecutors, defense attorneys, and judges to
elicit their opinions of the effectiveness, advantages, and disadvantages of the
antigang statutes and gang prosecution practices in Nevada's urban counties.

The Extent of Gang Crime in Las Vegas and Reno

Widespread claims by police and legislators in local media that gang crime is
epidemic in Las Vegas are inconsistent with court and prosecution data from
1989 to 1995. In only about 6 percent of all charges filed for violent crime was
the defendant a known gang member; this percentage was fairly stable over
time. And contrary to media reports that gangs were controlling drug trafficking,
only about 5 percent of all charges for drug trafficking in Las Vegas were filed
against gang members. However, media and police reports of widespread gang
crime in Las Vegas and, to a lesser extent, Reno, were pivotal in enactment of
the antigang legislation and increases in the organizational resources for the local
police and prosecution. Gang crime was clearly less prevalent in Reno, but
high-profile instances of gang crime and their perceived threat to the community
were also used in this jurisdiction to justify enactment of antigang legislation and
the establishment of gang units. Law enforcement officials' legislative testimony
was the basis for proclaiming the antigang statutes an important crime control
strategy for combating gangs in Nevada.

The Frequency of Use of Antigang Legislation

The most frequent charges in both counties under the antigang statutes (exhibits
1 and 2) are for random and reckless use of firearms. In Clark County,
between 1989 and 1995, more than 1,700 charges were filed for aiming a
firearm at a human being (see exhibit 1). Almost 200 charges were filed for
driveby shootings, the quintessential gang crime, but many of these charges
involved multiple counts against the same defendant in a shooting spree. During
this period, 69 charges for possessing a dangerous weapon on school property
were filed in this county. The gang sentencing enhancement statute, which
mandates a prison sentence and doubles the length of confinement, was
charged 263 times in Clark County between 1991 and 1995. The statutes
imposing additional penalties for specific offenses have not been applied in
either county. Charges under the antigang statutes were far less common in
Washoe County. The gang sentencing enhancement statute, for example, was
charged only 24 times in this jurisdiction from 1991 (the year it was enacted)
through 1995 (see exhibit 2).

Although charges under specific antigang statutes are common in some cases,
such as aiming a firearm, convictions for these charges are the exception. In
Clark County, only about 10 percent of the charges for aiming a firearm at
another person resulted in a conviction--similar to the conviction rates for
driveby shootings. Conviction rates were about 9 percent for weapon offenses
on school property and slightly higher (14 percent) for the gang sentencing
enhancement statutes. In Washoe County, convictions for charges under the
gang sentencing enhancement and driveby shooting statutes occurred in 17
percent and 4 percent of the cases, respectively.

Antigang legislation may be an important tool for prosecutors in plea bargaining
negotiations. By requiring a prison sentence and a doubling of the prison term,
the threat of conviction under the gang sentencing enhancement statute can be a
powerful enticement for a guilty plea. Interviews with judges and deputy district
attorneys provided anecdotal evidence of the importance of the gang sentencing
enhancement statute in plea bargaining. Discussions in more informal meetings
with prosecutors and defense attorneys further suggested that the threat of a
gang sentencing enhancement conviction is often brought up in a subtle, less
direct manner in these negotiations. However, no direct empirical evidence was
found to support the assertion that the gang sentencing enhancement statute
provided leverage during plea bargaining. No significant differences in
conviction rates were discovered when gang cases processed before and after
the statute's enactment were compared.

The Impact of Antigang Legislation on Other Felony Charges

Many States use existing criminal statutes to address gang crime rather than
enact separate gang legislation. These include accessory, aiding and abetting,
racketeering, habitual offender, harassment, and witness intimidation statutes.
Even in States with antigang legislation, prosecutors' familiarity with existing
criminal codes may predispose them to proceed with normal charging practices,
avoiding the complexities and difficulties of some antigang statutes.

Analysis of court monitoring data before and after passage of the Nevada
antigang statutes reveals no significant change in charging practices for other
felony charges in gang cases. Charges for being an accessory, aiding and
abetting, racketeering, harassment, witness intimidation, and habitual offending
are rare in both Clark and Washoe Counties and have neither increased nor
decreased since passage of antigang legislation. Thus, this legislation had no
appreciable impact on other charges leveled against gang members in Nevada.

The Effectiveness of Gang Prosecution Units

As discussed, Clark and Washoe Counties each established gang prosecution
units in response to reports of increasing gang activity. The Clark County Gang
Prosecution Unit began operations in the fall of 1991 with three prosecutors
and small caseloads. Washoe County's DYO unit was established in 1993, with
five full-time attorneys, one investigator, and a secretary to manage caseloads.
Vertical prosecution, whereby the same attorney follows a case through
successive stages of criminal processing, was the approach for prosecution
within each county's gang units.

To evaluate the effectiveness of gang prosecution units, conviction and
sentencing practices for defendants processed in these units were compared
with practice for defendants charged with similar offenses but processed in
other track units. These "matched" samples were drawn in each county. Exhibit
3 summarizes differences in conviction and sentencing practices.

Although the number of convictions is similar across units, the likelihood of a
conviction is higher in the gang unit than in track units in Clark County. Track
units in Washoe County, however, have a higher conviction rate than does the
DYO unit. For those convicted of any charge, a higher proportion of
defendants in the gang units in both counties was given a prison sentence than
were convicted offenders in the track units. Longer prison sentences were also
given to defendants processed in the gang units.

Logistic regression analyses[2] were performed to determine whether
differences between gang units and track units in conviction and sentencing
decisions remained after controlling for other factors (e.g., prior record, age,
and number of charges). These analyses revealed no statistically significant
difference between gang units and track units in the likelihood of conviction and
imprisonment upon conviction. This means that the observed differences in
conviction and imprisonment practices between gang units and track units are
explained by differences in offender and case attributes across these units. Gang
prosecution units do not enhance the success of criminal processing beyond that
provided by other nongang prosecution units. A comparison of conviction rates
for gang members before and after the implementation of gang prosecution units
in Clark County also reveals no significant differences (59 percent versus 61
percent).

Changes in Gang Prosecution Units Over Time

External forces often influence the nature of criminal justice decisionmaking. In
the current context, new district attorneys were elected in both Washoe and
Clark Counties, starting their terms in 1995. In Washoe County, replacements
and changes in the deputy district attorneys were the primary changes, whereas
the new district attorney in Clark County substantially revised prosecution
tracks. The former gang prosecution unit in Clark County was abolished in
1995 and merged with the Major Violators Unit. Under this revised structure,
several of the most seasoned criminal attorneys are available to prosecute gang
crimes, and one of the key prosecutors in the former gang unit was assigned to
the Major Violators Unit.

Comparisons of court monitoring data before and after the changes in each
gang unit revealed few differences. Charges for gang sentencing enhancement
and driveby shootings had decreased in Clark and Washoe Counties in 1995,
although conviction rates for gang sentencing enhancement were higher in both
jurisdictions. These findings suggest that the gang sentencing enhancement
statute is used less often and more effectively, resulting in higher conviction rates
under the new administrations. Conviction rates for other charges involving gang
members have remained fairly stable over time.

Attitudes of Criminal Justice Officials

Surveys were distributed to criminal justice officials in Clark and Washoe
Counties to elicit their opinions and experiences with gangs, gang prosecution,
and the antigang legislation. Respondents included 12 police officers in gang
divisions, 19 deputy district attorneys, 28 public defenders, and 12 criminal
court judges. Each group's attitudes about particular aspects of gang
prosecution are summarized in exhibits 4 and 5.

When asked to evaluate the impact of Nevada's antigang legislation, most
respondents favorably rated some aspects of these laws. The highest praise
came from police gang officers and the lowest from public defenders. The
majority of judges believed that the legislation was an effective tool, but
prosecutors had mixed feelings. Most prosecutors believed that the gang
sentencing enhancement statute was effective in addressing gang crime, but few
believed that enhancements linked to school-related crimes, firearm forfeitures,
or the use of a minor in criminal acts were reducing gang crime.

Except for public defenders, the vast majority of those surveyed viewed
securing cooperation from victims and witnesses as a "major problem" in gang
cases (see exhibit 5). The clear majority of gang and track district attorneys
also reported victim/witness credibility as a major problem. Most prosecutors,
judges, and police in the sample believed that the proof requirement for
demonstrating that an offense was committed to further a criminal gang was at
least a moderate problem in prosecuting gang cases.

Conclusions and Implications

By all indications, gang activity in large metropolitan areas has become a major
social problem. Driveby shootings, carjackings, drug trafficking, and other
predatory offenses by active gang members and their associates are common
occurrences in urban areas. Although social programs have been implemented
to reduce youths' risks and enhance their resistance to the pressures of gang
participation, the primary response of the criminal justice community has been
to increase the severity and certainty of punishment for gang-related offending.
In Nevada and its major cities, Las Vegas and Reno, the twin goals of severe
and certain punishment are promoted through enactment of antigang statutes in
the criminal code and creation of specialized gang prosecution units to increase
the conviction rate for gang-related offenses.

Even though national surveys of the prevalence of gang prosecution units have
been conducted and the largest U.S. cities often provide annual summaries and
internal reports about the operation of their gang units (see Genelin 1994; ILJ
1994; Reiner 1992), scant attention has been paid to process evaluations of the
onset of antigang legislation and impact evaluations of the effectiveness of
antigang statutes and gang prosecution units. In fact, no comprehensive
research has been conducted on the social, political, and economic obstacles to
enactment of antigang legislation and how competing interests influence the
structure and scope of the final statutes.[3] The frequency with which these
statutes are used in criminal court decisionmaking and the effectiveness of gang
prosecution units in smaller jurisdictions also have not been addressed in
previous research. This being the case, the results of the current study constitute
an empirical benchmark that fills a major void in the understanding of the
prosecution of gang cases.

The results indicate that some antigang statutes are widely used (especially
those that address aiming firearms) whereas others, such as provisions against
driveby shootings and gang sentencing enhancement, are less commonly
employed. Once adjustments are made for different offender and case
attributes, gang prosecution units in Clark and Washoe Counties yield
conviction and incarceration rates comparable to those for defendants
processed in other prosecution tracks. The results of this study have several
implications for other jurisdictions that have or are considering establishing
antigang legislation and gang prosecution units.

First, antigang legislation in Nevada was enacted with little social and political
opposition. The passage of such legislation in other States would probably
follow a similar path. Media coverage of gang crime and specific instances of
brutal and random attacks on citizens provide a strong background for the
mobilization of antigang control measures.

Second, criminal justice officials have used claims of escalating gang crime to
increase their organizational resources. This was especially true of the police
and prosecution in Clark County. Calls for increased resources were answered,
even though there was no evidence from court data in this jurisdiction that gang
crime had increased over time. If the experiences of Nevada are representative
of other States, local media coverage of gang crime will play a major role in the
development and support of antigang legislation and gang prosecution units in
other jurisdictions.

Third, antigang legislation provides district attorneys with additional leverage in
prosecuting gang members. Although most charges under the antigang statutes
in Nevada do not result in a conviction under the specific statute, the threat of
conviction under the gang enhancement statute may serve as a major enticement
for a guilty plea on other charges. Most jurisdictions across the country have
not used antigang legislation, relying instead on existing statutes for criminal
prosecution. However, other States may benefit from the implementation and
selective use of a gang sentencing enhancement statute similar to Nevada's,
which doubles the penalty for gang crime, because it mandates a largely
nondiscretionary prison sentence for these offenders.

Finally, specialized gang prosecution units are an important arena for processing
and adjudicating gang cases. Whether other jurisdictions should establish such
specialized units, however, depends on the gravity of the gang problem and
specific expertise of the district attorneys. Rather than establishing a separate
unit, designating one deputy district attorney as the "gang prosecutor" may be
sufficient for smaller jurisdictions in addressing gang crime, especially if that
person has had hands-on training in prosecuting gang members. Although
external changes in the structure and composition of gang prosecution units may
pose a serious threat to their effectiveness, this problem was minimized in
Nevada by making a special effort to retain the most qualified gang prosecutors
in spite of this changing political environment.

Notes

1. Under the Nevada Revised Statutes (193.168, Section 6), a "criminal
gang" is defined as any combination of persons, organized formally or
informally, so constructed that the organization will continue its operation even if
individual members enter or leave the organization, which (a) has a common
name or identifying symbol; (b) has particular conduct, status, and customs
indicative of it; and (c) has as one of its common activities engaging in criminal
activity punishable as a felony, other than the conduct that constitutes the
primary offense.

2. Logistic regression is a statistical technique used to assess the relative
importance of selected variables and whether the impact of a particular variable
remains significant after considering its relationship with other variables in the
model. A type of multiple regression, logistic regression is used primarily when
the outcome variable (like conviction or acquittal) has only two values.

3. The political and economic nature of the antigang legislation in Nevada is
addressed at length in the final report for this study (NIJ grant 94-IJ-
CX-0053).

References

Blumstein, Alfred, Jacqueline Cohen, Susan E. Martin, and Michael Tonry, eds.
1983. Research on Sentencing: The Search for Reform. Washington, DC:
National Academy Press.

Conly, Catherine H. 1993. Street Gangs: Current Knowledge and Strategies.
Issues and Practices. Washington, DC: U.S. Department of Justice. National
Institute of Justice. NCJ 143290.

Curry, G. David et al. 1994. Update on Gang Crime and Law Enforcement
Recordkeeping: Report of the 1994 NIJ Extended National Assessment of
Law Enforcement Anti-Gang Information Resources: Final Report,
Washington, DC: U.S. Department of Justice, National Institute of Justice. NCJ
161476.

Dahmann, Judith S. 1983. Prosecutorial Response to Violent Gang Criminality:
An Evaluation of Operation Hardcore. Washington, DC: U.S. Department of
Justice. National Institute of Justice. NCJ 91553.

Daley, Richard M. 1985. Gang Prevention Unit Report. Chicago: Cook
County State Attorney's Office.

Genelin, Michael. 1994. "Inter-Office Memorandum on the Operation of
Hardcore Gang Division." Los Angeles, CA. Los Angeles County District
Attorney's Office.

Institute of Law and Justice. 1994. "Gang Prosecution in the United States."
Draft of Final Report Submitted to the National Institute of Justice. Washington,
DC.

Klein, Malcolm W., Margaret A. Gordon, and Cheryl L. Maxson. 1987.
"Police Response to Street Gang Violence: Improving the Investigative
Process." Final Report Submitted to the National Institute of Justice.
Washington, DC.

Las Vegas Review-Journal. 1991. "Judge Drops Racketeering Indictments."
March 29, 1991.

Lucherini, Robert. 1991. "District Attorney Gang Unit Proposal." Proposal
Submitted to Clark County Board of County Commissioners. September 12,
1991.

Martin, Susan. 1983. "The Politics of Sentencing Reform: Sentencing
Guidelines in Pennsylvania and Minnesota." In Research on Sentencing: The
Search for Reform, ed. A. Blumstein, et al. Washington, DC: National
Academy Press: 265-304.

Maxson, Cheryl L., Malcolm W. Klein, and Margaret A. Gordon. 1987.
"Gangs in Smaller Cities." Paper presented at the Annual Meetings of the
American Society of Criminology, Montreal, Canada.

McKinney, Kay A. 1988. Juvenile Gangs: Crime and Drug Trafficking. Juvenile
Justice Bulletin. Washington, DC: U.S. Department of Justice. Office of
Juvenile Justice and Delinquency Prevention. September 1998.

Miethe, Terance D., and Charles A. Moore. 1987. Evaluation of Minnesota's
Felony Sentencing Guidelines. Washington, DC: U.S. Department of Justice.
National Institute of Justice. NCJ 107743.

-------. 1988. "Officials' Reactions to Sentencing Guidelines." Journal of
Research in Crime and Delinquency 25 (May): 170-187.

Office of Juvenile Justice and Delinquency Prevention. 1993. Arrests of Youth.
Washington, DC: U.S. Department of Justice. Office of Juvenile Justice and
Delinquency Prevention.

Reiner, Ira. 1992. Gangs, Crime and Violence in Los Angeles: Findings and
Proposals from the District Attorney's Office. Fairfax, VA: National Youth
Gang Information Center.

Spergel, Irving A. 1990. "Youth Gangs: Continuity and Change." In Crime and
Justice: An Annual Review of Research, Vol. 12, ed. Norval Morris and
Michael Tonry. Chicago: University of Chicago Press: 171-275.

-------. 1992. Youth Gangs: Problems and Response. Washington, DC: U.S.
Department of Justice. Office of Juvenile Justice and Delinquency Prevention.
NCJ 145313.

Zatz, Marjorie S. 1985. "Los Cholos: Legal Processing of Chicano Gang
Members." Social Problems 33(1): 13-30.

---------------------------

Chapter 7

Evaluation of a Task Force Approach to Gangs

Susan Pennell
Director, Criminal Justice Research Division
San Diego Association of Governments

Roni Melton
Research Analyst
Criminal Justice Research Division
San Diego Association of Governments

Acknowledgments

This project was supported by National Institute of Justice grant
91-DD-CX-0046 and by the Bureau of Justice Assistance.

---------------------------

Abstract

Jurisdictions United for Drug Gang Enforcement (JUDGE) is a
multijurisdictional task force that has operated in San Diego County since 1988.
Headed by the District Attorney's office, it also includes the police, probation
officers, and deputy district attorneys working together to target known gang
members involved in drug use and sales. The early focus was enforcing the
conditions of probation and drug laws, using undercover tactics, intensive
supervision, and vertical prosecution. This evaluation examined the first 2 years
of program operation, comparing the period before and after implementation.
The process evaluation concluded that JUDGE was implemented as designed
and had selected appropriate targets and methods for ensuring offender
accountability. The evaluation of JUDGE's impact concluded that, although
more than 80 percent of the targets were rearrested, the number of arrestees
declined and most arrests were for violations of probation conditions. The
proportion of arrests involving drugs also declined substantially. JUDGE cost
much more than traditional supervision, chiefly because of the expenses of
vertical prosecution and keeping arrestees in custody. The absence of a control
group, which necessitated the pre-post-study method, proved to be a major
limitation of the evaluation. Another limitation was that the relatively harsh
treatment of JUDGE youths may have been a function of age: As they grew
older and their contacts with the police increased, arrestees were more harshly
treated than when they were younger.

---------------------------

This chapter describes an evaluation of Jurisdictions Unified for Drug Gang
Enforcement (JUDGE), a gang task force program established in San Diego
County in 1988 with the support of Bureau of Justice Assistance (BJA) Byrne
program funds. An earlier report of this evaluation presented findings from the
first 2 years of the program, described the JUDGE program as it existed in
1995, and offered recommendations for future consideration (Pennell, Melton,
and Hoctor 1996). This chapter also includes information about the present
(2000) state of the program, its evolution over time, and the lessons learned
while conducting a retrospective evaluation.

Impetus for JUDGE in 1988

The upsurge of gangs and associated drug and criminal activity in the
mid-1980s, coupled with reductions in available resources, has led an
increasing number of States and local jurisdictions to develop multiagency task
forces. The 1987 grant application for Byrne funding to create the JUDGE
program identified more than 27 street gangs with more than 2,300 members,
most of whom were Hispanic and black. That early grant application further
stated that "the current situation of gang related narcotic control has created a
wave of violence involving several drive-by shootings and homicides. Street
gangs have begun to resemble modern organized crime operations in terms of
sophistication and tactics." According to the second-year grant application, a
task force was needed because "[t]he overall burden on the probation
department (of increased gang and drug activity) has resulted in many juvenile
gang and drug offenders going back on the streets and engaging in narcotics
activity without any real fear of supervision or accountability. It is difficult at
best for these probation officers to follow-up the conditions of probation on a
consistent basis" (Second Year Grant Application, San Diego County District
Attorney's Office).

Sponsored by the BJA through the Anti-Drug Abuse Act as a Byrne-funded
task force, JUDGE has been in operation since 1988. Initially, the program staff
included three deputy district attorneys, two senior probation officers, six police
officers representing three agencies, one sergeant, and one investigative
specialist. Early task force efforts, mandated by the State Office of Criminal
Justice Planning (the Byrne grant program administrator), focused on
coordinating the efforts of three agencies to strictly enforce probation conditions
and drug laws for juvenile gang members with a drug history. Two populations
were targeted: juvenile street gang members on probation for narcotics offenses
and street gang members not yet on probation but known to be involved in the
use, sale, and distribution of narcotics. In the second year, the focus expanded
to include selected adult gang members involved in drug dealing.

The JUDGE program adopted the criteria developed by the California
Department of Justice to define gangs and gang members. To be considered a
gang, a group had to meet all of the following criteria:

o Have a name and identifiable leadership.

o Claim a territory, turf, neighborhood, or criminal enterprise.

o Associate on a continuous or regular basis.

o Engage in delinquent or criminal behavior.

To be documented as a gang member, an individual had to meet at least one of
the following criteria:

o Admit gang membership.

o Have tattoos or wear or possess clothing or paraphernalia associated with a
specific gang.

o Be observed participating in delinquent or criminal activity with known gang
members.

o Be known to the police as having a close association with known gang
members.

o Be identified by a reliable informant as a gang member.

Program Description

The JUDGE program has three primary components: vertical prosecution by
the District Attorney's Office,[1] the Probation Department's narcotics task
force, and special enforcement operations carried out by the police. Vertical
prosecution had the following objectives:

o Reduce the average project caseload for prosecutors, compared with
nonvertical felony prosecution units.

o When appropriate, resist the release of defendants from custody before the
trial or hearing.

o Ensure that the most severe sentence possible is imposed on a convicted or
adjudicated defendant.

o Reduce the average amount of time between the arrest and disposition of
charges against a project defendant.

o Eliminate or reduce the use of plea bargaining in project cases.

o Increase the conviction and adjudication rate for project defendants.

The Probation Department's Narcotics Task Force had the following
objectives:

o Supervise a caseload of 20 to 30 targeted probationers.

o Coordinate with law enforcement through requests for assistance with Fourth
Amendment waiver searches[2] and meetings with task force members to
identify probation violators and develop arrest strategies for drug offenders and
probation violators.

o Assist local law enforcement in identifying and apprehending project
probation violators.

Special enforcement operations had the following objectives:

o Increase the number of warrants, indictments, and arrests of individuals
charged with drug offenses.

o Increase the number of case referrals and cases filed on individuals charged
with drug offenses.

o Increase the use of informants in drug investigations.

o Train law enforcement personnel on topics related to enhanced operation of
anti-drug-abuse projects.

As will be shown later in this chapter, the evaluation of the JUDGE program
could not determine whether some of these objectives had been met. No
specific or measurable performance targets had been set, such as increasing by
a certain percentage. The lack of baseline data for some indicators also made it
difficult to accurately assess the program's effectiveness.

Before examining the results of the evaluation, however, it is appropriate to
briefly describe the local and national context in which JUDGE was
implemented.

The Context

One year before the JUDGE program was established in 1988, the San Diego
Police Department (SDPD), the largest of San Diego County's 10 police
agencies, created a specialized gang detail to focus on identifying and
investigating known gang members and affiliates. The department began
identifying crimes as gang related (if either the victim or the suspect were
documented gang members), and its statistics showed that gang members were
involved in significant numbers of assaults and homicides. One year after the
JUDGE program was established, the San Diego Probation Department
developed a specialized team, the Gang Suppression Unit (GSU), to increase
supervision of gang members who were on probation. Like the JUDGE team,
the GSU team (which supervised most of the JUDGE targets) was located in
the Probation Department. Also, in the mid-1980s, the District Attorney
created a Gang Prosecution Unit to specialize in the vertical prosecution of
high-risk gang members. Each of these special units incorporated features
common to such units, including extensive training for investigators about gang
culture, undercover teams that conducted surveillance of gang members, greater
use of informants, and exclusive focus on a particular type of offender.

Taking a multiagency approach to a crime problem was not a new idea in San
Diego County. The Countywide Narcotics Task Force, representing all local
police agencies, as well as State and Federal agents, had been in operation
since 1955. In other areas of the country, multiagency task forces evolved,
primarily in response to organized crime and then, in the mid-1980s, in
response to the drug problem. The focus and purpose, the nature of the
cooperation, and the types of agencies involved have all changed significantly
over time (Chaiken et al. 1990). Studies have compared these task forces with
traditional narcotics investigative units, investigated several multijurisdictional
task forces in a single State, and examined task forces operating in conjunction
with other law enforcement entities (Ruboy and Coldren 1992). Most of this
research has been descriptive, with arrests, seizures, and convictions the
primary measures of task force activities (Justice Research and Statistics
Association 1993). A more recent study (Senese and Levinson 1994) involved
surveys of probation and parole agencies that focused on gangs. The authors
noted that strong programs usually--

o Have a lower than average client-to-agent ratio.

o Have a specialized gang unit.

o Monitor gang members' behavior.

o Prohibit gang members on probation from associating with other gang
members.

o Emphasize a law enforcement, control approach, including intensive
supervision.

o Ensure that all criminal justice entities cooperate and share information.

Earlier studies on task forces focused primarily on describing these activities
and did not include a rigorous methodology for measuring their success. The
JUDGE program incorporated all of these features, as well as others noted later
in this chapter that are purported to contribute to the success of task-force
efforts.

The JUDGE Program Evaluation

The researchers' original intent was to conduct both a process evaluation (to
determine whether the project had been implemented as proposed) and an
impact evaluation (to determine the effects of increasing accountability on
reoffending) of the first 2 years of the JUDGE program (1988 and 1989). This
evaluation did not take place as planned, however, and took longer than
expected to complete. The final report, which came out in 1996, was a more
qualitative, descriptive study than originally proposed and included an update
on JUDGE activities in 1995.

The evaluation, which was funded by the National Institute of Justice (NIJ),
began in 1991 and spanned a 3-year period. Since the evaluation began after
program implementation, a true experimental design could not be initiated.
Instead, researchers and program staff decided to use a quasi-experimental
design, retrospectively selecting a control group of youths identified in SDPD
and probation department files as gang members, to determine how the system
responded to this group of gang members before the JUDGE program began
and to compare their recidivism rate with that of youths targeted by the
program. Unfortunately, however, this control group could not be identified for
the following reasons:

o Criteria for defining gang members were not clearly defined before 1988.

o Some arrestees who seemed to fit the profile were never placed on
probation.

o Some arrestees did not have a documented history of using or dealing drugs.

o Most important, many of the youth identified for the control group eventually
became JUDGE targets.

It was therefore necessary to modify the original research design to include all
279 youths targeted by the JUDGE program in 1988 and 1989. Data were
collected on these youths for the 2 years before the program was created and
the 2 years after they had become targets. Thus, what was to be a rigorous
impact evaluation became a descriptive, one-group, pre-post study of youths
processed by the JUDGE program.

Data were compiled for every juvenile targeted by JUDGE during the first 2
years of the grant. Sources for data included arrest reports, probation reports,
criminal history records, and court files. Generally, all source documents were
incorporated in each individual's JUDGE file. Data elements included
sociodemographic information, age at first arrest and true finding (complaint),
types and levels of offenses, time in custody, drug test results, date of intake
and release from JUDGE, sentences imposed, and recidivism up to 2 years
after release. Additional data elements included gang affiliation, criminal history,
offenses that resulted in probation, probation conditions, contacts by probation
and JUDGE staff, and offenses committed during probation. For the process
evaluation, outcome data were compared to project objectives to measure
compliance (in terms of program implementation) and results (such as probation
violations and drug test results). The consequences of increased probation
supervision, including revocation, custody time, and use of vertical prosecution,
were evaluated by comparing how cases were processed before and after
JUDGE targeting.

In addition, to gauge the process and progress of JUDGE implementation,
interviews were conducted with criminal justice personnel in the JUDGE
program and the agencies that coordinated with JUDGE staff. Participants were
asked to identify successful program elements, barriers to coordination,
availability of resources, and training issues.

The analysis included a description of program implementation and operations
and a comparison of how these youths were handled by the criminal justice
system before and after JUDGE implementation. Exhibit 1 identifies the data
elements and the sources for the data.

Research Objectives and Results

The study was designed to achieve the following:

o Determine if the JUDGE program objectives were met during the grant
period.

o Assess the impact of such program activities as surveillance, special
enforcement, and vertical prosecution on the expected outcomes: increased
probation violations and arrests; more defendants held for pretrial custody; and
more probation revocations, convictions, and sentences.

o Evaluate the program's impact on offenders as measured by recidivism rates,
the need for probation intervention, and gang affiliation.

o Provide recommendations for similar programs in other jurisdictions.

Unfortunately, several research objectives could not be addressed
systematically. The lack of a control group put the conclusions regarding the
program's effectiveness into question. Documentation for several JUDGE
activities was either unavailable or inadequate, including numbers relating to
informants. One reason for the lack of sufficient data was the turnover of
program staff responsible for documentation. In addition, in the second year of
the evaluation, the investigative focus of JUDGE expanded to include adult
offenders who were not necessarily gang members or drug offenders. Research
was limited to the first 2 years of the JUDGE program and included only
youthful offenders targeted in 1988 and 1989.

Another limitation of the study is that the increase in law enforcement
supervision over time could simply be the result of the natural maturation of a
youthful population, rather than a direct effect of the JUDGE program. The
juvenile justice system historically has treated younger juveniles less harshly with
respect to formal processing. In San Diego, earlier studies have shown this to
be the case (Pennell and Curtis 1983, 1988). Juveniles tend to have several
contacts with police and probation before more formal action takes place,
generally when they continue delinquent activity as they age. Therefore, if the
JUDGE youths were in fact treated more severely after becoming JUDGE
targets (and they were), this may be a result of the system reacting overall to
several contacts rather than the result of specific actions taken by a special task
force.

On a process level, the results suggested that the JUDGE staff targeted the
intended offenders by using screening criteria to define gang members and drug
involvement, intervened at the appropriate phases in system processing, and
enforced probation conditions.

Offender Characteristics

Juveniles were identified as JUDGE targets if they met all of the following
screening criteria:

o Youths with a drug history (arrests for possession, sale, or trafficking or other
indication in the file).

o Wards of the court.

o Gang members.

More than half (56 percent) of the 279 JUDGE targets were 16 to 17 years old
when targeted. Targets were nearly equally divided between blacks (45
percent) and Hispanics (52 percent) (see exhibit 2). According to the SDPD's
gang unit, these figures correspond to the ethnic breakdown of gangs in San
Diego in 1988 and 1989.

All program targets were documented gang members (see exhibit 3). The vast
majority of them were involved with drugs: 73 percent had been convicted of a
drug offense, and 23 percent had an arrest for a drug violation. Only 3 percent
had no indication of drug use or sales in their files. The most frequent charge of
first referral to probation was a felony-level property offense. Slightly more than
half of the youths (55 percent) were targeted during the prosecution stage and
almost a quarter (23 percent) were targeted when on probation. At the time
they were targeted by the JUDGE unit, 89 percent of the juveniles were wards
of the court. Given the composition of the targeted group, youthful gang
members involved with drugs with a criminal history, the JUDGE program
followed the guidelines outlined in its grant proposal.

Police Objectives

Police responsibilities in the JUDGE program included special enforcement
(including undercover efforts), identification of gang members, and surveillance,
as well as maintaining profiles of suspects in the regional computer.

All youths targeted by JUDGE in 1988 and 1989 were documented gang
members. Before targeting, 52 percent of them had been placed in custody at
time of arrest (see exhibit 4). After targeting, the proportion rose to 68 percent,
consistent with JUDGE objectives. Two years before targeting, half of the
arrests for those subsequently to become JUDGE targets were for a felony and
7 percent were for violating probation. Two years after targeting, only 30
percent of the arrests were on felony charges, while 32 percent were for
probation violations, suggesting increased surveillance and more intensive
supervision. Before the youths were JUDGE targets, 61 percent had a petition
(complaint) requested, compared to 81 percent for the 2 years after being
targeted.

Prosecution Efforts

To determine whether prosecution under the JUDGE program had achieved its
objectives, proportionate filing rates, conviction rates, plea bargains, cases with
vertical prosecution, custody prior to disposition, sentence, and average time
from arrest to disposition were compared for cases involving youths before and
after targeting. These data were evaluated by case, not by individual; the results
are shown in exhibit 5.

Before JUDGE targeting, the 279 juveniles had a total of 1,098 arrests. In the 2
years following, the 279 youths had been arrested 841 times (a 23-percent
drop), perhaps related to time spent in custody and fewer opportunities to
reoffend. After targeting, 80 percent of the arrests resulted in petitions filed with
the court, compared with 61 percent of arrests filed before targeting, suggesting
a more severe approach (see exhibit 5). In addition, 49 percent of the cases
during the targeting period were prosecuted vertically, as opposed to 15
percent of the cases before JUDGE was implemented. More than two-thirds
(68 percent) of cases involving JUDGE targets led to conviction, compared
with less than half (44 percent) before the program was implemented. Cases
processed by JUDGE were more likely to result in sentences involving custody.
About a third (33 percent) of the cases in the JUDGE period were sentenced
to local custody, compared with 16 percent of the pre-JUDGE cases. About 7
percent of the JUDGE cases resulted in sentences to the California Youth
Authority (CYA), whereas only 1 percent of the cases processed before
JUDGE began resulted in this sentence. Time from arrest to disposition was
shorter after JUDGE was implemented; 61 percent of the JUDGE cases were
processed within 30 days, whereas only 47 percent of the pre-JUDGE cases
were prosecuted that quickly.

These differences suggest that prosecution under the JUDGE program was
indeed more severe than in the cases handled before the JUDGE unit was
created and that prosecution goals were met. Generally, however, juveniles are
processed more harshly as they penetrate further into the justice system, and
this may have also contributed to these differences.

Custody Decisions

The JUDGE program sought to hold youths accountable for their behavior.
One way to achieve this objective is to initiate graduated sanctions, such as
placing people in custody at varying levels of processing, when responding to
probation violations. A review of cases 2 years before and after JUDGE
intervention revealed that youths were more likely to have been placed in
custody after becoming JUDGE targets for similar sorts of offenses. As exhibit
6 shows, more than two-thirds (68 percent) of the post-JUDGE cases resulted
in custody upon arrest, compared with 52 percent before the program was
implemented. This pattern was consistent for several types of hearings. At
disposition, 63 percent of the post-JUDGE cases involved placement in
custody, in contrast with 47 percent of those processed before JUDGE. These
proportionate increases following implementation of the JUDGE program may
be the result of vertical prosecution, which is associated with a higher level of
preparedness by the deputy district attorney, since the attorney follows the case
from initial investigation to disposition.

Technical Violations

To keep offenders accountable, fairly strict probation conditions are ordered to
monitor behavior more closely. Intensive supervision generally involves
increased contact with defendants, which, in turn, increases the opportunities to
observe unacceptable behavior that includes violating probation conditions. The
data suggest that this was the case with the JUDGE targets. Before JUDGE,
according to the first-year grant application, probation resources were limited
and insufficient to respond to the increase in gang-related activity. By teaming
with law enforcement, JUDGE probation officers were able to enhance their
surveillance and supervision of identified youths.

When youths in San Diego County become delinquent wards of the juvenile
court, the probation officer prepares a social study that describes the activities
they have engaged in (drugs, gangs, truancy, etc.). The court has a "laundry list"
of behavior or rules to which the youths may be mandated to adhere, such as
attending school, obeying school rules and the rules of parents and guardians,
not associating with gang members, or going to drug treatment. From the
information compiled in the social study, including prior delinquent history, the
probation officer makes recommendations to the court regarding which
behavior or conditions should be mandated by the court. Before the JUDGE
program was implemented, the probation department did not have the
resources or time to devote to identifying appropriate conditions and following
up with supervision. The following exhibits suggest that after being targeted by
JUDGE, a higher proportion of youths had specific conditions ordered and
were subsequently found to be in violation of these conditions.

As exhibit 7 illustrates, before becoming JUDGE targets, 12 to 39 percent of
the study group had probation conditions ordered by the court that included no
contact with illegal narcotics, following probation officers' rules, not associating
with specific individuals, obeying school rules, submitting to Fourth Amendment
waiver searches, observing curfew, not carrying or using weapons, and
submitting to drug testing. After becoming JUDGE targets, 55 to 78 percent of
the juveniles had such conditions ordered by the court. While a JUDGE target,
60 percent or more were found to be in violation of such probation conditions
as not following rules, associating with specific individuals, possessing narcotics,
and violating curfew (see exhibit 8). Overall, nearly two-thirds of the JUDGE
targets (66 percent) were arrested for probation violations after JUDGE
intervention, compared with 17 percent before becoming a target (see exhibit
9).

Consistent with the features of intensive supervision, the individuals targeted by
JUDGE and later supervised by the Gang Suppression Unit were frequently
reported by their probation officers for violations (see exhibit 8). It is difficult to
state with confidence whether they were reported because they were serious
offenders and got into trouble frequently or because the increased supervision
made it more likely that they would get caught. (Data are available for the
3-year period of JUDGE tracking of the study group with respect to conditions
ordered and violated.) More than 50 percent of the JUDGE targets were
reported for violating the following conditions:

o Following probation officers' rules (64 percent).

o Obeying school rules (53 percent).

o Not associating with certain others (60 percent).

o No use or sales of illegal narcotics (99 percent).

o Curfew (69 percent).

Total Time in Custody

Offenders who are in custody have few or no opportunities to reoffend. The
data show that JUDGE targets spent a considerable amount of time behind bars
during their JUDGE tenure. More than 40 percent were in juvenile hall or some
other local institution for a total of a year or more, although this was not
necessarily consecutive time. Another 25 percent had 6 months to a year of
custody time. Unfortunately, no records had been kept of their custody time
before being targeted, so it is not certain if the time in custody was actually
greater after targeting. Other indicators, however, such as differences in the
number of technical violations, suggest that defendants probably spent more
time in custody after JUDGE, since they were reported more often for violating
probation conditions than before being targeted.

Recidivism

A typical measure of success in the criminal justice system is the recidivism rate,
the frequency with which offenders return to the system via arrest. The measure
has its limitations and is generally bolstered with other measures, such as
reintegration into mainstream society through employment, school attendance,
or reduction in drug use. Also, it is recognized that intensive supervision leads to
more arrests due to increased surveillance. Unfortunately, since social
integration information was not available for this study, analysis of recidivism
was limited to considering the number and nature of arrests 2 years before and
after JUDGE intervention.

Most of the 279 JUDGE targets (83 percent) were rearrested within 2 years
after JUDGE intervention, but the mean number of arrests dropped slightly (3.9
to 3.0), perhaps because offenders were in custody part of the time (exhibit 9).
Arrests for felony drug violations dropped to 28 percent of the total, compared
to 63 percent before JUDGE was implemented. The proportion of arrests for
violent felonies rose slightly, from 22 to 27 percent. As expected, most of the
arrests after JUDGE targeting (66 percent versus 17 percent) were for
probation violations. An arrest for a probation violation occurs when it
becomes apparent that a youth is not complying with the conditions of his or her
probation.

Analysis of the JUDGE program suggests that it was implemented as designed
with respect to appropriate targets and methods for ensuring offender
accountability such as complaints filed, custody time, vertical prosecution,
conviction, and sentencing. Outcome measures suggest that although more than
80 percent of the JUDGE targets were rearrested, the number of arrests
dropped, and most of these were for probation violations. Moreover, the
proportion that involved drug violations declined substantially.

When researchers reexamined the JUDGE targets in 1995, 64 percent of the
279 had had new court cases filed, with an average of 3 cases per individual.
These results suggested that many JUDGE targets remained involved in criminal
activity and that the task force had appropriately focused on isolating a small
segment of offenders who appear particularly crime-prone, monitoring their
behavior closely, and applying sanctions swiftly and with certainty.

This finding is difficult to interpret because not enough is known about this
group of offenders to determine whether they are particularly crime-prone or
the extent to which an enforcement approach actually contributes to reoffending
because it does not address other issues such as drug treatment, education, and
employment. The current juvenile justice strategy in San Diego County employs
both intervention and enforcement, based on the tenets of the comprehensive
juvenile justice strategy developed by the Office of Juvenile Justice and
Delinquency Prevention, which incorporates risk factors along with offender
accountability (Howell 1995). As discussed below, however, the JUDGE unit
remains in operation.

After the Evaluation

Upon revisiting the JUDGE program in 1995, researchers learned that its
mission had shifted direction periodically. After the 1994 three strikes
legislation, for example, the JUDGE unit began targeting "second strikers" who
could potentially commit a third offense that would result in prison time. The FY
1999-2000 grant application represents the 11th year of BJA Byrne funding.
The JUDGE program's current focus is on adult criminal offenders who are on
probation or parole and identified as habitual drug offenders (whether or not
they are gang-affiliated), drug-involved documented juvenile gang members,
and discretionary targets. The program still involves cooperation and teamwork
between law enforcement, probation, and the District Attorney. Six of the 10
law enforcement agencies in San Diego County are represented, as well as the
California State Parole Board, the California State Bureau of Narcotics
Enforcement, and the San Diego County Department of Social Services.
Because of money constraints and local agency policy, the other four law
enforcement agencies have elected not to participate in JUDGE.

Probation and parole officers conduct searches with police to identify
probationers and parolees involved in gangs or drugs. Targeted offenders
continue to be prosecuted vertically, and prosecutors accompany police
officers in the field for observation and training and to serve as an immediate
resource. Although the current emphasis is on adult offenders, juveniles are
frequently targeted and arrested as part of ongoing investigations. According to
its coordinator, JUDGE continues to report a conviction rate of more than 90
percent, with most of the juveniles committed to California Youth Authority and
most adults sent to prison.

The FY 1999-2000 JUDGE grant application included each of the 12 critical
elements that, according to a recent study (BJA 2000), have helped several
multijurisdictional task forces achieve their programmatic and organizational
objectives:

1. Written interagency agreement

2. Prosecutor involvement

3. Computerized information/intelligence databases and systems

4. Case planning and selection

5. Communication

6. Coordination

7. Budget planning

8. Dedicated and well-defined mission

9. Monitoring and evaluation

10. Staffing and recruitment policy

11. Effective asset seizure and forfeiture activities

12. Technical assistance and training.

In addition, the JUDGE program incorporates several of the practices
documented in a 1998 study on "what works" to prevent crime (Sherman et al.
1998), although some of these practices--including monitoring of gang
members,
intensive supervision, and units that focus on repeat offenders--have had mixed
results. The measure of success used by JUDGE today is "having a high
proportion of investigations that result in criminal prosecutions and maximum
incarceration time for targeted offenders." The effectiveness is determined by
the project coordinator, who reports to the San Diego Police Chiefs' and
Sheriff's Association.

Although both countywide FBI index crimes and gang crimes within the city of
San Diego have declined over the past few years (Allnutt and Pennell 2000;
San Diego Police Department Gang Unit 2000), it is difficult, if not impossible,
to determine how much of the decline can be attributed to the JUDGE unit's
efforts, particularly in light of all the other efforts taking place. These include the
U.S. Attorney's Violent Crime Task Force, which focuses on gang violence and
narcotics activity; the Probation Department's Gang Suppression Unit, which is
still in operation; the SDPD's Gang Unit and Narcotics Division; the District
Attorney's Office's Gang Prosecution Unit; and the North County Gang Task
Force. The ultimate reason for JUDGE's continued longevity may have little to
do with research-based documentation of effectiveness and more to do with the
interests of the local criminal justice administrators in continuing an approach
that allows the targets to change over time. As stated in a recent Bureau of
Justice Assistance report, if task forces are to be successfully sustained, they
must be able to integrate themselves into the existing criminal justice system and
adapt to changes in their environment (Bureau of Justice Assistance 2000).
Apparently, the JUDGE program has succeeded in this endeavor.

Lessons Learned

The experience of this study illustrates the problems in conducting a
retrospective study of a program that shifted focus during the evaluation period.
A number of issues were raised, including the availability, reliability, and validity
of the data. Foremost in this regard was the issue of selecting a comparable
group of subjects for a control group. Another factor that complicated the
impact assessment was that probationers could be targeted at any time in the
process, including before arrest. In addition, some process-level objectives,
such as increasing surveillance hours and the use of informants, could not be
assessed because the data were not maintained, and the existence of multiple
data sources in numerous locations hampered the efficiency of the data
collection effort.

The JUDGE assessment underscores the need for researchers and practitioners
to begin at the program's outset by linking the program goals and activities with
expected measures and outcomes and developing the indicators that will
demonstrate the program's effectiveness. It also highlights the reality that
programs are often implemented and maintained based on the interests of
policymakers and administrators. The task force concept has great appeal to
policymakers, as does the notion of accountability for offenders; both are
assumed to be efficient and effective.

Notes

1. In vertical prosecution, the same prosecutor handles a case all the way
from preliminary hearing to sentencing.

2. Convicted defendants can waive their Fourth Amendment protections
against unreasonable search and seizure as a condition of probation.

References

Allnut, Donna, and Susan Pennell. 2000. Crime in the San Diego Region:
Mid-Year Report. San Diego, CA: San Diego Association of Governments.

Bureau of Justice Assistance. 2000. Creating a New Criminal Justice System
for the 21st Century: Findings and Results From State and Local Program
Evaluations. Effective Programs Monograph No. 2. Washington, DC: U.S.
Department of Justice, Bureau of Justice Assistance, NCJ 178936.

Chaiken, Jan, Marcia Chaiken, and Clifford Karchmer. 1990.
Multijurisdictional Drug Law Enforcement Strategies: Reducing Supply and
Demand. Washington, DC: U.S. Department of Justice, National Institute of
Justice, NCJ 126658.

Howell, James C., ed. 1995. Guide for Implementing the Comprehensive
Strategy for Serious, Violent, and Chronic Juvenile Offenders. Washington,
DC: U.S. Department of Justice, Office of Juvenile Justice and Delinquency
Prevention, NCJ 153681.

Institute for Law and Justice. 1992. "Gang Prosecution Prosecutor Survey."
Draft prepared for the U.S. Department of Justice, National Institute of Justice,
Washington, DC.

Justice Research and Statistics Association. 1993. Multijurisdictional Drug
Control Task Forces: A Five Year Review, 1988-1993. Washington, DC:
U.S. Department of Justice, Bureau of Justice Assistance, NCJ 146395.

Pennell, Susan, and Christine Curtis. 1983. The Serious Juvenile Offender. San
Diego, CA: San Diego Association of Governments.

-------. 1988. Juvenile Violence and Gang-Related Crime. San Diego, CA:
San Diego Association of Governments.

Pennell, Susan, Roni Melton, and Darlanne Hoctor. 1996. Assessment of a
Multi-Agency Approach to Drug Involved Gang Members. Program
Evaluation. Washington, DC: U.S. Department of Justice, National Institute of
Justice, NCJ 185252.

Ruboy, M.A., and J.R. Coldren, Jr. 1992. Law Enforcement Task Force
Evaluation Projects: Results and Findings in the States. Washington, DC: U.S.
Department of Justice, Bureau of Justice Assistance, NCJ 136697.

San Diego Police Department Gang Unit. 2000. Personal communication.

Senese, Jeffrey D., and Robert B. Levinson. 1994. "A National Assessment of
Gangs on Probation and Parole: Control and Prevention of Gangs in
Community Corrections." Paper presented at American Society of Criminology
Annual Meeting, Miami, FL.

Sherman, Lawrence W., Denise Gottfredson, Doris L. MacKenzie, John Eck,
Peter Reuter, and Shawn D. Bushway. 1998. Preventing Crime: What Works,
What Doesn't, What's Promising. Research in Brief. Washington, DC: U.S.
Department of Justice, National Institute of Justice, July, NCJ 171676.

---------------------------

Chapter 8

Gang Prevention Programs for Female Adolescents: An Evaluation

Katherine Williams, Ph.D.
Development Services Group, Inc.

G. David Curry, Ph.D.
Department of Criminology and Criminal Justice, Department of Sociology
University of Missouri-St. Louis

Marcia I. Cohen
Vice President, Research and Evaluation
Development Services Group, Inc.

Acknowledgments

This research was supported by the National Institute of Justice, U.S.
Department of Justice, under grant 93-IJ-CX-0051; and by the Administration
for Children, Youth and Families, U.S. Department of Health and Human
Services.

---------------------------

Abstract

Until the most recent wave of gang violence, research treated gang crime as
almost exclusively a male phenomenon. Evidence indicates the number of
female gang members is rising, and the unique features of female gang
involvement reinforce the importance of prevention and intervention for females
at risk. Gang involvement has been shown to have more long-term effects on
females and more severe effects on their children, and crime victimization is
greater among female gang members than among their male counterparts. This
evaluation of youth gang and drug prevention and intervention for females
focused on programs in Boston; Pueblo, Colorado; and Seattle. All three
programs successfully provided a core mix of services, but the process
evaluation revealed some inadequacies in monitoring, recordkeeping, and staff
training. Staff turnover was high, and some difficulties were encountered in
involving parents and recruiting and retaining participants. All three programs
experienced problems with local evaluations. Because the impact evaluation
was limited in Seattle and Boston by data collection problems, only the findings
from Pueblo were reported. Pueblo was also the most successful in
implementation, effectively reducing five of the seven types of delinquency
measured. School performance was also measured. On entering the program,
participants tended to earn lower school grades than nonparticipants and to
have dropped out. After program completion, participants still earned lower
grades, but the gap was no longer statistically significant. Yet the self-esteem of
participants was no higher than that of nonparticipants as a result of the
program. One key to Pueblo's success may have been the comprehensiveness
of its services, which included project components that addressed key factors
correlated with female gang involvement.

---------------------------

Youth and gang violence increased substantially in the United States from 1985
until the mid-1990s. With the rising statistics, national concern over
gang-related crime increased (Huff 1990; National Institute of Justice 1992;
Spergel 1990). In earlier periods of rising gang activity, research had treated
gang-related crime as a solely male phenomenon. With the 1985-95 wave of
gang violence and public response, however, research and program practice
began to focus for the first time on the role of women in gangs (Campbell 1984;
Chesney-Lind and Hagedorn 1999; Miller 2001; Miller and Decker 2001;
Moore 1991).

Estimates of the level of female involvement in gangs vary with the sources of
available data. Surveys of law enforcement agencies have consistently reported
the contribution of female gang members to officially recorded crime to be
comparatively low. More recent information on law enforcement assessment of
gang problems has been provided by the annual surveys of the National Youth
Gang Center (NYGC) (1999). The 1998 NYGC survey of law enforcement
agencies (NYGC 2000) estimated that there were about 62,420 female gang
members nationwide (8 percent of the 780,233 estimated gang members).
These results suggest that law enforcement reports of female gang members
increased dramatically after 1991, when the estimated figure was 9,092 (Curry,
Ball, and Fox 1994). Surveys of at-risk youth in school populations have
produced larger estimates of female involvement in gangs (Bjerregaard and
Smith 1993; Esbensen and Huizinga 1993). The most recent and geographically
comprehensive survey of youth by Esbensen and Winfree (1998) estimated that
38 percent of gang members were female. Field studies of female gang
involvement may offer an explanation for the differences between law
enforcement and survey estimates (Chesney-Lind, Sheldon, and Joe 1996;
Fleisher 1998; Hagedorn and Devitt 1999; Moore 1991). Although female
gang members tend to engage in criminal behavior more often than males who
are not gang members (Esbensen, Deschenes, and Winfree 1999; Fagan
1990), they self-report involvement in less serious delinquency than male gang
members. There is also evidence that females cease participating in gangs at an
earlier age than males.

Though female gang involvement may contribute less to gang-related crime than
does male involvement, it has unique features that make it worth studying.
Moore (1991) emphasized that gang involvement has more long-term effects on
the lives of female gang members and a more severe impact on the lives of their
children (and, perhaps, consequently for the community and society) than it
does on the lives of their male counterparts. Miller (2001; also see her chapter
in this volume) observed that female gang members experience higher levels of
victimization as a result of their gang participation. Fleisher (1998) suggested
that gang-affiliated females contribute to the social processes through which
male members maintain group identity and cohesion. The complexity of female
participation in gang activity and the potential for longer term negative
consequences reinforce the importance of gang prevention and intervention
programs targeted to females at risk for gang involvement.

Federal Government sensitivity to the issue of female gang involvement was
demonstrated by the funding of seven female gang prevention programs by the
Family and Youth Services Bureau (FYSB) of the Administration for Children,
Youth, and Families (ACYF) of the U.S. Department of Health and Human
Services in 1990. Four more female gang prevention programs were funded by
the same agency in 1992 as part of the Youth Gang Drug Prevention Program,
but FYSB's focus on gang involvement as a program concern was phased out
in the late 1990s. By 1995, the program's gang prevention and intervention
components were discontinued. Today, few Federal demonstration programs
address this population (Curry 1999b).

This report presents the findings from an evaluation of three federally sponsored
youth gang and drug prevention and intervention projects designed specifically
for females. The evaluation represents a joint commitment of the National
Institute of Justice (NIJ) and FYSB, each of which contributed funds to support
the effort. NIJ managed the evaluation. The study was conducted by
Development Services Group, Inc. (DSG), in Bethesda, Maryland, and G.
David Curry of the University of Missouri-St. Louis. Dr. Curry was responsible
for collecting and analyzing the outcome data for the study, and DSG staff were
responsible for overall project management, process evaluation activities, and
production of the final report.

Study Sites

The three evaluated projects were chosen from four sites operating programs
for adolescent females funded by FYSB in 1992 as a part of the Youth Gang
Drug Prevention Program described above. Program selection was based on
geographic location, ethnicity of service population, differences in program
focus, and the availability within each program of adolescent females never
involved in gang activity and adolescent females either currently or previously
involved in gang activity. Under FYSB guidelines, projects could serve both
girls at risk of gang involvement and those who were already involved in gangs.
(The fourth site, in Washington, D.C., was not selected because it did not serve
gang-involved girls.) The projects had been funded for 3 years, beginning in
1992. Federal funding for each project was $150,000 per year, including small
(less than $15,000) budgets for local evaluation activities that were separate
from this evaluation. Each site also contributed in-kind resources. The projects
are described briefly below.

FORCE (Females Obtaining Resources and Cultural Enrichment): The Next
Level

One of several youth programs operated by the Community Initiatives Division
of the Boston Housing Authority (BHA) in Boston, Massachusetts, FORCE:
The Next Level was a continuation of a project for girls living in public housing
developments. The project had been established by Judge Baker's Children
Center in cooperation with BHA. The original FORCE project was funded by
FYSB in 1990 for 2 years. At the end of the second year, BHA applied for
additional funding from FYSB to enrich the program by adding specialists in
family support, leadership development, personal growth, and recreation and
by starting support groups for girls in the program. FORCE: The Next Level
was designed to expand the earlier program by including older girls, involving
participants' families, and providing a range of gender- and culture-specific
services matched to the developmental needs of urban, low-income females.
The program offered recreational, personal growth, leadership, and support
group services and was designed to serve 400 girls and their families in 6 BHA
housing developments. The project primarily served black females, with the
exception of one public housing development that had a majority of white
residents.

Movimiento Ascendencia (Upward Movement)

This program was operated by the Pueblo Youth Services Bureau (PYSB) in
Pueblo, Colorado, and served a primarily Mexican-American population. The
project was one of a continuum of service programs provided by PYSB to
youths in Pueblo and the surrounding county. It was established to provide
young females with positive alternatives to substance abuse and gang
involvement. (These programs included drug prevention programs, a youth
center for runaways and the homeless, community-based services for youth
involved with the juvenile justice system, a diversion program for first-time
felony offenders, a restitution and community services program, and a
transitional living program.) Movimiento Ascendencia was designed to serve
240 girls at risk of gang involvement and 120 gang-involved females and their
families. Project activities were designed around three main components:
mediation or conflict resolution, self-esteem or social support, and cultural
awareness.

The Adolescent Female Gang Prevention and Intervention Project

An initiative of the Seattle Team for Youth in Seattle, Washington, this program
was designed to address the needs of adjudicated and preadjudicated teenage
females of color and to prevent or reduce their participation in local gangs.
Project activities included substance abuse education and intervention services,
mentoring, self-esteem building, positive ethnic and cultural identification, social
skills training, and services addressing teen pregnancy, housing, parenting, and
other issues related to being a teen parent. The Seattle Department of Housing
and Human Services provided services via subcontracts with the Atlantic Street
Center to provide case management and substance abuse awareness and
education and, with Sisters in Common, to provide a 12-week support
group-based curriculum in self-esteem and culturally relevant activities. The
program was designed to provide 60 females with intensive case management
services and 175 females with education and gang prevention programs.

All three programs served populations primarily composed of girls who are
members of minority groups. Given the target populations of gang-involved and
at-risk populations, this is not surprising. Before the 1960s, gangs were most
often represented as primarily composed of young, white males and formed
along ethnic lines. Thrasher, studying gangs in the 1920s (1963 [1927]: 130),
estimated that only 7 percent of Chicago gangs consisted entirely of black
members. Mexican-American gangs did not receive attention until the 1940s
(Moore 1978). Since the 1970s, most field research, particularly research on
females (Campbell 1984; Horowitz 1983; Joe and Chesney-Lind 1995; Moore
1991), has focused on minority populations. While survey research on juvenile
populations has generally shown a greater proportion of white youths
self-reporting gang membership (Esbensen and Huizinga 1993; Esbensen and
Winfree 1998), law enforcement data since the 1970s have consistently shown
overrepresentation of minorities among police-identified gang members. For
1996, the NYGC (1999) survey reported gang membership in cities with
populations greater than 25,000 to be 44 percent Hispanic, 40 percent black,
and 10 percent white. It has been argued (Curry 1999a) that law enforcement
records and survey results show different ethnic composition of gangs because
older and more criminally involved gang members are coming to the attention of
law enforcement. It may have been these girls' greater difficulty in leaving gangs
and greater involvement with criminal justice agencies that led the agencies
applying for ACYF funding for female gang prevention and intervention to
direct their services to minority youths.

Evaluation Method

The research design included both process and impact evaluation components.
The study had the following objectives:

o Process evaluation objectives:

-- Describe the organization and implementation of three youth gang prevention
and intervention projects designed specifically for black women and Latinas.

-- Describe the services and activities of these prevention and intervention
projects and the females who participated in them.

-- Describe the implementation of the local evaluations.

o Impact evaluation objectives:

-- Describe background characteristics, family interactions, peer relationships,
school involvement, delinquent activities, and gang and drug involvement for
program participants and nonparticipants.

-- Examine each program's impact on participants.

The process evaluation addressed the design, implementation, operation, and
community context of the three projects. Data collection was designed to
provide information on a wide range of program-related variables from several
perspectives. For each project, data sources included proposals and quarterly
reports sent to FYSB, project-developed material, interviews with project staff
and knowledgeable community leaders, and direct observation of project
activities during site visits.

Each site was visited twice during the grant period by two senior DSG research
staff members. The first series of visits, which took place between the 20th and
22nd months of project operation, focused on startup activities, community
context, and activities. Because all three projects experienced significant startup
problems, it was not necessary to conduct site visits earlier. The second series
of visits took place between months 30 and 34 of project operation, and
focused more intensively on intervention and prevention activities and on
following up on issues raised during the first site visit.

The outcome part of the study was an interview-based survey of six groups of
adolescent females: gang-involved, formerly gang-involved, and
non-gang-involved program participants and nonparticipants. The study was
designed to add to the existing literature on female gang participation by
describing female participation in gang activity, exploring the extent to which
program participants differed from nonparticipants on the key variables
associated with the risk for gang involvement, and assessing the extent to which
program participation resulted in positive outcomes.

Sample

Researchers hoped to recruit 30 gang-involved participants, 30
non-gang-involved participants, 30 gang-involved nonparticipants, and 30
non-gang-involved nonparticipants from each site, for a total of 360 subjects
from the 3 sites. It was intended that program participants would be randomly
selected from program rosters developed by program staff at each site. These
initial plans had to be adjusted, however, and the result was three different
sample selection procedures.

The Pueblo program files were available for sampling, and it was possible to
select a random sample of participants. Nonparticipants were identified using
snowball samples[1] developed through school and juvenile justice contacts.

In Boston, problems were encountered in identifying program participants who
could be interviewed and finding female gang members, either in or out of the
program, to interview. FORCE records on individual girls lacked uniformity and
order, so two separate sets of sampling frames had to be developed, one
compiled from case records maintained by the specialists and the other
constructed from the sign-in sheets for program activities. Even with the
assistance of program staff, it was difficult to find the girls identified through the
sampling procedures. Most of the Boston sample is therefore a convenience
sample of girls involved in the FORCE program who were made available to
interviewers when girls randomly selected from program records could not be
found.

In Seattle, only girls receiving services from the Atlantic Street Center were
available for interview. The nonparticipants in Seattle came from three very
different convenience samples: a probation caseload of serious offenders, a
caseload of homeless adolescent females, and members of a church-affiliated
program for adolescent females at risk.

The final distribution of interviews by program and site is shown in exhibit 1.
The Pueblo site came the closest to achieving the numbers presented in the
original design. In Seattle, the Atlantic Street Center simply had very few girls in
its program. Program girls who were enrolled only in the drug education or
Sisters in Common programs were not included in the outcome study sample.
In Boston, few self-reported current or former gang members were in the
program.

Interviews

The interview, adapted from instruments used in three federally funded studies
of gang involvement (Cohen et al. 1994; Curry and Spergel 1992; Decker
1996), assessed gang involvement, delinquency, and program participation and
impact. Pilot testing of an early version indicated male biases in the instrument.
Consequently, it was revised to be more useful with a female population. In
addition to questions on their entry into and activity in gangs, items were added
on children and mothering, family attitudes, and female attitudes toward gangs.
The instrument was also revised to reflect the activities and attitudes of young
women who were not gang members by adding questions that asked such
women why they had not joined a gang, how they felt about gangs, and what
pressure they received from gangs. Hypothetical situations assessing their
attitudes about gang involvement were presented. The interview covered
delinquency, gang involvement, family patterns, academic performance,
employability and job history, self-esteem, and substance abuse.

Data collection took place at the three sites during the final 12 months of
scheduled program operation. At each site, local interviewers were responsible
for setting up and conducting the interviews. For all except the program
participants in Seattle, interviews were tape recorded.

Because of the differences among sampling techniques at each location and the
difficulties in completing the planned number of interviews in Boston and
Seattle, outcome data from each site were analyzed separately. Because the
outcome interviews for the Pueblo program represent the strongest research
design, only the impact findings from Pueblo are presented in this report.
Findings from all three sites are in the final report (Williams, Curry, and Cohen
forthcoming).

Extent of Local Gang Problems and Female Involvement

Boston

Police and project personnel reported that the Boston area began to have
serious youth gang problems beginning in 1987 and 1988. In an NIJ-funded
study of law enforcement antigang resources, Curry et al. (1992) identified 70
gangs with 2,200 members in the Boston area. BHA staff reported, however,
that the incidence of gang activity was probably higher, because official reports
included only activities that could be tied directly to gangs. They described most
gangs as locally formed and locally based; although gangs typically did not go to
other public housing developments to cause problems, gang activity had been
reported in both minority and white developments. Moreover, they reported
that the nature and extent of gang activities varied among the BHA
developments. For instance, drug selling appeared to be a specialized activity.

FORCE staff also reported hearing that women were becoming more involved
in gangs. Females were likely to support male gang members in criminal
activities in the developments, such as car theft and, especially, drug sales. Staff
believed that girls were being recruited into gang activity at an increasingly
young age. There had been an increase in the number of 14- to 16-year-old
females in the developments who had their own apartments. These young
women were often not in school and were sometimes living with a male partner
who was involved in drug sales and were seen as being at risk of becoming
involved in drug sales as well. FORCE staff also reported that there had been
an "incredible" rise in violence among younger youths.

Pueblo

According to local officials, gangs in the Pueblo area during the early 1990s
were locally organized and had few, if any, formal ties with nationally
recognized gangs. Many of the gangs were associated with the old barrios
(neighborhoods). Gang activities had largely been confined to intergang
violence and a variety of petty crimes. According to a local gang task force
report, there was no evidence that local gangs had developed into the
sophisticated, for-profit organizations that have evolved in some urban areas. A
1994 Pueblo Police Department Crime Analysis and Gang Unit summary
report showed 29 identified gangs and 630 members in their gang roster.
Fifty-seven percent of the individuals listed were Hispanic males, 15 percent
were white males, 10 percent were black males, 14 percent were Hispanic
females, 3 percent were white females, and less than 1 percent were black
females. Two-thirds of the group were between the ages of 16 and 20.
One-quarter were between the ages of 21 and 30. According to the police,
problems caused by gangs included graffiti and intimidation--both against each
other and against strangers. In the early 1990s, most of the intimidation was
taking place in the schools, but this activity decreased after police officers were
assigned to high schools. Violence increased during 1993 and 1994,
culminating in several high-profile driveby shootings, three deaths, and one
critical wounding.

According to Pueblo project staff, the female gangs were "extremely visible" in
the community, hiding weapons and drugs, organizing inhalant parties, and
sheltering "outlaws." Pueblo's female gang members were characterized as
being very much involved in exacerbating violence and crime, drug marketing,
and precipitating violence against cliques of opposing gangs. Their role in the
gang community was well known. More than a dozen "named" female cliques
had been identified at the time program planning began.

Seattle

Seattle's youth gang problem, like that of many cities in the Pacific Northwest,
escalated in the mid-1980s. Gangs from Los Angeles began moving north and
had been active in Seattle since 1987. Although the city had a history of youth
gangs before this migration, there was little gang violence. In the early 1990s,
police identified four major gangs: Crips, Bloods, Black Gangster Disciples,
and Southside Locos. According to the police, there were 125 documented
gang members, with an estimated additional 3,000 to 4,000 youths involved in
gang activity. Police estimated that two-thirds of drug sales in the area involved
gang-involved youths.

Female gang activity in Seattle also rose during this period. In 1990, 16 percent
of the youths referred for services to the Seattle Team for Youth were females.
In 1991, this proportion had increased to 19 percent. The Seattle Police
Department also reported that the young women were forming their own
groups, quasi-gangs, or actual gangs. Some of the girls had formed "auxiliary"
groups on their own. Moreover, they were often physically more aggressive
than male gang members.

Process Evaluation Findings: Summary

Project Implementation

The first objective of the process evaluation was to describe the organization
and implementation of these projects. The research questions for this objective
focused on the organizational structure and staffing patterns of the projects, the
types of staff training, the monitoring and recordkeeping systems used to track
services, and the barriers to implementation.

Organization and staffing. The three projects were organized in very different
ways. In two sites, the lead agency was a unit of local government. In Boston,
the Community Initiatives Department of the Boston Housing Authority was the
lead agency. In Seattle, the project was initiated by the Seattle Team for Youth,
a consortium of school, social service, and community agencies created in 1990
to prevent or intervene in local youth gang participation. The Seattle
Department of Housing and Human Services was the lead agency, and project
activities were provided by two subcontractors. In the third project, the lead
agency was the Pueblo Youth Services Bureau (PYSB), a private, nonprofit,
community-based organization that had been serving Pueblo and the
surrounding county since 1973.

The three projects also differed in their staffing patterns. In Pueblo, the
executive director of PYSB served as the project director and had daily
contact with program staff. A coordinator and three outreach workers were
hired specifically for the project. In Boston, funding provided salaries for
specialists in personal growth, leadership, and family as well as for the
coordinator's position. Additional part-time recreation and support-group
positions were never consistently filled. BHA provided a youth worker for each
project site as an in-kind contribution. In Seattle, the Atlantic Street Center
provided three case managers, each of whom managed the cases of 17 to 20
girls. There was one supervisor for the three case managers. In addition,
approximately 10 women volunteered as mentors for the Sisters in Common
portion of the program.

Staff training. In Pueblo, staff training focused on conflict mediation and
resolution skills, signs and symptoms of drug and alcohol abuse, and providing
information on sexuality, pregnancy, and sexually transmitted diseases. In
Boston, the FORCE staff also identified more basic training needs: stress
management, diversity, youth outreach strategies, and developing goals and
objectives. For the most part, FORCE service delivery staff were
paraprofessionals who typically lacked systematic training in the areas
necessary to carry out many of their job activities. Compared with the staff in
Pueblo, the FORCE staff felt that they had not been provided with adequate
training. This may have been due in part to the differences between the two
projects' organizational structures. In Pueblo, project staffers worked closely
with the project director. This type of supervisory relationship did not exist for
FORCE workers, who were much more isolated.

Monitoring and recordkeeping. For various reasons, monitoring and
recordkeeping presented a challenge for these projects, as they do for many
community-based prevention programs. In Boston, management information
and reporting systems for the FORCE program were the most rudimentary of
the three projects. Sign-in sheets to record name and activity were used by the
youth workers to document the daily activities at each housing development.
The coordinator tallied the sheets by hand each month to provide monthly BHA
and quarterly FYSB progress reports. Collecting data to track program
participants was most difficult in Seattle because the Seattle Team for Youth
used separate contractors, each with its own approach to recordkeeping. The
lead agency did not provide centralized project case tracking. In Pueblo,
program staff kept individual files on all participants once consistent program
operations began and maintained a detailed, accurate log on participant
attendance. Although a personal computer was available to project staff, none
of the case-file or service-delivery information was automated. Moreover, staff
lacked the training to design and maintain electronic tracking systems, and it
was unclear whether the software needed was available.

As a result, accurate, unduplicated counts of youths receiving program services,
hours of service, and length of stay in the program were difficult to obtain.
There was no consistency from site to site in the approach to keeping individual
case records and service plans. Furthermore, it seemed clear that, aside from
providing some numbers to satisfy Federal reporting requirements, none of the
three projects made systematic use of the information they recorded on project
participants. These findings are not unique to these three projects (Cohen et al.
1994) and illustrate the difficulty in conducting research in community-based
demonstration projects. The lack of systematic client and service information
makes it difficult to accurately describe the "service mix" or the extent of service
delivery each participant received, which in turn makes it difficult to discover
which prevention and intervention strategies and services are most effective.

Barriers to implementation. All three projects experienced implementation
problems common to many demonstration programs. These included staff
turnover, lack of transportation for participants, lack of parental participation,
problems in implementing planned program activities, participant recruitment
and retention problems, inadequate physical facilities, and inadequate local
evaluation.

Staff turnover and difficulty in hiring appropriate staff were the most serious
implementation challenges in each project. In Pueblo, the project director said
that the pressure to begin project activities in a timely fashion was such that the
original staff were hired in haste, so she was not able to find the "right" people.
The result was a complete turnover of the service delivery staff by the end of
the first year. The director believed that to be successful, staff needed to be
culturally sensitive and responsive, able to manage stress, have good
documentation skills, and possess a clear understanding of the youths with
whom they would be working and the problems these youths may face.
According to the director, it was also important for staff to be from (or known
in) the communities where they work and be willing to "get down and dirty" to
do whatever the job requires (i.e., staff shouldn't expect an 8-to-5 office job).

Similarly, in Boston, when several workers in the participating housing projects
left the program, the participants often left with them and did not return. In
Seattle, one of the subcontractors operating in the first 2 years of the program
experienced major staff turnover, including the resignation of the director. This
led to inadequate service delivery and cancellation of the subcontract. Staffing
problems led to delays in implementing program activities in all of the projects.
Thus, all of the projects had difficult startups and never fully implemented their
original program designs.

In each project, it was difficult to get parents involved. In Pueblo, although
parent involvement was seen as an important goal, attempts to engage parents
in program activities met with only mixed success. The more successful efforts
included one-to-one contact with parents in their own homes, having parents
transport their own children to and from activities, and scheduling quarterly
family activities. By the end of the second year, quarterly reports indicated that
parental involvement had improved dramatically. Boston's workers encouraged
parents to become involved by organizing talent shows, dinners, and
mother-daughter nights, and Seattle staff invited parents to the Sisters in
Common dinners held after the group meetings.

The poor physical condition of the buildings was an additional problem for two
of the projects. At the time of the first Pueblo site visit, the PYSB headquarters
was located in downtown Pueblo in a building that had significant structural and
operating problems. Similarly, most of Boston's FORCE projects were located
in rundown public housing projects. Some of the facilities could be described as
no less than squalid, others were being renovated, and several others were
poorly maintained and barely adequate.

Transportation for program participants was another obstacle. In Pueblo, the
targeted communities were spread out, program participants came from
different areas, and public transportation was deficient. Consequently, it was
very difficult to get participants to and from the areas where activities took
place. Similar transportation problems were reported in Boston and Seattle,
although they were not as severe because both cities had adequate public
transportation systems. In Boston, many project activities took place in the
housing developments where the girls lived. In addition, both projects gave bus
tokens to participants so they could get to services and activities. Staff from all
three projects reported transporting girls in their own vehicles as well.

Project Services and Recruitment

The second objective of the process evaluation was to describe the services
and activities of the projects and their participants. Research questions focused
on documenting intervention strategies and services, methods used to recruit
and retain youths, and the demographic characteristics of the participants.

Project services. Comparing prevention and intervention services among
projects is often a problem because commonly accepted definitions of services
have not been widely adopted (Cohen et al. 1994). To assess the services
delivered in the projects studied, definitions were developed for 30 services
commonly provided by youth prevention and intervention programs. These
services were grouped into five broad strategies--individual, family, school,
peer,
and community--reflecting common areas of risk for gang and drug
involvement.
(Exhibit 2 shows services delivered at each site.) The types of services offered
were identified by reviewing program records of the three projects, interviewing
staff, and observing services being delivered and then classifying activities
according to the definitions of service developed for this evaluation. Although
the three projects were organized differently, all focused primarily on providing
services to individual youths and (less often) their families. Even the
"community" strategies were focused primarily at the individual level (e.g.,
offering a safe haven, cultural enrichment for participants, and tutoring or help
with homework) rather than on communitywide changes such as education
reform or community mobilization. All three projects were successful in
providing a "core" mix of services, including social and life skills training,
alternative activities (generally recreation), informal counseling, tutoring or
homework support, mentoring and positive role modeling, and cultural
enhancement for project participants.

Social and life skills training were defined as interventions designed to help
youths develop ways to communicate, solve problems, and make decisions; to
find ways to control anger and aggressive impulses (including conflict
resolution); to identify and understand complex feelings and emotions; and to
acquire or refine basic household skills. All three projects were very active in
providing these services. Social and life skills training were the service areas
most likely to contain gender-specific content. Frequently, the topics chosen
were specifically designed to meet the concerns and needs of adolescent
females (e.g., self-esteem, personal grooming and hygiene, pregnancy
prevention, career goals, etc.).

Prevention programs typically included organized sports or recreational
activities in a structured, supervised setting as an alternative to gang and
drug-related activities. These activities, which also provided structure, furnished
positive role models and peers, and helped develop a sense of personal
competence and control, made up the major portion of the programs at all three
study sites. Frequently, they gave participants the opportunity to experience a
range of activities that would not normally be available to them.

Informal counseling, as defined in this evaluation, involved activities organized
by program staff who had not had formal training in counseling and therapy. It
often took place when a "teachable moment" occurred during other program
activities. In some cases, it included "crisis counseling," when a youth had an
immediate problem, and program staff helped her to explore solutions. This
type of counseling was an important staff role in all three projects. As happens
with many community-based prevention programs for youth, much of this
informal counseling took place in connection with other program activities. The
participants' interactions with one another were often used as teaching
opportunities by the project staff.

In prevention programs, teachers, parent volunteers, program staff, members of
the community, or older students often provide tutoring and homework support.
Both the Boston and Pueblo projects offered some of these types of support.
The Pueblo program had the most organized educational support component.
Staff were available twice a week after school, in participating schools, to
provide this service. The schools offered space, supplies, and media equipment.
School staff interviewed for the evaluation saw this as an especially valuable
service. In Boston, this activity was left to the discretion of the individual youth
workers and was not offered at every site. Some youth workers successfully
incorporated these activities into daily routines. Those who did not seemed to
feel that it was difficult to persuade girls to participate or that their meeting
space was not conducive to studying. Seattle also offered tutoring, primarily to
a small group of Southeast Asian youths. Tutors were recruited from the
University of Washington. None of the projects became involved in such school
strategies as policy development, reform, or advocacy.

All three projects provided some type of positive female role models to
participants, though this type of service was more formally organized in Pueblo
and Seattle than in Boston. Pueblo had the best developed formal mentoring
component. Participation was voluntary; girls were matched with a mentor only
if they wanted to be. All mentor candidates had to pass a background check
(including police checks and social service register check for child abuse) and
go through a formal training process before being "matched" with a program
participant. Each mentor sponsored one girl and contracted for a minimum of 2
hours per week for 9 months. Typical activities for mentors and their "matches"
included attending community events and events sponsored by PYSB,
shopping, talking on the phone, eating out, playing sports, and sometimes just
"hanging out" at home. One mentor explained that she was trying to expose her
"match" to new experiences, work-related activities, and alternative lifestyles
that she might otherwise not experience. PYSB also scheduled events for
mentors and "matches" once or twice a month, usually on weekends.

Cultural enhancement components of prevention programs focus on both
increasing youths' awareness of other cultures and reinforcing positive cultural
identity and pride through increasing their knowledge of their own culture's
history, traditions, and values. All of the projects attempted to broaden the
cultural perspectives of participants. In Pueblo, for example, staff organized a
variety of activities, such as attending cultural fairs and listening to speakers talk
about different cultures. Other activities focused on the richness of the Hispanic
cultural heritage. Exposing participants to other cultures was sometimes a
challenge, because of the isolation of the Pueblo community. Whenever
possible, staff would organize trips to Colorado Springs and Denver to provide
these experiences.

The service delivery model implemented in these projects differed from site to
site and within programs that had multiple service delivery sites. In addition, the
projects differed in the degree to which they had a coherent service delivery
program. In Seattle, three subcontractors provided services, and participants
received some, but not all, of the services. In Boston, though the personal
growth and leadership specialists presented their workshops each week in
every site, each youth worker tended to organize different activities, depending
on his or her own interests and those of the girls. Services were offered most
uniformly in Pueblo, with most participants receiving social and life-skills
training. This training was typically part of the afterschool activities, which also
included workshops, guest speakers, lectures, and group discussions. Pueblo
also had the best developed mentoring program.

Recruitment and retention. In Pueblo and Boston, participation was voluntary.
Outreach by youth workers was the primary means of recruitment in both
projects, although referrals were also received from other sources. In contrast,
participants in the Seattle project were required to participate as part of their
conditions of probation.

The national evaluation of the FYSB Youth Gang Drug Prevention Program
(Cohen et al. 1994) identified several barriers to recruiting and retaining youths
in prevention programs like the ones described here. New prevention programs
can take a long time to gain the level of community acceptance necessary to
maintain a consistent level of participation. With only 3 years of funding, lack of
such acceptance can be a significant problem for new projects. Program names
may also affect program acceptance. Prevention program names that include
such words as "gang" or "drugs" may deter parents (and some youths) from
participating because of the potential for stigma associated with these terms.
Transportation was cited as a major problem for recruiting and retaining youth
as was frequent staff turnover and the associated lapses in service delivery,
which can cause youths to drop out of prevention projects.

The three projects reviewed in this evaluation were able to avoid some, but not
all, of these barriers. The Pueblo Youth Services Bureau and the Atlantic Street
Center in Seattle were the best situated in terms of "name recognition" and
acceptance in the local community. Because both organizations had long,
positive connections to the communities they served, they were able to establish
a new program without investing the time usually needed to gain community
acceptance. In contrast, in some of the housing developments, the FORCE
project faced considerable resistance to establishing the program. A great deal
of effort was necessary to increase the project's visibility and community trust.

Both Boston and Pueblo avoided potentially stigmatizing the project
participants by using positive project titles that did not include the words
"gangs" or "drugs." The names Movimiento Ascendencia (Upward Movement)
and FORCE (Females Obtaining Resources and Cultural Enrichment) were
chosen for the positive images they projected, an important factor for some
parents in allowing their daughters to participate. Although the Seattle project
did not use an overall project name, its Sisters in Common component also
projected a positive name and image.

The projects were less successful in overcoming the lack of transportation and
staff turnover. Access to reliable transportation was a problem for all three
projects. This was especially serious in Pueblo because, as noted, the project
covered large distances, and public transportation was inadequate. Staff
turnover was a significant problem in Boston and Pueblo. Some programs
changed services and times of operation to better accommodate participants.
Hiring staff from the community being served proved to be a helpful tactic for
ensuring program participation.

Project participants. The interviews with clients revealed some differences in the
populations served at each site. The average age for participants in the three
program sites was 14. The girls served by FORCE in Boston were significantly
younger than those served in Pueblo, whereas the girls served in Seattle were
significantly older than those in Pueblo. In addition, the groups differed
ethnically; most participants in Boston and Seattle were black, and 95 percent
of the participants in Pueblo were Latina.

Each program was distinctive in the characteristics of its participants because,
as stated earlier, there was no uniform approach to recruitment. In Pueblo and
Boston, participation was voluntary and participants were recruited primarily by
youth workers. In Pueblo, the staff recruitment efforts focused on the wider
Pueblo community. In Boston, recruitment was more narrowly focused on the
girls in the housing developments where the program was located. By contrast,
in Seattle, program participation was mandatory; girls were referred to the
project as a condition of probation. Retention of participants was an issue in
both Boston and Pueblo when the projects experienced staff turnover, but was
less of a problem in Seattle, because of the mandatory nature of participation.

There was a clear difference between participants in Boston and those in the
other two sites in self-reported previous delinquency and gang involvement.
Program participants in Boston were significantly less likely than those in Pueblo
and Seattle to have had contact with the juvenile justice system and to have
been involved in gang activity. In fact, the Boston project had limited contact
with gang members or gang-involved girls. To a great extent, this variation was
deliberate. The FORCE program in Boston was essentially a prevention
program dealing with younger girls, whereas the Seattle project focused on a
small number of older girls with histories of serious and chronic delinquency,
and Pueblo served girls in need of both prevention and intervention services.

Local Evaluation

The third objective of the process evaluation was to describe the
implementation of the local evaluations carried out by each program. Research
questions focused on the approach taken by local evaluators at each project
and how effective the projects were in conducting their local evaluation. The
project applications submitted to FYSB presented evaluation plans with strong
designs that included both process and outcome components and plans for
conducting local evaluations. The Pueblo and Boston projects contracted with
outside evaluators throughout the funding period, whereas Seattle used the
services of the internal evaluator on staff at the Department of Housing and
Human Services.

The original proposal for FORCE outlined an ambitious evaluation plan calling
for both qualitative and quantitative data collection to assess the effectiveness of
program implementation and outcome. The Pueblo project also had the
potential for a strong local evaluation. Its proposal called for both process and
impact data collection and analysis. The impact evaluation was to include pre-
and post-testing of program participants and control group subjects from a
neighboring town. Finally, the Seattle project's plans for the outcome
component of its local evaluation consisted of a self-esteem questionnaire filled
out by staff for each participant before and after program completion, a
self-report questionnaire filled out by each participant at the beginning and end
of each group discussion session, and a decisionmaking questionnaire
completed by each participant at the beginning and end of each session. In
addition, attendance forms and quarterly report forms were to be submitted for
all participants.

Unfortunately, because of bureaucratic red tape, slow startup, and delays in
implementing the evaluation designs, none of the local evaluations produced
useful findings. In Boston, the original evaluator left at the end of the first year,
so no evaluation was completed that year. BHA then contracted out for the
second- and third-year studies, but signed contracts too late in the year for the
evaluator to complete interviews, so data were not systematically collected, and
no reports were produced. Pueblo's evaluators were very slow to develop and
implement data-collection instruments, so completion rates for outcome
instruments were low. Only 144 of 237 participants completed the pretests and
of those, only 28 completed both pre- and post-tests. In Seattle, the evaluator
from the Department of Housing and Human Services designed three surveys.
After the first year, program staff refused to use the surveys, and new survey
instruments were not developed, so only a year 1 report was produced. It was
of limited value because the contractors and the program design changed after
the first year.

Impact Evaluation Findings: Summary

The impact evaluation, which addressed the objectives of describing the
background, activities, and gang and drug involvement for participants and
nonparticipants and examining each program's impact on participants, was
seriously limited by the data collection problems at the Seattle and Boston
programs. As noted above, program and comparison populations were
obtained by different methods at each location. The comparison groups at the
sites were not matched in any way at the time of their selection, and the
evaluators were not in a position to "dictate" how the program and comparison
groups were to be selected by the sites. The number of gang-involved girls in
the Boston and Seattle programs was much smaller than anticipated by the
original impact evaluation design. Of the three projects, Pueblo's was the most
successful in implementing the original program design. For these reasons, the
impact evaluation portion of the report focuses on the findings from the Pueblo
program only. (For additional information on the Boston and Seattle sites, see
Williams, Curry, and Cohen forthcoming.)

Demographic and Social Characteristics of Participants

The evaluation's fourth objective called for a description of the demographic
and social characteristics of the participants and members of the comparison
groups. Research questions focused on describing demographics, history of
delinquency, family patterns and living situations, academic performance, and
self-esteem for both groups. Program participants in Movimiento Ascendencia
were randomly selected. An early count (after about 50 girls had been
interviewed) showed more gang than nongang girls in the program sample. A
chi-square test comparing the percentage of girls that reported being, or having
been, gang members with a 50-50 breakdown showed no statistically
significant difference. Therefore, it is reasonable to report that the program
served equal numbers of gang members and nonmembers. The snowball
sample of nonparticipants relied on schools and juvenile justice system contacts
to recruit a comparison group of young women who had not participated in the
program. The girls in the comparison group who had been gang members were
more likely to describe themselves as "former" gang members than did the girls
in the program (55.2 percent compared with 32.1 percent), but this difference
was not statistically significant.

The program and comparison groups of girls in the Pueblo sample did not differ
statistically on several potentially important variables, including age,
self-reported gang membership, ethnicity, household structure, and school
status (see exhibit 3). There were 64 girls in the two groups who reported
having never been gang members (32 in each), and there were no significant
differences between the two on risk for gang involvement. Gang involvement
short of being a gang member was measured by the girls' meeting any one of six
criteria identified by Curry and Spergel (1992) and Curry (2000). These
included wearing gang colors, "hanging out" with gang members, having gang
members as friends, engaging in drug use or vandalism with gang members,
flashing gang signs, or involvement in gang fighting. Neither group of girls was
statistically more likely to live in communities with gangs, have a gang member
in their family, or date gang members.

The three scales of self-esteem developed by Hare were among the key
variables analyzed in this evaluation (Hare 1977, 1980). These measures of
self-esteem was designed specifically for use with minority youths. It can be
argued that these measures can just as easily be interpreted as indexes of social
attachment or social efficacy in particular types of relationships. To avoid
ongoing arguments in the literature over the "true" definition of self-esteem, the
evaluation treated the Hare scales simply as a measure of how a girl felt about
her relationships with household members, teachers and students at school, and
her peer group. This made the measures important indicators of the girls'
relationships with key institutions and informal groups. The Hare scores were
computed as sums within each relational category, so that a higher score
represented a better feeling about the relationship. When the average scores for
program participants and nonparticipants were compared, no significant
differences were found (exhibit 4).

Interviewers in Pueblo had access to police and juvenile court worker
information on individual girls, so that official records for each girl could be
checked. Program participants were slightly more likely to have been arrested
than the comparison population (46.7 percent compared with 41.0 percent)
and were more likely to have been on juvenile probation than comparison
group members (31.7 percent versus 21.3 percent), but fewer had been
incarcerated (10.1 percent compared with 21.3 percent). This latter finding
may derive from the fact that currently incarcerated girls could be part of the
comparison group while there were no currently incarcerated program
participants (see exhibit 5). Official records also showed that the comparison
sample of girls had been arrested more frequently (3.9 arrests per person,
compared to 2.7). None of these differences, however, was statistically
significant at the .05 level.

Since it has long been assumed that most juvenile offending is not officially
detected, official contact with the juvenile justice system was supplemented with
self-reported data when possible. The interview schedule contained a number
of self-reported items that were translated into "ever used violence," "ever
committed a property offense," "ever used alcohol," and "ever used other
drugs." Program girls were more likely than members of the comparison group
to report having ever committed each kind of offense. The differences,
however, were statistically significant only for property and drug offenses
(exhibit 6).

From these comparisons (which, in the opinion of the evaluators, provide
adequate measures of risk for gang involvement), it is safe to conclude that girls
served by the Pueblo program were as much at risk of gang involvement and
delinquency as a comparable sample of girls from the community served.

Estimating Program Impact

The focus of the final evaluation objective was to understand the impact of the
services provided to the participants. The analyses address the following
research questions: Were the services provided by each program perceived as
helpful by the participants, and were the programs effective in reducing the
delinquency, increasing the educational attainment, and improving the
self-esteem of participants?

Helpfulness of services. Data on services received were extracted from
program files for 60 Pueblo participants. Services were recorded by program
name, type of service, dates, and number of hours. (Exhibit 7 shows Pueblo
services by type, average number of hours per participant, proportion
remembering the service, and perceived helpfulness.) Length of participation in
the program from first to last recorded service ranged from 1.3 months to 20
months, with the average length 8 months and the median 6 months.

Delinquency. It is useful to note limitations in available outcome measures. All
information was obtained directly from the girls. The interview instruments
included questions about a variety of delinquent behaviors. Most were worded
so that time could not be taken into account. Seven items, however, were
specifically worded to compare more recent behavior with past behavior.
Program participants were first asked whether they had engaged in any of
seven delinquent activities since becoming involved in the program. They were
then asked to recall if they had ever engaged in any of these activities before
becoming involved in the program. Nonparticipants were asked whether they
had engaged in any of these activities in the past year and at any previous time.
There are known methodological problems with such retrospective questions. If
the question format has a built-in bias, one would expect it to be in favor of the
program having an effect, especially since, in most cases, the interviewers
approached the program girls through their connection with the program.

The seven types of delinquent behavior the girls were asked about were
throwing objects (such as rocks, snowballs, or bottles) at cars or people;
purposely damaging or destroying someone else's property; running away from
home; knowingly buying, selling, or holding stolen goods (or trying to do any of
these things); stealing or trying to steal something worth less than $50; stealing
or trying to steal something worth more than $50; and carrying a concealed
weapon other than a plain pocketknife. In addition to these measures of
delinquency, information about school grades was obtained from the girls in the
same retrospective-recall format. The Hare measures of self-esteem (discussed
in the previous section) were also compared for both groups.

Pueblo's program did not exclude current or active gang members from
participating. Given the correlation between gang membership and other types
of delinquency, it is important to control for this variable in assessing the impact
of the program and its components on self-reported delinquency. There were
no differences between current and former gang members on whether they had
"ever" engaged in delinquent behavior. Therefore, gang involvement was
defined as "ever having been a gang member." It is also important to control for
previous delinquent behavior. Logistic regression models made it possible to
control for previous delinquency and gang membership in measuring program
impact. (See exhibit 8.)

In each analysis, represented by the rows (exhibit 8), the dichotomous
dependent variable is a type of self-reported delinquency. There are three
dichotomous independent variables in each analysis: self-reported previous
delinquency, ever having been a gang member, and being a program
participant. Setting the analysis up in this way makes it possible to control for
self-reported delinquency before enrollment in the program as well as for gang
involvement in measuring the program's impact on self-reported delinquency
after enrollment.

The regression coefficients for previous reporting of each type of delinquent
behavior were statistically significant. For all seven types of delinquent behavior,
self-reported delinquency was reduced, and for five, the coefficients were
statistically significant. Even when previous delinquency was controlled, gang
membership had a statistically significant impact on four delinquent behaviors.
Controlling for previous delinquency and gang membership, there were
statistically significant program effects in reducing five of the seven types of
delinquency. All program-effect coefficients had negative signs, indicating a
reduction in delinquent behavior associated with program participation.

The seven delinquency items for behavior since becoming a program member
had a Cronbach's alpha[2] of .801, and those for previous delinquent behavior
had a Cronbach's alpha of .793. This made it feasible to treat the seven items
as a measure of an underlying general dimension of delinquency. Girls were
given a score on a scale running from zero (for no postprogram self-reported
delinquency) to seven (if they had reported engaging in all seven types of
delinquency after leaving the program). Then, each girl's preprogram
self-reported delinquency on a one-to-seven scale was subtracted from her
postprogram score. This computation was used as a change score. For
nonparticipants, the average change score was -0.38. For program
participants, the average change score was -1.42. This difference was
statistically significant at the .01 level.

School success. As noted earlier, the Pueblo program excelled in reaching out
to girls at risk for gang involvement and delinquency. Participants in Movimiento
Ascendencia were somewhat more likely to have been high school dropouts
than girls in the comparison group. Program participants also reported lower
current and previous school grades than girls not in the program. Of program
participants, 46 percent reported earning grades of D and F the year before
entering the program and 34 percent reported C's and D's. For nonparticipants,
the comparable percentages for the previous year's grades were 36 percent
and 14 percent. If A's and B's are assigned a value of 4, B's and C's a value of
3, C's and D's a value of 2, and D's and F's a value of 1, participants in Pueblo
earned an average reported grade of 1.86 the year before entering the
program, compared with 2.42 for the comparison group. This difference was
statistically significant at the .05 level. The current-year mean grade of 2.70 for
participants and 2.97 for the comparison group still indicated lower grades for
participants, but the gap between the two groups had narrowed and was no
longer statistically significant.

Self-esteem. Correlations between the number of hours spent in self-esteem
enhancement activities and each of the three components of the Hare
self-esteem scale were computed. None was significantly related to the number
of hours spent in self-esteem programs.

Conclusions and Recommendations

Recent research on prevention shows that multicomponent programs that
intervene across risk factors and provide interventions across many domains
(e.g., community, family, school, peer, and individual) are most effective in
reducing delinquency (Kelley et al. 1997; Sherman et al. 1998). Moore (1991:
35-39) sketched an outline of intervention programs for Mexican-American
gang members dating from the 1940s that identified the most important
programs as being community-based and empowered residents to exercise
control over their social worlds. According to Moore, while they "rarely
transformed either the gang or any gang member . . . they provided important
links to conventionality--links that were missing for increasing proportions of the
young gang members."

Of the three projects included in this evaluation, Movimiento Ascendencia in
Pueblo offered the most comprehensive services. The project components
addressed the key factors correlated with female gang involvement and
provided the link to conventionality Moore described. In addition, the Pueblo
project specifically targeted gang members and gang-involved girls. For gang
intervention programs to be successful, Spergel and colleagues (1994)
recommended that they avoid either denying or exaggerating community gang
problems. Acknowledging the existence of gang and drug problems in the
community makes it easier to assess the risks of being involved in gangs and
promotes prevention. The findings of the present study suggest that future
program initiatives should be explicit in requiring that prevention and
intervention projects address the multiple risk factors for gang and drug
involvement. Funding levels should reflect the importance of these efforts.

Organizational structure and strong leadership can be key components of
successful program implementation. The most successful program in this
evaluation and others studied by two of the authors (Cohen et al. 1994) were
operated by nonprofit organizations with strong "hands-on" leadership, minimal
levels of bureaucracy, and close ties to the local community. The process
evaluation revealed that the directors of these projects were dedicated and
committed to their work; kept lines of communication open; and provided their
staff with adequate training, supervision, and support. These conditions were
observed most strongly in the Pueblo project. Projects operated or overseen
by organizations with large, cumbersome bureaucracies, such as those in
Boston and Seattle, where those involved in day-to-day operations had no
decisionmaking authority, suffered from mismanagement, lack of adequate
oversight, and poor methods of funding and subcontracting. Projects in smaller
organizations were more flexible and responsive to clients' needs. These
findings suggest that funding agencies could help ensure successful project
implementation by giving preference to projects proposed by organizations with
a strong leadership that have few levels of separation between decisionmakers
and program activities.

All three projects encountered major problems with their local evaluations,
ranging from contracting issues (too little funding, subcontracts signed too late)
to evaluators who were not familiar with the program and lack of adequate
recordkeeping and data. Projects had little difficulty cooperating with the
national evaluation effort because it imposed a minimal burden on them; local
staff were paid by the national evaluators, and their chief responsibilities were to
provide lists of participants and make participants available to be interviewed.
They were not asked to conduct any special data collection for the national
evaluation.

At the time these projects were funded, they generally set aside 5 percent of
their total budgets for evaluations. Recent experiences with other agencies (e.g.,
Cohen and Johnson 1998), showed that similar drug prevention demonstration
programs that committed a higher percentage of funds to evaluation had
evaluations of a significantly higher quality. In addition, several more recent
Federal demonstration programs (such as the Center for Substance Abuse
Prevention's "Starting Early, Starting Smart") required the participation of all
grant recipients on a cross-site evaluation committee, which planned the study
for a year before implementation. These evaluations are a generation improved
over the local evaluations of the projects reported here. Field evaluation
research has shown that a partnership between program and evaluation staff is
vital to successful implementation and evaluation (Resnicow et al. 2001). A
mutually beneficial relationship between program staff and evaluators must be
established early and nurtured. In none of the three projects reported here was
there this type of relationship.

Evaluation designs frequently change once a program begins operation, and
these projects were no exception. The outcome evaluation was hampered by
lack of adequate control groups at two of the three sites, and by being
compelled to use a retrospective pre-post design by default because the
evaluation was funded after the projects began. In addition, for a variety of
reasons local evaluators did not follow through on the proposed evaluation
plans, with the result being incomplete or missing local process and outcome
evaluations. This lack of followthrough, particularly in the local process
evaluations, not only deprives projects of the ability to make midcourse
corrections based on systematic research but also prevents the field from
improving over time.

In the future, sponsoring agencies should fund evaluations 6 months to a year
before projects are funded (or, preferably, with a year's planning before
implementation), require that grant recipients submit written plans that contain
outcome as well as process objectives, and develop a data-collection plan that
indicates how the data will be obtained to measure whether the outcome
objectives have been met. Demonstration project administrators should be
required to develop memoranda of understanding with all parties to ensure
cooperation; generate comparison or control groups; and collect a uniform core
dataset on program participants, operations, and activities. Funding agencies
should also require that local evaluators conduct both process and outcome
evaluations and that they submit quarterly evaluation status reports with detailed
updates on project implementation, on changes (if any) to the research design,
on progress in collecting data with treatment and control groups, and on any
problems encountered. Continued funding should be contingent on providing
this information to the funding agency (or its designee) and cooperating with
outside evaluators.

Such information would alert project monitors at funding agencies when there
are problems in project implementation and when the evaluation design or data
collection plans need to change because of problems in implementing the
original plan. The information would also be a red flag when local evaluations
are having difficulty in implementation and would allow for more immediate
intervention. Funding agencies could proactively assist grant recipients with their
evaluations by providing uniform instruments and sample data collection forms,
holding annual "findings" meetings, and offering technical assistance in measuring
whether annual objectives have been met.

Notes

1. Snowball sampling is a method often used in field research for drawing a
nonprobability sample in which each person interviewed is asked to suggest
additional people to interview.

2. Cronbach's alpha is a measure of how well a group of items focuses on a
single idea or construct. In most social science applications, .80 or higher is
considered an acceptable level of reliability.

References

Bjerregaard, B., and C. Smith. 1993. "Gender Differences in Gang
Participation, Delinquency, and Substance Use." Journal of Quantitative
Criminology 9(4): 329-355.

Campbell, A. 1984. The Girls in the Gang. New York: Basil Blackwell.

Chesney-Lind, M., and J.M. Hagedorn, eds. 1999. Female Gangs in America:
Essays on Girls, Gangs, and Gender. Chicago: Lake View Press.

Chesney-Lind, M., R.G. Sheldon, and K.A. Joe. 1996. "Girls, Delinquency,
and Gang Membership." In Gangs in America (2d ed.). ed. C.R. Huff.
Thousand Oaks, CA: Sage: 185-204.

Cohen, M., K. Williams, A. Bekelman, and S. Crosse. 1994. National
Evaluation of the Youth Gang Drug Prevention Program. Volume 1: Final
Report. Bethesda, MD: DSG, Inc.

Cohen, M., and J. Johnson. 1998. Findings From a Meta-Analysis of Violence
Prevention Grantees. Bethesda, MD: Development Services Group, Inc.

Curry, G.D. 1999a. "Race, Ethnicity and Gender Issues in Gangs: Reconciling
Police Data." In Problem-Oriented Policing: Crime-Specific Issues and Making
POP Work (vol. 2), ed. C.S. Brito and T. Allan. Washington, DC: Police
Executive Research Forum: 63-89.

-------. 1999b. "Responding to Female Gang Involvement." In Female Gangs
in America: Essays on Girls, Gangs, and Gender, ed. M. Chesney-Lind and
J.M. Hagedorn. Chicago: Lake View Press: 133-153.

-------. 2000. "Self-Reported Gang Involvement and Officially Recorded
Delinquency," Criminology 38: 1253-1274.

Curry, G.D., R.A. Ball, and R.J. Fox. 1994. Gang Crime and Law
Enforcement Recordkeeping. Research in Brief. Washington, DC: U.S.
Department of Justice. National Institute of Justice. June. NCJ 148345.

Curry, G.D., R.A. Ball, R.J. Fox, and D. Stone. 1992. National Assessment of
Law Enforcement Anti-Gang Information Resources. Draft final report.
Washington, DC: U.S. Department of Justice, National Institute of Justice. NCJ
147399.

Curry, G.D., and I.A. Spergel. 1992. "Gang Involvement and Delinquency
Among Hispanic and African-American Adolescent Males." Journal of
Research in Crime and Delinquency 29(3): 273-291.

Decker, S.H. 1996. "Collective and Normative Features of Gang Violence."
Justice Quarterly 13: 243-264.

Esbensen, F.A., E.P. Deschenes, and L.T. Winfree, Jr. 1999. "Differences
Between Gang Girls and Gang Boys." Youth and Society 31(1): 27-53.

Esbensen, F.A., and D. Huizinga. 1993. "Gangs, Drugs, and Delinquency in a
Survey of Urban Youth." Criminology 31: 565-587.

Esbensen, F.A., and L.T. Winfree, Jr. 1998. "Race and Gender Differences
Between Gang and Nongang Youths: Results from a Multisite Survey." Justice
Quarterly 15(3): 505-526.

Fagan, J. 1990. "Social Processes of Delinquency and Drug Use Among Urban
Gangs." In Gangs in America, ed. C.R. Huff. Newbury Park, CA: Sage
Publications: 183-219.

Fleisher, M.S. 1998. Dead End Kids: Gang Girls and the Boys They Know.
Madison, WI: University of Wisconsin Press.

Hagedorn, J.M., and M.L. Devitt. 1999. "Fighting Female: The Social
Construction of Female Gangs." In Female Gangs in America: Essays on Girls,
Gangs, and Gender, ed. M. Chesney-Lind and J.M. Hagedorn. Chicago: Lake
View Press: 256-276.

Hare, Bruce. 1977. "Racial and Socioeconomic Variation in Preadolescent
Area-Specific and General Self-Esteem." International Journal of Intercultural
Relations 1: 31-51.

-------. 1980. "Self-Perception and Academic Achievement: Variations in a
Desegregated Setting." American Journal of Psychiatry 137(6): 683-89.

Horowitz, R. 1983. Honor and the American Dream: Culture and Identity in a
Chicano Community. New Brunswick, NJ: Rutgers University Press.

Huff, C.R., ed. 1990. Gangs in America. Newbury Park, CA: Sage
Publications.

Joe, K., and M. Chesney-Lind. 1995. "Just Every Mother's Angel: An Analysis
of Gender and Ethnic Variations in Youth Gang Membership." Gender and
Society 9: 408-430.

Kelley, B.T., R. Loeber, K. Keenan, and M. DeLamatre. 1997.
Developmental Pathways in Boys Disruptive and Delinquent Behavior. OJJDP
Bulletin. Washington, DC: U.S. Department of Justice, Office of Juvenile
Justice and Delinquency Prevention. December. NCJ 165692.

Miller, J. 2001. One of the Guys: Girls, Gangs, and Gender. New York:
Oxford University Press.

Miller, J., and S.H. Decker. 2001. "Young Women and Gang Violence:
Gender, Street Offending, and Violent Victimization in Gangs." Justice
Quarterly 18(1): 87-114.

Moore, J.W. 1978. Homeboys: Gangs, Drugs, and Prison in the Barrios of Los
Angeles. Philadelphia: Temple University Press.

-------. 1991. Going Down to the Barrio: Homeboys and Homegirls in
Change. Philadelphia: Temple University Press.

National Institute of Justice. 1992. Action Plan Development for the Gangs
Initiative. Washington, DC: U.S. Department of Justice, National Institute of
Justice.

-------. 2000. 1998 National Youth Gang Survey: Summary. Washington,
D.C.: U.S. Department of Justice, Office of Juvenile Justice and Delinquency
Prevention. NCJ 183109.

National Youth Gang Center. 1999. 1996 National Youth Gang Survey:
Summary. Washington, DC: U.S. Department of Justice, Office of Juvenile
Justice and Delinquency Prevention. NCJ 173964.

Resnicow, K., R. Braithwaite, C. Dilorio, R, Vaughan, M.I. Cohen, and G.
Uhl. 2001. "Preventing Substance Use in High Risk Youth: Evaluation
Challenges and Solutions." The Journal of Primary Prevention 21(3): 399-415.

Sherman, L.W., D.C. Gottfredson, D.L. MacKenzie, J. Eck, P. Reuter, and
S.D. Bushway. 1998. Preventing Crime: What Works, What Doesn't, What's
Promising. Research In Brief. Washington, DC: U.S. Department of Justice,
National Institute of Justice. July. NCJ 171676.

Spergel, I.A. 1990. "Youth Gangs: Continuity and Change." In Crime and
Justice: A Review of Research (vol. 12), ed. M. Tonry and N. Morris.
Chicago: University of Chicago Press: 171-275.

Spergel, I.A., R. Chance, K. Ehrensaft, T. Regulus, C. Kane, R. Laseter, A.
Alexander, and S. Oh. 1994. Gang Suppression and Intervention: Community
Models. Washington, DC:, U.S. Department of Justice, Office of Juvenile
Justice and Delinquency Prevention. NCJ 148202.

Thrasher, F.M. 1963. The Gang: A Study of 1,313 Gangs in Chicago.
Chicago: University of Chicago Press (originally published in 1927).

Williams, K., D. Curry, and M. Cohen. Forthcoming. Final Report on the
Evaluation of Gang Programs for Female Adolescents. Washington, DC: U.S.
Department of Justice, Office of Justice Programs, National Institute of Justice.

---------------------------

Chapter 9

Reducing Gang Violence in Boston

Anthony A. Braga
David M. Kennedy
Program in Criminal Justice Policy and Management
Malcolm Wiener Center for Social Policy
John F. Kennedy School of Government
Harvard University

Acknowledgments

This research was supported under award 94-IJ-CX-0056 from the National
Institute of Justice, Office of Justice Programs, U.S. Department of Justice.
Funding was also received from the Office of Community Oriented Policing
Services, U.S. Department of Justice. The authors would like to thank Paul
Evans, James Jordan, Gary French, Paul Joyce, and the members of the
Boston Gun Project working group for their support in this research enterprise.
We also would like to thank Lois Mock of the National Institute of Justice for
her steadfast support, guidance, and patience in connection with this project.

---------------------------

Abstract

Operation Ceasefire is a problem-oriented policing intervention aimed at
reducing youth homicide and youth firearms violence in Boston. It represented
an innovative partnership between researchers and practitioners to assess the
city's youth homicide problem and implement an intervention designed to have a
substantial near-term impact on the problem. Operation Ceasefire was based
on the "pulling levers" deterrence strategy, which focused criminal justice
attention on a small number of chronically offending gang-involved youth
responsible for much of Boston's youth homicide problem. A rigorous
evaluation suggests that the Ceasefire intervention was associated with
significant reductions in youth homicide victimization, "shots fired" calls for
service, and gun assault incidents in Boston. Although the deterrence strategy
was the key element of the Operation Ceasefire intervention, the gang violence
reduction initiative was supported and strengthened by other program elements,
including social intervention, opportunity provision, and community organization
strategies.

---------------------------

Although overall homicide rates in the United States declined between the
1980s and 1990s, youth homicide rates, particularly incidents involving
firearms, increased dramatically (Fox 1996; Cook and Laub 1998). Like many
large cities in the United States, Boston experienced an epidemic of youth
violence beginning in the late 1980s and continuing through the early 1990s.
Measured as a homicide problem, the epidemic was confined almost entirely to
the city's poor black neighborhoods, was concentrated among youths ages 18
to 24, and was largely a firearms problem (Kennedy and Braga 1998). Youth
homicide in Boston (ages 24 and under) increased 230 percent--from 22
victims
in 1987 to 73 victims in 1990--and remained high even after the 1990 peak of
the epidemic. Between 1991 and 1995, Boston averaged about 44 youth
homicides per year.

As the United States was experiencing this epidemic of youth violence, the
capacity of police departments to design and implement creative, new
operational strategies also increased through the advent of "community" and
"problem-oriented" policing (Sparrow, Moore, and Kennedy 1990; Goldstein
1990). Beginning in 1995, an interagency working group of Harvard University
researchers, members of the Boston Police Department (BPD), and other
criminal justice agencies conducted research and analysis on Boston's youth
violence problem, designed a problem-solving intervention to reduce youth
violence, and implemented the intervention. The research showed that the
problem of youth violence in Boston was concentrated among a small number
of serially offending, gang-involved youths (Kennedy, Piehl, and Braga 1996;
Kennedy, Braga, and Piehl 1997). The problem-solving intervention that
evolved from the research diagnoses focused criminal justice attention on these
chronic gang offenders. The well-known Boston strategy to prevent gang
violence was implemented in mid-1996 and was associated with significant
decreases in youth homicide and other indicators of nonfatal serious violence
(Braga, Kennedy, Waring, and Piehl 2001).

The Boston Gun Project and Operation Ceasefire

Sponsored by the National Institute of Justice (NIJ), BPD and researchers
from Harvard University's John F. Kennedy School of Government formed a
partnership to design, implement, and evaluate an innovative problem-oriented
policing intervention to reduce youth violence. Known as the Boston Gun
Project, this enterprise involved the following:

o Assembling an interagency working group of largely line-level criminal justice
and other practitioners.

o Applying quantitative and qualitative research techniques to assess the nature
of, and dynamics driving, youth violence in Boston.

o Developing an intervention designed to have a substantial, near-term impact
on youth homicide.

o Implementing and adapting the intervention.

o Evaluating the intervention's impact (Braga, Kennedy, Waring, and Piehl
2001).

The Project began in early 1995 and implemented what is now known as the
"Operation Ceasefire" intervention, which began in May 1996.

Core participating agencies, as defined by regular participation in the Boston
Gun Project Working Group for its duration, included BPD; the Massachusetts
Departments of Probation and Parole; the Office of the Suffolk County District
Attorney; the Office of the U.S. Attorney; the Bureau of Alcohol, Tobacco and
Firearms (ATF); the Massachusetts Department of Youth Services (DYS)
(juvenile corrections); Boston School Police; and gang outreach and prevention
"streetworkers" attached to the Boston Community Centers program. Other
significant participants, involved either as regular partners later in the process or
episodically, included the Ten Point Coalition of activist black clergy, the Drug
Enforcement Administration (DEA), the Massachusetts State Police, the Office
of the Massachusetts Attorney General, and others (Kennedy, Piehl, and Braga
1996).

Boston Gun Project Research Results

The research activities of the Boston Gun Project provided important insights
on the nature of the city's youth violence problem. Project research showed that
the problem of youth homicide was concentrated among a small number of
chronically offending gang-involved youths. Boston's "gangs" were typically
small, relatively disorganized, neighborhood-based groups (Kennedy, Braga,
and Piehl 1997). As many scholars observe, defining the term "gang" is
complex (see, e.g., Ball and Curry 1995). Although Boston authorities use the
word "gang," it is in some sense a term of convenience, meaning in practice only
a "self-identified group of kids who act corporately (at least sometimes) and
violently (at least sometimes)" (Kennedy, Piehl, and Braga 1996: 158). Thus,
what "gang" means in Boston bears little resemblance to what it means, for
instance, in Chicago and Los Angeles. This common finding is critical to the
analysis and comprehension of gang- and group-related youth crime and
violence. The character of criminal and disorderly youth gangs and groups
varies widely both within and across cities (see, e.g., Curry, Ball, and Fox
1994).

In the summer of 1995, Boston had 61 gangs comprising only about 1,300
gang members (Kennedy, Braga, and Piehl 1997). Research on the age
distribution of gang members suggests a range between early adolescence and
young adulthood (Spergel 1995). Although it is difficult to set upper and lower
bounds on the age distribution of gang members in practice, most are between
14 and 24 years of age. According to 1996 U.S. Census estimates, these
1,300 gang members represented less than 1 percent of their age group
citywide and less than 3 percent of their age group in Boston neighborhoods
with gang turf (Kennedy, Braga, and Piehl 1997). The small proportion of
Boston's youths involved in gangs is consistent with estimates of youth
participation in gangs from other cities (Esbensen and Huizinga 1993; Spergel
1995). Project research revealed that Boston gangs were well known to
authorities and streetworkers; many gang members were also well known and
tended to have extensive and quite varied criminal histories (Kennedy, Piehl,
and Braga 1996).

Boston gang members were engaged in a wide range of illicit activity, ranging
from minor property and disorder crimes to serious and violent offenses. The
persistent involvement of Boston gang members in crime fits well with what is
known about the criminal behaviors of gang members in other jurisdictions.
Gang members are involved in far more delinquent and criminal behavior than
their non-gang counterparts (Fagan 1989; Hagedorn 1988; Thornberry et al.
1993) and engage in a wide variety of criminal behaviors, termed
"cafeteria-style" offending by Klein (1995).

Boston gangs also congregated in and defended "turf"--very small, well-defined
areas within their neighborhoods. Klein (1995:18) suggests that these areas are
"the life space of the gang" and that the most important role of turf is that it
offers the gang something tangible to identify with and possess (see also Tita
1997). The physical spaces occupied by Boston gangs were typically very
small; often, the turf occupied by a particular gang was no larger than a piece of
a single street block or the grounds surrounding an apartment building. The total
geographic expanse of the 61 gang areas was 1.7 square miles, only 3.6
percent of Boston's 47.47 square miles (Kennedy, Braga, and Piehl 1997).
Turf made up only 8.1 percent of the area of the Boston neighborhoods with
gangs--Roxbury, Dorchester, Mattapan, Jamaica Plain, Hyde Park, and the
South End (Kennedy, Braga, and Piehl 1997). Although these areas were
small, violence was disproportionately clustered in gang turf. About 25 percent
of incidents involving youth homicides, weapons offenses, simple and
aggravated assaults, armed robberies, and "shots fired" calls for service
occurred inside these gang areas (Kennedy, Braga, and Piehl 1997).

Conflicts between street gangs have long been noted to fuel much of the youth
violence in major cities (Block and Block 1993; Klein and Maxson 1989;
Curry, Ball, and Fox 1994; Miller 1977). Law enforcement agencies in
different cities use different definitions for "gang-related" crime, which affect the
number of gang-related crimes reported. For example, Los Angeles police
define homicide as gang related when gang members participate, regardless of
motive; Chicago police use a more restrictive definition and classify homicides
as gang related only if there is a gang motive evident (Maxson and Klein 1990).
In the Boston research, homicides were considered to be connected to gangs if
either the murderer or the victim was a gang member and the motivation behind
the murder was known or believed to be connected to gang activity. Thus, the
killing of a gang member by another gang member in a dispute over contested
turf would be considered gang related, as would the killing of a non-gang
bystander during the same dispute. However, the killing of a gang member by a
non-gang member during a robbery attempt was not considered to be gang
related. Although there were relatively few gang-involved youths in Boston,
Project research revealed that members of the 61 gangs were responsible for at
least 60 percent of all youth homicides in the city (Kennedy, Piehl, and Braga
1996).

Chronic disputes among gangs were the primary drivers of gang violence in
Boston (Braga, Piehl, and Kennedy 1999). The majority of Boston youth
homicides identified as gang-related were not about drugs, money, or other
issues for which the violence could be considered instrumental. In fact, only 6
percent of the gang-related homicides directly involved a drug dispute (Braga,
Piehl, and Kennedy 1999). Rather, Boston's gang-related youth homicides
were usually personal and vendetta-like. This does not mean that vendettas
might not have started over drug- or money-related issues; however, if the
vendettas had such a genesis, they had taken on an independent life by the time
the Boston Gun Project research was conducted. Also, the gang-related
homicides largely were not "senseless" or random, at least to the participants.
The incidents had a history, however absurd to observers, and the episode was
part of a larger history.

Most research on gang violence has found that violent behavior tends to be
expressive rather than instrumental (Decker 1996). However,
Sanchez-Jankowski's (1991) 10-year observational study of gangs in Los
Angeles, New York City, and Boston argued that gangs are purposive
organizations and much gang violence is instrumental in nature. The Boston
research suggests otherwise for Boston gangs in the early 1990s, and much of
the gang literature suggests that most gang violence is typically retaliatory in
nature (Block and Block 1993; Decker and Van Winkle 1996; Klein and
Maxson 1989). In a followup study of East Los Angeles gangs, Moore (1991)
cites the following reasons for gang fights: invasion of territory by a rival gang,
rivalry over dating, fights relating to sporting events, and personal matters in
which the gang is brought in to support an aggrieved individual.

Project research also showed that firearms associated with Boston youths,
especially with those in gangs, tended to be semiautomatic pistols, which often
were new and apparently had been recently diverted from retail (Kennedy,
Piehl, and Braga 1996). Many of these guns were first sold at retail in
Massachusetts, while others had been smuggled in from out of state (Kennedy,
Piehl, and Braga 1996). Research has shown that gang violence often involves
the use of firearms. Sheley and Wright (1995) found gang membership
associated with gun ownership, gun theft, gun sales, gun sharing, and other
crimes, and with norms promoting gun ownership and use. In surveys of
arrestees in several cities, Decker and Pennell (1995) found gun ownership
more highly correlated with gang membership than with drug dealing. Evidence
also exists that gang members seek more sophisticated and powerful weaponry
than do their rivals. Gang members attempt to acquire numerous guns and
greater firepower in the belief that winning the "arms race" will increase their
personal protection and reduce their likelihood of victimization (Horowitz 1983;
Stover 1986). These qualitative insights are supported by Block and Block's
(1993) quantitative research on the increasing use and lethality of firearms in
gang homicides in Chicago. Hutson and colleagues (1995) also reported that a
major cause of gang-related homicide was the increased use of firearms,
particularly semiautomatic handguns.

The Operation Ceasefire Strategy to Reduce Gang Violence in Boston

Research findings and the Working Group process led to development of the
"Operation Ceasefire" intervention. Operation Ceasefire included two main
elements: a direct law enforcement attack on illicit firearms traffickers supplying
youths with guns, and an attempt to generate a strong deterrent to gang violence
(Kennedy, Piehl, and Braga 1996). The Working Group framed a set of
activities intended to systematically address patterns of firearms trafficking
identified by the research (Kennedy, Piehl, and Braga 1996). These included
focusing enforcement attention on intrastate gun trafficking, traffickers of the
types of guns most used by gang members, and guns most likely to be
trafficked (e.g., new guns and guns with obliterated serial numbers). These
enforcement priorities were supported through analyses of results by BPD from
comprehensive tracing of crime guns and by the development of leads through
systematic debriefing of arrestees involved with gangs or violent crime (Braga,
Kennedy, Waring, and Piehl 2001).

The second element of Operation Ceasefire was the "pulling levers" focused
deterrence strategy. This approach involved deterring chronic gang offenders'
violence by reaching out directly to gangs, saying explicitly that violence would
no longer be tolerated, and backing that message by "pulling every lever" legally
available when violence occurred (Kennedy 1997, 1998). Boston gangs were
not subjected to increased law enforcement attention arbitrarily nor did the
Working Group develop a "hit list" of gangs. Rather, the Working Group's
enforcement actions occurred in response to outbreaks of gang violence. In
other words, Boston gangs selected themselves for focused law enforcement
attention by engaging in violence. When gang violence occurred, Working
Group members sent a direct message to gang members that they were "under
the microscope" because of their violent behavior (Braga, Kennedy, Waring,
and Piehl 2001). Police officers, probation officers, and DYS caseworkers
immediately flooded the targeted gang's turf and communicated to gang
members that their presence was in response to the violence. Streetworkers
and members of the Ten Point Coalition walked the streets and explained that
they wanted the violence to stop and supported the efforts of their law
enforcement counterparts to make that happen. They also offered services and
opportunities to gang members, including health and social services, educational
and recreational opportunities, substance and alcohol abuse intervention
programs, and food and shelter.

As Operation Ceasefire unfolded, the Working Group assessed the
enforcement levers available to stop violent gang activity (Kennedy 1997).
Enforcement responses were tailored to particular gangs and often included a
wide range of actions, such as conducting probation checks, changing
community supervision conditions, serving outstanding arrest warrants, focusing
special prosecutorial attention on crimes committed by violent gang members,
increasing disorder enforcement, and disrupting street-level drug markets. A
basic premise of the pulling levers approach was to take advantage of the
chronic-offending behaviors of gang members. The Working Group recognized
that gang members were vulnerable to a variety of criminal justice sanctions and
that targeted enforcement actions could be used to good effect in controlling
their violent behavior (Kennedy 1997). Although Operation Ceasefire
implemented varying enforcement tactics on particular gang members within a
group, the focus of the operation was to stop the group's violent activities.

The enforcement actions selected by the Working Group were only as harsh as
necessary to stop a gang from engaging in violence. For many gangs and gang
members, heightened levels of police, probation, and DYS enforcement were
sufficient to end the violence. Quickly shutting down drug markets, serving
warrants, enforcing probation and DYS supervision conditions, making
disorder arrests, and dealing more strictly with resulting cases as they were
prosecuted and adjudicated were generally powerful enough actions to stop the
violence. To curb violence among some hardcore gangs and gang members,
however, the enhanced enforcement capabilities of Federal authorities were
needed. In these instances, ATF, DEA, and FBI investigations, prosecutions by
the U.S. Attorney's Office, and the prospect of serving a substantial term in a
Federal prison far from home were used to quell the violence.

While carrying out enforcement actions, Working Group members continued
communicating with violent gang members. The Group believed it was crucial to
demonstrate cause and effect to gang members subjected to the Operation
Ceasefire intervention. The Working Group delivered a direct and explicit
message to violent gangs that violent behavior would no longer be tolerated and
that it would use all legally available means to stop the violence (Kennedy
1998). This message also was communicated to gangs not engaged in violence
so they would understand what was happening to the violent gang and why.

In addition to talking to gang members on the street, the Working Group
delivered the deterrence message through fliers explaining the enforcement
actions and through forums with gang members (Kennedy 1997, 1998).
Forums were usually held in a public facility, such as a courthouse or
community recreational center. Gang members under criminal justice system
supervision were required to attend a forum by their probation or parole
officers; gang-involved juveniles under DYS community supervision were
required to attend by their caseworkers. Streetworkers and members of the
Ten Point Coalition were able to bring other gang members to a forum by
persuading them that it was in their best interest to attend. At each forum,
representatives of the law enforcement agencies involved in Operation
Ceasefire explained their actions to the gang members. Streetworkers and the
Ten Point clergy voiced their support of the law enforcement actions, asked the
youths to stop the violence, and reiterated their offers of services and
opportunities. At the end of the presentation, gang members were encouraged
to ask questions on the anti-violence campaign.

The Impact of Operation Ceasefire on Youth Violence in Boston

Immediately after Operation Ceasefire was implemented in mid-1996, the
annual number of Boston youth homicides fell markedly. As mentioned earlier,
Boston averaged about 44 youth homicides per year between 1991 and 1995.
In 1996, that number fell to 26 homicides and further decreased to 15 youth
homicides in 1997, a trend that continued through 1998 (18) and 1999 (15).

Although these numbers show a sudden, large decrease in Boston youth
homicides, they do not establish whether Operation Ceasefire was associated
with that decrease. Thus, with the support of NIJ, researchers from Harvard
University's John F. Kennedy School of Government rigorously evaluated the
effects of Operation Ceasefire on youth violence in Boston (Braga, Kennedy,
Waring, and Piehl 2001). Using carefully constructed time series analysis
models that controlled for trends and seasonal variations, the Kennedy School
evaluation team found that implementation of Operation Ceasefire was
associated with the following monthly numbers for Boston: a 63-percent
decrease in youth homicides, a 32-percent decrease in calls about shots fired, a
25-percent decrease in Boston gun assaults, and a 44-percent decrease in
youth gun assault incidents in one high-risk district (Braga, Kennedy, Waring,
and Piehl 2001).

The youth homicide and gun violence reductions associated with Operation
Ceasefire could have been caused or meaningfully influenced by other causal
factors. Therefore, the evaluation team added control variables to the time
series analysis models, including changes in the unemployment rate, in Boston's
youth population, in violent index crimes, in older homicide victimization, and in
street-level drug activity as measured by BPD arrest data (Braga, Kennedy,
Waring, and Piehl 2001). The addition of these control variables did not
substantively change the findings. Operation Ceasefire was still associated with
significant decreases in the monthly number of youth homicides and other
indicators of nonfatal serious violence.

These analyses supported the conclusion that a large reduction in Boston's
youth homicide and gun violence was associated with Operation Ceasefire.
However, because many major cities in the United States have enjoyed
noteworthy reductions in homicide and nonfatal serious violence, the Kennedy
School research team also distinguished Boston's youth homicide trends from
national and regional trends. Using Supplementary Homicide Report data, the
team analyzed monthly numbers of homicide victims ages 24 and under for 29
major New England cities and 39 major U.S. cities. (For a complete
methodological description, see Braga, Kennedy, Waring, and Piehl 2001.)
Drawing on carefully constructed time series analyses, the research suggested
that the significant reduction in youth homicide associated with Operation
Ceasefire in Boston was distinct when compared with youth homicide trends in
most major New England and U.S. cities.

The Effectiveness of Operation Ceasefire

A central hypothesis within the Working Group was that a meaningful period of
substantially reduced youth violence might serve as a "firebreak" and result in a
relatively long-lasting reduction in future youth violence (Kennedy et al. 1996).
The idea was that youth violence in Boston had become a self-sustaining cycle
among a relatively small number of youths, with objectively high levels of risk
leading to nominally self-protective behavior, such as gun acquisition and use,
gang formation, and tough street behavior, all of which fueled the cycle of
violence (Kennedy et al. 1996). If this cycle could be interrupted, a new
equilibrium at a lower level of risk and violence might be established, perhaps
without the need for continued high levels of either deterrent or facilitative
intervention. The impact evaluation suggests that youth violence did indeed
settle at a new lower equilibrium after Operation Ceasefire was implemented.

Unfortunately, the Kennedy School research team was unable to collect the
necessary pre- and post-test data to shed light on the specific mechanisms
responsible for the significant reductions in violence associated with Operation
Ceasefire. The research team focused on problem analysis and program
development and, a priori, did not know what form the intervention would take
and whom the target audience would be. It seems unlikely that the abrupt, large
reduction in violence was due to Operation Ceasefire's focus on illegal firearms
trafficking. Gun trafficking cases associated with Operation Ceasefire followed
the significant reduction in violence rather than preceded it. Most of the guns
held by Boston gang members in May 1996 were still held by them several
months later, when violence fell to its new, lower equilibrium. Although it may
well be that antitrafficking efforts strengthened and prolonged the impact of the
intervention, the sudden reduction in violence was not due to a sudden change
in the supply of guns to Boston gang members or to the stockpile of guns held
by Boston gang members. Rather, that reduction was due to a change in the use
of guns by Boston gang members. The principal impact therefore was almost
certainly a demand-side, deterrence-based effect, rather than a supply-side
effect.

Because the necessary evaluation data are unavailable, it is necessary to draw
on the research literature on gang intervention programs to speculate on the
effectiveness of the Operation Ceasefire approach to controlling gang violence.
As part of the Office of Juvenile Justice and Delinquency Prevention (OJJDP)
National Youth Gang Suppression and Intervention Program, Spergel and
Curry (1990, 1993) surveyed 254 law enforcement, school, and community
representatives in 45 cities and 6 institutional sites on their gang intervention
programs. From these survey data, Spergel and Curry developed a typology of
interventions used to deal with gang problems; they grouped gang intervention
programs into four broad categories: suppression, social intervention,
opportunity provision, and community organization. Although Operation
Ceasefire is a problem-oriented policing project centered on law enforcement
interventions, the multidimensional and complex activities of the program fall in
all four categories. It appears that the pulling levers deterrence strategy was the
key program element most directly responsible for the noteworthy youth
violence reductions in Boston. However, other elements of Operation Ceasefire
that involved social intervention, opportunity provision, and community
organization strategies supported and strengthened law enforcement's ability to
reduce gang violence. Beyond deterring violent behavior, Operation Ceasefire
was designed to facilitate desired behaviors among gang members. As Spergel
(1995) observes, coordinated strategies that integrate these varied domains are
most likely to be effective in dealing with chronic youth gang problems. The
following sections discuss how Operation Ceasefire program activities in these
four intervention categories may have reduced gang violence in Boston.

Suppression

The typical law enforcement suppression approach assumes that most street
gangs are criminal associations that must be attacked through efficient gang
tracking, identification, and targeted enforcement strategy (Spergel 1995). The
basic premise of this approach is that improved data collection systems and
coordination of information across different criminal justice agencies lead to
more efficiency and to more gang members being removed from the streets,
rapidly prosecuted, and sent to prison for longer sentences (Spergel 1995).
Typical suppression programs included street sweeps in which police officers
round up hundreds of suspected gang members, special gang probation and
parole caseloads through which gang members are subjected to heightened
levels of surveillance and more stringent revocation rules, prosecution programs
that target gang leaders and serious gang offenders, civil procedures that use
gang membership to define arrest for conspiracy or unlawful associations, and
school-based law enforcement programs that use surveillance and buy-bust
operations (Klein 1993).

These suppression approaches are loosely based on deterrence theory (Klein
1993). Law enforcement agencies attempt to influence the behavior of gang
members or eliminate gangs entirely by dramatically increasing the certainty,
severity, and swiftness of criminal justice sanctions. Unfortunately, gangs and
gang problems usually endure despite these intensive operations. Klein (1993)
suggests that law enforcement agencies do not generally have the capacity to
"eliminate" all gangs in a gang-troubled jurisdiction, nor do they have the
capacity to respond powerfully to all gang offending in such jurisdictions.
Pledges to do so, though common, are simply not credible to gang members.
This study also observes that emphasizing selective enforcement by
deterrence-based gang suppression programs may increase the cohesiveness of
gang members, who often perceive such actions as unwarranted harassment,
rather than cause them to withdraw from gang activity. Therefore, suppression
programs may have the perverse effect of strengthening gang solidarity.

Beyond the certainty, severity, and swiftness of sanctions, effective deterrence
depends on communicating threats of punishment to the public. As Zimring and
Hawkins (1973) observe, "the deterrence threat may best be viewed as a form
of advertising" (142). The Operation Ceasefire Working Group recognized that
for the strategy to be successful, it was crucial to deliver a credible deterrence
message to Boston gangs. Therefore, Operation Ceasefire only targeted those
gangs engaged in violent behavior, rather than wasting resources on those that
were not. Spergel (1995) suggests that problem-solving approaches to gang
problems based on more limited goals, such as gang violence reduction rather
than gang destruction, are more likely to be effective in controlling gang
problems. Operation Ceasefire did not attempt to eliminate all gangs or
eliminate all gang offending in Boston. Despite the large reductions in youth
violence, Boston still has gangs that commit crimes. However, the city's gangs
do not commit violent acts as frequently as they did in the past.

The pulling-levers approach attempted to prevent gang violence by making
gang members choose to change their behavior to avoid consequences that
would follow from their violence and gun use. A key element of this strategy
was the delivery of a direct, explicit "retail deterrence" message to a relatively
small target audience about the kind of behavior that would provoke a special
response and what that response would be. In addition to any increases in
certainty, severity, and swiftness of sanctions associated with acts of violence,
the Operation Ceasefire strategy pursued deterrence through advertising its law
enforcement strategy and the personalized nature of its application. It was
crucial that gang youths understood the new regime that the city was imposing.
Beyond the particular gangs and gang members subjected to the intervention,
the deterrence message was communicated to a relatively small audience (all
gang-involved youths in Boston) rather than to a general audience (all Boston
youths), and it articulated explicit cause-and-effect connections between the
behavior of the target population and that of the authorities. Knowledge of what
happened to others in the target population was intended to prevent further acts
of violence by gangs in Boston.

In communicating the deterrence message, the Working Group also wanted to
reach a shared moral ground with gang members. The Group wanted gang
members to understand that most victims of gang violence were other gang
members, that the strategy was designed to protect both gang members and the
community in which they lived, and that the Working Group had gang members'
best interests in mind--even if their own actions required coercion to protect
them. The Working Group also hoped that face-to-face communication with
gang members would undercut any feelings of anonymity and invulnerability
they might have, and that a clear demonstration of interagency solidarity would
enhance offenders' sense that something new and powerful was happening.

Operation Ceasefire was also intended to provide gang youths with a way to
save face when confronted with the threat of violence. Youths in inner city
neighborhoods adopt street behavior that is often reinforced by their peers
(Anderson 1994). Street culture dictates that young men earn respect through
tough behavior. Youths may be expected by members of their groups to
respond to threats and challenges in violent, expressive ways (Wilkinson and
Fagan 1996). Failure to adhere to these "scripts" may result in loss of status and
exposure to a greater risk of victimization or isolation from peers (Wilkinson
and Fagan 1996). The Working Group recognized these dynamics in conflicts
between Boston gangs. A gang that suffered violence had to respond or it
would lose status and respect on the street and open itself to further
victimization. Operation Ceasefire was intended to give gang youths an "out"
from responding violently to these status challenges by swiftly addressing the
offending group, eliminating the need for a response and providing the
victimized group with a justification for inaction. Victimized gang youths could
save face by asserting that the authorities had removed the offending gang
youths from the street and that a violent response was unwise, given the
increased attention to the current situation.

Social Intervention and Opportunity Provision

Social intervention programs encompass social service agency-based and
detached streetworker programs; opportunity provision strategies attempt to
offer gang members legitimate opportunities and means to success that are at
least as appealing as illegitimate options (Curry and Decker 1998; Spergel
1995; Klein 1995). Boston streetworkers were key members of the Operation
Ceasefire Working Group and, along with DYS case workers, probation
officers, and parole officers in the group, added a much-needed social
intervention and opportunity provision dimension to Operation Ceasefire.

Boston's Mayor established the Boston Community Centers' Streetworkers
social service program in 1991. Streetworkers were charged with seeking out
at-risk youths in Boston's neighborhoods and providing them with services such
as job skills training, substance abuse counseling, and special education. Many
Boston streetworkers are themselves former gang members. Gang researchers
have suggested that meaningful gang crime prevention programs should recruit
gang members to participate in the program as staff and consultants (Hagedorn
1988; Bursik and Grasmick 1993). Beyond their important roles as social
service providers, streetworkers attempted to prevent outbreaks of violence by
mediating disputes between gangs. If the streetworkers had good relations with
members of two gangs in conflict, they would work to settle the dispute
between them through separate meetings with each group that would culminate
in a negotiation session with key members of both rival groups. Although this
tactic was not explicitly used in Operation Ceasefire, Boston's streetworkers
used these negotiating skills and relationships with gang youths to support law
enforcement efforts to stop outbreaks of violence. Streetworkers also ran
programs intended to keep gang-involved youths safely occupied and bring
them into contact in ways that might breed tolerance, including a Peace League
of gang-on-gang basketball games held at neutral, controlled sites.

With these resources, the Working Group was able to pair criminal justice
sanctions, or the threat of sanctions, with help and services. The premise behind
this strategy was straightforward: When the risk to drug-dealing gang members
increases, legitimate work becomes more attractive; when legitimate work is
more available, raising risks will be more effective in reducing violence. The
availability of social services and opportunities was intended to increase
Operation Ceasefire's preventive power by offering gang members any
assistance they might want, including protection from their enemies, drug
treatment, and access to education and job training programs.

Community Organization

Community organization strategies to cope with gang problems include attempts
to create community solidarity, networking, education, and involvement
(Spergel and Curry 1993). The Ten Point Coalition of activist black clergy
played an important role in organizing Boston communities suffering from gang
violence (see Winship and Berrien 1999). The Coalition formed in 1992 after
gang members invaded the Morningstar Baptist Church, where a slain rival gang
member was being memorialized, and attacked mourners with knives and guns.
In the wake of that incident, the Ten Point Coalition decided to respond to
violence in their community by reaching out to youths involved in drugs and
gangs and by organizing within Boston's black community. The Ten Point clergy
came to work closely with the Boston Community Centers' streetworkers
program to provide at-risk youths with opportunities. Although the Coalition
was initially critical of the Boston law enforcement community, it eventually
forged a strong working relationship. Ten Point clergy and others involved in
this faith-based organization accompanied police officers on home visits to the
families of troubled youths and acted as advocates for youths in the criminal
justice system. The clergy's home visits and street work were later incorporated
into Operation Ceasefire's portfolio of interventions. Ten Point clergy also
provided a strong moral voice at the forums held to present Operation
Ceasefire's anti-violence message to gang members.

Although it was not involved in Operation Ceasefire until after the intervention
strategy had been designed and implemented, the Ten Point Coalition played a
crucial role in framing the discussion that made it much easier to speak directly
about youth violence in Boston. Members of the Ceasefire Working Group
could speak with relative safety about the painful realities of minority male
offending and victimization, gangs, and chronic offenders. The Ten Point clergy
also made it possible for Boston's minority community to have an ongoing
conversation with the city's law enforcement agencies on legitimate and
illegitimate means to control crime. The clergy supported Operation Ceasefire's
tight focus on violent youths, but condemned indiscriminate, highly aggressive
law enforcement sweeps that put non-violent minority youths at risk of being
swept into the criminal justice system. Before the Coalition developed its role as
an intermediary, Boston's black community viewed past activities of law
enforcement agencies to monitor violent youths as illegitimate and suspicious.
As Winship and Berrien (1999) observe, the Ten Point Coalition evolved into
an institution that provides an umbrella of legitimacy for police actions. With the
Coalition's approval of and involvement in Operation Ceasefire, the community
supported the approach as a legitimate youth violence prevention campaign.

Conclusion

The Boston Gun Project was an attempt to apply problem-oriented policing to
one important problem, youth violence, in one city, Boston. The Project
assembled a working group with members from a variety of agencies
representing a range of law enforcement, social service, and other operational
capacities (Kennedy et al. 1996). Typical of problem-solving operations, the
issue definition, core participants, and particulars of the intervention evolved
over the course of the collaboration. The Project's core operational
intervention, Operation Ceasefire, was designed to be effective anywhere in the
city where youth violence was a serious problem and was intended to interrupt
the cycle the Gun Project hypothesized to be driving youth violence in the city
(Kennedy et al. 1996). The pulling levers deterrence strategy at the heart of
Operation Ceasefire was designed to influence the behavior and environment of
the core population of chronic-offender gang-involved youths the Gun Project
research found to be central to Boston's youth violence problem (Kennedy
1997).

The Boston Gun Project applied the basic principles of problem-oriented
policing to a substantial public safety problem. Addressing this problem
required the involvement of multiple agencies and the community and substantial
investments in analysis, coordination, and implementation. The Gun Project
experience suggests that deploying criminal justice capacities to prevent crime
can yield substantial benefits. Although Operation Ceasefire was primarily a
criminal justice intervention, its other elements--social intervention, opportunity
provision, and community organization strategies--supported and strengthened
the ability of criminal justice agencies to reduce gang violence. The Boston
experience suggests that a multidimensional mix of interventions is a desirable
way to approach complex and sensitive problems like urban youth gang
violence. It is also important to note that Operation Ceasefire did not impose
additional costs on the participating organizations, but was implemented by
using existing resources more strategically.

The Operation Ceasefire approach worked because it was designed to solve a
problem driven by specific local dynamics; however, the targeted activities of
the Working Group may not necessarily be effective in other communities. It
was customized to the goals of the Working Group, the nature of Boston's
youth violence, and the city's capacity to incorporate a strategic intervention.
Operation Ceasefire, therefore, is unlikely to be a highly specifiable,
transportable "technology." Nevertheless, certain process elements of the
Boston Gun Project, such as the central role of the line-level Working Group
and the use of qualitative and quantitative research to understand chosen
problems, should be applicable to other problem-solving efforts. To the extent
that youth violence in other jurisdictions emanates from gangs and criminally
active groups, it appears that many lessons can be learned from the Boston
experience. Applying the working group problem-solving approach, criminal
justice practitioners in other jurisdictions can develop a set of intervention
strategies that fits the nuances of their youth violence problem and their
operational capacities.

References

Anderson, Elijah. 1994. "The Code of the Streets." The Atlantic Monthly May:
81-94.

Ball, Richard A., and G. David Curry. 1995. "The Logic of Definition in
Criminology: Purposes and Methods for Defining 'Gangs.'" Criminology 33:
225-245.

Block, Carolyn R., and Richard Block. 1993. Street Gang Crime in Chicago.
Research in Brief. Washington, DC: U.S. Department of Justice. National
Institute of Justice. December 1993, NCJ 144782.

Braga, Anthony A., David M. Kennedy, Elin J. Waring, and Anne M. Piehl.
2001. "Problem-Oriented Policing, Deterrence, and Youth Violence: An
Evaluation of Boston's Operation Ceasefire," Journal of Research in Crime and
Delinquency 38(3): 195-225.

Braga, Anthony A., Anne M. Piehl, and David M. Kennedy. 1999. "Youth
Homicide in Boston: An Assessment of Supplementary Homicide Report
Data." Homicide Studies 3: 277-299.

Bursik, Robert, and Harold G. Grasmick. 1993. Neighborhoods and Crime:
The Dimensions of Effective Community Control. New York: Lexington
Books.

Cook, Philip J., and John H. Laub. 1998. "The Unprecedented Epidemic in
Youth Violence." In Youth Violence, ed. Michael Tonry and Mark H. Moore.
Chicago: University of Chicago Press.

Curry, G. David, and Scott H. Decker. 1998. Confronting Gangs: Crime and
Community. Los Angeles: Roxbury Press.

Curry, G. David, Richard A. Ball, and Robert J. Fox. 1994. Gang Crime and
Law Enforcement Recordkeeping. Research in Brief. Washington, DC: U.S.
Department of Justice. National Institute of Justice. June 1994. NCJ 148345.

Decker, Scott H. 1996. "Gangs and Violence: The Expressive Character of
Collective Involvement." Justice Quarterly 11: 231-250.

Decker, Scott H., and Susan Pennell. 1995. Arrestees and Guns: Monitoring
the Illegal Firearms Market. Research Preview. Washington, DC: U.S.
Department of Justice. National Institute of Justice. September 1995. NCJ
184205.

Decker, Scott H., and Barrik Van Winkle. 1996. Life in the Gang: Family,
Friends, and Violence. New York: Cambridge University Press.

Esbensen, Finn-Aage, and David Huizinga. 1993. "Gangs, Drugs, and
Delinquency in a Survey of Urban Youth." Criminology 31: 565-587.

Fagan, Jeffrey. 1989. "The Social Organization of Drug Use and Drug Dealing
Among Urban Gangs." Criminology 27: 633-669.

Fox, James Alan. 1996. Trends in Juvenile Violence: A Report to the United
States Attorney General on Current and Future Rates of Juvenile Offending.
Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics. NCJ
170379.

Goldstein, Herman. 1990. Problem-Oriented Policing. Philadelphia: Temple
University Press.

Hagedorn, John. 1988. People and Folks: Gangs, Crime, and the Underclass in
a Rustbelt City. Chicago: Lakeview Press.

Horowitz, Ruth. 1983. Honor and the American Dream. New Brunswick, NJ:
Rutgers University Press.

Hutson, H. Range, Deirdre Anglin, Demetrios Kyriacou, Joel Hart, and Kelvin
Spears. 1995. "The Epidemic of Gang-Related Homicides in Los Angeles
County From 1979 Through 1994." Journal of the American Medical
Association 274: 1031-1036.

Kennedy, David M. 1997. "Pulling Levers: Chronic Offenders, High-Crime
Settings, and a Theory of Prevention." Valparaiso University Law Review 31:
449-484.

-------. 1998. "Pulling Levers: Getting Deterrence Right." National Institute of
Justice Journal 136 (July): 2-8.

Kennedy, David M., and Anthony A. Braga. 1998. "The Youth Homicide
Epidemic in Boston." Paper presented at the Sloan Youth Violence Working
Group meeting, Emory University, Atlanta, Georgia, March 20.

Kennedy, David M., Anthony A. Braga, and Anne M. Piehl. 1997. "The
(Un)Known Universe: Mapping Gangs and Gang Violence in Boston." In
Crime Mapping and Crime Prevention, ed. David Weisburd and Tom
McEwen. New York: Criminal Justice Press: 219-262.

Kennedy, David M., Anne M. Piehl, and Anthony A. Braga. 1996. "Youth
Violence in Boston: Gun Markets, Serious Youth Offenders, and A
Use-Reduction Strategy." Law and Contemporary Problems 59: 147-196.

Klein, Malcolm. 1993. "Attempting Gang Control by Suppression: The Misuse
of Deterrence Principles." Studies on Crime and Crime Prevention 2: 88-111.

-------. 1995. The American Street Gang: Its Nature, Prevalence, and Control.
New York and London: Oxford University Press.

Klein, Malcolm, and Cheryl Maxson. 1989. "Street Gang Violence." In Violent
Crimes, Violent Criminals, ed. Neil A. Weiner. Beverly Hills, CA: Sage
Publications.

Maxson, Cheryl, and Malcolm Klein. 1990. "Street Gang Violence: Twice as
Great, or Half as Great?" In Gangs in America, ed. C. Ronald Huff. Newbury
Park, CA: Sage Publications.

Miller, Walter B. 1977. Violence by Youth Gangs and Youth Groups as a
Crime Problem in Major American Cities. Washington, DC: U.S. Department
of Justice. Office of Juvenile Justice and Delinquency Prevention. NCJ 137446.

Moore, Joan. 1991. Going Down to the Barrio: Homeboys and Homegirls in
Change. Philadelphia: Temple University Press.

Sanchez-Jankowski, Martin. 1991. Islands in the Street: Gangs and American
Urban Society. Berkeley: University of California Press.

Sheley, Joseph, and James D. Wright. 1995. In the Line of Fire: Youth, Guns,
and Violence in Urban America. New York: Aldine de Gruyter.

Sparrow, Malcolm, Mark H. Moore, and David M. Kennedy. 1990. Beyond
911: A New Era for Policing. New York: Basic Books.

Spergel, Irving A. 1995. The Youth Gang Problem: A Community Approach.
New York and London: Oxford University Press.

Spergel, Irving A., and G. David Curry. 1990. "Strategies and Perceived
Agency Effectiveness in Dealing with the Youth Gang Problem." In Gangs in
America, ed. C. Ronald Huff. Newbury Park, CA: Sage Publications:
288-309.

-------. 1993. "The National Youth Gang Survey: A Research and
Development Process." In Gang Intervention Handbook, ed. Arnold Goldstein
and C. Ronald Huff. Champaign-Urbana, IL: Research Press.

Stover, Del. 1986. "A New Breed of Youth Gangs is on the Prowl and A
Bigger Threat Than Ever." American School Board Journal 173: 19-24, 35.

Thornberry, Terrence, Marvin D. Krohn, Alan J. Lizotte, and Deborah
Chard-Wierschem. 1993. "The Role of Juvenile Gangs in Facilitating
Delinquent Behavior." Journal of Research in Crime and Delinquency 30:
55-87.

Tita, George E. 1997. "Where Violent Gangs Hang Out: An Examination of
'Set Space.'" Paper presented at the Annual Meeting of the American
Sociological Association, August, Toronto, Canada.

Wilkinson, Deanna L., and Jeffrey Fagan. 1996. "The Role of Firearms in
Violence 'Scripts': The Dynamics of Gun Events Among Adolescent Males."
Law and Contemporary Problems 59: 55-90.

Winship, Christopher, and Jenny Berrien. 1999. "Boston Cops and Black
Churches." The Public Interest Summer: 52-68.

Zimring, Franklin, and Gordon Hawkins. 1973. Deterrence: The Legal Threat
in Crime Control. Chicago: University of Chicago Press.

---------------------------

Chapter 10

Developing a GIS-Based Regional Gang Incident Tracking System

James W. Meeker, J.D., Ph.D.
Katie J.B. Parsons, Ph.D.
Department of Criminology, Law, and Society
University of California-Irvine

Bryan J. Vila, Ph.D.
Department of Political Science
University of Wyoming

Acknowledgments

Portions of the projects reported here were supported by National Institute of
Justice grants 96-IJ-CX-0030 and 98-IJ-CX-0072; Office of Community
Oriented Policing Services, U.S. Department of Justice, grant
96-CNSX-0019; the University of California-Irvine; the California Governor's
Office; Orange County, California; and the Pacific Mutual Foundation. The
authors wish to acknowledge the contributions of UC-Irvine Focused Research
Group on Orange County Street Gangs research associates Thomas E.
Fossate, Ph.D.; Jodi Lane, Ph.D.; Douglas Wiebe, Ph.D.; and Darcy Purvis as
well as the support of the members of the Orange County Chiefs' and Sheriffs'
Association, the Juvenile Justice and Gang Strategy Steering Committee and its
chair, Chief James Cook of Westminster, Marilyn MacDougall and the Project
No Gangs staff, and David Hartl, Ph.D.

---------------------------

Abstract

Orange County, California, uses a geographic information system (GIS) to
track and monitor gang activity. Called the Gang Incident Tracking System
(GITS), it was intended to help law enforcement officials countywide make
more informed decisions to counter gang activity, which had been on the rise in
the area in recent years. Introduced in 1993, GITS was evaluated to determine
the validity and reliability of the data collected and whether the goals had been
met. The university researchers who partnered with the Orange County Chiefs'
and Sheriffs' Association to conduct the study found that the GITS made it
possible to present a reasonably unbiased and complete picture of gang
incidents and to establish a baseline of gang activity. Among the components
evaluated were the OCCSA's definition of gangs, the form used to collect data,
the process by which gang information is reported, the validity of measures
used to count gang incidents, the reliability of data coding, and the process for
supervisory review of incident reporting. Of particular concern to community
activists was whether the police overestimated gang crime. The evaluation
revealed that, far from overestimating this type of crime, the police tended to
underreport it. The GITS system above all demonstrated the usefulness of
multijurisdictional efforts to understand and ultimately prevent gang crime.
Output has been used by law enforcement to deploy personnel, allocate
resources, and evaluate gang intervention activities. The GITS system has made
it possible to use GIS technology to create computerized maps that analyze the
spatial and temporal distribution of gang activity.

---------------------------

In recent years, many communities that previously considered themselves
insulated from inner-city problems have been forced to acknowledge that gang
violence can extend into their own neighborhoods (Curry, Ball, and Fox 1994;
Spergel and Curry 1995). Orange County, California is one such community.
Located 40 miles south of Los Angeles, Orange County is a highly
heterogeneous suburban county with 2.7 million people living in 31 cities and
unincorporated areas. Since 1980, it has experienced rapid growth, increasing
urbanization, and racial and ethnic change. Despite a few traditional Hispanic
"turf" gangs firmly entrenched in its less affluent areas (see Vigil and Long
1990), Orange County historically has enjoyed low crime rates and relative
tranquility.

During the past decade, however, gang activity appears to have been on the
rise in the county. In 1991, the Orange County Grand Jury reported that gang
problems were escalating at an alarming rate, a finding repeated by the 1995
Orange County Grand Jury. According to police and media reports, gang crime
in Orange County had become not only more frequent, but more violent. More
mobile Asian gangs and white "skinhead" gangs emerged in the county, along
with a growing number of more traditional turf-oriented gangs.

Orange County residents have also become increasingly concerned about gangs
and crime. A 1994 survey[1] found that 75 percent of residents were aware of
gang problems in their community, and 61 percent thought that gang activities
had increased in the past few years. In addition, for the first time in 1993 and
then again in 1994, the Orange County Annual Survey[2] found that residents
considered crime--once an issue of little concern to them--to be the county's
most serious problem.

In response to increases in violent gang activity in neighboring Los Angeles and
growing public concern with gang activity, the Orange County Chiefs' and
Sheriff's Association (OCCSA) established a countywide Gang Strategy
Steering Committee (GSSC) in 1992. Joining forces with school districts, local
government agencies, community groups, and businesses, the county's 22 law
enforcement agencies developed and implemented an unprecedented
community-based, multiagency effort to address gang violence.

Since then, following the recommendations of the California State Task Force
on Gangs (California Council on Criminal Justice 1989: 37, 57), OCCSA
launched a comprehensive set of programs:

o Project No Gangs, a countywide community education and awareness
prevention program aimed at mobilizing community resources to fight the
influence of gangs.

o TARGET, a suppression program in eight cities in which law enforcement,
probation, and prosecution staff target hardcore gang leaders and repeat
offenders through vigorous surveillance and prosecution (Kent and Smith
1995).

o The Gang Incident Tracking System (GITS), designed to document the extent
of gang-related crime in the county and provide information for strategic
planning and evaluation purposes (Vila and Meeker 1997).

Although OCCSA had laid the foundation for interagency coordination and
data collection, it lacked the analytical resources and expertise to fully evaluate
and monitor the effectiveness of these programs. Early in 1995, OCCSA asked
the University of California, Irvine (UCI) to enter into a long-term partnership
to enhance its analytical capabilities. In keeping with its tradition of bringing
innovative research techniques to bear on important community problems, the
UCI School of Social Ecology established a Focused Research Group on
Gangs (FRG) within its Department of Criminology, Law, and Society to help
OCCSA and GSSC resolve a number of previously unanswered questions
about gangs, gang crime, and their effects on the community and assist in the
development of strategies to prevent and control illegal gang activity. This
chapter will focus on the results of FRG's evaluation of the GITS program.

The GITS Program

The goals of Orange County's GITS were to accurately identify the extent of
gang-related crime in Orange County, establish a baseline against which to
identify future trends in gang-related crime over time, and determine regional
variation in gang-related crime patterns. This information then could be used by
county law enforcement agencies to facilitate strategic planning and improve
resource allocation for controlling gang activities.

GITS became operational on January 1, 1993, when county law enforcement
agencies began reporting all gang-related incidents to a centralized database.
By the end of 1993, all 22 independent cities and the Orange County
Sheriff-Coroner Department (which serves contract cities and unincorporated
areas) had established relatively consistent internal procedures for identifying
and tracking gang-related crime and were reporting to GITS. Training
programs and a short training videotape were used to teach patrol officers
countywide how to identify and report gang-related incidents to GITS. GSSC
declared publicly that the level of gang crime in 1994, as determined by GITS
data, would be the benchmark against which the success of future law
enforcement gang-control activities would be judged.

To help assure that different agencies collected gang data consistently, the
GSSC definition of the term "gang" closely followed that used in California's
Street Terrorism Enforcement and Prevention (STEP) Act:

[A] group of three or more persons who have a common identifying sign,
symbol or name, and whose members individually or collectively engage in or
have engaged in a pattern of criminal activity creating an atmosphere of fear and
intimidation in the community. (California Penal Code section 186.22)

For a crime to be considered gang related, one of the following criteria had to
be met:

o A suspect or suspects are identified as gang members or admit membership in
a gang.

o A person becomes a victim due to his or her gang association.

o A reliable informant identifies an incident as gang activity.

o An informant of previously untested reliability identifies an incident as gang
activity, and this identification is corroborated by other independent information.

Crimes that do not fit the above criteria but have strong indications of gang
involvement (e.g., suspects display gang hand signs or the incident fits the
profile of gang incidents, such as driveby shootings or home-invasion robberies)
may also be included.

Orange County's definition follows what is often called the Los Angeles or
"gang involved" model for defining gang crimes, which is based on the
assumption that "the character of an individual as a gang member, regardless of
the types of criminal situations in which he or she is found" (Spergel 1992: 137)
is what is important for defining gang crime. Under the Chicago, or "gang
motivated," model, for a crime to be considered gang related, it must "grow out
of gang motivation or interest and enhance the status or function of the gang as
a group or organization" (Spergel 1992: 137) and must be motivated by issues
of retaliation, territory, recruiting, or "representing"[3] (Maxson and Klein 1990:
26; See also Block et al. 1996: 11-12). The limited research available on this
issue suggests that both definitions produce similar patterns in factors associated
with the crime (see Maxson and Klein 1990, 1996). The advantage of
collecting gang-involved (rather than gang-motivated) data is that these data
give law enforcement a better idea of the types of criminal activities that gang
members are committing, not only in a group setting but also individually, and
allow researchers to compare gang crime with other types of crime, such as
juvenile crime. Perhaps the most important argument in favor of the Los
Angeles model is that gang-motivated crimes can be obtained from
gang-involved data but not the other way around. From a practical standpoint,
it is comparatively easy to get gang-related crime data from first reports of
officers in the field, whereas the motivation behind an act is often difficult to
ascertain. Finally, people in communities with substantial gang problems tend to
be more concerned about the crimes committed by gang members than about
the motivation for these crimes.

The GITS Data Form

The GSSC created the original GITS data form (exhibit 1), which had several
key elements:

o Data collected were acquired from police reports. Each incident in the
database could be traced back to its original police report if further information
or validation was needed.

o Rough geographical location data were collected for each incident, but at a
coarse scale (half-mile squares) that severely limited their usefulness for spatial
analysis or program evaluation.

o Only 21 crime categories were listed, many of which tracked offenses
enumerated in California's Street Terrorism Enforcement and Prevention
(STEP) Act (CPC section 186.22). This limited the ability of the database to
measure the amount of gang crime and made comparisons with such measures
of crime as the FBI's Uniform Crime Reports (UCR) difficult.

o Information about characteristics of crime incidents, such as weapons used
and motivating or precipitating factors, was collected, as was information about
drug and alcohol use.

o Information about victims was collected, but the categories were not mutually
exclusive, which made it impossible to code victim data reliably.

After reviewing the original GITS data collection and analysis system, the FRG
research team reviewed the 1994 and 1995 data, and address data were
added for all these incidents. During this review process, law enforcement
agencies were asked to conduct their own careful review of all police reports
involving shootings, robberies, assaults, and weapons-law violations to identify
possible gang involvement. This enabled the evaluators to estimate the
proportion of the most serious gang crime incidents captured by GITS.

A new GITS data form was developed that corrected a number of problems
with the original form; it was adopted for use in 1996. The new form (exhibit 2)
is streamlined and reformatted for greater clarity to improve the reliability with
which data are coded. It also includes detailed instructions (exhibit 3) and
collects more--and more detailed--information than the original. For example,
the
new form asks for the street address of each gang incident so crimes can be
mapped in fine detail, and up to four criminal code violations can be listed as
they appear on police reports. The new form also tracks whether victims were
gang members and whether their victimization was intentional or unintentional.

Not surprisingly, the gross number of gang incidents reported in 1996 increased
by one-third once the system began tracking all gang-involved crimes instead of
being limited to the 21 listed in the original form. Only 3,384 of the 4,500
incidents reported in 1996 fit within the original 21 crime categories. Moreover,
capturing data on multiple counts made it possible to collect information on
6,134 gang-involved crimes, whereas the old form would have tracked only 55
percent of these crimes.

The GITS Reporting Process

Information on gang incidents is entered into the GITS database in the following
way: When officers indicate on their field incident reports that they believe an
incident is gang related, it is reviewed for possible inclusion in GITS by the
person in each department responsible for reviewing such reports. If the review
process is working properly (see the section on Validity of Gang Incident
Measures, below), every field incident report, including those not marked by
officers as possibly gang related, is reviewed again for gang-relatedness.
Incidents judged to be gang related are then coded onto the GITS form by
individuals assigned that task in each department and forwarded to the research
team at UCI. There, the forms are reviewed for completeness and potential
errors. Forms that fail this screening are returned to the department for
correction; those that pass are entered into the GITS database. Once data have
been entered, a final check ensures that they were entered accurately.

Plans are beginning to be made to remove UCI from the data-entry and
management side of GITS by adding case-management capabilities to a
database currently being used to track gang members. This case-management
system would also provide basic reports on such data as number of crimes,
weapons used and recovered, and victim information. UCI would then be able
to focus on more sophisticated reports, analysis, and mapping.

Validity of Gang Incident Measures

Several issues associated with validity were addressed in this project. Since
GITS definitions of the terms "gang" and "gang related" followed California law,
their reasonableness (i.e., face validity) did not appear to be a serious concern,
even though the criteria are broader than the gang-motivated criteria used in
some jurisdictions. As discussed previously, the California definition
encompasses more information and offers the flexibility of extracting a narrower
subset at a later date. Another concern was measurement validity--whether
incidents reported by police agencies as gang related actually fit GITS
criteria--and whether the GITS system itself met the objectives originally set by
OCCSA and the participating law enforcement agencies.

As part of the process of validating the measures, four groups responsible for
collecting information on gang-related activity in Orange County were
identified. The procedures used by these groups--officers, reviewers, data
coders, and records departments or bureaus--were assessed via direct
observation, tests, and interviews.

The reliability and validity of GITS data were assessed at four key processing
stages:

o At the identification stage, an officer responds to an incident or crime. What
the officer sees and chooses to include in the incident report aids in the initial
identification of an incident as gang related.

o At the review stage, someone verifies whether previously identified incidents
were actually gang related and reviews unidentified incidents for possible gang
relatedness.

o Data coding is the final stage that takes place within a police agency. At this
stage, a GITS form is completed based on information contained in an offense
report or incident report.

o Actual data entry takes place in the FRG office at UCI.

Several threats to validity were identified at the identification, review, and
data-coding stages. Internal procedures, gang indicators, and definitions were
examined with special care because of their potential to bias results. The
primary threats to data-coding validity and reliability considered in the
evaluation were inconsistency and the lack of proper training.

After a review of each of the 22 participating law enforcement agencies'
reporting procedures, three reporting models were identified that described
how gang incidents were identified, reviewed, verified, and recorded: the
General Supervisor model, the Gang Unit model, and the Reviewer model.
Evaluators tried to determine whether these models differed with regard to the
likelihood of either incorrectly identifying a nongang incident as gang related or
failing to identify a gang-related incident during the review of incident reports.
The biggest difference among these models is in the position of the person
responsible for the review stage within the organization (e.g., watch
commander, gang-unit supervisor, records clerk). The review stage is critical
because GITS depends heavily on the ability of reviewers to identify
gang-related incidents accurately.

In all three models, an original crime report may remain in the patrol division or
be routed to many different units or divisions in a department; it will be routed
to the proper person for completion of a GITS form only if the reviewer
correctly identifies an incident as gang related. A strength shared by all models
included initial identification by patrol officers who are often highly qualified to
determine whether a gang member was involved because they have frequent
contact with gang members and firsthand experience with the incident. A
second strength was that all models included a review step that attempted to
verify the field officer's initial determination and check for missed cases.
Weaknesses shared by all three models include substantial variation among
officers in identifying incidents as gang related, report distribution and routing
problems within departments, inadequate training of data coders, and late
submission of GITS report forms.

The General Supervisor Model

In this model, a general supervisor or watch commander who is trained in law
enforcement, but generally has no specific training in gang enforcement, reviews
all crime reports for gang-related activity and then either completes a GITS
data report or forwards the report to someone else to complete the form. The
main strength of this model is that the reviewer has the authority to request
reports and to require that they be filled out completely and accurately. Its
weakness is that it is subject to distributional and procedural problems. General
supervisors have other priorities and sometimes fail to apply definitions and
follow reporting directions accurately. As in other models, reports are often
directed out of the loop by being assigned to other departments before they
have been reviewed to determine gang relatedness.

The Gang Unit Model

In this model, a supervisor from the gang unit who is trained in law enforcement
and is also a gang expert reviews all crime reports to determine whether any of
the incidents are gang related, and then someone in the gang unit completes a
GITS data report form. This model's strength is that the gang-unit supervisor
has up-to-date knowledge about local and regional gangs. Its weaknesses are
that the gang-unit supervisor may not receive all reports to review and that the
gang unit may have more pressing matters than reviewing cases and completing
GITS forms.

The Reviewer Model

In this model, an outside reviewer examines all crime reports for indicators of
gang-related activity and completes a GITS form. The reviewer is usually a
member of the department's support staff, a clerk, or a cadet. The main
strength of this model is that the reviewers tend to do a complete and thorough
job. Although the reviewer may not have any specific training, after working
with GITS for a time, he or she becomes quite consistent and accurate in
review procedures. The reviewer also is the same person who completes the
GITS forms. Distributional problems do not affect this model because the
review occurs at the place where all reports are usually collected--the records
section. The greatest weaknesses of the model are that the reviewers are
completely out of the power loop, often have many other responsibilities, and
have little, if any, law enforcement training and no specific gang training.

Of the three models, the gang unit model often provides the greatest validity.
Unfortunately, GITS responsibilities are often a low priority for gang-unit
supervisors, which tends to undermine the reliability of GITS reporting and the
timeliness with which reports are submitted. An experienced reviewer who is
attached to a gang unit or involved in gang enforcement duties for a prolonged
period is more timely in filling out GITS reports than a gang-unit supervisor and,
if trained properly, can make equally valid assessments regarding report
classification. Therefore, in spite of its inherent weaknesses, more and more
agencies are adopting the reviewer model.

Validity Assessment Methods

It was especially important to know whether personnel in each of the reporting
agencies understood and applied the administrative definitions of key concepts
in this project. Another concern was that internal procedures and directives that
affect when an officer should identify an incident as gang related might threaten
valid identification by officers. If the administrative definitions were not applied
uniformly across jurisdictional boundaries, data from different agencies would
not be comparable.

The accuracy and completeness of crime reports affected reviewers' ability to
carry out their responsibilities. In some agencies, reviewers were responsible
for identification as well as review. Departments designated people for this
responsibility in different ways; reviewers included supervisors, cadets, senior
volunteers, and interns. The validity of a reviewer's decisions may also be
limited because of internal procedures, gang indicators reported by police, and
definitions. The reviewer must depend on crime reports and narratives to
identify gang-related incidents correctly. Thus, it was imperative to know
whether reviewers were relying on officers to check a "gang related" box on
incident reports or taking the time to read incident report narratives to obtain
clues regarding gang-relatedness. Several research strategies--observation of
participants, review of a large sample of police reports, and experiments with
data coders--were used to assess how these threats affected GITS reporting.

Observations Regarding Use of Definitional Criteria

Project staff conducted more than 50 hours of ride-alongs with gang unit and
patrol officers to assess how officers distinguished an incident as gang related
and to identify the reporting criteria they actually applied. Ride-alongs began
with semistructured interviews with officers that enabled evaluators to assess
the extent of definitional inconsistencies among officers and agencies across the
county. Similar interviews were conducted with supervisors and others
responsible for review.

Field observations and discussions with police, supervisors, and chiefs indicated
a problem with how the definition of "gang related" was applied in various
jurisdictions. Although GSSC had officially adopted a "gang-involved"
definition, in some cases, officers and jurisdictions applied more restrictive
criteria. This became apparent after the report form was expanded to use State
Penal Code offenses rather than the original 21 crime categories. Many officers
and departments objected to including domestic violence and sexual assault
crimes on GITS reports because these crimes were not traditionally considered
gang motivated, so few are included in the database. Conversely, robberies by
known or suspected gang members are routinely entered even if they were not
motivated by gang interests. Although training sessions routinely stressed that
the "gang involved" definition was to be used, evidence suggests that it was
more consistently applied to the original 21 categories and to crimes
traditionally thought to be associated with gang behavior.

These ride-alongs also highlighted other cross-jurisdictional reporting
differences, especially for less serious gang-related crimes such as tagging and
vandalism. Although more serious violent crimes were reported quite
consistently across jurisdictions, jurisdictions with higher levels of violent gang
crime tended to be more lax in reporting less serious gang incidents.
Jurisdictions with less of a gang problem tended to be more diligent in reporting
minor property crimes.

After reviewing the data, no evidence was found to support minority activists'
allegations that the police were drastically overestimating the amount of
gang-related crime. From 1994 through 1997, the number of gang incidents
remained relatively constant--at approximately 3,500 incidents per year (for the
original 21 offenses). This is a relatively small number for an area with a
population of 2.7 million and is only a fraction of the UCR crimes recorded for
these categories.[4]

These interviews showed that records departments were common chokepoints,
from which gang-related cases were not circulated to the appropriate personnel
or departments for completion of a GITS coding sheet (see Souryal 1981). In
particular, cases sometimes were lost because of what Weston (1978) called
"horizontal communication barriers"; gang units, for example, sometimes failed
to receive robbery reports because the reports were assigned to robbery
detectives and no copy was forwarded to them. This problem could be
circumvented informally through "the grapevine," but the quality and amount of
information passed along in this manner varied greatly. For example, a gang
officer could hear that Joe Smith (a known Alpha gang member) was arrested
the previous night by the vice unit. Another gang officer could hear that Ronald
Smith (a known Beta gang member) was arrested the previous night by the vice
unit. But unless the gang officers followed up on the rumors or the gang unit
received an incident or arrest report, GITS would never have a record of that
gang-involved incident.

For the most part, only slight differences were found in how officers and
reviewers across the county defined "gang" and "gang related" crime. The
formal countywide definitions were followed quite closely and, without
exception, gang-unit officers could recite them word for word. Posters with the
definition were found in report-writing rooms, break rooms, and briefing rooms
in each agency. Yet several gang-unit officers expressed the opinion that patrol
officers probably did not understand the nuances of the definitions of "gang
involved" versus "gang motivated" as well as did experts like themselves.

This perception was troubling because analysis of internal procedures indicated
that agency review procedures were often not followed, and an officer's
preliminary classification of an incident sometimes tended to have more
influence on reviewers than was supposed to be the case. Especially in the
larger agencies, where hundreds of cases per day needed to be reviewed for
gang-related activity, review was often minimal and limited to cases marked by
police officers as possibly being gang related or cases requested by reviewers.
Some individuals with review responsibilities were unaware of GITS or of their
own role in collecting gang data for their agencies. Limited or missed reviews
could result in missed or misclassified cases. The results of the data validation
described below, however, appear to indicate that the number of missed cases
involving violent and weapons-related incidents was quite small, although there
may have been more missed cases involving property-related offenses.

How Incidents Were Classified

Ride-alongs and interviews with patrol and gang-unit officers and reviewers
were also used to identify the extent to which internal procedures, definitions,
and training affected the internal consistency with which these individuals
classified incidents as gang related. Gang officers were asked to explain the
criteria (e.g., dress, tattoos, age, ethnicity) they used to determine whether an
individual is a gang member (see Piliavin and Briar 1964; Quinney 1970; Smith
and Visher 1981). Common descriptors of gang members related by officers
and reviewers during interviews and ride-alongs included shaved heads, baggy
jeans, tennis shoes, and tattoos. More important than what they wore,
however, was how they wore it. Attitude, stance, and walk were all important
supporting indicators. All gang officers who described the appearance of gang
members were careful to note that these attributes do not necessarily mean that
an individual is a gang member but only give officers a reason to stop, observe
more closely, and ask questions. Officers seemed well aware that these types
of clothing are quite popular among young people and that juveniles who are
not involved in gangs often wear "gang attire."

Supervisors and those assigned to review incidents to identify gang involvement
were also interviewed to determine which indicators they employed when
screening and reviewing incident reports. These semistructured interviews
helped researchers judge whether supervisors and reviewers in different
agencies were applying similar criteria. It was found that those responsible for
review depended on quick glances at crime reports to classify an incident as
gang related and that such classification often required more than one indicator.
Supervisors and reviewers looked for such things as the location of the crime
(e.g., whether it took place in a known gang area) and clothing descriptions.
The names of the suspect and victim also were considered good indicators of
gang-relatedness. If gang-unit officers were performing the review, they often
quickly recognized the names of gang members. Other reviewers entered
names on a crime report in the Orange County District Attorney's Office's Gang
Reporting Evaluation and Tracking (GREAT) database, which lists the names
and affiliations of known gang members, to determine whether anyone involved
in an incident was a known gang member. The modus operandi of a particular
crime could also serve as an indicator. For example, driveby shootings and
home-invasion robberies were often attributed to gangs. In some departments,
suspect description was vital (e.g., if the victim described the suspect as a
"cholo" or "gang member type"). Some reviewers also took the time to read the
narratives and look for such indicators as admission of gang membership or the
question "Where are you from?" which is often asked of strangers by gang
members.

Officers and reviewers were found to be quite consistent in the criteria they
applied to classify incidents as gang related throughout the county and across
reporting models. This consistency may be due to the extensive training that
officers receive from the Orange County Gang Investigators' Association and
from gang seminars conducted by the District Attorney's Office.

Pre-Evaluation Attempts to Validate GITS Data

Before UCI became involved in the project, a consultant supervised a
"validation check" of GITS data for the first 6 months of 1994. Although the
process was more of a verification than a validation exercise, a brief description
is included here to provide a baseline against which to compare later data
quality. Four areas were addressed in the verification: data acquisition,
aggregation, analysis, and reporting. Agencies were asked to verify that GITS
data reflected both the original tracking form submitted for data entry and
original field incident reports submitted by patrol officers, and that the incident
fit California's gang-related criteria. No agency was asked to verify more than
30 incidents. The error rate reported was approximately 16 percent. Although
the consultant's report did not specify which type of problem contributed most
to the errors, subsequent interviews suggest that most of the problems were
with data entry. The data were never validated to determine whether police
were measuring what they thought they were measuring or whether GITS data
provided an accurate measure of the gang problem known to police. Only
cases already determined to be gang related were verified; cases that had not
been identified as gang related were not rechecked.

Validation of 1994 and 1995 Data

To improve the validity of GITS data and the reliability with which they were
reported, UCI evaluators took steps to validate 1994 and 1995 data by
examining a random sample of all police reports in participating agencies.
After a review of each agency's procedures, a validation strategy was
developed for each department. The evaluation focused on the four crime
categories--shootings, robberies, assaults, and weapons-law violations--for
which the most incidents had been reported up to 1995 in all of the
departments. Depending on the size of the agency, the extent of reported gang
problems, and the apparent reliability of tracking procedures, agencies were
required to review all gang-related cases, review a random sample of all reports
within the four crime categories (not just gang-related reports), count all
gang-related cases, or engage in a combination of two or more of these
activities for 1994 and 1995. The validation was completed in early 1996.

When UCI took over the data collection, the data had not been cleaned or
analyzed for 6 to 8 months. Agency staff performed the validation described
above with help from project staff. For 1994, 8,295 cases were reviewed and
only 283 gang-related cases were found that needed to be added to GITS--at
worst, an error rate of 3 percent. For 1995, 4,302 cases were reviewed and
only 216 were added--at worst, an error rate of 5 percent. These error rates
did not differ significantly by department, type of crime, gang, or any other
identifiable variable.

Data Coding Reliability Tests

Researchers sent gang-related incident reports to data coders in each agency
who were responsible for completing GITS reports and forwarding them to the
research team. As is evident in the original and revised GITS report forms
(exhibits 1 and 2, respectively), coders were required to extract a substantial
amount of information from each incident report and exercise a good deal of
judgment to decide such things as the category of crime, the relationship
between victim and offender, and the motivation for the incident.

Experiments devised to test data-coder reliability were administered in
February 1996, October 1996, and October 1997. Since the goal of the
experiments was to identify and rectify problems, data coders were advised
about problems after each experiment and trained to remedy them.

The experiments required that data coders review hypothetical narratives based
on actual crime reports and fill out GITS report forms. Each experiment
consisted of three narratives, one of which was retained in slightly modified
form for all three experiments. The retained narrative was used to test the
consistency of individual raters over time. The other narratives varied with each
iteration of the experiment and were designed to test for consistency among
raters, both within and among departments.

Coders from the same department often found it difficult to remain consistent
from one experiment to the next. Much of this inconsistency appears to be the
result of substantial turnover among data coders. Not surprisingly, experienced
coders were much more reliable than inexperienced ones, and departments with
at least one experienced coder tended to be more reliable because new coders
were trained by experienced coders. The importance attached to the GITS
project by a department also appeared to affect coder reliability: In
departments where GITS and gangs were a high priority, coder reliability was
high and coders tended to be retained for longer periods; where GITS was a
lower priority, coder reliability suffered. Many experimental GITS forms were
rejected because they contained errors--on average, 34 percent across the
three experiments. A similar proportion of the coders made one or more errors
in interpreting the narratives. Most of these errors, however, involved rather
minor distinctions in areas of the form dealing with detailed arrest and victim
information. On substantive recording issues likely to influence policy, such as
specifying the proper crime, time of day, and location, strong reliability was
found irrespective of coders, departments, and time.

Improving GITS Data Collection

These experiences with the collection of gang-incident data from many law
enforcement agencies in a large metropolitan region highlight the importance of
coder and reviewer training and regular retraining targeted at specific problems.
Another key problem was the need to increase the involvement of line officers
and gang units in data collection. If people in these positions were more
committed to GITS, it seems likely that undercounting would be diminished,
perhaps substantially. Thus, evaluators found it important to inform line officers
and gang units about GITS, how they fit into the project, and how it could
benefit them. These groups also needed feedback from the project, something
that some chiefs elected not to provide--usually because they either viewed
GITS reports as providing strategic management information beyond the
purview of line officers or feared that politically sensitive gang information might
find its way into the hands of political opponents or the media. The evaluation
team's position on this issue was that both research and management
imperatives justified increasing the stake that reviewers and line officers had in
GITS by making it more useful to them.

The Goals Evaluation

Another important evaluation issue was to determine how well GITS fulfilled the
original goals for the program set by GSSC. Such tasks are notoriously difficult
to accomplish because evaluators must be more like detectives than social
researchers, trying to unearth and sort out the original reasons for the program
(Rossi and Freeman 1993: 106). As previously explained, many agencies were
involved in creating GITS, including various branches of Orange County's
criminal justice system (law enforcement agencies, the District Attorney's
Office, and the Probation Department), Federal law enforcement
representatives, and local school superintendents. It seemed likely that some of
these agencies may have had different agendas in planning and approving a
multijurisdictional gang incident tracking system.

Understanding how the goals of these agencies were transformed into the goals
for the database was an essential step toward evaluating GITS. Understanding
whose goals and whose agenda the committee gave the most weight to was
also important. As one of the system's founders explained, "Keep in mind the
whole context of this being a compromise all the way around. . . . [A]ll of [the
agencies involved] have different priorities. In one respect, I think it is rather
remarkable that they all agreed to do something together" (personal interview,
July 6, 1995).

The formal goals established for GITS and the goals held by individual
participants early in the development process were both examined. Project
researcher Katie J.B. Parsons used snowball-sampling techniques[5] to identify
public officials who were involved in the initial stages of GITS development and
then interviewed them about their expectations for GITS, the goals they
believed GITS would accomplish, and their reasons for believing that a gang
tracking system was necessary. Interviews were taped and transcribed for later
analysis, although interviewees had the opportunity to refrain from being taped
for the entire interview or for specific questions that they did not feel at ease
answering on tape. Only one interviewee asked to go off the record for some
responses. Notes were also taken both as backup in case the tape recorder
failed and to provide context for untranscribable portions of the interview.
Parsons also made field notes on perceptions or feelings about interviews
immediately after they were completed. Other archival materials, including
planning documents, research memos, and meeting notes, were also cataloged,
inventoried, and analyzed to identify program goals and objectives. This was
done to track the evolution of goals and decisions about GITS and its mission.

The evaluation team also critiqued the quality of the goals themselves. GITS has
attracted much attention, and several jurisdictions have expressed an interest in
adopting something similar. Thus, evaluators believed that it was particularly
important to clarify and strengthen goal statements. Nakamura and Smallwood
(1980) argue that goals must be clear and specific, because vague goals can be
misinterpreted and manipulated. They should also be technically and logically
consistent and include detailed descriptions of how they are to be
accomplished.

Goal Findings

Nineteen police chiefs[6] and a research consultant who worked on the project
before the evaluation were interviewed individually, as were the GSSC
coordinator and the District Attorney.

Initially, local police chiefs who had worked in Los Angeles and were familiar
with that county's attempts to track gang crime were recruited by GSSC to be
involved in the planning process. The original planning committee included these
and other chiefs interested in research, gang experts, and representatives from
the Sheriff's and District Attorney's Offices. This committee made several
limited attempts to discuss the proposed project with Los Angeles and Los
Angeles County and other area agencies that were attempting to tackle similar
problems, to ascertain what those agencies were doing and what kinds of
information they were collecting. Because the committee members were
high-level administrators, GITS's goals and objectives reflected strategic and
administrative interests rather than tactical concerns. Tactical information was
already available to line personnel through the GREAT (later Cal/Gang) system,
an investigative database containing information about known gang members.
Other GSSC participants accepted these goals and objectives.

Interviews with GSSC participants show that the main goal was to create a
comprehensive countywide gang crime database that would help produce a
snapshot of the shape, size, nature, and scope of the gang problem in Orange
County so management could make more informed decisions concerning such
gang issues as personnel deployment, resource management, and informing the
public on countywide trends. As one chief stated:

I think one of the reasons the OCCSA became so involved in wanting to track
gang activity was that we weren't sure what was happening around us or why
gangs were doing what they were doing. We couldn't respond to our
community leaders and residents. We were guessing a lot until we got involved
with GITS and other gang initiatives that looked at what was happening and
why it was negatively impacting our communities. We recognized that we
needed more information to make better decisions. (Chief interview, February
18, 1997)

Written statements concerning the program's overall goal were found in archival
materials. The only formal written program description stated that GITS was to
function as a management tool to facilitate strategic planning and resource
allocation (Smith 1994). Regional GSSC meeting notes suggest that "the
purpose of the program [was] to develop a quality database from which we can
develop management decisions" (Hartl 1993). GSSC minutes call GITS a
"snapshot" statistical profile taken on a monthly basis for management
information purposes--not a system for continuous recordkeeping and reporting
(Hartl 1994a). In August 1994, GSSC updated the overall county gang
strategy policy to include "findings compiled and analyzed for purposes of
providing management information to law enforcement officials to combat gang
violence and gang activity" (Hartl 1994b).

This goal specifically described the intended uses of GITS and whom it was
designed to serve. It also clearly indicated that the database was intended to
provide strategic intelligence for law enforcement managers so they could make
better decisions. GSSC also discussed whether and how this system could be
used by individual gang units for investigative purposes. For example,
investigators and middle managers might have used GITS data to decide where
to focus gang unit patrols and which gangs to focus on (GSSC interview,
March 1, 1996).Yet because the timeliness of data submission varied
substantially from agency to agency (many were months late in submitting
reports), GSSC members considered the database to be of limited use for
tactical decisions requiring up-to-date information.

In early 1993, OCCSA accepted the following series of objectives for the
GITS program:

o Establish a centralized database into which all Orange County law
enforcement agencies would report gang-related crime.

o Accurately identify the extent of gang-related crime in Orange County by
identifying gangs that operate within the county, their membership, and the
crime related to their activities.

o Establish baseline data on gang activity against which to compare future
trends in gang-related crime.

o Determine regional variations in gang-related crime patterns.

Originally these objectives were taken from several documents and drafted by a
consultant to the GSSC based on minutes from several committee meetings
(personal interview, July 7, 1995). Other objectives found in the files included a
research proposal stating: "The purpose of this procedure is to establish a
uniform method of reporting gang related crimes to a central statistical database
in order to provide a clear profile of the gang crime problem to Orange County
law enforcement agencies" (Smith 1993).

As often happens in real-world planning, formal goals for GITS were
established well after the development process began. This can be less than
optimal because without clearly articulated goals to guide the development
process, it is easy for a project to diverge from its original purpose. As the
following discussion indicates, however, there was no evidence of substantial
problems with GITS goals.

The formal objectives were examined, both on their own and as a group, for
technical and logical inconsistencies.

Objective 1: Establish GITS

The first objective was to "establish a centralized database into which all of
Orange County law enforcement agencies would report gang-related crime."
Although this objective was achieved--GITS was established--it neglected to
address an important issue: maintenance of the database. Another minor
inconsistency was that not all the county's law enforcement agencies contribute
to GITS, because some of them, such as the Marshal's Office, the Probation
Department, and the FBI, seldom handle gang incidents.[7] A better
restatement of this objective would be to "establish and maintain a centralized
database into which all of Orange County policing agencies will report
gang-related crime."

Objective 2: Determine the Extent of Gang Crime

The second objective of GITS was to "accurately identify the extent of
gang-related crime in Orange County by identifying gangs that operate within
the county, their membership, and the crime related to their activities." GITS, as
ultimately established, tracks only crime related to gangs and the activities of
gang members. This objective includes duties that are currently performed by
another system. The GREAT and (more recently) Cal/Gang systems are
charged with "identifying gangs that operate within the county, their
membership, and (more recently) the crime related to their activities." GITS
objectives should focus only on GITS duties and ideally should state: "Identify
the extent of gang-related crime in Orange County by tracking gang-related
crime through official police offense or incident reports."

Objective 3: Establish a Baseline for Trend Comparison

The third objective was to "establish baseline data on gang activity against
which to compare future trends in gang-related crime." Had this specified the
trends to be tracked, it would have been easier to develop reporting forms. It
also seems likely that the forms would have tracked fewer types of
information--something that would have tended to improve reporting accuracy.
Still, the database unarguably has been useful for trend analysis. The research
consultant conducted a preliminary analysis of gang-crime trends shortly after
GITS began collecting data. More recently, the evaluation team has conducted
elaborate statistical and geographic analyses using the data. As the previous
sections of this report demonstrate, GITS data make it possible to identify local
and regional trends in the nature and distribution of gang incidents
geographically and over time. Overall, this objective is acceptable as written,
but other agencies may want to consider limiting the scope of their collection to
economize and, perhaps, improve reliability.

Objective 4: Determine Regional Variation in Gang Activities

The previous objective required that GITS "determine regional variation in
gang-related crime patterns." Gang-related crime patterns refer to groups of
gang offenses that share common attributes or characteristics: geographic
location, gang identifiers, or crime measures (see Chang et al. 1979). GITS
produces countywide information semiannually on crime patterns, juvenile and
adult arrests, the relationship between victims and offenders, and weapon
involvement. Yet the phrase "regional variation" has proven to be a source of
substantial contention among the chiefs of police--and between the chiefs and
project researchers. Some interpret "regional" to mean city-by-city analysis,
whereas others consider groups of cities or sections of the county as
appropriate units of analysis within the region.

If any single issue manages to undermine the astounding level of cooperation
among the jurisdictions involved in GITS, it will be the question of how data are
to be reported. Members of OCCSA who are most concerned about this issue
fear that publication of maps or other data that make it possible to compare
levels of gang incidents within or between different jurisdictions could have
grave political, economic, and research implications. Some chiefs are
concerned that their cities will be improperly or unfairly compared with
neighboring jurisdictions with lower reported rates of gang incidents. Such a
comparison, they fear, could have dire consequences for tourism, real estate
values, and retail businesses in their cities. They are also concerned that such
comparisons would tend to penalize chiefs whose departments meticulously
reported gang incidents, while casting a favorable light on those that
underreported.

Other chiefs, even several whose jurisdictions had substantial amounts of gang
crime, respond that the only appropriate strategy is openness about the extent
of the problem. They argue that every analytical tool available should be used to
understand the causes of gang problems and determine which strategies under
their control are most effective and efficient. As researchers, of course, the
authors of this study favor this latter stance, although they are bound by the
memorandum of understanding between OCCSA and UCI, which restricts the
release of any city-specific data without the permission of that city's chief of
police.

Although the fears of inappropriate jurisdictional comparisons have not been
realized, the chiefs' unwillingness to be more open with the data has diminished
the data's utility. On one occasion, it even prevented the sharing of data
between two jurisdictions that were involved in a political conflict. Eventually
the conflict ended and the gang data were shared.

It is important, however, to recognize that GITS may never have been
implemented if this objective had been stated more clearly, given the various
chiefs' widely divergent views on how the data should be controlled. Other
jurisdictions implementing similar cross-jurisdictional programs should do their
best to develop a compromise acceptable to all participants early in the
planning process, to avoid the problems encountered in Orange County.

Expected and Unexpected Benefits

As expected, GITS data were being used by law enforcement to deploy
personnel, allocate resources, and evaluate gang prevention, intervention, and
suppression activities. GITS also provides a measure of gang activity in targeted
neighborhoods and enables managers to determine whether programs such as
intensive supervision and prosecution have any effect on gang crime.
Subregional analyses have also been used by several smaller cities to examine
gang crime in their section of the county and help them coordinate strategic
activities.

Other unforeseen benefits from GITS were its utility for grant applications and
public education. The GITS project helped attract grant funds that supported a
number of antigang programs in several cities. "It [GITS] has brought us a
tremendous amount of money to work on problems in a very real way" (chief
interview, January 29, 1997). Even though GITS was designed to provide a
countywide perspective, the data allow city-level analysis. Several chiefs of
police have taken advantage of this rich data source by using GITS output to
bolster proposals for grants submitted to private foundations and State and
Federal agencies. They also have used reports generated using GITS data to
educate city councils and citizen groups. One chief said that the data show that
the gang problem is not as large as he thought and there is hope that some of its
negative effects in his city can be relieved (Police chief interview, January 29,
1997).

Structuring Interagency Cooperation

There are at least two ways to structure interagency cooperation to create a
shared database like GITS: from the top, based on a cooperative agreement
among the heads of the various law enforcement agencies involved, or from the
bottom, based on cooperative sharing of data from those in the law
enforcement agencies (usually, crime analysts) who are actually managing and
analyzing the data. For ease of discussion, the first will be called the "Chief
Model," and the latter, the "Analyst Model." GITS is an example of the Chief
Model, while the Regional Crime Analysis System in the Baltimore-Washington
area exemplified the Analyst Model. Both have strengths as well as
weaknesses.

The Chief Model

When GITS was created, OCCSA already had a long history of interagency
cooperation as a forum in which the heads of Orange County's law enforcement
agencies met regularly to discuss issues of shared interest. When OCCSA
created GITS as a strategic tool to measure the level of gang crime in the
county and to plot its variation over time, it declared GITS as a priority for the
organization, thus ensuring the cooperation of all members. Absent OCCSA's
clout, it is highly unlikely that all of the law enforcement agencies would have
independently agreed to cooperate and create a shared database. With
OCCSA behind the project, peer pressure could be used to encourage chiefs
who otherwise might not have participated to cooperate. Since there is wide
variation in the reporting forms and records-management systems used by the
22 departments, GITS had to rely on their willingness to fill out a separate
standardized gang incident reporting form. Only police chiefs had the power to
make this happen.

Another benefit of the Chief Model was that OCCSA could speak on behalf of
all law enforcement in the county, putting it in a much stronger bargaining
position to lobby for external funding from the U.S. Department of Justice and
to persuade UCI to provide researchers to evaluate the database. OCCSA's
strategy of using UCI to collect and analyze the data lent the database
additional credibility and afforded the departments access to more sophisticated
analytical capabilities. This was particularly important in the application of
geographic information system (GIS) technology. At the beginning of the
project, none of the departments had this advanced capability (although more
departments are acquiring it) and even today, their ability to use GIS analyses
varies widely.

Yet the GITS experience also demonstrated some of the Chief Model's
drawbacks. The program's original goal was to create a strategic tool to
provide a countywide summary of gang crime. Because different chiefs have
different management styles and philosophies concerning access to data, the
ability to use the database on a more tactical level has been limited. Not all
chiefs have been willing to allow tactical personnel direct access to the data.
Because of incompatible records-management systems, GITS relies on a
separate reporting form that requires additional effort from the departments.
Different departments place different priorities on keeping their part of the
database up to date, so timely use of the data is limited. So far, this has limited
the reporting ability of the database to twice a year. Because each department
maintains control over the use of its data, sharing data among departments
requires permission of the respective chiefs. This can cause delays and, in some
cases, has constrained analysis. Because officers on the street who collect the
data do not have direct access to GITS, their stake in the database is limited.
As discussed previously, this issue has the potential to weaken the validity and
completeness of GITS data.

The Analyst Model

At the second annual Crime Mapping Research Conference (held on
December 10-12, 1998), sponsored by the National Institute of Justice's Crime
Mapping Research Center, the Regional Crime Analysis System for the
Baltimore-Washington area made a presentation of its interagency efforts to
share crime-mapping data. This effort originated with crime analysts in
contiguous jurisdictions, who were trying to improve their ability to recognize
and respond to evolving crime patterns. This cooperative effort involves fewer
than 10 departments, whose analysts send one another updated geocoded
crime data formatted according to standards specified by an oversight
committee. Although the Regional Crime Analysis System operates with
permission of the chiefs of the participating agencies, it is designed and
operated by analysts. The system includes the most current data and has been
used successfully in the tactical tracking of crimes across jurisdictions.

Yet the Analyst Model's focus on tactical concerns may limit its strategic value.
Not all agencies in the geographic area can participate, because not all have the
analytical ability to provide data or analyze it in the format required by the
oversight committee. The ability of smaller departments without GIS capability
to participate is also limited. Because there is no cooperative area wide
organization of chiefs to back the analysts' efforts, their ability to secure outside
funding and support has been limited. Finally, all the analysis is conducted
in-house, so the advantages of outside analysis, such as credibility and
expertise, which can be useful for strategic and publicity purposes, have not
been realized.

In sum, both types of interagency cooperation have advantages as well as
drawbacks and it is clear that how interagency cooperative efforts are
structured and evolve affects the utility, flexibility, and durability of such efforts.
It may well be worthwhile to combine the strengths of both models in a hybrid
approach.

Summary and Discussion of Findings

The GITS project clearly demonstrates the usefulness of multijurisdictional
efforts to understand, combat and, ultimately, prevent street gang activities. Just
as clearly, it demonstrates the value of partnerships between criminal justice
practitioners and university researchers. A summary of key findings from this
research is presented below, followed by a discussion of interesting
opportunities for future collaborative research.

One of the most gratifying findings of this project is that several dozen law
enforcement and community agencies can collaborate successfully with one
another and with a team of university researchers. OCCSA and GSSC provide
an excellent model for regions struggling with the reality that crime is often
multijurisdictional in nature. The findings reported here show how this type of
cooperative endeavor can benefit practitioners. (As discussed above, other
jurisdictions interested in adopting or adapting the GITS system may do so
while avoiding some of the difficulties that plagued OCCSA's early efforts.)
They also reveal opportunities for fruitful scholarly research.

The GITS database has enabled law enforcement agencies and public officials
to identify the nature and extent of gang crime in the county. It also provides
them with a useful tool for evaluating gang prevention, intervention, and
suppression programs. It has enabled them to establish a baseline against which
future changes in gang activity can be measured. Moreover, the collection and
analysis of the data by the university, independent of direct law enforcement
control, helps increase public confidence that policymakers' assessments about
gang-related crime are reasonable and accurate. This is likely to help bring
more balance to public perceptions of a problem that is often exaggerated. A
look at the GITS data shows that the number of incidents has been declining
slowly since 1994, along with the number of violent and property incidents (see
exhibit 4). Vandalism incidents, however, have been increasing since 1996.

Applying such powerful tools as GIS technology and multivariate statistical
analysis has made GITS data more useful. Examples include the preparation of
crime maps that make complex statistical and GIS analyses accessible to
criminal justice practitioners, elected officials, and the public. The two maps
(exhibits 5 and 6) are representative of this type of analysis.

Westminster's Gang Territories and Home Addresses

A map of gang territories in the city of Westminster, based on information
obtained by the Orange County District Attorney's Office and verified by the
Westminster Gang Unit, is displayed in exhibit 5. (The gangs listed on the map
are not the only gangs active in the city, but only the more traditional Hispanic
gangs claim territory.) Data from Westminster Police Department field
interviews were used to obtain home addresses of identified members of these
five territorial gangs. This allowed evaluators to visually compare gang
territories with members' home addresses. As the map shows, some gang
members live outside their gang's territory--sometimes even in another gang's
territory. Most members of the West Trece and Orphans gangs appear to live
inside their gang's territorial boundaries, whereas most Varrio Midway City and
King Kobra members live outside their territories. This type of visual
information can help police departments and researchers understand where
territories are located and identify new gang rivalries using hot-spot data.

Home Addresses of Gang Members Contacted in Westminster

A map displaying the home addresses of gang members contacted by
Westminster police officers and TARGET teams inside Westminster city
boundaries (exhibit 6) highlights the need for multijurisdictional gang efforts.
Gang members frequently cause trouble outside their own communities; some
of the gang members who were field interviewed in Westminster had come
from 30 miles away. Without countywide cooperation, it would not be possible
to ascertain the gang affiliation and past criminal activities of gang members
once they crossed into another city. Gangs and their members can move
between jurisdictions with little trouble; the gang information system now has
that same advantage.

Other Benefits of GITS

GITS also makes it possible to identify spatial and temporal patterns of criminal
offending to target resources more carefully (e.g., temporal analyses indicate
that "midnight basketball" programs that recently were in vogue are
inappropriate in Orange County--and perhaps other regions as well).

Theoretically, the development of geographic and statistical models that can
predict more than 70 percent of the variation in violent gang incidents at the
census tract level (Vila and Meeker 1999; Fossati 1999) should enable
researchers to make much more valid evaluations of gang control programs
because reliable and valid gang crime measures can be obtained before and
after program implementation.

Geocoded data also have made it possible to compare people's perceptions
and fears about gangs and gang crime with the prevalence of gang incidents in
their communities. Minority activists and others have voiced substantial concern
about whether the increase in gang activity in areas like Orange County reflects
biases within the criminal justice system rather than reality (see Collins 1997;
Munoz 1997; Schrader 1997; Larkin 1996; Nunciato 1996; Sizgorich 1996;
and Lopez and Mirande 1990). Ride-alongs, interviews, field observation, and
evaluation of official records, however, showed that law enforcement agencies
tend to underreport gang incidents to GITS.

It was also found that the concerted effort to train officers about legal criteria in
California for defining gang members appeared to pay off. Contrary to activists'
claims, there was no evidence that officers were classifying young people as
gang members merely because of their mode of dress, ethnicity, or place of
residence when they reported gang incidents for use in the countywide
database. On a similar note, the data collected by GITS appeared to present a
reasonably unbiased and complete picture of gang incidents handled by the
police.

GITS has met the goals set for it by law enforcement managers. Annual
announcements by OCCSA regarding countywide trends in gang crime enable
the public to judge progress on gang-related issues. GITS data are being used
by law enforcement policy decisionmakers to deploy personnel, allocate
resources, and evaluate gang prevention, intervention, and suppression. GITS
output also has been used by a number of chiefs to help educate local residents
and leaders and strengthen requests for additional resources from granting
agencies. Finally, GITS reports have helped practitioners keep gang problems
in perspective; gang incidents represent a relatively small--if especially
troublesome--portion of the overall crime problem faced by most jurisdictions
in
Orange County.

Notes

1. The survey, conducted for Drug Abuse is Life Abuse, was conducted
through random telephone interviews with 600 adult Orange County residents
(Mark Baldassare and Associates 1994).

2. A random telephone survey of 1,000 adult county residents conducted
annually by Mark Baldassare and Cheryl Katz since 1982 (Baldassare and
Katz 1993, 1994, 1995a).

3. Communique from Spergel included information about affiliation, street
fights, personal conflicts within gangs, extortion, and vice offenses as gang
motivated. (Fax from Irving Spergel, February 27, 1998).

4. In 1999, the FBI's Uniform Crime Report listed 92 homicides, 2,842
robberies, and 5,386 aggravated assaults in Orange County. GITS reported 17
homicides, 687 robberies, and 299 aggravated assaults. Gang crime in Orange
County accounts for 18 percent of the homicides, 24 percent of the robberies,
and 5 percent of the aggravated assaults.

5. Snowball sampling is a method often used in field research for drawing a
nonprobability sample in which each person interviewed is asked to suggest
additional people to interview.

6. Interviews were conducted with 19 chiefs of police. There are 22 chief
positions in Orange County. Several chiefs have been replaced since the initial
interviews took place. No attempt was made to interview new chiefs because
their knowledge of GITS was limited.

7. These agencies are part of GSSC but their functions and interactions with
gang crime limit their ability to participate. Although they deal with gangs, they
rarely handle new gang incidents, focusing most of their attention on following
up previous incidents, providing security during court appearances, or offering
investigative assistance. Thus, these agencies' data generally are not
comparable to police data. The Probation Department is somewhat of an
exception; it contributes data that are maintained by GITS in a separate
database.

References

Baldassare, Mark, and Cheryl Katz. 1993. Orange County Annual Survey:
1993 Final Report. Irvine, CA: University of California, Irvine. Department of
Urban and Regional Planning.

-------. 1994. Orange County Annual Survey: 1994 Final Report. Irvine, CA:
University of California, Irvine. Department of Urban and Regional Planning.

-------. 1995a. Orange County Annual Survey: 1995 Final Report. Irvine, CA:
University of California, Irvine. Department of Urban and Regional Planning.

-------. 1995b. "The Impacts of Suburban Context on Crime Perceptions,
Mobility Fears, and Policy Preferences." University of California, Irvine,
Department of Urban and Regional Planning Working Paper #1995-25.

Block, Carolyn Rebecca, Antigone Christakos, Ayad Jacob, and Roger
Przybylski. 1996. Street Gangs and Crime: Patterns and Trends in Chicago.
Research Bulletin. Chicago: Illinois Criminal Justice Information Authority.

California Council on Criminal Justice, State Task Force on Gangs and Drugs.
1989. Final Report.

Chang, S.K., W.H. Simms, C.M. Makres, and A. Bodnar. 1979. Crime
Analysis System Support: Descriptive Report of Manual and Automated Crime
Analysis Functions. Gaithersburg, MD: International Association of Chiefs of
Police.

Collins, Jeff. 1997. "Panel to Probe Police Labels of Gangsters." Orange
County Register. May 25, 1997.

Curry, G. David, Richard A. Ball, and Robert J. Fox. 1994. Gang Crime and
Law Enforcement Recordkeeping. Research in Brief. June. Washington, DC:
U.S. Department of Justice, National Institute of Justice, NCJ 148345.

Fossati, Thomas E., 1999. "Modeling Regional Variation in Urban Ecology and
Violent Gang Crime." Ph.D. diss., University of California, Irvine.

Hartl, D. 1993. "Regional Gang Strategy Steering Committee Meeting Minutes
(November 15)." Unpublished manuscript.

-------. 1994a. "Gang Strategy Steering Committee Meeting Minutes (July
28)." Unpublished manuscript.

-------. 1994b. "Orange County Chiefs of Police and Sheriff's Association
Gang Strategy Steering Committee: Briefing Points (August)." Unpublished
manuscript.

Kent, Douglas R., and Peggy Smith. 1995. "The Tri-Agency Resource Gang
Enforcement Team: A Selective Approach to Reduce Gang Crime." In The
Modern Gang Reader, ed. Malcolm W. Klein, Cheryl L. Maxson, and Jody
Miller. Los Angeles: Roxbury: chapter 25.

Larkin, Patrick. 1996. "UCI Gang Crime Study Called Revealing, but Trackers
Releasing Little Information for Now." Weekender (Los Angeles Times Orange
County Supplement) April 20, 1996, A2.

Lopez, Jose, and Alfredo Mirande. 1990. "The Gangs of Orange County: A
Critique of Social Policy." Aztlan 19(1): 125-146.

Mark Baldassare and Associates. 1994. "Drug Abuse is Life Abuse Survey:
Third Annual Study." Final Report Prepared for Drug Abuse is Life Abuse
Foundation.

Maxson, Cheryl L., and Malcolm W. Klein. 1990. "Street Gang Violence:
Twice as Great, or Half as Great?" In Gangs in America, ed. C. Ronald Huff.
Newbury Park, CA: Sage: 71-100.

-------. 1996. "Defining Gang Homicide: An Updated Look at Member and
Motive Approaches." In Gangs in America, 2d. edition, ed. C. Ronald Huff.
Newbury Park, CA: Sage: 3-20.

Munoz, Lorenza. 1997. "Gang Database Raises Civil Rights Concerns." Los
Angeles Times. July 14, 1997, B1.

Nakamura, Robert T., and Frank Smallwood. 1980. The Politics of Policy
Implementation. New York: St. Martin's Press.

Nunciato, Christian. 1996. "UCI Gang Study Faces Criticism." New University
News. October 14, 1996, 3.

Piliavin, Irving, and Scott Briar. 1964. "Police Encounters with Juveniles."
American Journal of Sociology 70: 206-214.

Quinney, Richard. 1970. The Social Reality of Crime. Boston: Little, Brown
and Company.

Rossi, Peter H., and Howard E. Freeman. 1993. Evaluation: A Systemic
Approach, 5th ed. Newbury Park, CA: Sage.

Schrader, Esther. 1997. "Minorities Tell of Mistrust, Fear of Police." Los
Angeles Times. May 10, 1997, B1.

Sizgorich, Thomas. 1996. "Garbage In, Garbage Out." Orange County
Weekly. July 5-July 11, 1996, 8.

Smith, D.S., and C.A. Visher. 1981. "Street-Level Justice: Situational
Determinants of Police Arrest Decisions." Social Problems 29(2): 167-177.

Smith, P. 1993. "Gang-Related Crime in Orange County: A Research
Proposal." Unpublished manuscript.

-------. 1994. "The Orange County Chiefs of Police and Sheriff's Association
Gang Incident Tracking System: Program Description." Unpublished
manuscript.

Souryal, Sam S. 1981. Police Organization and Administration. New York:
Harcourt Brace Jovanovich.

Spergel, Irving A. 1990. "Youth Gangs: Continuity and Change." In Crime and
Justice: A Review of Research, Vol. 12, ed. Michael Tonry and Norval Morris.
Chicago: University of Chicago Press: 171-275.

Spergel, Irving A., and G. David Curry. 1995. "The National Youth Gang
Survey: A Research and Development Process." In The Modern Gang Reader,
ed. Malcolm W. Klein, Cheryl L. Maxson, and Jody Miller. Los Angeles:
Roxbury: 254-265.

Vigil, James D., and John M. Long. 1990. "Emic and Etic Perspectives on
Gang Culture: The Chicano Case." In Gangs in America, ed. C. Ronald Huff.
Newbury Park, CA: Sage: 55-68.

Vila, Bryan, and James W. Meeker. 1997. "A Regional Gang Incident
Tracking System." Journal of Gang Research 4(3): 23-36.

-------. 1999. Gang Activity in Orange County, California. Final Report to the
National Institute of Justice, grant number 96-IJ-CX-0030, Washington, DC:
National Institute of Justice (August). NCJRS, NCJ 181242.

Weston, Paul B. 1978. Supervision in the Administration of Justice: Police,
Correction, Courts. 2d. ed. Springfield, IL: Charles C. Thomas.

Interviews conducted by Dr. Katie J.B. Parsons:

Personal interviews:

o July 6, 1995

o July 7, 1995

GSSC interview:

o March 1, 1996

Chief interviews:

o January 29, 1997

o February 18, 1997

---------------------------

About the National Institute of Justice

NIJ is the research and development agency of the U.S. Department of Justice
and is the only Federal agency solely dedicated to researching crime control
and justice issues. NIJ provides objective, independent, nonpartisan,
evidence-based knowledge and tools to meet the challenges of crime and
justice, particularly at the State and local levels. NIJ's principal authorities are
derived from the Omnibus Crime Control and Safe Streets Act of 1968, as
amended (42 U.S.C. sections 3721-3722).

NIJ's Mission

In partnership with others, NIJ's mission is to prevent and reduce crime,
improve law enforcement and the administration of justice, and promote public
safety. By applying the disciplines of the social and physical sciences, NIJ 

o Researches the nature and impact of crime and delinquency.

o Develops applied technologies, standards, and tools for criminal justice
practitioners.

o Evaluates existing programs and responses to crime.

o Tests innovative concepts and program models in the field.

o Assists policymakers, program partners, and justice agencies.

o Disseminates knowledge to many audiences.

NIJ's Strategic Direction and Program Areas

NIJ is committed to five challenges as part of its strategic plan: 1) rethinking
justice and the processes that create just communities; 2) understanding the
nexus between social conditions and crime; 3) breaking the cycle of crime by
testing research-based interventions; 4) creating the tools and technologies that
meet the needs of practitioners; and 5) expanding horizons through
interdisciplinary and international perspectives. In addressing these strategic
challenges, the Institute is involved in the following program areas: crime control
and prevention, drugs and crime, justice systems and offender behavior,
violence and victimization, communications and information technologies,
critical incident response, investigative and forensic sciences (including DNA),
less-than-lethal technologies, officer protection, education and training
technologies, testing and standards, technology assistance to law enforcement
and corrections agencies, field testing of promising programs, and international
crime control. NIJ communicates its findings through conferences and print and
electronic media.

NIJ's Structure

The NIJ Director is appointed by the President and confirmed by the Senate.
The NIJ Director establishes the Institute's objectives, guided by the priorities
of the Office of Justice Programs, the U.S. Department of Justice, and the
needs of the field. NIJ actively solicits the views of criminal justice and other
professionals and researchers to inform its search for the knowledge and tools
to guide policy and practice.

NIJ has three operating units. The Office of Research and Evaluation manages
social science research and evaluation and crime mapping research. The Office
of Science and Technology manages technology research and development,
standards development, and technology assistance to State and local law
enforcement and corrections agencies. The Office of Development and
Communications manages field tests of model programs, international research,
and knowledge dissemination programs. NIJ is a component of the Office of
Justice Programs, which also includes the Bureau of Justice Assistance, the
Bureau of Justice Statistics, the Office of Juvenile Justice and Delinquency
Prevention, and the Office for Victims of Crime.

To find out more about the National Institute of Justice, please contact:

National Criminal Justice Reference Service
P.O. Box 6000
Rockville, MD 20849-6000
800-851-3420
e-mail: askncjrs@ncjrs.org

---------------------------

NCJ 190351