Title: Toward a Drugs and Crime Research Agenda for the 21st Century

Series: Special Report

Author: National Institute of Justice

Published: September 2003

Subject: Drugs and crime and drug policy

271 pages

637,000 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).



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



Toward a Drugs and Crime Research Agenda for the 21st Century



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



This and other publications and products of the U.S. Department of

Justice, Office of Justice Programs, National Institute of Justice can be

found on the World Wide Web at the following site:



Office of Justice Programs

National Institute of Justice

http://www.ojp.usdoj.gov/nij



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



July 03



Toward a Drugs and Crime Research Agenda for the 21st Century



NCJ 194616



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



Sarah V. Hart

Director



Findings and conclusions of the research reported here are those of the

authors and do not 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.



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



CONTENTS



Introduction



At the Intersection of Public Health and Criminal Justice Research on

Drugs and Crime



Research on Drugs-Crime Linkages: The Next Generation



The Drugs-Crime Wars: Past, Present, and Future Directions in Theory,

Policy, and Program Interventions



Appendix A: Summary of Proceedings



Appendix B: Forum Agenda



Appendix C: List of Participants



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



INTRODUCTION



Henry H. Brownstein with Christine Crossland



For criminal justice practitioners who deal with drugs and crime day in

and day out, the reality of the drugs-crime nexus is indisputable. In a

manual designed to help police chiefs and sheriffs control drug abuse, the

International Association of Chiefs of Police (IACP) stated unequivocally

its belief in "a significant though complex" relationship between drug

abusers and criminal offenders. Change one group, IACP proposed, and

you change the other: "If there is a reduction in the number of people who

abuse drugs in your community, there will be a reduction in the

commission of certain types of crime in your community."[1]



When IACP released its manual more than a decade ago, researchers

already were confirming what practitioners believed and documenting the

relationship between drugs and crime.[2] Public policy and programs were

and continue to be developed on the basis of this knowledge.[3] But

although researchers and practitioners alike knew the relationship existed,

the nature of that relationship eluded them then and continues to elude

them today.[4] 



To shed light on the drugs-crime link requires research, and the first step is

to specify the research topics to be covered. Taking the lead, the National

Institute of Justice (NIJ) and the National Institute on Drug Abuse (NIDA)

brought together academics and other researchers and asked them to

answer three questions: What do we know about drugs and crime, what do

we not know, and, most important, what do we need to know? Both

agencies see this knowledge not as an end in itself but as a means to

accurately define the problem of drugs and crime and promote future

research. The agenda for research was developed under NIJ and NIDA

sponsorship at a forum held in Washington, D.C., in April 2001. The

findings of the Drugs and Crime Research Forum are presented here.



In Pursuit of the Drugs-Crime Link



If we are going to make progress toward solving the problem of drugs and

crime, we need to shed light on the nature of the drugs-crime link by

designing effective responses. Developing a research agenda on drugs and

crime means tackling the central issue of the drugs-crime link. Is the

link a matter of cause and effect or is it something far more complex?



There is no lack of theories. The direct cause model of the drugs-crime

relationship has attracted its share of supporters. It states simply that either

drug use leads to crime or crime leads to drug use. The simplicity is

appealing. Who would not find it tempting to believe that reducing drug

use can lower the crime rate? In fact, some policies and programs have

been developed on the basis of the direct cause model or the belief in a

significant relationship between drugs and crime. As IACP recognized, the

relationship is real enough. And NIJ's Arrestee Drug Abuse Monitoring

program has demonstrated year after year that among people apprehended

and charged with a crime, a large percentage uses drugs.[5]



However, as sociologist Erich Goode has cautioned, "Even the fact that

drugs and crime are frequently found together or correlated does not

demonstrate their causal connection."[6] The consensus among researchers

who study the issue confirms Goode's observation. The evidence for the

direct cause model is just not there.[7]



We seem more willing today to accept the complexity of the drugs-crime

relationship, more open to the notion that "[t]here is considerable

uncertainty . . . about the degree to which drug use causes crime or the

degree to which criminal involvement causes drug use."[8] In a recent

review of the literature, sociologists Helene Raskin White and Dennis M.

Gorman definitively dismissed the direct cause model. They concluded

instead that the drugs-crime link is best explained by the common cause

model, in which any association of drugs and crime has a cluster of

causes.[9]



Those who subscribe to the common cause model believe that to

adequately understand the relationship of drugs to crime requires attention

to many issues, social, cultural, chemical, and biological among them.

What the model means for policy and practice is that any response to drugs

and crime that works in one set of circumstances may not work in another.

For researchers, it means the research agenda is vast. Policy and practice

can be informed by what we know up to this point, but progress in

responding to the drugs-crime problem requires knowing more.



Building on the Past: The Drugs and Crime Research Forum



NIDA and the National Institute of Law Enforcement and Criminal

Justice, NIJ's predecessor organization, were asked by Congress in 1976 to

find out what was known about drugs and crime. The product of the

agencies' collaboration was Drugs and Crime: A Survey and Analysis of

the Literature. Though not strictly a research agenda, the survey was a first

step "to identify where the gaps in our knowledge lie and to direct research

to fill those gaps."[10] It was intended to "set the stage for more focused

future research."[11]



In 2000, NIJ's call for the development of a research agenda was another

step toward meeting that need. The authors of Drugs and Crime noted at

the time that "few if any [studies] directly address the drugs-crime nexus

issue."[12] This report on the development of the research agenda will

demonstrate that although much has been learned in the intervening years

about drugs, drug use, drug abuse, drug markets, and drug law

enforcement, much work is needed to shed light on the complexities of the

drugs-crime link.



Three papers were commissioned for the research forum. Each addressed

the questions of what we know, what we do not know, and what we need

to know about the drugs-crime link. Prepared by experts in epidemiology,

public policy, social work, and allied disciplines, the papers served as

the focal point and framework for discussions by forum participants. (The

forum summary, agenda, and a list of the participants are presented in

appendixes A, B, and C.) After the forum adjourned, NIDA created a

listserv for participants to continue to exchange their thoughts.



The discussions did not all fit the same mold. Roundtables were generated

from one-sentence statements by participants about drugs and crime.

What we do not know about the drugs-crime relationship was treated at

length. The many strands of thought, lines of discussion, and themes came

together when Forum participants addressed the final question: What

future research is most important, and what research is needed most

urgently? Mindful that the next generation of researchers will be tackling

the problem of drugs and crime, Forum participants recommended topics

for research by their graduate students.



Probing Drugs and Crime: Three Perspectives



"At the Intersection of Public Health and Criminal Justice Research on

Drugs and Crime" was commissioned by NIDA from James C. Anthony

with Valerie Forman. Anthony asked such questions as--



o Have we made effective and adequate use of recent developments in

science and technology to advance the study (and hence the understanding)

of the relationship between drugs and crime? Given the vast literature

generated during past decades on this subject, have we adequately,

appropriately, and effectively integrated research from both the public

safety and public health perspectives?



o Are the tensions between the two perspectives greater than our ability to

overcome them? Given what we know now and the current tension

between researchers in public safety and public health, how can we

conceptualize and organize our thinking and research to enhance our

knowledge and understanding of the relationship in the most productive

ways?



o What do we really know about the suspected causal connection between

drugs and crime? In looking at drugs and crime, what is the intersection at

which public health and public safety meet? How can we achieve the goals

of greater understanding and definitive evidence and greater mastery in

design and application in policy, programs, and techniques to prevent and

reduce harmful health and safety consequences of drug use?



o What do we need to do to integrate molecular biology, genetics, and

neuroscience into discussions of drugs and crime? What do we need to

do to place discussions of the drugs-crime nexus in the context of history?

How can we clarify the question of causal inference? How can we use the

notions of scale and rubrics to help understand the relationship between

drugs and crime? 



"Research on Drugs-Crime Linkages: The Next Generation" was

commissioned by NIJ from Robert MacCoun, Beau Kilmer, and Peter

Reuter. Among the questions asked by MacCoun and his colleagues

were--



o Are our conceptualizations of the relationship between drugs and crime

adequate to move forward in our understanding of that nexus? How must

we conceptualize the relationship to be able to address questions not only

of concomitance and statistical correlation, but also of social significance

and causality?



o To the extent that the drugs-crime relationship is causal, to what extent

do we understand the nature of the causal influences? How can we use

Paul Goldstein's tripartite taxonomy to build on work already done, and

how can we move beyond the taxonomy? How can we use notions such as

Bruce Johnson's conduct norm analysis or Alfred Blumstein's drugs-gun

diffusion hypothesis?[13]



o How does the considerable heterogeneity of users, substances,

locations, and situations as well as differences in market dynamics affect

what we need to have and to do to address the drugs-crime nexus? How do

we address the question of causal influences? How will research in the

coming decade deal with the heterogeneity of effects across users,

substances, cities, neighborhoods, and situations?



"The Drugs-Crime Wars: Past, Present, and Future Directions in Theory,

Policy, and Program Interventions" was commissioned by NIJ from Duane

C. McBride, Curtis J. VanderWaal, and Yvonne M. Terry-McElrath. In

this paper, McBride and his colleagues raised the following questions:



o In the past two or three decades, what progress has been made in our

knowledge and understanding of the relationship between drugs and

crime? Does knowledge of the statistical relationship help us understand

the nature of the relationship? What do we know about the nature of the

nexus and what do we need to do now to advance the state of our

knowledge? In the past century, how have we used that knowledge to

guide public policy? Could we do a better job of linking what we think and

what we know about drugs and crime to what we do to address individual

and social problems in the realms of public health and public safety? Do

we know enough about what has been tried (for example, programs and

program evaluations) to know what works?



o How is the idea of social capital important to our understanding of the

drugs-crime nexus? What is the significance of the dynamic tension

between drug policy as it shifts and the drugs-crime connection as policy

changes? What is the value of interventions and treatment when dealing

with drug-using offenders?



What Do We Need to Learn?



It will come as no surprise that the question of what we know about drugs

and crime was eclipsed by that of what we do not know and what we have

yet to learn. The papers and accompanying discussions yielded an

abundance of ideas on research topics for the coming decades. The major

themes included the following:



o Drug-related crime.



o Drug enforcement.



o Drug markets.



o Drug offenders.



o Drug policy.



o Treatment and intervention.



o Drug use and abuse.



o Ethnographic studies.



o Health sciences perspectives.



o Minority research.



o Research methods.



o Victimization studies.



Categorized more broadly, the topics proposed for research are the

drugs-crime nexus, the social contexts of drug use and crime, and refining

study methods and designs.



What Explains the Drugs-Crime Nexus?



We know that drugs and crime are related. We also know something about

the different ways they might be related, and perhaps something about the

ways they may be related in time and space. What we have yet to learn is

how they are related. In other words, we need to probe the underlying

dynamics of the relationship. We do not know, for example, why so many

people who commit crime also use drugs or why some people who use

drugs commit crime but others who use drugs do not commit crime.

Research in this broad area might take several directions.



Find new ways to conceptualize the drugs-crime nexus. Several years ago,

Goldstein proposed a tripartite framework as a way to disentangle the

relationship between drugs and crime, specifically violent crime. Violence

could be the direct outcome of ingesting drugs, the result of a user's

compulsion to obtain drugs or money for drugs, or a product of the

disorganization and violence inherent in the social systems in which drugs

are manufactured and exchanged.[14] Over the years, this framework has

been useful for studying drugs and violent crime but of limited value for

studying drugs and other types of crime. Beyond what has been learned

from this model, how can researchers conceptualize the way or ways drugs

and crime--not just violent crime--are related?



Combine research perspectives. Research on drug use illustrates how

different disciplines can combine forces. Social science research is

beginning to merge with biological research, particularly genetic research.

Questions include the following:



o How can the study of genes, the social environment, and behavior help

us better understand the link between drugs and crime?



o Are there physiological propensities for drug using? If so, what is the

impact of the user's environment?



o Are alcohol and marijuana complements of or substitutes for other

drugs?



o Do different drugs have different effects on groups of people who are

genetically different?



o How can we address the ethical concerns of such research?



Give more attention to minorities. A disproportionate number of the

people arrested, charged, and in custody for drug and other criminal

offending are from minority groups. The reason is unknown but needs to

be probed. What can we learn about the involvement of various ethnic

and racial groups in the drugs-crime link? What can we learn about gender

and the drugs-crime link? Some answers might be found in comparative,

multisite studies of drug use and drug markets in different ethnic

communities. What is the relationship of gender, age, race, and culture to

drug involvement and crime? What is the effect of disparity (in income,

for example), prejudice, and discrimination on the distribution of resources

used for treatment and prevention? How can we explain racial and ethnic

differences in drug use and involvement in crime? Do people view the

drugs-crime link differently because of their race, gender, or age?



From these general research areas on racial and ethnic diversity, it is

possible to derive many specific topics. What can we find out about the

relationship between drugs, crime, and the increase in the number of

women of color who are incarcerated? Have changes in the economy

affected the involvement of disadvantaged black and Hispanic/Latino

males in drugs and crime? If so, how? What is the impact of drug-related

incarceration on families and children or on prospects for education and

employment in minority communities?



What Do We Need to Know About the Social Context of Drugs and

Crime?



It is widely believed that drug use adversely affects users. But drug use and

crime are affected by and in turn affect forces operating in society at large.

Drug users interact with many people: sellers with buyers, buyers with

sellers, criminal offenders with their victims. There is a social context of

drug use.



Social patterns in the drug world. There have been many studies of drug

users and some studies of drug markets. But what do we need to learn

about the social relations and interactions of the people whose lives are

affected by drugs?



The commerce of drugs and crime: drug markets. Theories about and the

operations and institutional arrangements of drug markets are plentiful, but

not enough research has been done to test them. How stable are drug

markets, and how do they change over time? For example, has the

maturation of the crack cocaine market in some cities affected those cities'

crime rates? What is the connection between local market activity and

fluctuations in supply and demand at the national level? What influences

the relationship between sellers and buyers? How and why do new markets

emerge, and what impact do they have on existing markets? How are

prices set in local drug markets, and how are wages set?



Patterns of use and abuse. We know something about the demographics of

drug use, but what do we know about intergenerational patterns? How do

use patterns vary with social or biological differences? How do patterns of

alcohol use compare with patterns of use of other drugs? Can drug use

help explain juvenile involvement in crime or violence? Are patterns of

use of certain drugs, such as club drugs, designer drugs, or inhalants,

different from patterns of use of other drugs? What can we find out about

how and why people start or stop using drugs?



Criminal offending by drug users. There are some studies of drug

offenders, but how much do we know about how or why drug offenders

commit crime? Are some people genetically predisposed to drug use?

Is there a relationship between drug use and social status, and if so, how

might social status in turn be related to involvement in crime? Beyond

using illicit drugs, to what extent are drug users and sellers involved in

other crime? Can we realistically estimate how much other crime is

committed by drug offenders? What risk do these people pose to their own

health and safety? 



Victims of drug users and drug use. Drug users are in some ways their own

victims, but are there other victims? What do we know about other people

with whom drug users relate? How can we define for research and policy

purposes what we mean by "victims of drugs"? How do we define

victimization in this context? Are there indirect victims, such as families

and communities, as well as direct victims?



The public's response to drug use and drug-related crime. Society

considers drug-related crime and illicit drug use as affronts and responds

accordingly. Enforcement strategies are one example. To what extent are

the responses based on a real understanding of these problems? Are the

responses making a difference?



Enforcing drug laws. What is the effect of enforcement policies, programs,

and practices on drug use, drug dealing, and drug-related crime? What is

the relationship between street-level enforcement and street-level drug

market activity, particularly violent activity? What impact do drug

seizures, drug arrests, and asset forfeiture, among other interdictions, have

on drug and drug-related crime? What is the impact of public concern

about racial profiling and police corruption on the ability of law

enforcement to respond to drugs and crime?



Treating drug use and abuse. With so many different drugs and so many

different types of users, what can we say about the efficacy of drug

treatment in addressing drugs and crime? What is the nexus of drug

treatment and criminal justice? For example, what are the results of

treatment in correctional settings and what do evaluations reveal? Do

incentives or disincentives help drug users to succeed in treatment? How

do we define success? How important are aftercare programs and family

interventions? What is the best way to treat drug users who are dually

diagnosed (for example, those who are also mentally ill)? What are the

dropout rates for treatment, and what does it matter? What treatments 

work best with what types of drug use? How do we distinguish users

from abusers? What difference does that distinction make for treatment

planning?



Intervening to prevent drug use or crime. Although relatively little is

known about preventing drug use, the topic receives a great deal of 

attention. To what extent can media campaigns help prevent drug use? Is

the impact of prevention programs the same for all social categories of

users or irrespective of type of drug? How can we educate young people

about the impact that drugs can have on their lives? Should more attention

be paid to problem behavior, norm violations, and rule breaking than to

drug prevention?



Public policy. When we think about public policy on drugs, we typically

do not think about policy in general but rather about specific aspects, such

as interdiction, enforcement, treatment, and prevention. But can we step

back and think broadly and measure the impact of drug policy over the

past decade, or even the past century? Can we learn from policy

simulations that examine past and prospective views of drug use? Can we

learn from comparative studies of different countries? What is the impact

of different directions in drug policy? What policies have worked or not

worked with adult and juvenile drug offenders? Can research examine

drugs, crime, and public policy together? Can we find out from

policymakers and practitioners what decisions they need to make and

what questions they need to answer about drugs and crime? How can we

move drug policy analysis beyond econometrics (supply and demand, for

example) and begin to study drug use from the perspective of politics,

criminal justice, public health, and social work?



Methods of Studying Drugs and Crime



Research methods are dictated by the questions researchers ask. Some of

the questions already explored indicate that certain methodological

concerns might need to be addressed.



Attention to measurement and design. What are the best measures

currently available to study drug use and involvement in drug markets and

drug treatment? How can they be improved? How can we construct

integrated data collection measures? What is the best way to design

measures and procedures to evaluate drug control programs? What are the

best measures for assessing drug treatment outcomes? What is the role of

cost-benefit analysis in drug studies? What can we learn from longitudinal

studies about the long-term effects of drug use and abuse? How can we

introduce randomization to long-term studies of drug treatment? How can

statistical techniques developed by other sciences be adopted by the 

social sciences? How can we encourage multidisciplinary teams of

researchers to work together to study drugs and crime?



Ethnography. There is a long, distinguished tradition of ethnographic

research in the field of drug studies. Ethnographic studies, however, are

almost by definition limited to a single area or a small group of people.

What might we learn by secondary analyses of ethnographic studies?



What might we learn by replicating ethnographic studies in other

communities or among other groups of drug users? How useful might it be

to link ethnographic studies of community structure with studies of drug

users and dealers in their communities? What can we learn from studying

communities of sellers and users? Would it be useful to establish

prospective, qualitative field sites in various communities as a type of

surveillance system to monitor changing drugs and drug-use patterns?



Using available data and studies. How can we make better use of available

data to study drugs and crime? Are there obstacles to making better use of

available data to learn from them what we can? What can we learn from

meta-analyses of previously conducted research studies of drugs and

crime?



What Is In This Report?



Following this introduction are the three papers commissioned for the

forum and appendixes containing a summary of the forum proceedings, the

agenda, and a list of the names and organizational affiliations of the

participants.



Notes



1. International Association of Chiefs of Police, Reducing Crime by

Reducing Drug Abuse: A Manual for Police Chiefs and Sheriffs,

Gaithersburg, MD: International Association of Chiefs of Police,

1989:5.



2. See Tonry, M., and J.Q. Wilson, eds., Drugs and Crime, vol. 13 of

Crime and Justice: A Review of Research, Chicago: University of Chicago

Press, 1990; and R. Weisheit, ed., Drugs, Crime and the Criminal Justice

System, Cincinnati: Anderson Publishing, 1990.



3. See, for example, Forcier, M.W., "Substance Abuse, Crime and

Prison-Based Treatment," Sociological Practice Review 2 (1991):

123-131; Office of National Drug Control Policy, National

Drug Control Strategy--2000 Annual Report, Washington, DC: The White

House, 2000 (and earlier ONDCP annual reports); and Longshore, D., F.

Taxman, S. Turner, A. Harrell, T. Fain, and J. Byrne, "Operation Drug

TEST Evaluation," final report submitted to the National Institute of

Justice, U.S. Department of Justice, 2000 (grant 97-IJ-CX-0041).



4. Examples of studies that explored the nature of the drugs-crime

relationship in that period include Brownstein, H.H., and P.J. Goldstein,

"A Typology of Drug Related Homicides," in Weisheit, ed., Drugs, Crime

and the Criminal Justice System: 171-192; Chaiken, J.M., and M.R.

Chaiken, "Drugs and Predatory Crime," in Tonry and Wilson, eds.,

Drugs and Crime: 203-239; Fagan, J., "Intoxication and Aggression," in

Tonry and Wilson, eds., Drugs and Crime: 241-320; Goldstein, P.J., H.H.

Brownstein, P.J. Ryan, and P.A. Bellucci, "Crack and Homicide in New

York City, 1988: A Conceptually Based Event Analysis," Contemporary

Drug Problems 16 (1989): 651-687; and Johnson, B.D., T. Williams, K.A.

Dei, and H. Sanabria, "Drug Abuse in the Inner City: Impact on Hard-Drug

Users and the Community," in Tonry and Wilson, eds., Drugs and Crime:

9-67. 



5. Arrestee Drug Abuse Monitoring, 1999 Annual Report on Drug Use

Among Adult and Juvenile Arrestees, Washington, DC: U.S. Department

of Justice, National Institute of Justice, June 2000 (NCJ 181426).



6. Goode, E., Between Politics and Reason--The Drug Legalization

Debate, New York: St. Martin's Press, 1997: 119.



7. See, Brownstein, H.H., "What Does 'Drug-Related' Mean? Reflections

on the Problem of Objectification," The Criminologist 18 (1993): 1, 5-7;

Chaiken and Chaiken, "Drugs and Predatory Crime," in Tonry and Wilson,

eds., Drugs and Crime: 203-239; Fagan, J., "Intoxication and Aggression,"

in Tonry and Wilson, eds., Drugs and Crime: 241-320; Goode, E.,

Between Politics and Reason--The Drug Legalization Debate; White,

H.R., and D.M. Gorman, "Dynamics of the Drug-Crime Relationship," in

G. LaFree, ed., The Nature of Crime: Continuity and Change, vol. 1 of

Criminal Justice 2000, Washington, DC: U.S. Department of Justice,

National Institute of Justice, July 2000 (NCJ 182408); and Wilson, J.Q.,

"Drugs and Crime," in Tonry and Wilson, eds., Drugs and Crime: 521-

545.



8. Office of Justice Programs, Office of Justice Programs Fiscal Year 2000

Program Plan: Resources for the Field, Washington, DC: U.S. Department

of Justice, Office of Justice Programs, 2000: 23 (NCJ 182238).



9. White and Gorman, "Dynamics of the Drug-Crime Relationship," in G.

LaFree, ed., The Nature of Crime: Continuity and Change: 193. 



10. Gandossy, R.P., J.R. Williams, J. Cohen, and H.J. Harwood, Drugs

and Crime: A Survey and Analysis of the Literature, Washington, DC:

U.S. Department of Justice, National Institute of Justice, 1980: xi (NCJ

159074).



11. Ibid.: 122. The survey covered five "crucial areas": patterns of drug use

and criminal behavior; "life cycle" characteristics (age of onset of drug use

and crime, for example); "economic issues" (price and supply/demand, for

example); treatment; and methods (sampling, for example).



12. Ibid.: 122.



13. These works are discussed in this report.



14. Goldstein, P., "The Drugs/Violence Nexus: A Tripartite Conceptual

Framework," Journal of Drug Issues 15 (1985): 493-506.



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



AT THE INTERSECTION OF PUBLIC HEALTH AND CRIMINAL

JUSTICE RESEARCH ON DRUGS AND CRIME



James C. Anthony with Valerie Forman



About the Authors



James C. Anthony, Ph.D., is a professor in the Department of Mental

Hygiene at Johns Hopkins University. Valerie Forman is a National

Institute of Mental Health predoctoral fellow with the Psychiatric

Epidemiology Training Program at Johns Hopkins University.



Introduction



This paper discusses intersections of public health research and criminal

justice research on the topic of drugs and crime. The drugs of interest

mainly are marijuana, heroin, and other internationally regulated

compounds of illegal origin, and such internationally regulated products of

legal origin as pharmaceutical cocaine hydrochloride, codeine, and

oxycodone, which also may be consumed on an extraordinary basis (i.e.,

outside the bounds of accepted medical practice). An important point of

departure for this paper is a widely held assumption about two goals of

research on this topic. The first goal is to achieve greater understanding

and develop a body of definitive evidence on drugs and crime. The second

goal is to achieve greater mastery of the design and application of policies,

programs, and techniques to improve public health and public safety by

preventing and reducing harmful consequences of drug use. 



The outline for this paper corresponds with assignments delegated at a

planning meeting held at the National Institute of Justice (NIJ) in January

2001. This introductory section provides some background notes on the

literature reviewed for the paper and describes an organizing conceptual

framework that can be used to assess gaps in the current evidence. The

next section identifies some tensions that merit discussion as we try to

forge a new research agenda on drugs and crime. We then address the

central question posed in our planning meeting for the drugs-crime

research forum: "What do we know about the drugs-crime

interrelationship?" We cannot provide a comprehensive answer to this

question in a relatively short paper, but we will offer a starting point for

discussion, focusing on suspected causal relationships between drugs and

crime. We also present a few concluding statements that were designed to

facilitate discussion at the forum on drugs-crime research held at NIJ in

April 2001.



A Burgeoning Literature on a Variety of Fronts



A scholar interested in the topic of drugs and crime has much to read.

Some of the classics of the field include Terry and Pellens's The Opium

Problem (1928); early papers on drug taking and sociopathy by Kolb and

Pescor, who were two of the early clinical leaders in research at the facility

that ultimately became the National Institute on Drug Abuse's (NIDA's)

Intramural Research Program and Addiction Research Center; and work

by Dunham and Lindesmith, whose surprisingly contemporary remarks

and observations started to systematize some of the field's research

questions on the social psychology of the drugs-crime relationship.

Many of the issues that confront the drugs-crime researcher today were

articulated by Terry and Pellens (1928), Kolb (1925), Pescor (1939),

Dunham (Faris and Dunham, 1939), Lindesmith (1938), and their

contemporaries in the first half of the 20th century.



These issues were re-articulated and a new set of themes was clarified in

subsequent research, such as The Road to H investigations led by Chein

(1964), the work of Preble and Casey as described in "Taking Care of

Business--The Heroin User's Life on the Street" (1969), Cohen's

Delinquent Boys (1955), and Robins' Deviant Children Grown Up (1966).

Two of the most important emerging themes from this research offer a

challenge to conventional thinking about the drugs-crime relationship:



o There is no single drugs-crime relationship. Rather, there are

drugs-crime relationships, most of which are complex rather than simple.



o There is no simple solution to the complex challenges faced when

drugs-crime relationships come into play.



By way of illustration, Brownstein and Goldstein offered and refined a

tripartite conceptualization of drugs-crime relationships, which serves as a

useful guide to some of the surrounding issues. Within this tripartite

framework, one set of criminal offenses is described as

psychopharmacologically induced (e.g., responses to intoxication states

after drug taking). A second set of offenses is described as

economic-compulsive in nature (e.g., instrumental income-producing

criminal acts as needed to stave off symptoms of withdrawal states that

appear once drug use has stopped). A third set of offenses is described as

"systemic" and might be understood best as a consequence of a drug user

entering or living within a social context in which extraordinary drug use

is just one of a set of often intercorrelated criminal behaviors. That is,

we do not need an appeal to drug intoxication, drug withdrawal states, or

drug-induced compulsive behavior to account for offenses observed in this

third category (Goldstein, 1985; Brownstein and Goldstein, 1990).



The tripartite framework clarifies three separate types of drugs-crime

relationships, none of which is simple. As for analysis of simple solutions

for these complex problems, a therapeutically oriented drug maintenance

program might reduce the economic-compulsive type of offending without

influencing the occurrence of crimes determined by poor judgment or

other manifestations of intoxication states. A successful supply-side drug

eradication program might reduce both pharmacological and

economic-compulsive types of offending, but not offending of the

systemic variety. Imprisonment of the drug user within a drug-free prison

environment might extinguish today's crimes but might not influence

tomorrow's offending when the prisoner is released back to the home

community. Even if the prisoner remains drug free during the immediate

postrelease period, the history of incarceration and a criminal record

might constrain job opportunities and economic success to the point of

inducing crimes that otherwise would not have been committed if the drug

user never had been incarcerated in the first place.



Illuminated in this manner, the facets of multiple drugs-crime relationships

become more clear; new opportunities for research open up. As these

opportunities have been recognized, there has been a tremendous growth

in scholarship and research activity on the topic of drugs and crime (see

exhibit 1).



Scholars may benefit from an assembled listing or bibliography of this

literature, now available in electronic form as a technical report from the

Electronic Collaboratory for Investigations about Drugs at Johns Hopkins

University (Forman, 2001). Readers interested in a recent comprehensive

review of these publications can turn to the Harrison and

Backenheimer-edited issue of Substance Use & Misuse on the drugs-crime

nexus in the United States, which was published by Stanley Einstein and

Marcel Decker, Inc., in 1998.



A Conceptual Framework for Drugs-Crime Research



Confronting the accumulated body of evidence and new literature, we have

attempted to sort each element of evidence in relation to a conceptual

framework originally devised for the field of psychiatric epidemiology and

epidemiological research in general (Anthony and Van Etten, 1998). This

conceptual framework is used as we train public health scientists for

advanced research on drug dependence and related conditions. The

framework may prove to be useful in the domain of criminal justice

research as well, perhaps with suitable amendments by interested teachers

and scholars.



The rubrics. Early in their public health research training, we ask our

predoctoral and postdoctoral fellows to master the epidemiology of drug

dependence. Here, drug dependence is defined as a syndrome or "running

together" of clinical features, and sometimes is called drug addiction,

especially when the focus is on such clinical features as obsession-like

cravings and compulsion-like repetitive behaviors in which drug

taking is central. The clinical features of the drug dependence syndrome

include pharmacological tolerance, characteristic withdrawal signs and

symptoms, almost obsessional thinking about drugs and drug-related

behavior, and other observable mental, behavioral, and social adaptational

manifestations of neuroadaptational processes that get started and progress

with repeated drug taking.



This epidemiology of drug dependence is a subject matter to be mastered

by the public health research fellows, just as they master the concepts,

principles, and techniques used as methodological tools in the public

health sciences. Mastery of this subject matter begins with study of the

just-mentioned clinical features, the history of diagnostic criteria or case

definitions used in public health research on drug dependence, and what

has been learned about its neuroadaptational and genetic substrates. In

the process, research fellows learn of patterned variation in drug

dependence syndromes, some of which can be understood in relation to the

pharmacology and pharmacokinetics of different drugs, such as cocaine

versus heroin or methamphetamine versus oxycodone. Research fellows

also learn about different measurement techniques used in laboratory,

clinical, and field studies of the drug dependence syndromes. For example,

under certain conditions, an appropriate dose of a narcotic antagonist can

be used as a bioassay to check for the presence of dependence on heroin or

other opioid drugs (e.g., via precipitated withdrawal). Nonetheless, in

general, the measurements of drug dependence rely heavily on self-report

information obtained under specially protected confidential circumstances.

To the extent that subjectively felt experiences such as "craving" and

obsessional thinking about drugs are central clinical features for drug

dependence, we cannot substitute human urine, saliva, or sweat samples

for self-reports (Anthony, Neumark, and Van Etten, 2000).



Once issues of definition and measurement have been mastered, research

fellows move on to what we call the "rubrics" of epidemiology--its main

subheadings and associated research questions. These main rubrics and

primary associated research questions are displayed in exhibit 2.



Successful research fellows learn these rubrics and use them to master not

only the state of currently available evidence on each form of the drug

dependence syndrome, but also the current gaps in evidence and the

research concepts and tools needed to fill the gaps in evidence.



The relationship of each rubric to an associated set of research concepts

and tools sometimes helps to clarify and differentiate the rubrics. Links

between each rubric and corresponding research concepts and tools are

presented in exhibit 3.



Quantity. Under the rubric of quantity, the main associated research

question is How many in the population are becoming affected, have

become affected, and are now affected?" In this context, "becoming

affected" can refer to becoming a drug user, developing drug dependence,

initiating criminal behavior, or some combination thereof (e.g., see

Gfroerer and Brodsky, 1992; Kosterman et al., 2000; Golub and Johnson,

2001a).



As reflected in the published scientific literature and technical reports

made available by NIJ, the Substance Abuse and Mental Health Services

Administration, and other Federal agencies, a substantial fraction of the

Nation's research expenditures on drugs and crime is directed toward the

"report card" function of public health and criminal justice research under

the rubric of quantity. A recent National Research Council report (Manski

et al., 2001) tallied more than 60 Federal agencies with data systems

designed to keep track of estimates on the number of drug users in

households, among school-attending youths, among arrestees, among

patients seen in emergency rooms, and in various other segments of

American life.



On the U.S. Department of Health and Human Services (HHS) side, we

have the National Household Survey on Drug Abuse (NHSDA) with

growing national probability samples of adolescents and adults (now with

a sample size of more than 70,000 subjects per year); the Monitoring the

Future (MTF) study, which started as a way to track drug use among

graduating high school seniors through a national probability sample each

year and now encompasses 8th and 10th graders; and a less intensive but

more massive Centers for Disease Control and Prevention (CDC)

surveillance of drug use and other health risk behaviors of teenagers in

school. On the U.S. Department of Justice side, we have other ambitious

survey operations, such as the Arrestee Drug Abuse Monitoring (ADAM)

program (formerly Drug Use Forecasting [DUF]), which monitors drug

taking among arrestees through both self-reports and bioassay techniques,

and the National Crime Victimization Survey.



Evaluated as part of the public health and criminal justice research

enterprise, these substantial efforts may be understood best as examples of

surveillance operations. The label "surveillance" does not trivialize the

important work of the professionals and scientists whose daily labors, year

in and year out, yield the hard-won surveillance data. In fact, many of our

country's surveillance operations in this domain of inquiry truly are gems

and tend to be regarded as the best of the best in the world. In some

respects, they are the envy of the world. Nonetheless, by definition,

surveillance activities are designed with timeliness and practicality in

mind, sometimes with deliberate decisions to sacrifice validity of

measurement in favor of enhanced survey response rates.



For example, NHSDA, MTF, and the CDC survey all use self-report

methods to measure drug-taking and crime-related behavior (e.g., weapon

carrying). The option of bioassays to confirm self-report data now is

regarded as impractical or too costly for surveys on a mass population

scale, and there has been concern expressed that bioassays might reduce

survey participation rates below acceptable values. NIDA is engaging in

survey research experimentation with bioassays to complement self-report

data to assess practical questions of this type. In the meantime, serious

concerns have been expressed about the capacities of these data systems to

provide evidence for policy evaluation (see, e.g., Manski et al., 2001).

Nevertheless, evaluated from the standpoint of original plans for the data,

these criticisms are somewhat impertinent as surveillance indicators. The

criticisms are asking the surveillance operations to do far more than they

originally were designed to offer.



The first rubric of epidemiology also encompasses studies of birth cohorts

that are intended to estimate risks of adversity, plot trajectories of

normative development, or quantify important population characteristics

such as rates of officially recognized offending. The concept of a cohort

study is familiar to criminal justice researchers and public health scientists

alike. Prominent examples in the criminal justice research arena include

Robins' classic nonconcurrent cohort study of children seen by child

guidance workers in the early 20th century (1966), and the work of Tracy,

Wolfgang, and Figlio entitled Delinquency Careers in Two Birth

Cohorts (1990).



The fact that the rubric of quantity is mentioned first does not mean that

research under this rubric is easy or a methodological snap. Not at all.

From the standpoint of data gathering, those of us who have recruited,

trained, and supervised teams of 60 or more field worker-interviewers and

quality control staff for data entry, documentation, and management can

appreciate the operational challenges in surveillance work. From a

statistical vantage point, the nature of the surveillance operations often

includes interdependent observations within samples (e.g., sampled

students within samples of schools; sampled household residents within

neighborhoods; multiple respondents within sampled households,

emergency rooms, or criminal justice facilities). These interdependencies

motivate solutions that call on the calculus (e.g., in Taylor series

linearization for variance estimation). In some estimation applications,

there is a need for Bayesian statistics not yet taught widely in graduate

research training programs.



As to the importance of these "counting" operations, we may turn to a

recent research contribution by Cohen, who incorporated values from

these surveillance operations in his attempt to estimate the monetary value

of rescuing a high-risk youth from a life of delinquency, crime, and other

socially maladaptive behavior (1998). To complete this work, Cohen had

to turn to an array of previous results from counting operations that

ranged from the National Institute of Mental Health Epidemiologic

Catchment Area surveys we conducted in Baltimore during the early 1980s

with colleagues at four other university-based sites to work that Blumstein

and his group completed to estimate basic parameters of criminal justice

research, e.g., an estimated 6 percent of all boys account for more than half

of all arrests (Blumstein et al., 1986). 



There can be little doubt that investigators in the drugs-crime arena should

be interested in Cohen's conclusions about varying programmatic

investments and the monetary returns from programs to intervene with

high-risk youths. Nevertheless, it is somewhat startling to know that

Cohen had to turn back to counting evidence gathered in the early 1980s

and before to produce estimates to be used for policy and programmatic

decisions almost two decades later. These quantitative estimates are not

Avogadro's number; rather, they are values expected to change over time,

if not from place to place. If we value probing quantitative criminological

research exemplified by Cohen's work, then we must ensure that the

drugs-crime research agenda includes periodic repetition of surveys to

yield the required estimates.



Studying the accumulated evidence on the drugs-crime relationship, we

have been able to sort much of it into the rubric of quantity. Quite clearly,

NIJ and NIDA now make a considerable investment in the basic counting

tasks required to estimate and quantify such parameters as how many

adult arrestees and juvenile offenders are taking drugs each year. Each

repetition of these surveillance operations provides evidence on variation

in the estimates from time to time and from place to place. The study of

this type of variation falls under the second rubric, which is called

location.



Location. Our second rubric is location, and the main associated research

question is "Where in the population are affected individuals more or less

likely to be found, with variation in occurrence and frequency

differentiated by characteristics of time, place, and person?" On occasion,

work under this rubric is guided by theory, but more often the research has

a more descriptive character.



James et al. (2002) provide an illustration of the nature of research and

evidence about location. The research team set out to plot geographic

variation in the occurrence of drug purchase opportunities experienced by

young adults in the United States. In this figure, a "drug purchase

opportunity," a special form of drug-related crime opportunity, is

measured by a survey response to a standardized assessment in interviews

conducted for NHSDA. As depicted in exhibit 4, and substantiated with a

univariate response regression model for description, there is variation in

the occurrence of these drug purchase opportunities across locational

regions of the country and for young men versus young women. In this

context, the statistical methods are not intended to probe the causes of the

observed variation from place to place, nor the observed male-female

differences. Rather, the methods are used simply to help quantify the

uncertainty in the survey-based estimates and substantiate the presence of

variation from place to place and the male-female differences (James et al.,

2002).



This illustration is useful because it reminds us that location refers not

only to geographic variation but also to variation in relation to

individual-level characteristics (e.g., sex, age, socioeconomic status,

ethnicity). For example, Fendrich et al. (1995) studied juvenile and older

murderers to understand varying degrees of drug involvement in murder.

Locational research also plots temporal changes, as illustrated in a recent

NIJ report on the possibility of new marijuana epidemics, to be described

below (Golub and Johnson, 2001b).



Estimates of the consistency of association between drug use and various

arrest and criminal behavior types also serve to illustrate analyses

focused on location within population experience: Crime was found to be

more common among drug users than among nondrug users. As Harrison

and Gfroerer (1992) make clear in their NHSDA analyses on this topic, the

research questions they were trying to answer concerned the number of

drug users, the number of individuals engaged in criminal behavior, and

the overlap in these numbers. With respect to location, their work clarified

the proportion of drug users who were engaged in criminal behavior and

the prevalence of criminal behavior in relation to drug use. As is true in

the work of James et al., these investigators did not draw on the apparatus

of causal inference, matching, or other scientific maneuvers to disentangle

whether the criminal behavior was a response to the drug use or vice versa.

Nonetheless, taking a step beyond studies of officially recognized crimes,

arrestees, and convicted criminals, Harrison and Gfroerer helped confirm

links between drugs and criminal behavior, but they did not seek to

produce definitive evidence about the causes of drug use or criminal

behavior.



Much of our current research enterprise at the interface of drugs and crime

has this type of descriptive character. Substantial HHS investments in the

MTF study and NHSDA already have been mentioned. On the NIJ side,

we call on ADAM to help clarify variation in the occurrence of drug use

among arrestees across multiple jurisdictions, not only in the United States

but also overseas. For the most part, we do not require these investments

to yield definitive evidence that might be central in causal inference.

Nonetheless, the evidence from these studies helps to describe the location

of drug taking, criminal behavior, and the intersections of these behaviors,

and sometimes to describe or predict the co-occurring and separate

patterns of drug use and criminal behavior.



Analyses conducted under this rubric without a special push toward causal

inference can be especially important in identifying hot spots within

geocoded areas as well as health disparities that might differentially fall on

one or another racial or ethnic minority group. Here, it is a predictive as

well as a descriptive purpose that can be achieved. However, when the

task is to predict and not to explain, there is no special calling for the

methods required for firm causal inferences, as depicted in exhibit 3.



Within the drugs-crime arena, there are many different examples of

surveillance operations under the rubric of location, such as we can see

in recent work by Golub and Johnson (2001b) in which they used

DUF/ADAM data as evidence to advance their claims about a new and

possibly expanding epidemic of marijuana use in the United States. True

to the descriptive character of locational research, Golub and Johnson

present evidence of the new and possibly expanding epidemic among

offenders in some areas (e.g., Atlanta) and evidence of no epidemic in

other areas (e.g., Miami), but they do not seek to explain why there should

be an epidemic in one place but not in another. Because these data are

from incarcerated individuals, an important set of complications arises in

their interpretation. One suspects that the observed time trends and

variation from place to place might reflect operations of local police

departments as much or more than it reflects any underlying change in the

dynamics of marijuana epidemiology.



This rubric of location also encompasses studies in which the investigators

may be striving toward causal explanation, but they fall short, often

demonstrated in a shift toward the language of "prediction" and away from

the language of "explanation." Two different hypothetical concluding

statements can illustrate this point. When the research team falls short of

its goal, the researchers may summarize their work by saying something

like "Based on this study's evidence, the level of drug use in early

adolescence predicted later delinquency and criminal behavior in the

young adult years." A different verb is selected for the alternative, stronger

form of concluding statement: "Based on this study's evidence, the levels

of delinquency and criminal behavior in the young adult years depend on

levels of drug use in early adolescence." 



As the focus shifts from description or prediction toward explanation and

causal inference, we move from the rubric of location to the third rubric of

causes. The shift in focus calls into play a new set of research concepts,

principles, and tools, as outlined in exhibit 3.



Many scholars will appreciate that a single study may contribute evidence

under several rubrics at once. For example, the periodic reports of

NHSDA, MTF, and ADAM routinely present evidence that falls under the

rubric of quantity as well as the rubric of location. Rarely, the authors of

these reports seek to make causal inferences from their surveillance data.



The yield of a study often is not clear at the outset or in the stages of study

planning, and the study orientation to theory is not always a discriminating

feature. Some theory-based studies have started as investigations of causes

but have ended up making contributions solely in the domains of

prediction and description. Other atheoretic studies end up making useful

contributions in our studies of cause. Consider the first conclusive study

on the topic of age-related risk of Down syndrome (DS) and associated

mental retardation, completed some 50 years ago. The investigators sought

to plot the risk of DS by the age of the mother at the time of delivery. An

exponential increase in risk after age 40 was clear in the first graphs. We

still do not know what causes the chromosomal trisomies that give rise to

DS, nor do we know why these trisomies and DS are more common when

older mothers give birth. But even in the absence of firm causal theory and

evidence, it has been possible to reduce the occurrence of DS in human

populations by encouraging mothers to bear their children before age 40.

Hence, a strictly descriptive study provoked an effective intervention to

reduce the occurrence of an important genetic condition.



It is regrettable that our studies of disparities affecting racially and

ethnically defined subgroups of the American population generally fall

under the rubric of location, as do our studies of the changing dynamics of

household and family composition in the United States. For example, we

now can say with some certainty that African-American males experience

rates of arrest, prosecution, and incarceration for drug possession offenses

that cannot be explained by their rates of drug taking, but we do not have

good evidence on the causes of this racial disparity. Initial inquiries

suggest differential law enforcement and judicial practices, which

sometimes encompass racial profiling, but rigorous scientific evidence

on these practices is scarce.



With respect to the dynamics of household and family composition, the

phenomena of youthful drug taking and related criminal offending have

links back to the families of origin, now often characterized by absence or

infrequent appearance of the father in many of our population groups.

This is not to say that female-headed households are homogeneous or

uniformly deleterious with respect to socially maladaptive behavior of

young people. It would be a mistake to presume that the traditional

mother-father household always and in all contexts is more protective than

a female-headed household with respect to the risk of youthful drug taking

or delinquent behavior (e.g., see Chilcoat, 1992). Mothers often mobilize

family resources or draw on assets that in some measure may help

compensate for absent fathers (e.g., by involving grandparents, neighbors,

church groups), as described by Kellam, Ensminger, and Turner (1977),

Pearson et al. (1990), and others.



A research agenda on race, ethnicity, and family or household composition

can be motivated by an awareness that the drugs-crime relationships will

depend to some extent on demographic trends. Against the backdrop of

demographic trends such as these, including an increased prominence of

Hispanic children and families in the United States, it will be important to

sustain the research agenda in the domain of locational variations of this

type. Important steps in this direction have been taken in the Federal

agencies responsible for surveillance of drug-related behaviors, including

increased attention to measurement of ethnic self-identification (e.g.,

with respect to Cuban origin, Puerto Rican origin, and other subgroups of

the Hispanic population; with respect to Chinese origin, Samoan origin,

and other subgroups within the Asian-Pacific Islander category). Similar

attention is required in criminal justice research such as ADAM and

I-ADAM (International ADAM) and in administrative statistics compiled

on operations of the criminal justice system in this country.



Whereas the human genome project is challenging conventional views

about "race" as a scientific concept, studies on self-identified

race-ethnicity will have a sustained importance in the NIJ-NIDA research

agenda on the topic of drugs and crime. 



This evaluation of importance can be grounded in an awareness of the

demographic trends described above, but it also draws on an appreciation

of what studies of self-identified race-ethnicity may teach us about the

influence of cultural contexts and socially learned behaviors with respect

to drug taking and criminal behavior. 



Finally, a note on ethnographic studies should be added here. In general,

the sample size and "sample space" characteristics of these studies do not

make ethnography an especially fertile discipline with respect to the first

rubric of quantity, except when the characteristic under study has

extremely limited dispersion. For anyone who looks to ethnographic

studies for quantitative values, there often are some unanswerable

questions about generalizability and precision of the study estimates. In

some respects, ethnography might be characterized as a search for the

boundaries of no variation in a socially shared human characteristic.



This is not to say that ethnography is barren when it comes to quantitative

data. To the contrary, the small scale of ethnographic research makes it

possible for ethnographers to shift directions more quickly than is possible

in ordinary surveillance operations. As a result, ethnographic field workers

helped in the early identification of crack cocaine, methamphetamine, and

oxycodone outbreaks--years before these outbreaks could be identified in

large-sample surveillance data.



Under the rubric of location, ethnographic field workers were among the

first to note inner-city adolescents whose drug taking started with

marijuana rather than with the more normative experiences with alcohol

and tobacco. They also were the first to characterize a growing use of

"blunts"--tobacco cigars hollowed out and filled with marijuana for a

combined tobacco-marijuana intoxication (Golub and Johnson, 1999). In a

recent round of observations, there is a suggestion that for some youths,

the typical "gateway" drugs have been skipped--an example of subgroup

variation in the more typical developmental sequences running through

alcohol, tobacco, and marijuana to drugs such as heroin, stimulants, and

hallucinogens. Large-sample epidemiological surveillance data now seem

to confirm the initial ethnographic observations on this topic (e.g., see

Golub and Johnson, 1999; Golub and Johnson, 2001a).



One of the reasons ethnographic research is important under the rubric of

location is that it can open our eyes to new conceptions of time, place, and

personal characteristics that impinge on the drugs-crime relationship.

These ethnographic studies are especially useful in descriptions of the

cultural context and socially learned behaviors described above. Their

evidence can add depth and insight to otherwise superficially understood

intersections of drug taking and criminal behavior.



Causes. The third rubric of epidemiology pertains to the study of causes

and draws on the research apparatus required for causal inference (exhibit

2). On occasion, this research apparatus can be quite simple in concept.

For example, a relatively small sample of monozygotic (MZ) twins

discordant for an important outcome is sufficient to provide definitive

evidence about environment with respect to the causes of that outcome.

These MZ twins are genetically matched: If they are discordant for

outcome, one may look for gene-environment interactions, but more often

one looks for differences in environmental conditions in utero (e.g.,

dichorionic versus monochorionic sacs), perinatally (e.g., insults at the

time of delivery), or in later development (e.g., head trauma for one twin

but not the other during infancy or childhood). The National Institutes of

Health (NIH) investment in recent twin research to estimate heritability of

different forms of drug use now generally is paying off in two ways:



o Each study is indicating at least some degree of heritability of drug

dependence, and sometimes heritability of drug use, especially legal drug

use (e.g., tobacco). 



o Each study is indicating ample room for gene-environment interaction

or for influence of environmental conditions and processes. 



These results from causal research help substantiate a case for a future

research agenda on the genetic sources of variation and on environmental

modulation of these genetic sources of variation.



Randomized trials with relatively simple structure also can be used to

probe causal hypotheses with definitive results. For example, these trials

may offer our best avenues toward definitive evidence on whether

cessation of illegal drug use is followed by reductions or elimination in

criminal behavior. An alternative is to nest the study of causes within a

more expanded agenda of systems research on drugs-crime relationships

(Manski et al., 2001). To the extent that systems research entails a finely

detailed specification of mechanisms that link events and processes within

a system, this type of research falls more clearly under the rubric of

mechanisms, as described below.



Outside of the simplicity of research on discordant MZ twins and

randomized controlled trials, a complex apparatus of study design and

statistical method is required to extract definitive evidence in research on

drugs-crime relationships. Given the importance of inferences about

causes in the drugs-crime relationship, it may be understandable that

graduate research training programs have become increasingly

methodological in their orientations.



It may be appropriate to discuss the potential contribution of ethnographic

research in relation to the causes of the drugs-crime relationship. To

date, most ethnographic research on drugs and crime has been descriptive

in character. It has provided leads for more probing causal investigations,

but it has not produced definitive evidence on the links between drugs and

crime. In this respect, ethnography's contribution may be most important

under the rubric of location. Before anyone could mobilize large-sample

surveillance operations to study the new drugs-crime phenomena

connected with crack cocaine (e.g., crack and prostitution), it was possible

for ethnographers to move in and make headway. To some extent,

ethnographers have been pioneers in research on methamphetamine and

club drugs such as MDMA (Ecstasy), and we can expect more of the same

in relation to our first new drugs-crime outbreaks of the 21st century,

which involve sustained release oxycodone.



An NIJ-NIDA investment in ethnographic research on drugs-crime

relationships of this type will continue to be important--if only to help us

begin to understand the unusually circumscribed geographic distributions

of methamphetamine and oxycodone use in the United States and the

patterns of criminal behavior associated with use of these drugs.

Ethnography can be used to produce a catalog of causal explanations for

methamphetamine's emergence as a threat to public health and public

safety in rural sectors of the American Midwest and for oxycodone's

emergence in small cities and towns of the Appalachian mountain range,

especially from West Virginia southward. It is not clear that ethnography

or any other scientific field will be capable of producing definitive

evidence about specific explanations in this catalog of causes.

Nonetheless, there is value and importance in the attempt to do so, and

ethnographers can bring rigor and scientific discipline to this process of

investigating these causes. The alternative seems to be to leave these

investigations to the field of journalism.



Mechanisms. Within epidemiology generally, mechanisms refer to

linkages of states and processes that lead toward expressions in clinical

features of health and illness or disease. As applied to the drugs-crime

relationship, one might ask about the mechanisms of linked states and

processes leading to or away from an association between illegal drug use

and criminal behavior.



For an illustration of these mechanisms, one may turn to the coercive

process and deviancy training models introduced in the work of Patterson

and Dishion. Their Oregon Boys study has provided longitudinal evidence

on what surely must be central linkages in the mechanisms underlying

drugs-crime relationships (e.g., see Patterson, Dishion, and Yoerger 2000).

For example, studying these school-based samples of boys through ages

17-18, and using standardized coding of a 30-minute free

discussion-interaction between best friends, they found substantial

over-time correlation of deviant friendship process (e.g., duration of

rule-breaking talk bouts as coded from videotape). Dishion also has

reported on a link from initial drug use to increased affiliation with deviant

peers and onward to initiation of criminal behavior that is more consistent

with the delinquency-to-drugs link that emerged in the longitudinal

research of Johnston and colleagues based on MTF analyses published

more than 20 years ago (Dishion et al., 1996; Johnston, O'Malley, and

Eveland, 1978), as well as on more recent studies (e.g., Elliott and

Huizinga, 1989).



The use of multiwave longitudinal study designs to probe into suspected

causal mechanisms is well known in both public health and criminal

justice research circles. The Alcohol, Drug Abuse, and Mental Health

Administration (the precursor to SAMHSA), and more recently NIH and

OJJDP have maintained support for a series of important longitudinal

studies over the years (e.g., see the work of Jessor and Jessor, Kellam and

Ensminger, Block and Block, McCord, Bachman, Kandel, Robins,

Elliott and Huizinga, Hawkins and Catalano, and many other studies of

this type, as listed in compendiums such as Verdonik and Sherrod, 1984).

Advantages of long-term investment in these longitudinal studies can be

seen in the research articles from many of the research projects with

multiwave assessments, for example, the Pittsburgh Youth Study (e.g.,

Loeber et al., 1998), the Denver Youth Study, and the Rochester Youth

Development Study (Loeber et al., 1999); and the research groups led by

the Brooks, Newcomb, and Bentler (e.g., see Brook et al., 1996, Brook et

al., 2000; Newcomb and Bentler, 1988; Newcomb 1992).



One of the questions in the design of an agenda for future research on

drugs and crime is how the evidence from large- and medium-sized

samples from longitudinal studies of this type might be linked with

evidence from the generally much smaller intensive studies of cases. Until

there is consensus about effective interventions to disrupt the drugs-crime

relationship, possibilities for a linkage exist through the concept of natural

history.



In the history of medicine and medical research, the first natural historians

of disease were clinicians and clinically oriented observers who made

careful observations at the bedside of patients, in the absence of effective

interventions. They watched, measured as best they could (e.g., body

temperature), and described change in relation to the passage of time from

the first recognition of clinical features. Within the realm of drugs and

crime research, ethnographers and social scientists generally have taken

over the responsibilities of careful clinical observers in relation to illegal

drug use and criminal behavior. During the last 50 years, thanks to the

work of Robins (1966), Winick (1962), Preble and Casey (1969), Agar

(1973), Waldorf (1998), Nurco (Nurco et al., 1975, 1996; Nurco, 1998),

and their successors, we have learned much about the natural history of

drug use, drug dependence, and associated criminal behavior through

ethnographic and social science research. 



The natural history of a disease proves to be an important element under

the rubric of mechanisms. In the past, a careful description of a disease's

natural history often has guided investigators toward underlying causal

mechanisms.



In the years before effective drug treatments, Winick and others drew

attention to the maturing out process for drug addicts, and there is a

parallel literature on maturing out with respect to criminal behavior in

general (Winick, 1963). The maturing out process continues to be an

important locus for new research on the drugs-crime relationship.



Other clues about causal mechanisms are being produced in observational

and longitudinal studies of individual cases or families characterized by

some feature of the drugs-crime relationship. For example, we have

Dunlap's intensive studies of families in which one of the members is a

crack cocaine dealer (Dunlap and Johnson, 1996; Dunlap, Johnson, and

Manwar, 1994); research such as Spunt's study of adolescent offenders

with a history of violent crime (Spunt et al., 1990), Longshore's linkage of

DUF and California Bureau of Criminal Statistics data (Longshore 2000),

and the earlier related studies started by Hser, Anglin, and McGlothlin

(1987); and investigations led by Inciardi, Johnson, and Goldstein or

members of their research groups (e.g., see Inciardi and Russe, 1977;

Inciardi 1990; Inciardi and Pottieger, 1998; Johnson, Dunlap, and Maher,

1998; Goldstein 1998; Spunt et al., 1990, 1994, 1995).



Several interesting elaborations of these intensive case studies have

developed in the realm of criminal justice research. For example, Logan

(2001) has added bioassays for metabolites of the neurotransmitter

serotonin as well as testosterone assays as part of his intensive followup

studies of crack users. This example serves to illustrate a potential

intersection of public health and criminal justice research that should be

explored in more depth as we work through a future agenda for research

on drugs and crime.



A conceptual shuttling back and forth between these intensive smaller

sample studies and the generally larger sample longitudinal studies would

seem to have advantages for investigators who work in one or another of

these arenas, and there are a few investigators who conduct both types of

studies (e.g., see Dishion and Loeber, 1985; Dishion, Patterson, and Reid,

1988; Dishion et al., 1996). This type of bridgework between the

microsocial and ethnographic research traditions and large-scale

longitudinal sample research deserves to be a deliberate focal point on the

future drugs-crime research agenda. This focal point is important because

the study of causal mechanisms and processes can draw attention to

potentially vulnerable links where new interventions might be directed.



In epidemiology generally, the focus of research on causal mechanisms is

shifting to genes and encoded gene products, as displayed in our most

recently emerging subspecialties of genetic epidemiology and molecular

epidemiology. To some extent, Elliott has a head start in a potential

cross-fertilization between criminal justice research, genetic epidemiology,

and molecular epidemiology. He already has introduced harvesting of

DNA samples in the context of his national longitudinal study (Elliott,

2001). Opportunities for case-control studies and other genetically

informative designs, including whole genome scans nested in a

case-cohort study design, will become possible as this research evolves.

Eventually, this type of work should lead us toward more definitive

evidence on causal mechanisms underlying the drugs-crime relationship,

including gene-environment interactions.



It is possible to make a forecast of likely integrations of genetic research,

cognitive sciences, and the more traditional disciplines of behavioral and

social sciences for a future agenda for NIJ and NIDA research on

drugs-crime relationships. For example, exhibits 5-7 represent an

elaboration of conceptual models our research group has developed as an

aid to our study of transitions from drug use to drug dependence. Exhibit 5

expresses a suspected causal influence of drug use on criminal behavior. It

also expresses a separate influence of drug dependence on criminal

behavior. These two specifications are consistent with the

Goldstein-Brownstein distinctions between drug-related crimes that

might arise from acute drug intoxication states versus crimes that are

rooted in the economic-compulsive behavior of an individual who suffers

withdrawal states as a result of sustained drug use and neuroadaptation.

There are many law-abiding drug dependent individuals who do not

commit crimes, even when they are suffering from withdrawal pains.

Hence, exhibit 5 includes a speculative causal pathway that runs directly

from withdrawal to the occurrence of criminal behavior, over and above

the separately specified role of the drug dependence syndrome for which

withdrawal serves as a manifest indicator.



We speculate that an individual's genome can contribute to the

drugs-crime relationship in different ways. Exhibit 5 concentrates on a

possibility that some genetic polymorphisms or mutations may be

intercorrelated manifestations of an underlying diathesis or vulnerability to

make the transition into drug dependence from a state of nondependent

drug taking, as reflected in pathway 1. It also specifies a possibility that a

specific polymorphism (or mutation) has an additional influence on this

transition, as reflected in pathway 2. As indicated by pathway 3, we may

hope for development of effective intervention techniques that can disrupt

what otherwise might be an expression of the diathesis. If effective, these

interventions will slow or disrupt the natural history of drug dependence at

a step in the process that links nondependent drug taking and the

subsequent transition into drug dependence. This effect of intervention, by

itself, may be sufficient to alter the drugs-crime relationships depicted to

the right of the exhibit.



The potential role of the cognitive sciences is expressed in the

intermediate pathways that link nondependent drug taking and drug

dependence to later criminal behavior. Here, aggression may be

conceptualized in a generic sense as rowdy misbehavior or social

maladaptation secondary to drug taking, which can occur with or without

criminal behavior. Executive dysfunction refers to impairments in the

cognitive processes that subserve human capacity to plan, direct, and

control one's future behavior within adaptational boundaries and may

encompass more generalized planning behavior (e.g., see Tolman,

Edleson, and Fendrich, 1996).



As depicted in exhibit 5, during states of acute drug intoxication, there

may be a release of aggressive behavior and a disruption of regulatory

executive functions. As levels of drug dependence increase, levels of

aggressive behavior can change in an upward or downward direction and

executive dysfunctions can occur. The complexity of interrelationships

between aggression and executive dysfunction is reflected in the reciprocal

causal paths between these two constructs. Increased executive

dysfunction translates as inept decisionmaking about aggression and the

subjective utility functions that govern decisions about whether to commit

a crime. As part of generally adaptive fight-flight responses and

modulation of monoamine neurotransmitter signaling pathways during

bouts of aggression, there can be a cascade of executive dysfunctions:

Mere rowdiness can be transformed into aggravated assault.



To be sure, exhibit 5 is only a model that represents little more than an

oversimplified representation of the complexities that link an individual's

genome with cognition and behavior. Models by definition are

oversimplified representations. It is fair to ask whether the model requires

additional specifications, such as the possibility that religious convictions

might tend to modulate the relationship between drug taking and

aggression or criminal behavior. In this oversimplification, exhibit 5 does

not convey all such possibilities. These possibilities for elaboration of the

longitudinal model should help the reader understand some of the

complexities faced in observational studies of causal mechanisms that

account for observed drugs-crime relationships.



Exhibit 6 presents even more simplification to sharpen focus on the

drugs-crime relationship specifically. The genetically based diathesis and

other covariates of exhibit 5 are set into the background (i.e., presumed to

exist but not explicitly depicted). In exhibit 6, we see a readily appreciated

reciprocity between the level of drug taking and the level of drug

dependence: (a) the more drug taking, the more we find increased drug

dependence levels, and (b) the more drug dependence, the more we find

increased levels of drug taking. We also see the level of criminal behavior

expressed as a function of levels of drug taking and drug dependence, as

shown in exhibit 5. An additional elaboration involves the longitudinality

of this model. We have levels of criminal behavior at one point in time

influencing levels of criminal behavior at future points in time, but in

exhibit 6, we do not specify a link from levels of criminal behavior to

subsequent drug taking or drug dependence levels. At least in theory, and

in some prior suggestive evidence, this omission represents a potentially

important misspecification of our model for the drugs-crime relationship

(e.g., see Johnson et al., 1995).



Exhibit 7 adds a level of complexity to the model depicted in exhibit 6 and

poses a substantive question for the agenda of action research: "How

might an intervention lead to change in this system of interrelationships?"

We introduce the possibility that social status (e.g., status in the

community, socioeconomic status, lawful income) depends on criminal

behavior and also on the level of drug dependence, and that criminal

behavior influences the subsequent level of drug dependence by its

intermediate influence on social status. The model depicted in this exhibit

also provides for a plausible link from the level of drug dependence to

subsequent criminal behavior. Namely, as drug dependence increases and

social status (e.g., lawful income) falls short, criminal behavior may

increase (as in the Goldstein-Brownstein tripartite model). In addition,

subsequent levels of drug dependence may be influenced by the changes in

social status, either upward or downward.



The model in exhibit 7 also introduces a conglomerate concept of

"assortative peering," expressing a well-known truism: "birds of a feather

flock together." The occurrence of drug taking is linked to later formation

of peer group relationships, as is the occurrence of criminal behavior. To

some extent, we can say that past drug use and past criminal behavior

influence current peer group affiliations, and to some extent, we can say

that past peer group affiliations influence future drug use and future

criminal behavior. These complexities are expressed by hypothesized

causal paths in exhibit 7.



Conceptual models of this type are incomplete representations of the

causal mechanisms that lie beneath observed drugs-crime relationships, yet

they are elaborations of the Goldstein-Brownstein tripartite model.

Nonetheless, most readers will agree that these representations are

oversimplified. If they have value, it is to highlight some future directions

for the joint NIJ-NIDA research agenda on drugs-crime relationships. 



We do not yet have a longitudinal research program to investigate the

relatively simple model of interrelationships between levels of drug use,

drug dependence, and criminal behavior as depicted in exhibit 6, let alone

the more complex model of exhibit 7, with its sociological construct of

social status and the social-psychological construct of affiliation with

behaviorally similar peers (assortative peering, homophily, etc.).

Fortunately, there already is a cadre of criminologists and drug researchers

who are trained in sociology and social psychology and can readily

incorporate the biomedical and clinical concepts of drug dependence into

their research plans, if supported to do so.



It will be more difficult to forge a research agenda that integrates the

genetics research and cognitive sciences constructs depicted in exhibit 5.

For the most part, genetics and cognitive sciences are unknown territories

for most NIJ and NIDA investigators who have made important

contributions in past research on the drugs-crime relationships. For most

drugs-crime researchers, it would not be difficult to integrate concepts and

measurements of aggressive behavior and the clinical syndrome of drug

dependence within their existing research plans. Far more difficulty will

be encountered during the process of integrating genetics and the

neuropsychological and neurophysiological measurements of the cognitive

sciences. 



We can learn a lot about the drugs-crime relationship simply by replicating

and refining important longitudinal research on drugs-crime relationships

that was initiated during the second half of the 20th century. Many of these

longitudinal studies have cohorts that still are intact, and followup studies

are now underway to learn more as these cohorts mature into adolescence

and make the transitions into young and middle adulthood. There are

mountains of data from 20th-century studies that have not yet been fully

exploited through careful analysis.



Nonetheless, as we look forward through the next decades of research, the

NIJ-NIDA agenda must go beyond what has developed as the best

20th-century research on the drugs-crime relationship. Ten decades from

now, if we are to leave the 21st century with an enhanced understanding of

the drugs-crime relationship and with a greater capacity for effective

action to improve public health and safety in this domain, we cannot

continue to work within the narrow paradigms and methodologies of the

traditional scientific disciplines mastered by drugs-crime investigators

of the 20th century. If we are successful, then in a few decades, the

biomedical, genetic, and cognitive science substrates of the drugs-crime

relationship will no longer be a matter of mere speculation, as depicted in

exhibits 5 and 7. There will be definitive evidence, solid understanding,

and effective action-plans based on what we learn from the pioneers who

move into that now-unexplored territory.



Prevention and control. The long-term value of research on causal

mechanisms depends on identifying potentially vulnerable linkages in the

sequence of states and processes that lead to illegal drug use and criminal

behavior. It may go without saying that increasingly definitive evidence

about causes and causal mechanisms will help us achieve our goals in the

domain of effective prevention and control. Nonetheless, a reminder may

be useful with respect to a dynamic interrelationship between etiological

studies (of causes) and the emergence of effective interventions. As

illustrated in the circumstance of DS and maternal age, with limited

evidence on the underlying causal mechanisms of DS, by manipulating

maternal age we have a very effective instrument to prevent and reduce the

risk of DS. As explained in our original paper on the rubrics of

epidemiology, many effective public health preventive interventions

emerged before firm knowledge about causes and causal mechanisms

became available (Anthony and Van Etten, 1998).



A related concept involves the use of randomized preventive trials to

provide increasingly definitive evidence about suspected causal

relationships. Some readers of this paper will know of work that Kellam

and our Johns Hopkins research team have completed, using randomized

field trials to probe the interrelationship between early aggressive and

rule-breaking behavior and later drug involvement among boys (e.g.,

Kellam and Anthony, 1998). In essence, we decided that more

observational research on the link from early aggression or deviance and

later drug use would be less important than an experimental test. Within

the framework of a randomized field trial, we tried to and succeeded

in reducing aggressive and deviant behavior of first-graders using an

experimental intervention assigned at random. For the boys assigned to

experimental intervention, we have found later reduced occurrence of drug

involvement, and we have replicated these results in a second cohort of

first graders (Kellam and Anthony, 1998). More replications along these

lines are needed before anyone can claim that early aggression or deviance

is a "cause" of later drug use, but this experimentation illustrates how

experimental research in the domain of prevention and control can yield

benefits in the form of improved evidence to test causal theories. This idea

is not new. Hawkins, Catalano, Offord, and others have noted it as well

(e.g., Hawkins, Von Cleve, and Catalano, 1991; Hawkins, Catalano, and

Miller, 1992; Jones and Offord, 1989). But it is an idea that often is

overlooked by investigators more interested in theory testing and who

orient themselves toward goal 1, described in this paper's first paragraph.

Under the fifth rubric, we try to orient the research to serve both goal 1

and goal 2.



Because elements of this rubric of prevention and control are being

covered in the companion papers that accompany this working manuscript,

I will close this section more quickly than might be customary. Before

doing so, I would like to mention the contributions of operations research

and systems research in this domain, which often have been neglected in

epidemiology. Over the years, the thoughtful and quantitatively sharp

work of Blumstein and colleagues has continued to inspire an important

line of research on prevention and control that is pertinent to the

drugs-crime relationship. Although I am not confident about all of the data

or assumptions of the underlying analysis approaches, I have been

especially impressed by the directions taken by Blumstein colleagues

Caulkins and Cohen in this regard.



For example, Cohen (1998) discusses potential synergy of programs and

distinguishes the aggregate benefits of programs designed to reduce

crime versus the sum of the benefits of individual programs. It is possible

that no single program would help city residents feel safe enough to derive

lifestyle-related expenditure benefits (e.g., walking a mile through a rough

neighborhood versus taking a taxicab because of concerns about safety).

Combinations of programs might do so. This distinction ties into the

concept of marginal costs versus average costs associated with drug-using

and delinquent youths or criminals, where the marginal costs exclude fear

of crime and private security expenditures because these costs are largely

unaffected by any one criminal's actions.



Caulkins and his colleagues developed a challenging line of systems

research that can ultimately yield new ideas and evidence about policy

instruments in relation to the drugs-crime relationship. The evolution of

this work toward selection of policies and programmatic instruments at

different stages of a drug-taking epidemic is especially important

(Caulkins, Crawford, and Reuter, 1993; Behrens et al., 1999). 



A Selective Overview of Tensions Faced in Research on Drugs and Crime



Numerous tensions are faced at the intersection of public health and

criminal justice research on the drugs-crime relationship. This section

identifies and describes a selection of these tensions, and in some

instances recommendations are offered for NIJ and NIDA action to help

resolve the tensions.



Tensions in Theoretical Perspectives, Concept, and Definition



Heterogeneity at the intersections of public health and criminal justice

research is not limited to differences of opinions and judgment about

empirical observations, the inferences we can draw from these

observations, and the uses to which we apply the observations (e.g., cost

analyses of alternative programs). There are some fundamental tensions

within and across theoretical perspectives and also approach.



The concept of scale. Ecologists work with a concept of scale that may

help us understand some of the tensions mentioned above and may serve

as an axis of orientation as we turn to future directions for research (e.g.,

see Brown, 1995; Wiens et al., 1986). As a concept, scale resonates with

what educational researchers and behavioral and social scientists often

have termed multilevel or hierarchical models, as in Ennett's and the

Duncans' research with young people nested within ecological niches of

higher order such as classrooms, schools, or families (e.g., see Ennett

et al., 1997; Duncan et al., 1997; Duncan, Duncan, and Hops, 1998) and

our own research group's nesting of individual drug users and collections

of drug users in their neighborhoods of residence (e.g., Bobashev and

Anthony, 1998; Petronis and Anthony, 2000). Parker and Toth (1990) also

have appealed to related macro versus micro concepts in their research on

alcohol and homicide, as have Patterson and colleagues in their research

on peer groups (Patterson, Dishion, and Yoerger, 2000). Bronfenbrenner's

ecological systems theory for human developmental research slices scale

into macro, meso, and micro divisions that many investigators have found

useful (Bronfenbrenner, 1979, 1986).



Although not with any direct reference to a formal ecological concept of

scale, we can see resonance of this concept in Markowitz and Grossman's

studies of taxes and regulations on alcohol and their hypothesized effects

on criminal behavior (Markowitz and Grossman, 2000), the research of

Caulkins and colleagues on national drug policy and programmatic

initiatives (e.g., Behrens et al., 2001), and Holder's research on preventive

interventions directed toward communities in the United States (Holder

1993, 2001; Holder et al., 1999, 2000). Scale is worked outward from the

individual organism in the direction of larger social groups, organizations,

and geopolitical units. In public health and criminal justice research, we

often refer to pre-established institutional or geopolitical boundaries

(schools, census tracts, nations) when we work at higher scale. In ecology,

mathematical models and methods such as advanced wavelet analysis are

used to allow the empirical data to inform scale--as in research on

landscape ecology (Anthony and Bradshaw, 2001).



Some tensions arise in research when investigators ignore scale in their

theoretical perspectives or empirical research reports. For example, most

of us work within a conceptual framework that leads us to comprehend

estimates of the drugs-crime relationship without reference to scale.

However, one should expect the drugs-crime relationship to be of one

order of magnitude when we are investigating individuals who all reside in

the same local area (e.g., as in much of the ethnographic research on the

drugs-crime nexus), a different order of magnitude when we work with

individuals and data from multiple neighborhoods, but with matching on

neighborhood in the analysis, and a different order of magnitude when our

data are from individuals across the Nation, with no analytical attention to

who lives near whom, except perhaps during the process of estimating

variances for confidence limits and standard errors (e.g., see Bobashev and

Anthony, 1998). 



Although not clearly within the scope of the original ecological concept of

scale, an investigator may work inward from the boundaries of the whole

organism toward subunits, ultimately leading to the signaling pathways

between neurons, messenger systems originating from genetic material,

and the simplest proteins and the encoding genes themselves. This

elaboration of the concept of scale creates yet another tension, in part

because the concepts of genetics and signaling pathways for

neurotransmission are more familiar in the public health research

community but are not yet in the mainstream of graduate or postdoctoral

research training in the criminal justice research community. To illustrate,

when I have talked with my criminal justice research colleagues about

Elliott's inclusion of DNA assays in the most recent waves of his National

Youth Survey, many of them have asked, "Why?" To be sure, some

skeptical behavioral genetics colleagues also have asked, "Why?" but this

is an instance in which the same verbal behavior has origins in

substantially different theoretical models. My point is that tension can

arise when concepts of scale are not made explicit.



Some of the work at the intersections of public health and criminal justice

research will be to make our concepts of scale explicit. In some respects,

this will be more readily accomplished as we work from the whole

organism outward, and the task may be more difficult as we try to integrate

molecular biology, genetics, and neuroscience into our discussions.

Nonetheless, this hard work will be essential as we make a 21st-century

science of drugs-crime relationships.



Orienting definitions and constructs. The literature also displays

considerable heterogeneity in orienting definitions and constructs. On the

public health side, there often has been an orientation toward drug use and

drug dependence or addiction as useful constructs in their own right. One

orientation often has been called the "medical model," but it amounts to

little more than an analysis of empirical syndromes (i.e., co-occurring

manifestations of the neuroadaptational processes that get started when

drug use begins, followed by a cascade of secondary and tertiary

adaptations, some of them occurring in the domain of social adaptational

roles and responsibilities). In a later section of this paper, I will return to

this syndrome concept. On the criminal justice side, drug use and the drug

problems associated with drug use often are treated as if they are not

interesting in their own right but are something akin to interchangeable

observable manifestations of something else that is more fundamental,

such as the "problem behavior syndrome" construct first elucidated by

Jessor and Jessor (1977) some 30 years ago. A more recent version of this

concept is a general deviance construct used by Scheier, Botvin, and others

in empirical studies (e.g., see Scheier and Botvin, 1996), and there also is a

recent respecification of the Jessor and Jessor model, with elaborations

that encompass the epidemiological concepts of risk factors and protective

factors (Jessor, 1998).



The literature also shows heterogeneity in the typologies of criminal

behavior or social maladaptation. Notions of childhood conduct disorder

followed by Antisocial Personality Disorder appear prominently in some

formulations, but are absent elsewhere (Loeber and Schmaling, 1985;

Stevens, Kaplan, and Bauer, 2001; Langbehn and Cadoret, 2001).



These definitions and constructs in our theoretical perspectives demand

work at the intersection of public health and criminal justice research. If

we cannot bridge these different approaches or marry them to produce

adaptive offspring, they will prove to be an unending source of unresolved

tension with implications for research progress. Unresolved tensions slow

down our progress in research that depends on a peer review process,

whether the peer review occurs at the stage of reviewing proposals or of

vetting journal articles.



At NIJ and NIDA, an important part of the research agenda can be a series

of meetings or technical workshops. The charge to workshop participants

is to bridge these orienting concepts and definitions across disciplines or

create an articulation between concepts that will accelerate research 

progress on drugs-crime relationships rather than slow it down.



Orienting conceptual frameworks and theories. The originating biomedical

branches of public health research sometimes take theory as a given or

work with theory in the background when there are emergent problems of

human suffering and disease to be solved. For example, the important

20th-century line of investigations required to identify lung cancer as an

adverse consequence of tobacco smoking was guided more by implicit 

concepts of carcinogenesis secondary to tobacco smoking. Strongly

articulated, explicit theories, if any, would have been mis-specified and

incomplete in that they could not possibly have incorporated the

postsmoking DNA adducts, protein adducts, and gene-encoded

metabolizing enzymes now prominent in the models of carcinogenesis. In

criminal justice research, true to its origins in the social and behavioral

sciences, the theoretical underpinnings are made more explicit (e.g., see

Thornberry, 1997; Kaplan, 1995). One might say that without explicit

theory, the research in this domain stands little chance in peer review, no

matter how important the empirical contribution.



This is another source of tension at the intersections of and sometimes

within the domains of public health and criminal justice research. In

Public Health Service study sections, I have seen study section members

be less than enthusiastic about proposed epidemiological research on drug

use and Antisocial Personality Disorder because the applicants had not

oriented themselves to the major theories of deviance well known in

criminal justice circles: "inadequate conceptual model" is the phrase that

comes to mind. I also have observed major differences of opinion about

scientific priority among experts in the criminal justice and social science

world, some of whom are comfortable with "psychologizing" constructs

within their theories (e.g., the self-derogation models developed by

Kaplan), and others who are more focused on constructs with a behavior

analytic origin (e.g., coercive process and deviancy training models

developed by Patterson, Dishion, and their research groups in Oregon).



Tension in relation to theoretical models has been readily apparent in this

NIJ-NIDA collaboration, which has offered a chance to step back and look

over a broad expanse of scientific progress in public health and criminal

justice research on the drugs-crime association. This broad perspective

creates germs of ideas that might be useful in a synthesis or integration of

various theoretical perspectives that range from the disciplines of

molecular or behavioral genetics to those of econometrics and the other

social sciences. However, there clearly is diversity and tension even within

fields as narrow as behavior genetics, where some work is oriented toward

developmental family processes (e.g., as advanced in the recent work of

Neiderhiser and colleagues), and other work is not (e.g., see Neiderhiser et

al., 1998, 1999; Neiderhiser, 2001; Brennan, Mednick, and Jacobsen,

1996; Tehrani et al., 1998; Kotler et al., 1999).



These tensions surface most clearly in debate and discussion of an

intersection of genetics research and studies of the drugs-crime

relationship. Many investigators from social science backgrounds are

hesitant to take part in discussions of genetics, gene expression, and

mechanisms of inheritance that might account for covariation of

drug-taking behavior and criminal offending. This hesitation can be traced

in part back to serious and important concerns about ethical issues,

eugenics, and the like. Some of the hesitation can be traced back to a gap

in graduate education: Graduates of social science training programs often

have not mastered the basics of human biology and genetics.



Looking from a different perspective, an observer can see other sources of

tension in relation to conceptual framework and theories. Graduates of

human biology and genetics programs often have not mastered the basics

of behavioral and social sciences research.



The intersection of the Human Genome Project, gene expression, and

proteomics with research on drugs-crime relationships merits close

attention at NIDA and NIJ. To some extent, this intersection can be

cultivated in a gradual process of shaping new investigators. NIDA's peer

review of its portfolio of research training programs and individual career

development awards can specify requirements for cross-discipline mastery.

On one side, new social science investigators can be required to master the

basics of human biology and genetics. On the other side, new human

biology and genetics investigators can be required to master the basics of

behavioral and social sciences.



NIDA already is sponsoring a series of training workshops for new

investigators to expose them to the different disciplines that now

contribute to its research mission. The initial workshops have focused on

epidemiology, pharmacology, and neuroscience and introduced

participants to those fields. Future workshops are planned, with a broad

agenda that cuts across the behavioral and social sciences, including

ethnography and behavior genetics, as well as domains of medical sciences

such as proteomics, drug development, and NIDA's clinical trials network.



Sustained investment in research education of these types will be needed at

NIJ and NIDA. Without attention to pharmacology, neuroscience, and

pharmacogenetics, it will be difficult for future investigators to develop a

fundamental understanding of the pharmacological and

economic-compulsive categories of offending in the Goldstein-Brownstein

tripartite conceptual framework. Without grounding in the social sciences,

it will be difficult for them to develop a fundamental understanding of the

systemic categories.



There now are investigators who can bridge the gaps that appear as

canyons between disciplines. Elliott's attempt to articulate his work with

the NIH human genetics initiative provides one example. In a primate lab

run by Steve Suomi at NIH, research on gene-environment interactions as

substrates of aggressive behavior, social maladaptation, and drug use

provides another example. This research is especially useful because the

environmental conditions experimentally manipulated in this lab have

conceptual linkages back to the deviancy training, inept parenting, and

parent-infant relationship models developed by Patterson, Dishion, Brook,

and others (Higley et al., 1996a, 1996b; Higley, Suomi, and Linnoila,

1996a, 1996b; Patterson, Dishion, and Yoerger, 2000; Dishion et al., 1996;

Brook et al., 1996; Brook, Tseng, and Cohen, 1996).



More examples of this type of bridgework are emerging from the work of

the research pioneers who try to keep pace with evolving contributions

from the NIH Human Genome Project. The NIJ-NIDA research agenda

can be enriched by a technical report series that brings examples of this

type to the community of investigators and research trainees.



Tension That Involves Approach or Methods



Review of the drugs-crime literature creates an opportunity for developing

new insights about the sometimes different approaches and methods that

have been developed in public health and criminal justice research work

groups. For example, ethnography with small groups has expanded to

almost large-sample ethnographic research that bears some resemblance to

large-sample survey research, but in many ways is different. To an

outsider, this expansion is a puzzle to be solved and has not yet been

grasped. In the public health research domain, the original role of an

ethnographer bore some resemblance to the role of the medical practitioner

as a student of the natural history of disease. The original natural history

studies were intensive case studies, with the doctor at the bedside of

individual sick patients making careful systematic observations about this

individual case and then that individual case, in the days when there might

have been symptom palliation (e.g., cold cloths for fever), but no effective

curative interventions to change the clinical course of disease. This has

some resonance with Agar's concepts of the ethnographer's attention to the

details of behavior and verbal expression and of writing the narrative and

taking down the stories of drug users in their own words (Agar, 1973). The

link from this role of the ethnographer to large-sample ethnography

remains unclear.



Measurement methods pioneered in behavioral sciences research and

introduced to studies of drug taking by Larson, Kaplan, and Schiffman

have started to surface in criminal justice research as well. Experience

Sampling Methods (ESM), originally developed to study the daily lives of

high school students, have now been introduced in research on drug use

(e.g., see Csikzentmihalyi and Larson, 1987, 1992). Their ESM procedure

requires study participants to wear an electronic pager device that beeps at

randomly scheduled intervals, signaling the participant to record some

predetermined aspects of his or her present feelings, activities, and/or

surrounding environmental conditions. Usually, dozens of self-reports are

collected over a week or more to capture as much of participants' daily

living as possible. One advantage of this method is the ability of the

researcher to examine drug use specific to each individual, given the

assessment of his or her baseline characteristics for comparison. An

additional benefit is the possibility of taking into account measured social

context of the behavior (e.g., see Farnworth, 2000). ESM also creates new

opportunities to investigate the determinants of drug-taking behavior that

might be unique to each individual and each situation (e.g., see Kaplan and

Lambert, 1995). 



These evolving ESM procedures require a number of conditions if

reliability and validity are to be enhanced. Kaplan and Lambert (1995)

identified the following prerequisites: having a favorable and trusting

relationship between study participants and researchers, ensuring complete

confidentiality of responses, meeting labor or equipment and programming

costs associated with beeping the participants several times per day, and

addressing difficulties faced when the participants are illiterate or

challenged by technology. 



Several recent studies of delinquent and antisocial behavior may help

clarify the utility of ESM procedures in research on the drugs-crime

relationship. For example, Farnworth (2000) studied a group of young

Australian offenders on probation and found that these respondents were

engaged in such productive activities as employment or education an

estimated 10 percent of the time. Compared with reference norms for

Australian adolescents, offenders spent 30 percent more time on passive

leisure activities. An estimated 42 percent of the time, offenders on

probation reported being bored, while 62 percent of the time they were

involved in unchallenging activities. The use of ESM to integrate studies

of drug-taking and criminal behaviors will provide new and important

evidence on relationships that generally have been studied via

retrospective reconstruction of behavior over long spans of developmental

time.



On another measurement front, there is a related tension that involves the

use of bioassay methods to study recent and past drug taking. Wish has

been a pioneer in the use of these methods for research on arrestees, and

recent studies by Harrison and Fendrich are extending this reach into

general household population samples of the type surveyed for the

National Household Survey on Drug Abuse (Wish, 1988; Yacoubian,

Wish, and Perez, 2001; Fendrich, 2001; Harrison, 2001). In future

research, one may anticipate these differences in approach to sustain a

tension until a general consensus has evolved.



With respect to approach in the domain of statistical methodology,

computational advances have contributed to an acceleration of innovation.

There is a resulting air of optimism for what might be accomplished, as in

the domains of longitudinal latent transition modeling, multilevel or nested

models, approaches to nonignorable missing data, and alternative methods

of research on directed acyclic pathways with mediation versus cyclical

pathways with reciprocities. At the same time, there is a tension because

these new statistical approaches have not become integrated in most

research training programs, and there remains certain skepticism about

heavily modeled data. 



Limitations on numeracy keep many of us from probing the assumptions

of complex models, whether these are models of behavior in individual

studies, econometric models, or operations and systems research models to

probe alternative program and policy decisions. Tension may be inevitable

in the face of such complexities.



Tension Involving Research Ethics



NIJ can play an important role in relation to investigations that probe

drugs-crime relationships. At present, a good part of the NIJ role has been

ceded to HHS and its new Office of Human Research Protections (OHRP).

True to its origins in NIH, OHRP is oriented toward the standards of

experimental medical intervention research (e.g., randomized trials to test

safety and efficacy of new drugs). OHRP specifications for informed

consent procedures and disclosure statements share this orientation.



Many behavioral and social sciences researchers have expressed concern

that the standards and specifications of experimental medical research

are not appropriate for studies of the drugs-crime relationship. For

example, in ethnographic and observational survey research, different

specifications for informed consent procedures and disclosure statements

are required. 



NIJ officials can initiate a useful dialogue with OHRP on this important

research topic. Perhaps more than any other government agency, NIJ can

help to stimulate a dialogue and negotiate a reorientation of current

practices in a manner that fosters new and creative research on the

drugs-crime relationship without a lapse in research ethics or slippage in

the protection of human subjects in this research.



Outside the Federal Government, researchers now face increasingly thorny

challenges in the protection of their research participants and the assurance

of confidentiality in relation to research data. For example, research that

includes assessments of tobacco smoking now requires special handling as

a result of legal action by the tobacco industry. These requirements

apparently extend to criminal justice research in which tobacco smoking is

approached as a self-reported indicator of deviance. The integration of

molecular biology and genetics into these research agendas, and even the

introduction of experience sampling methods or bioassays for drug testing,

raise new questions in the domain of research ethics, some of which have

been scrutinized in randomized experimental designs. These challenges

deserve the close attention of these research communities, with OHRP and

its NIJ counterpart in suitable roles.



Does Drug Use Cause Crime? A Focal Point



Each author of working papers for the drugs-crime research forum was

asked to identify a circumscribed set of research issues and probe what we

really know about them. Mindful of other sections to be written, we have

been able to organize these research issues in relation to a single focal

point, expressed in the question, "Does drug use cause crime?" One

advantage of this specific focal point is that it has a broad range and can

encompass many different strands of evidence developed in public health

and criminal justice research. Another advantage is that it is a crucial open

question for research on crime and drugs. As characterized by Harrison

and Backenheimer (1998), "Research has not been able to validate a causal

link between drug use and criminal behavior."



When confronted with an etiological research question such as "Does drug

use cause crime?," a public health scientist typically might turn to a

20th-century elaboration of the 19th century Henle-Koch postulates or

conditions for judging whether a specific disease might be caused by

specific bacteria. For a time, this 20th-century elaboration was known as

Hill's postulates (after Sir Austin Bradford Hill, a medical statistician)

and also as Evans's postulates (after A.S. Evans, an epidemiologist; Evans,

1976; Hill, 1965). Today, students of epidemiology learn them as criteria

for judging whether an association is causal or guidelines for evaluating

the evidence of a causal relationship, together with an analysis of the

relative strengths and weaknesses of evidence from randomized trials,

prospective and longitudinal studies, retrospective studies, and

case-control comparisons. Exhibit 8 presents these criteria and guidelines.



Before reviewing these criteria, four clarifications may be in order. First,

the criteria for evaluating causal significance of observed associations

represent standards of scientific evidence that are substantially different

from the standards used to judge causal evidence in civil and criminal

proceedings. For some segments of this paper's readership, the question,

"Does drug use cause crime?" may sound silly: "Of course drug use causes

crime. My grandmother could tell you that." (This was the type of reaction

TV/radio personality Rush Limbaugh gave to some of the early work that

Howard Chilcoat, Tom Dishion, and I published on the topic of whether

inner-city mothers and fathers might be able to help protect their children

against risk of early-onset drug use if they maintained levels of parental

vigilance generally associated with good parental supervision and

monitoring.)



Our response to these gentle readers is to beg forbearance. Of course, some

of what our grandparents learned to be true is not true, and the analysis of

responsibility for negligent or criminal acts in the individual case (as in a

court of law) necessarily has a different set of standards of evidence. For

example, evidence beyond a reasonable doubt is not the same as the

definitive evidence referenced in the first paragraph of this paper.



Our second clarification is that we acknowledge a possibility that

delinquent or criminal behavior might be a cause of drug use, the

chicken-egg problem referenced by Inciardi and advanced with evidence

by others. This possibility surfaces when one considers earlier sociological

models of deviance (e.g., Sutherland, Matza) or later sociopsychological

developmental models for youthful deviance, antisocial behavior, and

delinquency, such as the coercive interaction and deviancy training models

introduced by Patterson, Dishion, and colleagues; Coie and his colleagues

at Duke; and Kaplan at Texas A&M University (e.g., see Patterson, Reid,

and Dishion, 1992; Patterson, Dishion, and Yoerger, 2000; Dishion et al.,

1996; Coie and Lenox, 1994; Sandstrom and Coie, 1999; Bagwell et al.,

2000; Hubbard et al., 2001; Kaplan, 1995). For example, minor rule

violations in early childhood, well before the years of starting drug use,

might be followed by general peer rejection, differential association or

affiliation with other rejected and deviant peers, and subsequent

group-fostered delinquency and norm violations, including illegal drug

use. We note that this possibility, and the more advanced idea of

reciprocities between drug use and criminal behavior, do not necessarily

undermine inferences about drug use as a cause of crime. We face a

problem of slightly different conformation in our research on drug

dependence: The use of a drug is an absolutely necessary but not sufficient

condition for development of clinical syndromes of drug dependence, but

once the drug dependence process has started, the drug dependence takes

on a life of its own and becomes a determining influence for subsequent

drug use (i.e., drug use causes drug dependence, and then drug dependence

causes drug use).



Our third clarification is to ask first whether it is plausible that there is no

association between drug use and crime or criminal behavior or whether

there might be an inverse association (the more crime, the less drug use).

In our review of available evidence, we must acknowledge the possibility

that in some subsegments of human experience, there well may be a

negative association between drug use and criminal behavior (e.g., in the

highly disciplined and controlled environments of industrial espionage),

just as we must acknowledge the fact that some 90-year-olds have smoked

a pack or more of tobacco cigarettes virtually each day of adult life and

have not developed lung cancer. We also acknowledge the high probability

that in certain times and places or in certain subsegments of population

experience, there is no association between drug use and criminal behavior

(e.g., see Blum and Associates' studies of clinicians and professionals who

used LSD before it was regulated by the Food and Drug Administration;

Blum and Associates, 1964).



Notwithstanding these exceptional circumstances, there is a generally

consistent overall pattern of positive and sometimes quite strong

associations between illegal drug use and criminal behavior of other types.

These associations are observed not only in samples of offenders in the

criminal justice system (e.g., DUF and ADAM), but also in general

household population samples. This evidence has some vulnerability due

to constraints on methods (e.g., refusals by study participants to give

informed consent for participation), but recent consistent evidence from

general population surveys indicates that the observed association extends

beyond officially recognized crimes and does not suffer the transition bias

that is present in DUF, ADAM, and other criminal justice samples (e.g.,

perhaps the arrested or incarcerated offenders were caught because of

impairments from drug use, or the drug use of an offender is a

manifestation of a more general characteristic of carelessness that might

lead more readily to apprehension by the authorities).



Fourth, a "cloud of confusion" sometimes descends when people begin

talking about causes and causation. We will try to be clear. Although we

are asking whether drug use causes crime, we are not saying that there are

no other causes of crime. This issue sometimes is subject to

misinterpretation. For this reason, it might be more sensible to express the

question in a different way: "Under what conditions, if any, does criminal

behavior, as a response variable, depend in any substantive way on drug

use, such that we might be able to shape criminal behavior by shaping

drug use?" This question is not as pithy as, "Does drug use cause crime?"

but it might help us escape the cloud of confusion when we try to review

available evidence pertinent to this issue of causal inference.



Criterion/Guideline 1: Temporal Relationship



If illegal drug use is believed to be a cause of criminal behavior, then we

require evidence that illegal drug use has preceded the onset of that

criminal behavior. Judgments about this criterion or guideline can become

difficult when there are potential reciprocities. For example, when

sustained medicinal use of phenacetin and acetaminophen compounds

(e.g., Tylenol, Datril) was being investigated as a cause of interstitial

nephritis and end-stage renal disease (ESRD), one of the complications

was the possibility that the earliest clinical features of ESRD include

headaches. Of course, headaches can promote the sustained use of

pain-relieving medicines, including the acetaminophen compounds.



The drugs-crime relationship presents this type of temporal complexity, as

was seen in exhibits 6 and 7. Earlier aggression, conduct problems, and

criminal behavior may function as a direct cause of illegal drug use (e.g.,

see Kellam and Anthony, 1998), and possibly as an indirect cause (e.g., by

promoting affiliation with other delinquent and drug-using peers). Earlier

drug use also may function to promote later growth of conduct problems

or criminal behavior (e.g., see Johnson et al., 1995).



Criterion/Guideline 2: Biological or Other Theoretical Plausibility



Carrying books of matches is associated with the risk of developing lung

cancer, tends to precede rather than follow the onset of lung cancer, and

has at least a moderately strong association with lung cancer. However,

except with respect to the associated characteristic of tobacco smoking, we

have no biological or other theoretical plausibility to link carrying matches

per se with the etiology of lung cancer. Even if the matches-cancer

association were to withstand the challenges posed by the other criteria for

evaluating causal significance of an association, we would be inclined to

ask about the underlying theory and its plausibility and coherence in

relation to known relationships and facts.



The tripartite model for the drugs-crime nexus represents a substantiation

of plausible causal links from illegal drug use to criminal behavior. Other

related strands in the fabric of plausibility have been mentioned (e.g.,

differential crime opportunity, differential association).



The plausibility of a link between drug use and aggressive or violent

crimes rests to some extent on neuroscience theory and observed

clinico-pathological associations, as in contemporary thinking about

cocaine's influence on limbic-hypothalamic substrates of aggression

(Davis, 1996). In addition, there is a line of preclinical and clinical

laboratory experiments that has helped to solidify the plausibility of a

link from drug use to aggressive or violent behavior, and possibly to the

types of norm violations associated with nonviolent crime. The evidence

on links from the use of psychostimulant drugs (e.g., methamphetamine,

cocaine) and aggression is noteworthy in this respect. Administration of

cocaine to hamsters during adolescence increased the number of bites and

attacks indicative of a surge of offensive aggression (Harrison et al.,

2000). Moore and Thompson (1978), studying pigeons, found that high

doses of cocaine elicited aggressive behavior. In some species, increased

levels of aggression also have been observed with the administration of

amphetamine stimulant drug--not only when a large single dose (e.g.,

Melega et al., 1997), but also after sustained lower doses (Haber, Barchas,

and Barchas, 1981) are administered. These psychostimulants also may

increase the risk of self-directed aggression (e.g., see Peffer-Smith et. al.,

1983).



Experimental laboratory research with human subjects also has produced

supportive evidence along these lines, often with computerized

point-subtraction methods used to evoke aggression after the drug has

been administered and under control (no drug) conditions. For example,

Licata et al. (1993) administered a high dose, low dose, and no dose of

cocaine and found that subjects in the high-dose group expressed

significantly greater aggression than subjects in the control group; the

low-dose group did not differ from the control group.



Notwithstanding these strands of plausibility, there also is a considerable

amount of inconsistency in the observed data and some complexity in

relation to dose-response analyses. For example, Crowley et al. (1992)

found no increase in aggression when cocaine was administered in primate

lab research; Darmani and colleagues (1990) found increased aggression

among mice that were given relatively low doses of cocaine but not when

the mice were given higher doses of cocaine. Moro et al. (1997) found

reductions in the total number and length of aggressive activities in mice

after amphetamine administration. Cherek et al. (1989), studying humans,

examined the relationship between d-amphetamine on aggression using

point subtractions and found an increase in aggression among those

receiving 10 mg per 70 kg of body weight but a decrease in aggression

when 20 mg per 70 kg of body weight was administered.



Police experience on the street implicates dissociative drugs such as

phencyclidine (PCP) in relation to violent and aggressive behavior and

crime. We have been able to find some supportive experimental laboratory

evidence consistent with this streetwise experience (e.g., Burkhalter and

Balster, 1979; McCardle and Fishbein, 1989). Nevertheless, even with

PCP, there is a complex pattern of inconsistent evidence that does not

ring true with the experience on the street and common wisdom about

PCP. Tyler and Miczek (1982), Emley and Hutchinson (1983), and Miczek

and Haney (1994) reported no increase in aggressive behavior after

experimental administration of PCP and an erratic increase in aggression

only in a subgroup of animals receiving low doses. Hence, it may be

that PCP promotes aggression only in certain subgroups of the population

(e.g., see McCardle and Fishbein, 1989); and in some experiments,

animals receiving high doses of PCP were more likely to be victims of

aggression by nondrugged animals (e.g., see Russell, Greenberg, and

Segal, 1984; Tyler and Miczek, 1982).



In sum, there is some plausibility to the idea that drug use might promote

criminal behavior, with strands of plausibility coming from neuroscience

theory, the common wisdom and experience of criminal justice officials

and drug users, and laboratory experiments. The links between being a

drug user and becoming a crime victim represent an understudied

phenomenon, and the inconsistent patterns of laboratory evidence provoke

us to investigate the possibility that there might be substantial

heterogeneity within the population with respect to links from drug use to

aggressive behavior or to crime (e.g., see Parker and Rebhun, 1995).



History demonstrates one of the difficulties with this criterion for judging

causal significance of associations. Time and time again, new evidence

has contradicted what appeared to be a biologically plausible or

theoretically pleasing link between a suspected cause and a suspected

response. Today's biologically plausible or theory-driven causal inference

may be tomorrow's "old wives' tale." As is true for the other criteria and

guidelines, by itself this one counts for little.



Criterion/Guideline 3: Consistency of the Association



We already have clarified the possibility that no association or a negative

association might exist for certain subsegments of population experience.

For example, at some point, drug taking may incapacitate an individual

who otherwise would be involved in criminal behavior. Despite examples

of this type, and notwithstanding contrary evidence, the drugs-crime

research literature now includes a generally consistent replication of

positive associations between illegal drug use and criminal behavior (e.g.,

see Harrison and Backenheimer, 1998).



The body of laboratory experiments on drugs and aggressive or violent

behavior is not as consistent as one might expect. As described under

criterion/guideline 2, under some circumstances, laboratory experiments

have established a small set of drugs as causal agents in relation to

aggression and violence. However, for most drugs and many

circumstances, there are negative findings, and the evidence is not

consistent with causal links from drug taking to aggressive and violent

behavior.



Given the multiplicity of drugs, types of crimes, and varieties of social

contexts, it may be inevitable that the accumulated body of evidence on

the drugs-crime relationship appears inconsistent. Variation in the quality

of the research also has a bearing on consistencies or inconsistencies in the

evidence. As every first-year graduate student learns, research with

imprecise measurements will tend to yield evidence of no relationship

even when a relationship exists; research with measurements of limited

validity will tend to yield evidence of relationships where none exists.



Although not generally introduced as a feature of studying consistency of

relationships between causes and effects, a developmental perspective

may help to lead the reader to a greater appreciation of inconsistencies and

complexities faced in research on the drugs-crime relationship. That is, the

timing of the onset of the drug taking may condition the later expression of

criminal behavior and may lead to greater heterogeneity in the population

with respect to the drugs-crime relationship. For example, we have some

evidence on the possibility that earlier-onset drug use is associated with

later risk of developing drug problems (e.g., see Anthony and Petronis,

1995). We also have evidence that prompts us to conceptualize

earlier-onset drug use as a type of precocious adolescent development that

may disrupt normative developmental trajectories (e.g., see Newcomb,

1992; Dawes et al., 2000). There may be a tendency to interpret these

disruptions as sources of increased levels of later criminal behavior,

consistent with the idea that risk of drug problems are increased for

early-onset drug users; this has been the perspective our research group

has taken in its studies of this topic (e.g., Johnson et al., 1995; Anthony

and Petronis, 1995). Nonetheless, it is possible that precocious (i.e.,

early onset) drug taking is followed by disproportionately greater increases

in frequency of drug use and in risk of drug problems but that the

early-onset drug use dampens the level of criminal behavior that otherwise

might occur if the drug use had not started so early.



Our study of early-onset alcohol use and the later developmental trajectory

of conduct problems represents a case in point. In that study, cited above

under criterion/ guideline 1 (Johnson et al., 1995), we found that baseline

levels of conduct problems were greater for boys who had started drinking

alcohol before the adolescent years without parental permission and that

growth of conduct problems was greater for these early-onset alcohol

users--when compared with boys whose drinking did not start until later.

Similar relationships were observed for girls with early-onset alcohol

use--when compared with girls whose drinking did not start until later.

However, a discussion of this research with Blumstein has prompted us to

re-approach this problem with a different comparison in mind. Using

random effects regression, we are seeking to hold constant the baseline

level of conduct problems and study boys who have a high initial level of

conduct problems but who start drinking alcohol early on and compared

them with boys with an equally high initial level of conduct problems but

for whom alcohol consumption is delayed until adolescence. Approaching

the contrast in this manner, we may discover that early-onset drinking

dampens the growth trajectory for conduct problems; the steepest

trajectory for growth of conduct problems may be observed for boys with

high initial levels of conduct problems but without the impairments

associated with early drinking. The early drinking might lead to retardation

in the growth of conduct problems for boys who otherwise would escalate

to very high levels of conduct problems in adolescence.



This is a somewhat counterintuitive proposition, and it may run counter to

common wisdom and experience with respect to the effects of early-onset

drinking or drug use and the later lifecourse of young people. However,

our intuitions and common experience about these circumstances tend to

reflect a type of population-averaged summary of developmental

trajectories and generally do not encompass all varieties of human

experience. We mention this open research question as an example of the

complexities faced in developmental research on the drugs-crime

relationship and as a possible explanation for the inconsistencies observed

in drugs-crime research. The timing of the drug use may induce subgroup

variation in the drugs-crime relationship, which then is interpreted as

inconsistency in and a challenge to causal significance of the observed

associations.



Fortunately, complete consistency of evidence is not required. What is

required is a focused probing of the circumstances under which the

drugs-crime relationship is a causal relationship, with a deliberate effort to

ferret out situations in which there is no causal linkage between drug

use and criminal behavior. Deliberate scientific pursuit of these

circumstances and situations may require investigators to look overseas,

where use of such drugs as marijuana, cocaine, and heroin are not treated

as criminal behaviors. In social contexts of this type, by studying the

developmental trajectories of criminal behavior among young people with

and without early drug-taking experiences, we may be able to illuminate

some of the inconsistencies now observed in the drugs-crime evidence

available to us. For example, longitudinal studies of children growing up

in the Netherlands are underway. The recent effective decriminalization of

marijuana use in the Netherlands creates a social context for research on

this drug and later criminal behavior that merits attention on the NIJ-NIDA

research agenda. 



Criterion/Guideline 4: Alternative Explanations Ruled Out



This criterion or guideline represents the Achilles heel for much of the

prior research on the possible causal links between illegal drug use and

criminal behavior and represents a general difficulty for observational

research in general. Observing a possible causal relationship between

antecedent A and response B, the skeptical critic always can ask, "Isn't

there some unrecognized background factor C that can account for the A-B

relationship that you have observed in this study?" If so, "Isn't this a poorly

developed conceptual model?"



To some extent, these are a coward's questions about the drugs-crime

relationship in specific and about empirical research in general. Of course,

there might be some unrecognized background factor in empirical research

plans and in completed studies; if not the hand of the mischievous Norse

god Loki, then something else of a less celestial nature.



The challenge for the courageous skeptical critic is to assert a specific

background factor or set of background factors that might account for the

observed A-B relationship and that have not been considered explicitly or

taken into account in a study plan or description of completed work. For

example, observing the suspected causal association between tobacco

smoking and risk of lung cancer, the statistician Sir Ronald Fisher posed a

question of the following type: "Can't we explain the observed association

as a manifestation of an underlying predisposition or liability that

determines both the tobacco smoking and the lung cancer?"



In relation to the drugs-crime relationship, the most plausible background

factors seem to be of the variety named by Fisher, namely, unmeasured

predispositions; in this instance, the predispositions might involve who

abides by the conventional rules of society, who is willing to run afoul of

the law by taking a drug illegally, and who is willing to commit crimes

other than the crime of drug possession for personal use. To some extent,

these predispositions may be a manifestation of family heritage, a

manifestation of early experiential conditions and processes, or a synthesis

of both. Nevertheless, no matter what their origin, until these

predispositions are taken into account, they represent a plausible form of

alternative explanation whenever a drugs-crime relationship is found in

our empirical studies.



One line of response to this criticism has been to measure personality or

facets of temperament in observational studies and to re-estimate the

drugs-crime association with personality or temperament held constant

(e.g., via stratification or statistical adjustment in a regression model). But

this response always is subject to the criticism that the wrong facets of

personality or temperament were measured or that the measurement of

personality or temperament was not as good as it should have been.



It is in relation to this criterion that we now have new opportunities for

research at the intersection of public health and criminal justice research

on the drugs-crime relationship. Three important opportunities at this

intersection involve (1) genetics, twin, and family research; (2)

longitudinal studies with "subjects as their own controls" designs; and (3)

controlled experimental trials.



Future genetics, twin, and family studies can help to narrow the alternative

explanations in a useful manner. For example, in an earlier section we

described a design that exploits the genetic matching of monozygotic

twins to search for environmental conditions that contribute to the

occurrence of disease. Discordant MZ twin designs also can be used to

hold constant predispositions or liabilities linked to the individual genome

of the twins, while studying differences in the trajectory of criminal

behavior for the MZ twin whose illegal drug use starts first versus the MZ

twin whose illegal drug use starts later (or not at all).



Alternative twin and family research designs can be used to narrow other

explanations of the observed drugs-crime relationship (e.g., studies of

discordant siblings, studies based on the transmission disequilibrium test

when specific polymorphisms are under investigations). Cadoret and

colleagues have offered recent illustrations of the power of twin studies in

which some twins have been separated at birth, but these "natural

experiments" have become scarce in the United States and other parts of

the world where twins now generally are kept together in their new

adoptive families (e.g., see Cadoret et al., 1986, 1995; Cadoret, Leve, and

Devor, 1997). Tsuang et al. (2001) provide a recent useful overview of

pertinent findings from the Harvard Twin Study.



Longitudinal subjects-as-their-own-controls research designs also can help

rule out alternative explanations in the sense that each individual

participant is carrying forward a within-individual set of propensities to

become engaged in illegal drug use and other criminal offending. In these

longitudinal designs, in an otherwise law-abiding individual, if we were to

observe that criminal offending occurs only in the immediate aftermath of

a drug intoxication experience or only in the stages of withdrawal after

drug dependence, we would have additional evidence of a drugs-crime

association at the individual level. These longitudinal designs remain

vulnerable to a possible counterclaim that there is an underlying

predisposition that links earlier illegal drug use to later criminal offending

only during the context of drug intoxication or withdrawal states. That is,

the observed association between illegal drug intoxication or withdrawal

and the later criminal offending is a spurious artifact of uncontrolled

confounding: There is something else in the background, a third variable

that explains the observed sequence.



Medical and public health research is host to a variety of examples of this

type of spurious confounding. One of them involves the connection

between chickenpox and shingles. For most people, chickenpox occurs

early in life and shingles occurs late in life. There sometimes is an

exceptional case of shingles occurring with no prior history of chickenpox

in childhood, but these exceptional cases might be understood as instances

of "clinically inapparent" infections (i.e., with mild or minimal symptoms

in childhood, so mild as to pass without notice). A longitudinal research

design on this topic can lead to the impression that chickenpox causes

shingles, in the sense that shingles rarely or never occurs unless

chickenpox occurs first. This observed longitudinal link between

chickenpox and shingles satisfies the requirement described above:

Criminal behavior occurs only after a bout of illegal drug use. The fly in

the ointment in our chickenpox-shingles example is that we now know

that chickenpox does not cause shingles. Rather, it is an underlying virus

that causes both of these clinical phenomena. Exposure to the chickenpox

virus (herpes zoster) is the cause of chickenpox in childhood and is the

cause of shingles in later life when the virus emerges from an otherwise

dormant or latent state of no activity. The apparent linkage from earlier

chickenpox to later shingles is due to an underlying third variable, the

herpes zoster infection, which accounts for the appearance of both

outcomes.



The analogy to research on illegal drug use and later criminal offending

should be clear. Even when longitudinal research shows us examples of

participants who become engaged in criminal behavior only in the context

of drug intoxication or withdrawal states, we cannot be confident that

the illegal drug use is the cause of the associated criminal offending. Some

unknown underlying cause may be accounting for both outcomes.



The third approach, involving randomized trials, offers a way to bring

these unknown underlying variables into check. This approach already has

been described in relation to our research group's studies of an alternative

explanation for the drugs-crime relationship. Namely, we advanced the

hypothesis that a predisposing characteristic in the form of early

aggression or rule breaking is a potentially modifiable determinant of both

later illegal drug use and criminal behavior or other sorts. This hypothesis

does not reject the possibility that illegal drug use causes later criminal

behavior, but it introduces one alternative explanation for the observations

association between illegal drug use and criminal behavior (i.e., the earlier

aggression or tendency to break rules and social norms). As described

earlier, we sought to test this hypothesis by constructing an experimental

trial in which we disrupted the development of early aggression and rule

breaking (e.g., Kellam and Anthony, 1998). We used the power of

randomization to hold constant the profile of alternative explanations that

might account for later illegal drug use and criminal behavior. In a current

followup study of the youths who participating in this trial, we will be

testing whether the primary school intervention had a sustained impact on

illegal drug use and criminal behavior. If so, we might expect a weakened

association between illegal drug use and criminal behavior in the subgroup

of youths exposed to the active behavioral intervention arms of our study.



A related opportunity to test the drugs-crime relationship and to use

randomization to rule out alternative explanations involves controlled

trials of new therapeutic interventions directed toward illegal drug use and

drug dependence of adolescents. Observational studies now suggest

that entry into drug treatment reduces the rate of criminal offending, but

these studies leave open possibilities for alternative explanations (e.g.,

selection biases in the assignment of subjects to treatment, imbalances in

the other determinants of criminal offending). Randomization in the

setting of controlled trials of new therapeutic interventions creates an

opportunity to constrain these selection biases and bring into balance the

alternative sources of variation in criminal offending (e.g., see Manski et

al., 2001). 



By adding followup measurements of posttreatment criminal behavior to

current and newly emerging randomized controlled trials of therapeutic

interventions, NIJ and NIDA can help foster new evidence on the degree

to which illegal drug use is a cause of criminal offending. Alternative

explanations for the observed drugs-crime association and other

determinants of the offending behavior can either be brought into balance

by randomization or held constant as measured covariates in statistical

models of analysis. Some examples of past research along these lines are

described under criterion/guideline 7.



Criterion/Guideline 5: Dose-Response or Gradient Relationship



Absence of a dose-response or gradient relationship does not rule out

causal associations; there are good examples of threshold relationships

with no clear gradient. Nonetheless, there are examples in which the

probability or rate of criminal behavior is observed to be lower with lower

levels or frequencies of illegal drug use and is observed to be greater as

levels or frequencies increase.



In one recent and especially informative longitudinal cohort study, Brook

et al. (2001) studied the developmental trajectory of marijuana use from

childhood into adulthood and found that behavioral and attitudinal

indicators of unconventionality (e.g., attitudes tolerant of norm violations)

had a gradient relationship with later increases in marijuana involvement.

The research team also found that as levels of unconventionality increased,

so did marijuana involvement. These gradient relationships between

unconventionality and marijuana use help to substantiate a possible causal

link between earlier unconventionality and later developmental trajectories

of marijuana involvement. However, as in the circumstance of research on

the drugs-crime relationship, this research report leaves us with

unanswered questions of the following variety:



a. What about the predisposition that links unconventionality to the earliest

marijuana use? Where does the unconventionality come from, and is this

predisposition the same as the predisposition to smoke marijuana?



b. What about the reverse causal pathway and the possibility that

increasing marijuana use might promote later increases in

unconventionality?



c. As levels of marijuana use increase, are there later dose-response or

gradient-like increases in unconventionality?



In light of the population heterogeneity mentioned above, this

dose-response criterion might be especially troublesome in research on the

drugs-crime relationship. For example, consider the drug user whose

increasing bouts of intoxication yield less criminal behavior than

otherwise might occur and whose intoxication-associated carelessness

leads to apprehension and detoxification and outpatient treatment prior to

a bench appearance. The detoxification and treatment might be followed

by a return to the baseline level of criminal behavior (i.e., a higher level of

criminal behavior than was observed during the period of intoxication) and

an impression that treatment was ineffectual with respect to the frequency

of criminal behavior.



Criterion/Guideline 6: Strength of Association



Weak associations seem especially vulnerable to sources of spuriousness

and bias. One benchmark standard for strength is the association between

tobacco smoking and lung cancer: The risk of dying from lung cancer is

estimated to be 10 times or greater for persistent tobacco smokers than for

nonsmokers. Toward the other end of the spectrum of magnitude is a

widely appreciated but quite modest strength of association between being

male and illegal use of drugs: The risk of becoming an illegal drug user is

an estimated 1.5 to 3.0 times greater for an American male than for an

American female (Anthony and Helzer, 1995). Examining the range of

study estimates on the drugs-crime relationship, there are some studies

with extremely large relationships, but when a positive relationship is

observed, the strength of relationship tends to be quite modest. This

generally modest relationship may imply that alternative explanations

(e.g., predispositions) are sufficient to account for the observed

relationship.



Criterion/Guideline 7: Cessation Effects



Cessation effects already have been mentioned in the context of our

discussion of alternative explanations under criterion/ guideline 4. There

are many studies of co-occurring maturation processes that lead to fading

of both illegal drug use and other criminal behavior, especially since the

work of Winick. The observational studies of McGlothlin, Anglin, and

Hser in California and the work of Nurco, Lerner, and colleagues in

Baltimore also shed light on declines in criminal behavior during periods

of abstinence or reduced illegal drug use. The literature includes numerous

studies of what has happened to crime involvement after cessation of drug

use, based on observational studies.



As noted under criterion/guideline 4, some of the strongest evidence about

cessation effects can come from randomized experiments in which drug

treatment or other interventions are used to disrupt illegal drug use, with

subsequent evaluation of crime as an outcome of treatment. As noted

under criterion/guideline 6, for some segments of the drug-using

population, the cessation of drug use is followed by increases in frequency

of criminal behavior (i.e., once impairments associated with intoxication

are reduced).



Reprise: Does Drug Use Cause Crime? What Do We Not Know?



This review of a specific hypothesized causal relationship was intended to

highlight some of what we know about the drugs-crime relationship.

Its main purpose was to provoke discussion and help in a process of

identifying weaknesses and gaps in evidence that might be used to guide a

future research agenda. 



Evaluated in relation to these conventional criteria or guidelines for

judging the causal significance of observed associations, the reader

may have a better appreciation for the uncertainty conveyed in a recent

summary statement cited above: "Research has not been able to validate a

causal link between drug use and criminal behavior" (Harrison and

Backenheimer, 1998). The available evidence is ambiguous with respect to

temporal relationships. 



Instead, we offer a series of discussion points about what we might not yet

know.



Is the evidence on a temporal relationship compelling? Illegal drug use

precedes formal criminal behavior in some of these studies, but what about

the earlier antecedents of both drug use and crime in the form of rule

breaking, misbehavior, and minor norm violations? One can imagine a

co-occurrence process that begins with expression of irritable temperament

or aggression in the preschool years, followed by rule breaking or norm

violations in the primary school years, and then later co-occurrence of

illegal drug use and delinquent or criminal offending. Our own research

group and others have added some evidence on the possibility that drug

taking that starts by age 11 might promote growth trajectories for later

conduct problems among both boys and girls. The pattern of co-occurrence

of conduct problems and drug use is a centerpiece of Jessor's problem

behavior theory, and there is reason to look to experiments that will help

us differentiate these problem behaviors (e.g., differential response of

each form of problem behavior to different interventions, as suggested in

Dishion's early Adolescent Transitions experiment).



Plausibility? Our focus has been oriented toward the individual, but there

is a perspective on the drugs-crime relationship that is more ecological

or contextual in orientation. For example, a social environment

characterized by illegal drug use of individuals might give rise to norm

violations and criminal offending of other sorts, and not necessarily in the

form of offending by the drug users but rather in the form of offending

by others. The mugging of a heavily intoxicated drug user by a group of

nonusing passersby serves as one example of aggregate effects of illegal

drug use on crime that would not be apparent in individual-oriented

studies but would require multilevel studies of interrelationships between

individuals.



Consistency? What about the exceptions to a general pattern of

observations? It seems likely that the drugs-crime association varies from

time to time, place to place, and subgroup to subgroup. The study of

variation in these patterns of association will help to disclose the boundary

conditions and mechanisms that give rise to strong, weak, and possibly

inverse associations. Research across borders and in settings such as the

Amsterdam cannabis environment can help illuminate these boundary

conditions.



Alternative explanations? Several lines of research have been started on

the common causes for both illegal drug use and other criminal behavior,

some of them originating in family history studies and the clever adoption

paradigm adapted by Cadoret and his colleagues, some with a sharper

focus on mechanisms of inheritance (e.g., assays of genetic

polymorphisms), and some with a focus on personality and early social

environment.



It is not clear that studies to date have provided adequate control over

these sources of covariation. Nonetheless, the longitudinal study of

individuals over time has provided evidence from

subjects-as-their-own-controls designs, and the randomized trials of

interventions provide some evidence that, despite common causes, an

intervention directed toward illegal drug use can reduce frequency of

criminal behavior. Even if there are common causes (e.g., inherited traits),

for many observers, the longitudinal evidence coupled with experimental

evidence is sufficient to draw the inference that illegal drug use causes

criminal behavior. Reasonable people will disagree about this inference

from available observations, and the points of disagreement will lead us to

specific experiments or new studies to gather evidence that will be more

compelling.



Gradient? Is it possible that some of the inconsistency in observations

about the drugs-crime relationship can be traced to (a) selective attention

either to the lower end of drug involvement (e.g., among children,

adolescents, or high school seniors followed through the college years; see

Schulenberg et al., 1994) or to the higher end (e.g., among arrestees or

clients in drug treatment programs); (b) possible thresholds in the gradient

relationship, with between-sample heterogeneity with respect to the

effective threshold; or (c) an uncertain metric for assessing the type or

level of drug involvement? As described in the prior section entitled

"Criterion/guideline 2: Biological or other theoretical plausibility," we

have noted some inconsistent pharmacological effects across dosage levels

of the same drug and across different drugs. If we are to appropriately sort

the consistent and inconsistent findings of field studies on the drugs-crime

relationships, it may be necessary to reach for greater specificity with

respect to dosage levels or intensity of drug use and also with respect to

the pharmacological differences observed in laboratory experiments. It

no longer is enough to sort drugs into the non-scientific colloquial "soft"

and "hard" categories, nor to lump all "illegal drug use" as if there were no

heterogeneities of effect across the various forms of internationally

regulated drugs. The best field studies of the 21st century will abandon

these relatively crude classifications and will not carry forward an obsolete

tradition from the earlier ground-breaking days of drugs-crime research.



Strength of association? Due to uncertainties about reciprocal and dynamic

interrelationships between drugs and crime, it would be advantageous

to look closely at studies with fine-grained temporal analysis of the

drugs-crime relationship and to estimate strength of association

prospectively. This should be done in a manner that allows change in the

level of criminal behavior to be gauged in relation to change in the level of

drug use and vice versa, or with an expression of the relative risk of

criminal acts with and without antecedent illegal drug use. 



Cessation? Our recent National Research Council committee expressed

concern that selection effects, transition biases, or other artifacts might

lead to a spurious inference that criminal behavior declines or stops when

illicit drug use is ended, either with or without intervention. The evidence

on this criterion might require special scrutiny in light of concerns such as

these.



Conclusion



In the final section of this working paper, I would like to integrate the

organizing conceptual framework presented in the section on the rubrics

with the ecological concept of scale described previously. Here, there is an

adaptation of the formal ecological concept of scale that includes the

microcosm and an extension of the concept that reaches to the macrocosm

of the international regulatory environment. 



The integration of the five rubrics and the concept of scale is depicted in

exhibit 9. The result is a two-dimensional grid with the rubrics on one axis

and scale on the other axis and showing the conceptual domain where

research on drugs and research on criminal offending intersect. Each

rubric-scale intersection or subunit in the grid can be populated by past

and current examples of research on the drugs-crime relationship. In some

subunits, density of past and current research is quite high; work in these

domains may require strengthening, or perhaps these investigators should

be left alone to do their work. In other subunits, we have done little or no

past research activity; these subunits might warrant attention in a new

agenda for drugs-crime research.



Starting in the upper left-hand corner of this framework, we have the

intersection of quantity research with the microcosm represented by the

genes we inherit from our forebears. We may expect one day to have an

investigation that produces quantitative estimates of the frequency of

homozygotes and heterozygotes with respect to genes that are implicated

in the drugs-crime relationship, just as we now have these estimates for the

frequency of alleles mapped to apolipoprotein E4 and other genes or

polymorphisms implicated in the risk of developing Alzheimer's disease.



Working our way to the far upper right-hand corner, we stop at the level of

Nations. To the best of my knowledge, we have a limited set of

quantitative estimates for rates of drug-taking behaviors and criminal

justice statistics at the level of Nations; but definitive evidence on

variation across regions of the globe is lacking and represents a current gap

in knowledge. To some extent, this gap can be filled by cross-national

studies now underway, such as the World Health Organization's (WHO's)

recent World Mental Health 2000 research initiative being led by Ron

Kessler at Harvard and T. Bedirhan Ustun at WHO, with collaborators in

more than 20 countries around the globe (Kessler, 1999); the European

School Survey Project on Alcohol and Drugs (www.ipdt.pt/

investigacao/espad99/indice.htm); and cross-national studies supported by

NIDA in Latin America (e.g., Brook et al., 2001), including our own

PACARDO Project (Anthony, 2000).



In the middle range, between the microcosm of the gene and the

macrocosm of global regions, we have collections of estimates for various

social and geopolitical groups. In aggregate, these estimates can help us to

draw generalizations about the relative magnitude of problems associated

with drug taking of one sort or another (e.g., marijuana use versus cocaine

use) or with criminal behavior of one sort of another (e.g., aggravated

assault versus shoplifting or vandalism).



Even within the rubric of quantity, there are many gaps. For example, our

quantitative estimates often are based strictly on officially recognized

offending and do not encompass unrecognized offending. With respect to

drug taking, there is a plethora of evidence on the prevalence of drug use

and drug dependence but not much evidence on the incidence or risk of

becoming drug dependent. Here, also, we have big gaps in the evidence

that warrant some attention as we design an agenda for future research.



The intersections of the location, causes, and mechanisms rubrics with the

scale dimension brings us closer to evidence on variation from place to

place, time to time, or in relation to personal characteristics. A few

investigators have started to integrate genetic variation in their studies on

such topics as drugs and crime, and soon we may have more definitive

evidence on the relative frequency of different polymorphisms or

gene-encoded protein products for different subgroups of the population or

in different geopolitical zones. We can expect ecological analyses of the

between-subgroup and between-zone rates, with new evidence on location. 



Similarly, working outward from the simplest gene products to more

complex products of gene-environment interaction, the sex hormones

research of separate research groups led by Logan and by Angold,

Costello, and others should provide us with more evidence on rates of

antisocial behavior, drug use, and offending in relation to levels of

testosterone and other hormones before and after drug use (Logan, 2001;

Federman et al., 1997). The initial evidence is not expected to allow causal

inference, but the understanding of locational variation will allow us to

sharpen our causal theories and to integrate new biological perspectives on

the drugs-crime relationship.



In a middle position in this framework, somewhat overlapping the

different segments, the important line of research being conducted by

Higley, Suomi, and their colleagues in relation to gene-environment

interactions merits attention. This research, already mentioned in one of

the preceding sections, touches on aggressive behavior, social conditions

of child rearing, and drug use. Using a primate model, this research group

has been able to extend the line of research on infant-mother relations that

Harlow initiated. The group is engaged in experimental manipulation of

the early conditions of infant rearing, crossed with genetic predispositions

that in the wild have been found to be related to aggressive behavior and

excess mortality. The evidence from this research serves as an important

example of how the effects of an apparently noxious inherited

predisposition might be modulated by a change in child-rearing

environments. Does this animal model of gene-environment interaction

also hold for aggressive children, with later implications for their

drug-taking behavior? Questions such as this one merit discussion in

relation to the proposed drugs-crime research agenda, if only to choose not

to pursue these lines of research.



Turning to the last row of the framework, I offer some speculations about

gaps in research on prevention and control. At the level of scale that

reaches from microcosm to the whole organism, I see a gap in research on

underlying brain structures that subserve neuropsychological functioning

of clear importance in the choice behavior of drug users and offenders. To

the extent that drug users and offenders are making choices about various

elements in their behavioral repertoires, we may be able to understand

variations in response to prevention and control interventions as a function

of neuropsychological performance (e.g., with respect to direction, control,

and planning). Our comprehension of this variation can increase through a

program of research on fMRI brain imaging and neuropsychological

testing under experimentally controlled paradigms (e.g., aggression evoked

through computerized point subtraction or other procedures). In time, we

should be able to evaluate the degree to which response to these

interventions depends on brain structure and function as manifest in

neuropsychological tests as well as in response to genetic predispositions

of the type now being characterized in Suomi's primate laboratory and

elsewhere.



Working our way to the bottom right-hand side of this matrix, we find the

intersections with social groups and contexts of increasingly larger scale,

not only the peer group and family of origin or procreation, but also the

larger neighborhood, employment context, the community at large, and

across national boundaries. As we plot examples of intervention research

in this two-dimensional framework, it is easier to find examples of

individual investigations with narrow breadth of scale. For example, we

can find an intervention focused on the community but without elements

of intervention directed toward specific individuals in the community. We

can find many interventions directed toward individual arrestees but not

toward the social groups of which the arrestees are members.



One of the challenges for those who seek to shape the future research

agenda on drugs and crime will be to encourage broadband research that

cuts across multiple levels of scale. This is not to say that we should

eliminate narrow-band research because it often is necessary to solve a

research problem through focus, and focus is one of the defining

characteristics of narrow-band research. Nonetheless, as we look over

some of the more exciting research projects now underway, we can see

that the excitement is coming from the investigators' attempts to

encompass more than one level of scale. These attempts deserve

encouragement.



Before closing, we must turn to the empty spaces created in the circle but

not included as part of the two-dimensional grid. Within these spaces, we

have important drugs-crime research that does not fall neatly into the

two-dimensional conceptual framework. I am thinking of the recent

ethnographic studies of the gangs involved in drug sales (e.g., Levitt and

Venkatesh, 1998), and some of the other recent innovative qualitative

research on drug trafficking (e.g., Natarajan, Clarke, and Belanger, 1996;

Natarajan and Belanger, 1998), which shed new light on the structure and

organization of the criminal organizations that sustain drug supply and

influence drug-related criminal offending around the world. There also are

good recent examples of operations research focused on the organization

and administration of criminal justice agencies and the deployment of law

enforcement, prosecution, and judicial resources (e.g., see Maltz, 1996).

To the extent that these investigations guide us toward useful evidence

about prevention and control, and to the extent that they focus on

individuals or small groups of individuals (e.g., in a city or State), they

may be placed in the space on the left-hand side of the figure, between the

grid and the surrounding circle. To the extent that these control-oriented

investigations are directed to-ward international drug trafficking (e.g., see

Montagne, 1990), they may be placed in the space on the right-hand side

of the figure.



There are other research programs and initiatives that do not fall neatly

within the two-dimensional grid presented in exhibit 9. Methodological

research constitutes one set of examples (e.g., Wish et al., 2000; Harrison,

1997, 2001; Fendrich, 2001). Proposed new research on drug prices and a

consumer product index for illegal drugs represents another set (e.g., see

Manski et al., 2001).



As we move toward a new drugs-crime research agenda for NIJ and NIDA

at the intersection of public health and criminal justice studies, it is

important to remember the two major themes mentioned in the

introduction to this paper:



o There is no single drugs-crime relationship. Rather, there are

drugs-crime relationships, most of which are complex rather than simple.



o There is no simple solution to the complex challenges faced when

drugs-crime relationships come into play.



The two-dimensional grid encircled in exhibit 9 offers no simple solutions

to the complex challenges faced when drugs-crime relationships come into

play. That grid is only a tool that may help us identify important gaps in

the research evidence, gaps that must be filled as we work toward a more

complete understanding of the drugs-crime relationship and more effective

action plans that apply new understanding in the service of public health

and safety. In an important sense, the empty spaces encircling the

two-dimensional grid also can be useful tools as we try to identify and fill

the gaps in evidence. These empty spaces can serve to remind us that no

conceptual framework is all encompassing. We must "think outside the

box" in this regard. If we organize our scientific resources simply to

continue our current lines of research, we will not achieve lasting

reductions in illegal drug use and drug-related crimes, and we will never

lay claim to great victories in the service of public health and safety. 



References



Agar, M. (1973). Ripping and running. New York: Seminar Press.



Anthony, J.A., and Bradshaw, G.A. (2001). Wavelet analysis as an

approach to investigate the reciprocal relationship between ecological

pattern and process. The premises and problems with spatial analysis in

landscape ecology. Paper presented at Pattern, Process, Scale, &

Hierarchy: Interactions in Human-Dominated and Natural Landscapes, the

16th Annual Symposium of the U.S. Chapter of International Association

of Landscape Ecology, Memorial Union, Arizona State University, Tempe,

Arizona, April 25-29.



Anthony, J.C. (2000). Cross-national studies on clusters of drug use. NIH

grant award 5 R01 DA10502-02. 



Anthony, J.C., and Helzer, J.E. (1995). Epidemiology of drug dependence.

In M.T. Tsuang, M. Tohen, and G.E.P. Zahner (eds.), Psychiatric

epidemiology (pp. 361-406). New York: Wiley & Sons.



Anthony, J.C., Neumark, Y., and Van Etten, M.L. (2000). Do I do what I

say? A perspective on self-report methods in drug dependence

epidemiology. In A.A. Stone, J.S. Turkkan, C.A. Bachrach, J.B. Jobe, H.S.

Kurtzman, and V.S. Cain (eds.), The science of self-report: Implications

for research and practice. Mahwah, NJ: Lawrence C. Erlbaum Publishers. 



Anthony, J.C., and Petronis, K.R. (1995). Early-onset drug use and risk of

later drug problems. Drug and Alcohol Dependence, 40(1), 9-5. 



Anthony, J.C., and Van Etten, M.L. (1998). Epidemiology and its rubrics.

In A. Bellack and M. Hersen (eds.), Comprehensive clinical psychology

(vol. 1, pp. 355-390). Oxford, UK: Elsevier. 



Bagwell, C.L., Coie, J.D., Terry, R.A., and Lochman, J.E. (2000). Peer

clique participation and social status in preadolescence. Merrill-Palmer

Quarterly, 46(2), 280-305. 



Behrens, D.A., Caulkins, J.P., Tragler, G., and Feichtinger, G. (2000).

Optimal control of drug epidemics: Prevent and treat--But not at the same

time. Management Science, 46(3), 333-347. 



Behrens, D.A., Caulkins, J.P., Tragler, G., Haunschmied, J.L., and

Feichtinger, G. (1999). A dynamic model of drug initiation: Implications

for treatment and drug control. Mathematical Biosciences, 159(1), 1-20. 



Blum, R., and Associates (1964). Utopiates: The use and users of LSD 25.

New York: Atherton Press. 



Blumstein, A., Cohen J., Roth, J.A., and Visher, C.A. (eds.), and the

National Research Council (1986). Criminal careers and career criminals.

Washington, DC: National Academy Press.



Bobashev, G., and Anthony, J.C. (1998). Clusters of marijuana use in the

United States. American Journal of Epidemiology, 148(12), 1168-1174. 



Bobashev, G., and Anthony, J.C. (2000). Use of alternating logistic

regression in studies of drug use clustering. Substance Use & Misuse,

35(6-8), 1051-1073. 



Brennan, P.A., Mednick, S.A., and Jacobsen, B. (1996). Assessing the role

of genetics in crime using adoption cohorts. CIBA Foundation

Symposium, 194, 115-123. 



Bronfenbrenner, U. (1979). The ecology of human development:

Experiments by nature and design. Cambridge, MA: Harvard University

Press. 



Bronfenbrenner, U. (1986). Ecology of the family as a context for human

development: Research perspectives. Developmental Psychology, 22, 723-

742. 



Brook, J.S., Brook, D.W., Arencibia-Mireles, O., Richter, L., and

Whiteman, M. (2001). Risk factors for adolescent marijuana use across

cultures and across time. Journal of Genetic Psychology, 162(3), 357-374. 



Brook, J.S., Tseng, L.J., and Cohen, P. (1996). Toddler adjustment: Impact

of parents' drug use, personality, and parent-child relations. Journal of

Genetic Psychology, 157(3), 281-295. 



Brook, J.S., Whiteman. M., Finch, S.J., and Cohen, P. (1996). Young adult

drug use and delinquency: Childhood antecedents and adolescent

mediators. Journal of the American Academy of Child Adolescent

Psychiatry, 35(12), 1584-1592. 



Brook, J.S., Whiteman, M., Finch, S.J., Morojele, N.K., and Cohen, P.

(2000). Individual latent growth curves in the development of marijuana

use from childhood to young adulthood. Journal of Behavioral Medicine,

23(5), 451-464. 



Brown, J.H. (1995). Macroecology. Chicago: University of Chicago Press. 



Brownstein, H.H., and Goldstein, P.J. (1990). Research and the

development of public policy: the case of drugs and violent crime. Journal

of Applied Sociology, 7, 77-92. 



Burkhalter, J.E., and Balster, R. (1979). Effects of phencyclidine on

isolation-induced aggression in mice. Psychological Reports, 45(2),

571-576. 



Cadoret, R.J., Leve, L.D., and Devor, E. (1997). Genetics of aggressive

and violent behavior. Psychiatric Clinics of North America, 20(2),

301-322. 



Cadoret, R.J., Troughton, E., O'Gorman, T.W., and Heywood, E. (1986).

An adoption study of genetic and environmental factors in drug abuse.

Archives of General Psychiatry, 43(12), 1131-1136.



Cadoret, R.J., Yates, W.R., Troughton, E., Woodworth, G., and Stewart,

M.A. (1995). Adoption study demonstrating two genetic pathways to drug

abuse. Archives of General Psychiatry, 52(1), 42-52. 



Caulkins, J., Crawford, G., and Reuter, P. (1993). Simulation of adaptive

response: A model of drug interdiction. Mathematical and Computer

Modeling, 17, 37-52. 



Chein, I., Gerard, D.L., Lee, R.S., and Rosenfeld, E. (1964). The road to

H: Narcotics, delinquency, and social policy. New York: Basic Books. 



Cherek, D.R., Steinberg, J.L., Kelly, T.H., and Sebastian, C.S. (1989).

Effects of d-amphetamine on human aggressive responding maintained by

avoidance of provocation. Pharmacology Biochemistry and Behavior,

34(1), 65-71. 



Chilcoat, H.D. (1992). Parent monitoring and initiation of drug use in

elementary school children. Doctoral thesis. The Johns Hopkins

University, Baltimore, MD. 



Cohen, A.K. (1955). Delinquent boys. Glencoe, IL: Free Press. 



Cohen, M.A. (1998). The monetary value of saving a high-risk youth.

Journal of Quantitative Criminology, May, 19-33. 



Coie, J.D., and Lenox, K.F. (1994). The development of antisocial

individuals. Progress in Experimental Personality and Psychopathology

Research, 45-72. 



Crowley, T.J., Mikulich, S.K., Williams, E.A., Zerbe, G.O., and Ingersoll,

N.C. (1992). Cocaine, behavior, and alcohol-solution drinking in monkeys.

Drug and Alcohol Dependence, 29(3), 205-223. 



Csikzentmihalyi, M., and Larson, R. (1987). Validity and reliability of the

experience sampling method. Journal of Nervous and Mental Diseases,

175, 526-536. 



Csikzentmihalyi, M., and Larson, R. (1992). Validity and reliability of the

experience sampling method. In M.W. De Vries (ed.), The experience of

psychopathology: Investigating mental disorders in their natural settings.

New York: Cambridge University Press. 



Darmani, N.A., Hadfield, M.G., Carter, W.H., Jr., and Martin, B.R.

(1990). Acute and chronic effects of cocaine on isolation-induced

aggression in mice. Psychopharmacology (Berlin), 102(1), 37-40. 



Davis, W.M. (1996). Psychopharmacologic violence associated with

cocaine abuse: Kindling of a limbic dyscontrol syndrome? Progress in

Neuropsychopharmacology and Biological Psychiatry. 20(8), 1273-1300.



Dawes, M.A., Antelman, S.M., Vanyukov, M.M., Giancola, P., Tarter,

R.E., Susman, E.J., Mezzich, A., and Clark, D.B. (2000). Developmental

sources of variation in liability to adolescent substance use disorders. Drug

and Alcohol Dependence, 61(1), 3-14. 



Dishion, T.J., and Loeber, R. (1985). Adolescent marijuana and alcohol

use: The role of parents and peers revisited. American Journal of Drug and

Alcohol Abuse, 11, 11-25. 



Dishion, T.J., Patterson, G.R., and Reid, J. (1988). Parent and peer factors

associated with drug sampling in early adolescence: Implications for

treatment. In E.R. Rahdert and J. Grabowski (eds.), Adolescent drug

abuse: Analyses of treatment research (NIDA Research Monograph 77,

pp. 69-93). Washington, DC: U.S. Department of Health and Human

Services, National Institute on Drug Abuse. 



Dishion, T.J., Spracklen, K.M., Andrews, D.W., and Patterson, G.R.

(1996). Deviancy training in male adolescent friendships. Behavior

Therapy, 27, 373-390. 



Duncan, T.E., Duncan, S.C., and Hops, H. (1998). Latent variable

modeling of longitudinal and multilevel alcohol use data. Journal of

Studies on Alcohol, 59(4), 399-408. 



Duncan, T.E., Duncan, S.C., Hops, H., and Alpert, A. (1997). Multi-level

covariance structure analysis of intra-familial substance use. Drug and

Alcohol Dependence, 46(3), 167-180. 



Dunlap, E., and Johnson, B.D. (1996). Family and human resources in the

development of a female crack-seller career: Case study of a hidden

population. Journal of Drug Issues, 26(1), 175-198. 



Dunlap, E., Johnson, B.D., and Manwar, A. (1994). A successful female

crack dealer: Case study of a deviant career. Deviant Behavior, 15(1),

1-25. 



Elliott, D.S. (2001). NYS family study: Problem alcohol use & problem

behavior. NIH grant award 1 R01 AA11949-01A2. 



Elliott, D.S., and Huizinga, D. (1989). The relationship between

delinquent behavior and ADM problems. In C. Hampton (ed.), Juvenile

offenders with serious drug, alcohol and mental health problems.

Washington, DC: U.S. Government Printing Office. 



Emley, G.S., and Hutchinson, R.R. (1983). Effects of phencyclidine on

aggressive behavior in squirrel monkeys. Pharmacology Biochemistry and

Behavior, 18(2), 163-166. 



Ennett, S.T., Flewelling, R.L., Lindrooth, R.C., and Norton, E.C. (1997).

School and neighborhood characteristics associated with school rates of

alcohol, cigarette, and marijuana use. Journal of Health and Social

Behavior, 38(1), 55-71. 



Evans, A.S. (1976). Causation and disease: The Henle-Koch postulates

revisited. Yale Journal of Biology and Medicine, 49, 175-195. 



Faris, D.E.L., and Dunham, H.W. (1939). Mental disorder in urban areas.

Chicago: Hafner/Chicago University Press. 



Farnworth, L. (2000). Time use and leisure occupations of young

offenders. American Journal of Occupational Therapy, 54(3), 315-325. 



Federman, E.B., Costello, E.J., Angold, A., Farmer, E.M., and Erkanli, A.

(1997). Development of substance use and psychiatric comorbidity in an

epidemiologic study of white and American Indian young adolescents: The

Great Smoky Mountains Study. Drug and Alcohol Dependence, 44(2-3),

69-78. 



Fendrich, M. (2001). Feasibility of biological measurement in drug

surveys. NIH grant award 3 R01 DA12425-01A1S1. 



Fendrich, M., Macksey-Amiti, M.E., Goldstein, P., Spunt, B., and

Brownstein, H.B. (1995). Substance involvement among juvenile

murderers: Comparisons with older offenders based on interviews with

prison inmates. International Journal of the Addictions, 30(11),

1363-1382. 



Forman, V. (2001). ELCID Technical Report 2001.1. Baltimore: Johns

Hopkins University [Online]. Available: www.jhu.edu/

~janthony/elcid/techrep.0801.pdf. 



Gfroerer, J., and Brodsky, M. (1992). The incidence of illicit drug use in

the United States, 1962-1989. British Journal of Addiction, 87(9),

1345-1351. 



Goldstein, P.J. (1985). The drugs/violence nexus: A tripartite conceptual

framework. Journal of Drug Issues, 15(4), 493-506. 



Goldstein, P.J. (1998). Drugs, violence, and Federal funding: A research

odyssey. Substance Use & Misuse, 33(9), 1915-1936. 



Golub, A.L., and Johnson, B.D. (1999). Cohort changes in illegal drug use

among arrestees in Manhattan: From the Heroin Injection Generation to

the Blunts Generation. Substance Use & Misuse, 34(13), 1733-1763. 



Golub A., and Johnson, B.D. (2001a). Variation in youthful risks of

progression from alcohol and tobacco to marijuana and to hard drugs

across generations. American Journal of Public Health, 91(2), 225-232. 



Golub, A.L., and Johnson, B.D. (2001b). The rise of marijuana as the drug

of choice among youthful adult arrestees (Research in Brief, June, NCJ

187490). Washington DC: U.S. Department of Justice, National Institute

of Justice. 



Haber, S., Barchas, P.R., and Barchas, J.D. (1981). A primate analogue of

amphetamine-induced behaviors in humans. Biological Psychiatry, 16(2),

181-196.



Harrison, L. (1997). The validity of self-reported drug use in survey

research: An overview and critique of research methods. In L. Harrison

and A. Hughes (eds.), The Validity of Self-Reported Drug Use: Improving

the Accuracy of Survey Estimates (NIDA Research Monograph 167, pp.

17-36). Rockville, MD: U.S. Department of Health and Human Services,

National Institute on Drug Abuse.



Harrison, L. (2001). Validity of self-reported drug use in population

surveys. NIH grant award 5 R01 DA10930-03. 



Harrison, L., and Gfroerer, J. (1992). The intersection of drug use and

criminal behavior: Results from the National Household Survey on Drug

Abuse. Crime and Delinquency, 38, 422-443. 



Harrison, L.D., and Backenheimer, M. (1998). Editors' introduction:

Evolving insights into the drug-crime nexus. Substance Use & Misuse,

33(9), 1763-1777. 



Harrison, R.J., Connor, D.F., Nowak, C., and Melloni, R.H. (2000).

Chronic low-dose cocaine treatment during adolescence facilitates

aggression in hamsters. Physiology and Behavior, 69(4-5), 555-562. 



Hawkins, J.D., Catalano, R.F., and Miller, J.Y. (1992). Risk and protective

factors for alcohol and other drug problems in adolescence and early

adulthood: Implications for substance abuse prevention. Psychological

Bulletin, 112(1), 64-105. 



Hawkins, J.D., Von Cleve, E., and Catalano, R.F. (1991). Reducing early

childhood aggression: Results of a primary prevention program. Journal

of the American Academy of Child and Adolescent Psychiatry, 30(2),

208-217. 



Higley, J.D., Mehlman, P.T., Higley, S.B., Fernald, B., Vickers, J.,

Lindell, S.G., Taub, D.M., Suomi, S.J., and Linnoila, M. (1996a).

Excessive mortality in young free-ranging male nonhuman primates with

low cerebrospinal fluid 5-hydroxyindoleacetic acid concentrations.

Archives of General Psychiatry, 53(6), 537-543. 



Higley, J.D., Mehlman, P.T., Poland, R.E., Taub, D.M., Vickers, J.,

Suomi, S.J., and Linnoila, M. (1996b). CSF testosterone and 5-HIAA

correlate with different types of aggressive behaviors. Biological

Psychiatry, 40(11), 1067-1082. 



Higley, J.D., Suomi, S.J., and Linnoila, M. (1996a). A nonhuman primate

model of type II excessive alcohol consumption? Part 1. Low

cerebrospinal fluid 5-hydroxyindoleacetic acid concentrations and

diminished social competence correlate with excessive alcohol

consumption. Alcoholism: Clinical & Experimental Research, 20(4),

629-642. 



Higley, J.D., Suomi, S.J., and Linnoila, M. (1996b). A nonhuman primate

model of type II alcoholism? Part 2. Diminished social competence and

excessive aggression correlates with low cerebrospinal fluid

5-hydroxyindoleacetic acid concentrations. Alcoholism: Clinical &

Experimental Research, 20(4), 643-650. 



Hill, A.B. (1965).The environment and disease: association or causation?

Proceedings of the Royal Society of Medicine, 58, 295-300.



Holder, H.D. (1993). Prevention of alcohol-related accidents in the

community. Addiction, 88(7), 1003-1012. 



Holder, H.D. (2001). Prevention of alcohol problems in the 21st century:

Challenges and opportunities. American Journal on Addictions, 10(1),

1-15. 



Holder, H., Flay, B., Howard, J., Boyd, G., Voas, R., and Grossman, M.

(1999). Phases of alcohol problem prevention research. Alcoholism:

Clinical & Experimental Research, 23(1), 183-194. 



Holder, H.D., Gruenewald, P.J., Ponicki, W.R., Treno, A.J., Grube, J.W.,

Saltz, R.F., Voas, R.B., Reynolds, R., Davis, J., Sanchez, L., Gaumont, G.,

and Roeper, P. (2000). Effect of community-based interventions on

high-risk drinking and alcohol-related injuries. Journal of the American

Medical Association, 284(18), 2341-2347. 



Hser, Y.I., Anglin, M.D., and McGlothlin, W. (1987). Sex differences in

addict careers. 1. Initiation of use. American Journal of Drug and Alcohol

Abuse, 13(1-2), 33-57. 



Hubbard, J.A., Dodge, K.A., Cillessen, A.H., Coie, J.D., and Schwartz, D.

(2001). The dyadic nature of social information processing in boys'

reactive and proactive aggression. Journal of Personality and Social

Psychology, 80(2), 268-280. 



Inciardi, J.A. (1990). The crack-violence connection within a population

of hard-core adolescent offenders. In M. De La Rosa, E.Y. Lambert, and

B. Gropper (eds.), Drugs and violence: Causes, correlates, and

consequences (Research Monograph 103, pp. 92-111). Rockville, MD:

U.S. Department of Health and Human Services, National Institute on

Drug Abuse. 



Inciardi, J.A., and Pottieger, A.E. (1998). Drug use and street crime in

Miami: An (almost) twenty-year retrospective. Substance Use & Misuse,

33(9), 1839-1870. 



Inciardi, J.A., and Russe, B.R. (1977). Professional thieves and drugs.

International Journal of the Addictions, 12(8), 1087-1095. 



James, K.E., Wagner, F.A., and Anthony, J.C. (2002). Regional variation

in drug purchase opportunity among youths in the United States,

1996-1997. Journal of Urban Health, 79(1), 104-112. 



Jessor, R. (1998). New perspectives on adolescent risk behavior. In R.

Jessor (ed.), New perspectives on adolescent risk behavior (pp. 1-12).

Cambridge, UK: Cambridge University Press. 



Jessor, R., and Jessor, S.L. (1977). Problem behavior and psychosocial

development: A longitudinal study of youth. New York: Academic Press. 



Johnson, B.D., Dunlap, E., and Maher, L. (1998). Nurturing for careers in

drug use and crime: Conduct norms for children and juveniles in

crack-using households. Substance Use & Misuse, 33(7), 1511-1546. 



Johnson, E.O., Arria, A.M., Borges, G., Ialongo, N., and Anthony, J.C.

(1995). The growth of conduct problem behaviors from middle childhood

to early adolescence: Sex differences and the suspected influence of early

alcohol use. Journal of Studies on Alcohol, 56(6), 661-671. 



Johnston, L.D., O'Malley, P.M., and Eveland, L.K. (1978). Drugs and

delinquency: A search for causal connections. In D.B. Kandel (ed.),

Longitudinal research on drug use (pp.137-156). Washington, DC:

Hemisphere Publishing Co. 



Jones, M.B., and Offord, D.R. (1989). Reduction of antisocial behavior in

poor children by nonschool skill-development. Journal of Child

Psychology and Psychiatry and Allied Disciplines, 30(5), 737-750.



Kaplan, C.D., and Lambert, E.Y. (1995). The daily life of heroin-addicted

persons: The biography of specific methodology. In E.Y. Lambert, R.S.

Ashery, and R.H. Needle (eds.), Qualitative methods in drug abuse and

HIV research (NIDA Research Monograph 157, pp. 100-116). Rockville,

MD: U.S. Department of Health and Human Services, National Institute

on Drug Abuse. 



Kaplan, H.B. (ed.) (1995). Drugs, crime, and other deviant adaptations.

New York: Plenum Press. 



Kellam, S.G., and Anthony, J.C. (1998). Targeting antecedents to prevent

tobacco smoking: Findings from an epidemiologically based randomized

field trial. American Journal of Public Health, 88(10), 1490-1495.



Kellam, S.G., Ensminger, M.E., and Turner, R.J. (1977). Family structure

and the mental health of children. Concurrent and longitudinal

community-wide studies. Archives of General Psychiatry, 34(9),

1012-1022. 



Kessler, R.C. (1999). The World Health Organization International

Consortium in Psychiatric Epidemiology (ICPE): Initial work and future

directions. Acta Psychiatrica Scandinavica, 99(1), 2-9. 



Kolb, L. (1925). Drug addiction in its relation to crime. Mental Hygiene,

9, 77. 



Kosterman, R., Hawkins, J.D., Guo, J., Catalano, R.F., and Abbott, R.D.

(2000). The dynamics of alcohol and marijuana initiation: Patterns and

predictors of first use in adolescence. American Journal of Public Health,

90(3), 360-366. 



Kotler, M., Barak, P., Cohen, H., Averbuch, I.E., Grinshpoon, A.,

Gritsenko, I., Nemanov, L., and Ebstein, R.P. (1999). Homicidal behavior

in schizophrenia associated with a genetic polymorphism determining low

catechol O-methyltransferase (COMT) activity. American Journal of

Medical Genetics, 88(6), 628-633. 



Langbehn, D.R., and Cadoret, R.J. (2001). The adult antisocial syndrome

with and without antecedent conduct disorder: Comparisons from an

adoption study. Comprehensive Psychiatry, 42(4), 272-282. 



Levitt, S.D., and Venkatesh, S.A. (1998). An economic analysis of a

drug-selling gang's finances. NBER Working Paper No. 6592. Cambridge,

MA: National Bureau of Economic Research. 



Licata, A., Taylor, S., Berman, M., and Cranston, J. (1993). Effects of

cocaine on human aggression. Pharmacology Biochemistry and Behavior,

45(3), 549-552. 



Lindesmith, A.R. (1938). A sociological theory of drug addiction.

American Journal of Sociology, 43, 593-613. 



Loeber, R., Farrington, D.P., Stouthamer-Loeber, M., Moffitt, T., and

Caspi, A. (1998). The development of male offending: Key findings from

the first decade of the Pittsburgh Youth Study. Studies on Crime and

Crime Prevention, 7(2), 141-172. 



Loeber, R., and Schmaling, K.B. (1985). The utility of differentiating

between mixed and pure forms of antisocial child behavior. Journal of

Abnormal Child Psychology, 13(2), 315-335. 



Loeber, R., Wei, E., Stouthamer-Loeber, M., Huizinga, D., and

Thornberry, T.P. (1999). Behavioral antecedents to serious and violent

offending: Joint analyses from the Denver Youth Survey, Pittsburgh Youth

Study and the Rochester Youth Development Study. Studies on Crime and

Crime Prevention, 8(2), 245-264. 



Logan, T.K. (2001). HIV risk behavior and violence among crack users.

NIH grant award 5 R29 DA11578-03. 



Longshore, D.Y. (2000). Linking DUF and criminal history data. NIH

grant award 5 R03 DA11009-02. 



Maltz, M.D. (1996). From Poisson to the present: Applying operations

research to problems of crime and justice. Journal of Quantitative

Criminology, 12(1), 3-61. 



Manski, C.F., Pepper, J.V., and Petrie, C.V. (eds.), and the Committee on

Data and Research for Policy on Illegal Drugs (2001). Informing

America's policy on illegal drugs: What we don't know keeps hurting us.

Washington, DC: National Academy Press. 



Markowitz, S., and Grossman, M. (2000). The effects of beer taxes on

physical child abuse. Journal of Health Economics, 19(2), 271-282.



McCardle, L., and Fishbein, D.H. (1989). The self-reported effects of PCP

on human aggression. Addictive Behaviors, 14(4), 465-472. 



Melega, W.P., Raleigh, M.J., Stout, D.B., Huang, S.C., and Phelps, M.E.

(1997). Ethological and 6-[18F]fluoro-L-DOPA-PET profiles of long-term

vulnerability to chronic amphetamine. Behavioural Brain Research,

84(1-2), 259-268. 



Miczek, K.A., and Haney, M. (1994). Psychomotor stimulant effects of

d-amphetamine, MDMA and PCP: Aggressive and schedule-controlled

behavior in mice. Psychopharmacology (Berlin), 115(3), 358-365.



Montagne, M. (1990). The social epidemiology of international drug

trafficking: Comparison of source of supply and distribution networks.

International Journal of the Addictions, 25(5), 557-577. 



Moore, M.S., and Thompson, D.M. (1978). Acute and chronic effects of

cocaine on extinction-induced aggression. Journal of the Experimental

Analysis of Behavior, 29(2), 309-318. 



Moro, M., Salvador, A., and Simon, V.M. (1997). Effects of repeated

administration of d-amphetamine on agonistic behaviour of isolated male

mice. Behavioural Pharmacology, 8(4), 309-318. 



Natarajan, M., and Belanger, M. (1998). Varieties of drug trafficking

organizations: A typology of cases prosecuted in New York City. Journal

of Drug Issues, 28(4), 1005-1025. 



Natarajan, M., Clarke, R.V., and Belanger, M. (1996). Drug dealing and

pay phones: The scope for intervention. Security Journal, 7(4), 245-251. 



Neiderhiser, J.M. (2001). Understanding the roles of genome and

envirome: Methods in genetic epidemiology. British Journal of Psychiatry

Supplement, 40, S12-7. 



Neiderhiser, J.M., Pike, A., Hetherington, E.M., and Reiss, D. (1998).

Adolescent perceptions as mediators of parenting: Genetic and

environmental contributions. Developmental Psychology, 34(6),

1459-1469. 



Neiderhiser, J.M., Reiss, D., Hetherington, E.M., and Plomin, R. (1999).

Relationships between parenting and adolescent adjustment over time:

Genetic and environmental conditions. Developmental Psychology, 35,

680-692. 



Newcomb, M.D. (1992). Understanding the multidimensional nature of

drug use and abuse: The role of consumption, risk factors and protective

factors. In M. Glantz and R. Pickens (eds.), Vulnerability to drug abuse.

Washington, DC: American Psychological Association. 



Newcomb, M.D., and Bentler, P.M. (1988). Consequences of adolescent

drug use: Impact on the lives of young adults. Newbury Park, CA: Sage

Publications. 



Nurco, D. (1998). A long-term program of research on drug use and crime.

Substance Use & Misuse, 33(9), 1817-1839. 



Nurco, D.N., Bonito, A.J., Lerner, M., and Balter, M.B. (1975). Studying

addicts over time: Methodology and preliminary findings. American

Journal of Drug and Alcohol Abuse, 2(2), 183-196. 



Nurco, D.N., Kinlock, T., O'Grady, K., Lerner, M., and Hanlon, T.E.

(1996). Perceptions of social pathology in the neighborhood and the

etiology of narcotic addiction: A retrospective study. Journal of Nervous

and Mental Disease, 184(1), 35-42. 



Parker, R.N., and Rebhun, L.A. (1995). Alcohol and homicide: A deadly

combination of two American traditions. Albany: State University of New

York Press. 



Parker, R.N., and Toth, A.M. (1990). Family, intimacy, and homicide: A

macro-social approach. Violence and Victims, 5(3), 195-210. 



Patterson, G.R., Dishion, T.J., and Yoerger, K. (2000). Adolescent growth

in new forms of problem behavior: Macro- and micro-peer dynamics.

Prevention Science, 1(1), 3-13. 



Patterson, G.R., Reid, J.R., and Dishion, T.J. (1992). Antisocial boys.

Eugene, OR: Castalia Publishing Co. 



Pearson, J.L., Hunter, A.G., Ensminger, M.E., and Kellam, S.G. (1990).

Black grandmothers in multigenerational households: Diversity in family

structure and parenting involvement in the Woodlawn community. Child

Development, 61(2), 434-442.



Peffer-Smith, P.G., Smith, E.O., and Byrd, L.D. (1983). Effects of

d-amphetamine on self-aggression and posturing in stumptail macaques.

Journal of the Experimental Analysis of Behavior, 40(3), 313-320. 



Pescor, M.J. (1939). The Kolb classification of drug addicts. Public Health

Reports, Supp. 155. 



Petronis, K.R., and Anthony, J.C. (2000). Perceived risk of cocaine use

and experience: Do they cluster within U.S. neighborhoods and cities?

Drug and Alcohol Dependence, 57(3), 183-192. 



Preble, E., and Casey, J.J. (1969). Taking care of business--the heroin

user's life on the street. International Journal of the Addictions, 4, 1-24.



Robins, L.N. (1966). Deviant children grown up: A sociological and

psychiatric study of sociopathic personality. Baltimore, MD: Williams and

Wilkins.



Russell, J.W., Greenberg, B.D., and Segal, D.S. (1984). The effects of

phencyclidine on spontaneous aggressive behavior in the rat. Biological

Psychiatry, 19(2), 195-202. 



Sandstrom, M.J., and Coie, J.D. (1999). A developmental perspective on

peer rejection: Mechanisms of stability and change. Child Development,

70(4), 955-966. 



Scheier, L.M., and Botvin, G.J. (1996). Purpose in life, cognitive efficacy,

and general deviance as determinants of drug abuse in urban black youth.

Journal of Child and Adolescent Substance Abuse, 5(1), 1-26. 



Schulenberg, J., Bachman, J.G., O'Malley, P.M., and Johnston, L.D.

(1994). High school educational success and subsequent substance use: A

panel analysis following adolescents into young adulthood. Journal of

Health and Social Behavior, 35(1), 45-62. 



Spunt, B., Brownstein, H., Goldstein, P., Fendrich, M., and Liberty, H.J.

(1995). Drug use by homicide offenders. Journal of Psychoactive Drugs,

27(2), 125-134. 



Spunt, B.J., Goldstein, P.J., Bellucci, P.A., and Miller, T. (1990).

Race/ethnicity and gender differences in the drugs-violence relationship.

Journal of Psychoactive Drugs, 22(3), 293-303. 



Spunt, B., Goldstein, P., Brownstein, H., and Fendrich, M. (1994). The

role of marijuana in homicide. International Journal of the Addictions,

29(2), 195-213.



Stevens, M.C., Kaplan, R.F., and Bauer, L.O. (2001). Relationship of

cognitive ability to the developmental course of antisocial behavior in

substance-dependent patients. Progress in Neuropsychopharmacology and

Biological Psychiatry, 25(8), 1523-1536. 



Tehrani, J.A., Brennan, P.A., Hodgins, S., and Mednick, S.A. (1998).

Mental illness and criminal violence. Social Psychiatry and Psychiatric

Epidemiology, 33 Supp. 1, S81-85.



Terry, C.E., and Pellens, M. (1928). The opium problem. New York:

Committee on Drug Addictions and Bureau of Social Hygiene.



Thornberry, T.P. (ed.) (1997). Developmental theories of crime and

delinquency. New Brunswick, NJ: Transaction Publishers.



Tolman, R.M., Edleson, J.L., and Fendrich, M. (1996). The applicability

of the theory of planned behavior to abusive men's cessation of violent

behavior. Violence and Victims, 11(4), 341-354.



Tracy, P., Wolfgang, M.E., and Figlio, R.M. (1990). Delinquency careers

in two birth cohorts. New York: Plenum Press.



Tsuang, M.T., Bar, J.L., Harley, R.M., and Lyons, M.J. (2001). The

Harvard twin study of substance abuse: What we have learned. Harvard

Review of Psychiatry, 9(6), 267-279. 



Tyler, C.B., and Miczek, K.A. (1982). Effects of phencyclidine on

aggressive behavior in mice. Pharmacology Biochemistry and Behavior,

17(3), 503-510.



Verdonik, F., and Sherrod, L.R. (eds.) (1984). An inventory of

longitudinal research on childhood and adolescence. New York: Social

Science Research Council.



Waldorf, D. (1998). Misadventures in the drug trade. Substance Use &

Misuse, 33(9), 1957-1991. 



Wiens, J.A., Addicott, J.F., Case, T.J., and Diamond, J. (1986). Overview:

The importance of spatial and temporal scale in ecological investigations.

In J. Diamond and T.J. Case (eds.), Community Ecology (pp. 145-153).

New York: Harper and Row. 



Winick, C. (1962). Maturing out of narcotic addiction. U.N. Bulletin on

Narcotics, 14(1), 1-7. 



Wish, E.D. (1988). Identifying drug-abusing criminals. In C.G. Leukefeld

and F.M. Tims (eds.), Compulsory treatment of drug abuse: Research and

clinical practice (NIDA Research Monograph 86, pp. 139-159).

Washington, DC: U.S. Government Printing Office. 



Wish, E.D., Gray, T., Sushinsky, J., Yacoubian, G.S., and Fitzgerald, N.

(2000). An experiment to enhance the reporting of drug use by arrestees.

Journal of Drug Issues, 30(1), 55-76.



Yacoubian, G.S., Wish, E.D., and Perez, D.M. (2001). A comparison of

saliva testing to urinalysis in an arrestee population. Journal of

Psychoactive Drugs, 33(3), 289-294.



Author's Note



Valerie Forman coauthored the following sections: "Tension that involves

approach or methods" and "Criterion/guideline 2: Biological or other

theoretical plausibility." Correspondence concerning this paper should be

addressed to James. C. Anthony, Electronic Collaboratory for

Investigations About Drugs, JHU School of Hygiene and Public Health,

624 North Broadway, Room 893, Baltimore, MD 21205; e-mail

janthony@jhu.edu.



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



RESEARCH ON DRUGS-CRIME LINKAGES: THE NEXT

GENERATION



Robert MacCoun, Beau Kilmer, and Peter Reuter



About the Authors



Robert MacCoun is with the Goldman School of Public Policy and

Jurisprudence and Social Policy Program at Boalt Hall School of Law,

University of California at Berkeley. Please email any comments to

maccoun@socrates.berkeley.edu. Beau Kilmer is with the Drug Policy

Research Center at RAND. Peter Reuter is with the School of Public

Affairs and Department of Criminology at the University of

Maryland.



Introduction



The association between drugs and crime in the public mind is so strong

that a recent psychology experiment showed the word "drug" tightly linked

to such words as "choke," "knife," "fight," and "wound" in participants'

associative memory networks (Bushman, 1996). Although it is routine in

academia to deride public ignorance of all things criminological, in this

case the public is hardly deluded. Consider the following facts:[1]



o Across 35 cities in 1998, between 40 and 80 percent of male arrestees in

the Arrestee Drug Abuse Monitoring (ADAM) Program tested positive for

at least one drug at arrest (Arrestee Drug Abuse Monitoring Program,

1999).



o Nearly one-quarter (22 percent) of Federal prison inmates and one-third

(33 percent) of State prison inmates--nearly 40 percent of State inmates

convicted of robbery, burglary, or motor vehicle theft--reported being

under the influence of drugs at the time of their offense (Bureau of Justice

Statistics, 1997a, 1997b).



o Among State and Federal prison inmates, 27 percent of those serving

sentences for robbery and 30-32 percent of those serving sentences for

burglary said they committed their offense to buy drugs (Bureau of Justice

Statistics, 1991a, 1991b).



o In the 70 percent of cases in which the victim formed an opinion, 31

percent believed the offender was under the influence of drugs or alcohol

(National Crime Victimization Survey, 2000).



o A recent estimate of the economic costs of drug abuse reported that 60

percent were associated with crime and criminal justice (Harwood,

Fountain, and Livermore, 1998). 



Considerable complexities and nuances underlie these associations.

Although many of these subtleties were anticipated by astute observers in

the 1970s (see Gandossy et al., 1980), the past decade has seen a solid

scholarly consensus form around the following principles (see Fagan,

1990; Parker and Auerhahn, 1998; White and Gorman, 2000):



1. Many different data sources establish a raw correlation between drug

use and other criminal offenses. But correlation does not equal

causation: In principle, drug use might cause (promote, encourage) crime;

criminality might cause (promote, encourage) drug use; and/or both might

be caused (promoted, encouraged) by some set of "third

variables"--environmental, situational, dispositional, and/or biological. In

fact, all three pathways have empirical support in at least some settings

and populations.



2. These causal influences are probabilistic, not deterministic. Most drug

users are not otherwise criminally active, and the vast majority of

drug-using incidents neither cause nor accompany other forms of

criminality. Nevertheless, drugs clearly play an important causal role in

violent and property crime.



3. These causal influences are contingent, not unconditional. There is little

evidence that drug use per se directly causes people to become aggressive

in some direct and unconditional manner or that criminality per se causes

someone to use drugs. The drugs-crime link varies across individuals, over

time within an individual's development, across situations, and possibly

over time periods (as a function of the dynamics of drug epidemics and,

possibly, drug control policies).



4. That drug use can causally influence criminality does not necessarily

implicate the psychopharmacological properties of the drug. Intoxication,

the need or desire to raise money to buy drugs, and the nature of illicit

markets are distinct mechanisms by which drugs can cause crime. Thus,

drug prohibition cannot be only a response to drug-related crime, but it

may also be a causal antecedent to some drug-related crime. 



5. Alcohol is a drug, and it stimulates or augments a great deal of criminal

behavior, almost certainly more than the street drugs combined. We expect

that understanding the considerable heterogeneity of effects across users,

substances, cities, neighborhoods, and situations--and the interactions

among these factors--will be the central focus of drugs-crime research

during the remainder of this decade. This paper reviews the existing

literature, focusing particular attention on Goldstein's (1985) taxonomy,

the temporal dynamics of drug markets, and the consequences of

prohibition. These highlight some of the questions that should drive this

research. 



Drugs-Crime Linkages: Expanding the Goldstein Taxonomy



Goldstein's Framework



Paul Goldstein's (1985) conceptual essay offered a tripartite classification

of drugs-violence connections:



o Psychopharmacological: Violence due to the direct acute effects of a

psychoactive drug on the user.



o Economic-compulsive: Violence committed instrumentally to generate

money to purchase expensive drugs.



o Systemic: Violence associated with the marketing of illicit drugs, such

as turf battles, contract disputes, and so on. 



Goldstein and his colleagues (Brownstein et al., 1992; Goldstein et al.,

1989; Goldstein, Brownstein, and Ryan, 1992) applied this scheme

empirically to homicides in New York State (1984) and New York City

(1988). They found that drugs and alcohol were important causes for a

large share of all homicides in both samples. For 1988, near the height of

the crack epidemic, they classified 53 percent of 414 homicides as drug or

alcohol related; there was also a substantial percentage whose

drug-relatedness could not be determined. Of those homicides that could

be determined to be drug or alcohol related, 14 percent were

psychopharmacological (68 percent alcohol, 16 percent crack), 4 percent

were economic-compulsive, and 74 percent were systemic (61 percent

crack, 27 percent powder cocaine). By contrast, in 1984, before the crack

surge, only 42 percent of homicides were drug or alcohol related; 59

percent of those were psychopharmacological (79 percent alcohol), 3

percent were economic-compulsive, and 21 percent were systemic.

The difference between the findings of the two years might reflect

differences in geography to some extent (New York State versus New

York City), but it also reminds us that these numbers are not eternal

verities; they result from complex and historically dependent market

dynamics. 



Subsequent Applications



The generalizability of Goldstein et al.'s (1989) original findings was

limited by their location (New York) and timing (the height of the crack

explosion; see U.S. Sentencing Commission, 1995, 106).[2] Many studies

have tried to determine whether crimes were drug related, but few have

assessed whether the offender's drug need, drug use, or role in the drug

market was directly responsible for the crime. Although most of the

studies that used this framework were conducted by Goldstein and his

colleagues in New York (Parker and Auerhahn, 1998), there are others

worthy of attention, especially given their unique approaches. General

findings include the following: 



1. Non-NDRI (National Development and Research Institutes, Inc.) studies

of New York City in the mid- to late 1980s found that crack sellers are

more violent than other drug sellers and that their violence is not confined

to the drug-selling context (U.S. Sentencing Commission, 1995, citing

Fagan and Chin, 1990).



2. Studies of juvenile delinquents in Miami in the mid- to late 1980s found

that they were much more likely to commit a drug-related

economic-compulsive crime than a psychopharmacological or systemic

crime (Inciardi, 1990).[3]



3. The per capita drug-related homicide rate remained fairly stable in

Chicago from 1973 to 1984 and fluctuated from 1985 to 1995 (data are

from the Chicago Homicide Dataset; Block, Block, and Illinois Criminal

Justice Information Authority, 1998). Despite the fluctuations, the 1995

homicide rate was strikingly similar to the 1985 rate for all drug-related

motives except for homicides that resulted from a drug transaction;

the latter increased tenfold from 1985 to 1995. 



4. Results from Lattimore et al.'s (1997) homicide study of eight cities,

which included surveys of local officials and ADAM/UCR (Uniform

Crime Reports) analyses for 1985-94, suggest that drugs other than

cocaine and crack were not associated with homicide trends "in any

discernible way." They also found that the drug market structure was

less associated with violence than was expected. The Lattimore et al. study

questioned the role of crack and systemic crime because the crack markets

were described as highly competitive in cities where the homicide rate was

declining, increasing, or remaining the same (1997, p. 89). It is not clear,

however, that the same conclusions could be drawn if disaggregated

homicide rates (by circumstance) were considered. (Additional discussion

and methodological descriptions of these studies are reported in appendix

A.)



Limitations of Existing Research on the Goldstein Framework



The Goldstein tripartite framework has been a boon to drug research

reviewers--it is invaluable as an organizing scheme--but still, we are struck

by the relative rarity of actual empirical applications. Existing applications

overrepresent New York, and they overrepresent the crack epidemic at its

height relative to earlier and later periods. In fairness, the taxonomy was

not proposed until 1985, but it could be applied retrospectively to earlier

homicide case files. In our view, such comparisons would be invaluable.

There has been little consistency in the methods used to implement the

scheme (e.g., Goldstein's trained coders versus Inciardi's survey approach).

Little has been learned from that methodological diversity because, to our

knowledge, no two methods have ever been applied to the same sample of

cases for comparative purposes. Indeed, if one imagines a

three-dimensional matrix of major cities by time periods by methods,

almost every cell is empty and there are almost no vectors with more than

one cell occupied. This spotty record makes it hard to identify either

temporal trends or the influence of local variations on drug popularity,

drug market structures, or policies and enforcement practices. Finally, the

scheme has been applied mostly to homicide and less often to other, more

prevalent violent crimes.[4] 



Parker and Auerhahn (1998) complain that Goldstein's categories are not

mutually exclusive. This critique presumes a classical set-theoretic

approach that, in our opinion, is neither feasible nor scientifically useful

for drugs-violence research. Mutually exclusive categories are not

necessary for scientific classification (Meehl, 1995), and they are usually

impossible to achieve using sparse and noisy archival data (Ragin, 2000).

But we agree with Parker and Auerhahn's (1998) contention that "the

Goldstein tripartite framework . . . is not treated as a set of testable

propositions but rather as a set of assumptions about the nature of drug-

and alcohol-related violence."



In our view, an understanding of the taxometric properties of drug-related

violence ought to emerge inductively from more fine-grained coding of the

underlying features of these events--whether various drugs were found as

evidence, the results of toxicology on the offender and the victim, various

features of witness reports, prior record information, and so on. Because

each property or attribute would be coded separately, there would be no

effort to force events into a single classification. Psychometric analysis

could be used to test the hypothesized latent structure.[5] Such analyses

pose enormous logistical difficulties, but the payoffs for advancing our

understanding of drug violence would surely justify the effort.



In the remainder of this section, we will examine other ways in which

Goldstein's taxonomic scheme might be expanded and refined.



Psychopharmacological Violence



The prevailing view about psychopharmacological (as opposed to

economic-compulsive or systemic) violence is that it is rare and

attributable mostly to alcohol rather than illicit drugs (Fagan, 1990; Parker

and Auerhahn, 1998; White and Gorman, 2000). According to Fagan

(1990, p. 243):



[I]ntoxication does not consistently lead to aggressive behavior . . . only

limited evidence that consumption of alcohol, cocaine, heroin, or other

substances is a direct, pharmacologically based cause of crime.



According to Parker and Auerhahn (1998, p. 306):



Our review of the literature finds a great deal of evidence that the social

environment is a much more powerful contributor to the outcome of

violent behavior than are pharmacological factors associated with any of

the substances reviewed here.[6]



The Goldstein et al. (1989) analysis provides some support for these

claims; only 14 percent of the drug-related homicides appeared to be

psychopharmacological, and these largely involved alcohol either alone or

in combination with other drugs. But one in seven is hardly a trivial

fraction, and those results reflect the peak of the crack market wars, when

systemic homicides were occurring in unprecedented numbers, inflating

the denominator. 



Moderators. Examining the literature cited in many recent review essays, it

is difficult to avoid the suspicion that some authors hold

neuropharmacological factors to a stricter standard of proof than the

sociological factors under study. If the psychopharmacological claim is

that marijuana, heroin, or cocaine ingestion directly promotes violent

behavior absent any situational provocation or stressors, then that

claim is probably false. But evidence for Drug x Situation and Drug x

Psychology interaction effects hardly exonerates drug use as a causal

factor. It may be that no drug is sufficient to produce aggression in

isolation from psychological and situational moderators. But it seems clear

that some drugs--certainly alcohol--can amplify the psychological and

situational facilitators of aggression. Relevant moderators (see Bushman,

1997; Fagan, 1990; Ito, Miller, and Pollock, 1996) include:



o Situational stressors and frustrators (see Ito, Miller, and Pollock, 1996).



o Expectancy effects: personal and cultural beliefs about the effect of the

drug on behavior, and local norms about tolerable versus unacceptable

conduct when under the influence (e.g., Critchlow, 1986; Stacy, Widaman,

and Marlatt, 1990).



o Disinhibition (e.g., Parker and Auerhahn, 1998; but see Fagan, 1990).



o Impaired cognitive functioning, including reduced executive functioning

(self-control and decisionmaking ability; Fishbein, 2000; Giancola, 2000),

reduced attention to situational cues (Steele and Josephs, 1990), and

reduced self-attention (Ito, Miller, and Pollock, 1996).



o Social threats to self-identity or self-esteem (Baumeister, Smart, and

Boden, 1996) that seem particularly relevant in "cultures of honor" (see

Anderson, 1994; Bourgois, 1996; Cohen et al., 1996).



Moreover, the absence of evidence does not equal evidence of absence; the

laboratory literature on drugs and aggression is simply too spotty at present

to permit any firm conclusions. Almost the entire experimental literature

on moderators of the drugs-aggression relationship has examined alcohol

rather than illicit drugs. 



Comorbidity: Drugs in association with mental illness or alcoholism. A

second potential class of moderators of the drugs-aggression link involve

comorbid conditions--substance abuse in tandem with schizophrenia or

other psychoses, personality disorders, or alcoholism. Numerous studies

have identified a high prevalence of illicit substance abuse among

individuals diagnosed with psychiatric disorders (e.g., Compton et al.,

2000; Kessler et al., 1996; Mueser et al., 2000).[7] The causal nexus of

these comorbid conditions is unclear. The MacArthur Violence Risk

Assessment Study (Steadman et al., 1998), a prospective followup study of

clients admitted to acute psychiatric inpatient facilities, found that

substance abuse increased the probability of violent behavior, but this was

true for both psychiatric patients and matched community controls.

Neither drug dependence nor psychiatric illness predicted subsequent

violent crime in a 6-year followup of released jail detainees (Teplin,

Abram, and McClelland, 1994). 



Drug Use and Victimization[8]



Increased victimization provides another mechanism by which drugs can

become linked with violence. Although this category can be subsumed

under Goldstein's psychopharmacological category, treating it as a fourth

category might have merit because the causal mechanisms differ and it has

been largely neglected by researchers. There are a number of reasons to

expect that drug users ought to be particularly vulnerable to criminal

victimization, especially when intoxicated. First, intoxicated people often

appear (and sometimes are) more vulnerable than other targets for such

offenses as robbery, rape, or hate crimes. Second, intoxicated people are

often obnoxious, annoying, and/or offensive in their appearance, conduct,

and speech. Third, intoxication makes people's conduct unpredictable and

ambiguous--intoxication impairs the perception of signals, but it also

impairs the transmission of clear signals to others. Finally, in an active

illicit drug market, drug sellers are sometimes both intoxicated and flush

with cash. 



Fagan (1990) notes that the vulnerability of drug users to victimization has

been long recognized. For example, Wolfgang (1958) studied 

"victim-precipitated homicides" by assessing the incidence of intoxication

among victims. And Fagan (1990) reviews evidence from animal studies

showing that "substances that induce changes in an opponent's behavior

might result in increased aggression by a drug-free attacker . . ." (p. 251). 



Although Goldstein (1985) acknowledged that the victimization of drug

users constituted a distinct drugs-violence linkage, he did not include it as

a separate category in his classification scheme. Since then, the

victimization of drug users has received little attention in the drugs and

crime literature. This is not surprising given how difficult it is to assess the

relationship. First, as Goldstein (1985) argues, it is difficult to obtain this

information because victims do not want to talk to the police while

intoxicated and often do not remember the details of the offense; thus, it

may go unreported. Second, the victimization surveys that ask about

substance use usually include it as a predictor but do not ask whether it

contributed to a specific event. Third, many of these surveys only ask

about (or report) general drug use, not about specific drugs or the

circumstances of their use. Finally, the label "victim" is often problematic

when the participants are codisputants; indeed, the "victim" may have

initiated the provocation. In our view, these concerns are valid, but they do

not undermine the importance of victimization as a research topic.



The ubiquity of alcohol has made it the subject of victimization work for

50 years, and there is general agreement about its role in victimizations,

especially sexual assaults. The research on drugs is not as robust, but there

are some important findings that should be addressed in future works on

drugs and crime. The following sections provide insight about this

relationship by examining existing victimization studies of the general

population, women, and hard drug users. 



The general population. The Nation's largest victimization survey, the

National Criminal Victimization Survey (NCVS), does not ask about

victim drug use, but it is used in conjunction with other data to provide

insight about drugs and crime. Using NCVS, Markowitz's (2000)

multivariate analysis of almost 450,000 observations found that marijuana

decriminalization (a proxy for lower marijuana prices) will result in a

higher incidence of robbery and assault while higher cocaine prices will

decrease these crimes.[9] Neither measure was significantly related to rape

or sexual assault. When victims' perceptions of offender drug and alcohol

use during assaults were used as the dependent variable, the significance

of marijuana decriminalization and cocaine prices was ambiguous

(significance depends on model specification). For perceived use during

robberies, neither was significant. Although Markowitz suggests the

perception variable is questionable because of underreporting, these

findings raise questions about the causal relationship and the role of drug

use by victims, especially marijuana.



Based on an instrument similar to NCVS, Fisher et al.'s (1998) survey of

3,472 randomly selected college students found that regularly taking

recreational drugs predicted an increased likelihood of a violent

victimization but not of a theft victimization. For the general population,

Cottler et al.'s (1992) survey of a probability sample of 2,663 household

residents found that those who had used cocaine or heroin more than

five times in their lives were more than three times as likely to have

experienced a physical attack than nonusers. Those who used marijuana

more than five times (no use of other drugs) and those who used pills or

hallucinogens more than five times were no more likely to have

experienced a physical attack than nonusers. This is one of the few studies

that presents its results by drug and raises questions about the situations in

which hard drug users put themselves.



Women. Much of the victimization research focuses on women because

many of the studies are about sexual assault. Fisher, Cullen, and Turner

(2000) randomly selected 4,446 college women to participate in their

National College Women Victimization Study. That study did not find that

marijuana use was a significant predictor of sexual victimization and

stalking.[10] These findings are consistent with Markowitz's claim that the

price of cocaine and marijuana are not significant predictors of sexual

victimization. Beyond using prices and self-reports, some researchers have

drug-tested rape victims to assess their drug use. Hindmarch and

Brinkmann (1999) found that 41 percent of the 1,033 participants tested

negative for alcohol and other drugs, 37 percent tested positive for alcohol,

19 percent tested positive for cannabinoids, and 0.6 percent tested positive

for flunitrazepam (Rohypnol); however, the lack of information about

participant characteristics and site locations would prevent researchers

from creating the necessary control groups. 



Drug users. Tardiff et al. (1994) found that 31 percent of one sample of

homicide victims tested positive for cocaine metabolites. This rate did

not vary for firearm deaths versus nonfirearm deaths. McElrath, Chitwood,

and Comerford (1997) surveyed 308 intravenous drug users who were

receiving methadone and/or inpatient drug treatment about their

victimization and drug use in the previous 6 months. Those reporting

heroin use were significantly less likely to be victims of violent and

property crimes. McElrath et al. argue that heroin users sometimes have

"running partners" who may also look out for each other, thus decreasing

victimization. Crack cocaine users were four times as likely to be victims

of property crime than nonusers, leading the authors to suggest, "it is

possible that the drug-seeking behavior associated with crack-cocaine

places users in contact with a larger pool of motivated offenders."



Drug-user-on-drug-user crime was also addressed in Inciardi's delinquency

study (1990). Respondents were asked about not only drug-related

offenses they committed but also drug-related victimizations; 4.6 percent

reported being victims of psychopharmacological-related crimes, 39.9

percent reported being victims of drug robberies, and 9.0 percent reported

being victims of systemic violence.



Although every youth in the survey used at least one drug daily, it is not

clear whether the victimizations occurred while the victim was under

the influence. 



Crime victim surveys and offender surveys require respondents to make

attributions about the causes of offenders' behavior. Such causal

attributions are susceptible to numerous well-documented biases (e.g.,

Nisbett and Ross, 1980), but to date there has been little methodological

work validating these survey responses.



Economic-Compulsive Violence



Arrested and incarcerated offenders report that they committed their

offenses to raise money to purchase drugs. Of course, this might be a

convenient rationalization or excuse for antisocial behavior. Should we

believe them?



At least for heroin addiction, the answer is probably yes. Studies of heroin

"careers" show that the frequency of criminal activity tends to covary with

periods of intense use (see Fagan, 1990, for review), and addicts

significantly reduce their criminal involvement during periods of

methadone maintenance (see review in Rettig and Yarmolinsky, 1995).

But in studies applying the Goldstein taxonomy (see above),

economic-compulsive criminality has been relatively rare. White and

Gorman (2000) argue, "[B]ecause there is more money in crack

distribution than in previous illegal drug markets, drug dealing may have

obviated the need to commit property crimes and income-generating

violent crimes" (p. 189). Indeed, in our survey of drug sellers in

Washington, D.C., in the late 1980s (Reuter, MacCoun, and Murphy,

1990), more than 40 percent reported keeping some drugs for their own

consumption--39 percent of crack sellers and 69 percent of heroin sellers.

However, the claim about the high returns for crack selling is probably no

longer correct. Bourgois (1996) reports that proceeds from crack sales by

experienced users who could not maintain legitimate jobs were less than

minimum wage.[11]



But the argument that drug selling has replaced other income-generating

crime might reflect limitations of recent work. First, as we have noted,

most studies applying the Goldstein framework were conducted at the

peak of the crack epidemic, when the sheer prevalence of street drug sales

was probably at an all-time high (see Saner, MacCoun, and Reuter, 1995).

Second, most studies have largely examined crimes with violent outcomes

rather than robberies or burglaries in which no homicide occurred. One

exception is the Caulkins et al. (1997) study, which attributed a substantial

fraction of robberies and burglaries to economic-compulsive crime, and a

sizeable fraction of those economic-compulsive crimes to cocaine. 



The ADAM Program provides some opportunities for studying these

issues (e.g., Arrestee Drug Abuse Monitoring Program, 1999). The

ADAM/DUF (Drug Use Forecasting) instrument was modified in 1995 to

include a question asking whether the arrestee needed drugs or alcohol at

the time of the offense.[12] Appendix B summarizes data for the period

1995 to 1999 for this survey item. As one would expect, these attributions

are more common for income-generating offenses (14 percent) than for

non-income-generating offenses (10 percent)--a reliable but quite modest

difference. 



Our understanding and interpretation of economic-compulsive crime ought

to evolve as the scientific understanding of drug dependence evolves.

Recent decades have seen great progress in the understanding of such

phenomena as tolerance, withdrawal, reinforcement, and drug craving (see

Science, 1997). Leshner (1997, pp. 45-46) notes that many assume the

following:



[T]he more dramatic the physical withdrawal symptoms, the more serious

or dangerous the drug must be. This thinking is outdated . . . many of the

most addicting and dangerous drugs do not produce severe physical

symptoms upon withdrawal. . . . What does matter tremendously is

whether or not a drug causes what we now know to be the essence of

addiction: compulsive drug seeking and use, even in the face of negative

health and social consequences.



There are also intriguing new findings from behavioral economics research

on the price elasticity of demand for cocaine and opiates--the percentage

decline in demand for a 1-percent increase in price. The conventional

wisdom is that addicts are relatively insensitive to price, at least in the

short run, because they are enslaved to their drug and must find ways to

obtain it to avoid withdrawal symptoms. If addicts were relatively

insensitive to price, one would expect price increases to produce increased

economic-compulsive crime. But recent studies (reviewed in Caulkins and

Reuter, 1996) suggest considerable price sensitivity, with elasticities for

cocaine ranging from -0.7 to -2.0. A possible explanation for the high

elasticity among heavy users is that they spend most of their earnings on

the drug and may respond to the increased difficulty of maintaining

desired consumption levels (i.e., avoiding withdrawal) by seeking

treatment.



Systemic Violence



The third of Goldstein's categories is systemic violence. This has been

narrowly interpreted as referring to struggles for competitive advantage.

We suggest here that drug markets generate violence in a variety of ways

and that market violence varies systematically over time and place. 



A brief history of the markets. There was an epidemic of initiation into

heroin use in the 1970s; after that, heroin initiation rates remained low

until the late 1990s. The number of heroin addicts (a function of the

number of initiates and the length of their addiction careers) remained

fairly stable at about 750,000 from 1981 to 1997.[13] During that period,

most heroin purchases were made by an aging cohort of experienced users. 



Powder cocaine and crack had a similar dynamic, only with different

parameters. Powder cocaine initiation rates were high from about 1975 to

1988; the number of dependent users has been quite stable since about

1988. The crack epidemic came later, from about 1982 to 1990 (depending

on the city; see Blumstein and Cork, 1996). Estimates of the number of

dependent users of either crack or powder cocaine range from 600,000 to

3,600,000 (see Rhodes et al., 2000).



Many retailers are now also frequent users (Arrestee Drug Abuse

Monitoring Program, 1999). Selling seems to be opportunistic for many

users; sudden access to an unusually large source of cash may lead a

regular buyer to become a seller for a day. Thus, at the low end of the

market, it may be difficult to distinguish systemic from

psychopharmacological violence.



Enforcement against these markets, as measured by years of jail time per

ton of drugs, probably declined through the early 1980s but then

intensified from 1985 onward. In 1990, the Colombian government

aggressively attacked the principal exporters of cocaine from Colombia.

There are a number of indications that this led to a temporary tightening of

the cocaine market; otherwise, prices have declined throughout the

period, while consumption has been declining modestly since 1988.



Conceptual issues. The markets for illegal drugs operate without the usual

protections against fraud and violence offered by the civil tort system. The

state, instead of attempting to facilitate transactions, aims to disrupt them.

Contracts cannot be enforced through written documents and the legal

system; agreements are made hurriedly, sometimes in ambiguous code,

and orally.[14] Territories cannot be allocated through bidding for

desirable locations because there is no enforceable ownership of property

for these purposes.



Yet the illegality itself is insufficient to generate high levels of violence in

the market. Prostitution, although frequently unsightly and sometimes a

nuisance, does not generate much by way of additional violence.

Bookmaking, notwithstanding the drama of the film "The Sting," was also

a generally peaceful affair; bookies were more likely to die in bed than on

the job. Even for some drugs, the markets generate little violence;

marijuana in general does not spark much injury as the result of

competitive or transactional disputes.[15]



Some drug markets, however, are clearly violent; many participants are at

risk of being killed or seriously wounded by others in the same business,

either as buyers or sellers, and there are unintended shootings of innocent

bystanders. The crack market is thought to be particularly prone to

market-related violence.



Why are these drug markets, particularly for crack, so violent? We suggest

that four factors contribute:



1. The youth of participants. Rates for violent crime peak early, at about

ages 18-22. The young are particularly likely to lack foresight and thus

engage in violence to settle disputes. The crack market was the first mass

drug market in which most of the sellers were young.



2. The value of the drugs themselves. The cocaine that fills a plastic 

sandwich bag is worth thousands of dollars. The return to sudden,

situational violence could be very high.



3. The intensity of law enforcement. Transactions are conducted under 

considerable uncertainty as a consequence of increased law enforcement.

Intensified enforcement increases the incentives for violence by raising

the adverse consequences of identifying someone as a potential informant.



4. The indirect consequence of drug use. Users are more violent and

aggressive, and this encourages dealers to prefer selling out of doors

or in highly protected settings. It also promotes unreliable behavior among

user/dealers and thus more retaliation by their suppliers.



It is probably the combination of these factors, rather than any one of

them, that accounts for the extraordinary violence associated with crack

markets in the late 1980s. That violence seems to have fallen substantially

in the late 1990s, perhaps reflecting the aging of participants in crack

markets (Golub and Johnson, 1997), although violence itself, as well as

enforcement, may also have selected out the most violent participants;

Taylor, Caulkins, and Reuter (2000) present a model in which violence

declines with more intense enforcement as a consequence of selective

incarceration. 



Competitive and internal violence. Attention has been given to violence

generated by competition between sellers. Less attention has been given to

violence within selling organizations, although the older literature on

organized crime and illegal markets reported a great deal on this (e.g.,

Block, 1980).



Criminal organizations are hindered internally by lack of access to the civil

courts. Employment contracts cannot be enforced except privately.

Managerial succession is complicated by the specificity of reputation

within the organization; a promising midlevel manager cannot readily

provide evidence of performance to another potential employer so higher

level managers get weaker market signals and may withhold deserved

promotions or merit increases. This gives incentives to lower level agents

to use violence for upward mobility.



Symmetrical with successional violence is disciplinary violence. Managers

have reason to fear subordinates who can provide evidence against them;

the longer lasting the relationship, the greater the potential for harm from

informing. Thus, managers may use violence as a tool to reduce risks of

informing. They have more incentive for doing so than do high-level

dealers in transactions with low-level dealers because the information

about these acts will spread more rapidly and extensively.[16] There are

numerous stories of this kind of violence in Colombian drug-dealing

organizations.



Thus, the violence in atomistic markets has different sources than that in

markets serviced by larger selling organizations. Which generates greater

violence from a given set of participants cannot be determined

theoretically, but some of the decline in market-related violence may

reflect changes in organizational structure.



Other market characteristics and violence. Exhibit 1 presents a simple

classification of markets according to whether buyers and sellers come

from the neighborhood or elsewhere. We believe that this taxonomy,

originally identified for purposes of analyzing vulnerability to enforcement

(Reuter and MacCoun, 1992), may also be useful in the study of violence.

Markets characterized by mostly resident dealers and customers are

labeled local markets. Export markets are ones in which residents of the

neighborhood sell drugs to nonresidents. Markets in which mostly

nonresident dealers sell to local residents are characterized here as import

markets. Finally, markets in which both sellers and customers are mostly

nonresidents are labeled here as public markets because they tend to occur

at such large public locations as parks, train or bus stations, or

schoolyards. 



Each class of market differs in the potential for violence. Local markets,

precisely because they involve buyers and sellers who know each other, do

not lend themselves to territorial competition. At the other extreme are

public markets, in which buyers and sellers cannot readily find each other

except at specific locations; the incentives for territoriality are

consequently greater.



Transactional violence may also vary in these dimensions. Local markets

discourage cheating of buyers as a consequence of the ongoing

connections between buyers and sellers; a local customer is more likely to

spread information effectively to other potential customers than one who

has little connection to other buyers. It is not clear whether much of the

transactional violence comes from buyers, as opposed to associates and

rival sellers.



If this is correct, then the maturation of cocaine and heroin markets will

tend to reduce market-related violence by reducing the size of all but local

markets. Moreover, as a result of the dissemination of beepers and cell

phones, an increasing share of cocaine transactions may be occurring in

locations (apartments, restaurants, offices) that are agreed on by the buyer

and seller for their mutual convenience. Johnson, Golub, and Dunlap

(2000, p. 191, table 6.1) report that in New York City in the 1990s, the

"seller style" included phone and delivery services as well as freelancers.

Poor and socially isolated cocaine users still frequently conduct

transactions in exposed locations, chosen precisely because they facilitate

the coming together of buyers and sellers. So probably do many heroin

addicts, given their generally impoverished state. The ability to choose

locations on the basis of specific situational need not only reduces

territorially motivated violence but also reduces the vulnerability of

buyers to robbery and other victimization because fewer of them need to

congregate at specific locations, which thus become less attractive to

predators.



The Temporal Dynamics of Drug Markets



In the past several years, numerous authors have examined the emergence

and decline of crack markets as a key factor in the steep rise in American

violence from 1985 to 1990, and the even steeper drop since 1993 (see

Blumstein and Wallman, 2000). In our view, the case for crack's role in the

crime rise is quite compelling; its role in the post-1993 decrease is more

subtle and by no means an open-and-shut case.



Many discussions of the crime drop fail to distinguish between a decline in

the crack market and a decline in the linkage between crack and crime--but

a decline in the crack-crime link is part of the crime drop outcome to be

explained. It is true that DUF (and now ADAM) data show declines

in positive cocaine tests among arrestees in many cities (e.g., Arrestee

Drug Abuse Monitoring Program, 1999). And the reduced violence

attributable to crack selling has made crack markets less visible. But

nationwide, hardcore cocaine use remained surprisingly stable during

the 1990s (Rhodes et al., 2000). Indeed, from 1990 to 1998, there were

rising cocaine mentions in emergency rooms (Substance Abuse and

Mental Health Services Administration, 2001) and rising cocaine seizures.

Nevertheless, recent multicity comparisons (Baumer et al., 1998;

Lattimore et al., 1997) indicate reliable positive correlations between

various indices of crack use and homicide and other offense rates.



Various experts have suggested that the changing dynamics of drug

markets may matter as much or more as any decline in total market activity

(e.g., Ousey and Lee, 2000). Below, we consider a few more complex

accounts of the link between crack market dynamics and violence.



NDRI'S conduct norm account. Johnson and his colleagues at NDRI

(Lipton and Johnson, 1998) have produced a valuable interdisciplinary,

multimethod program of research on street drug markets in New York,

spanning several decades. They recently offered an account of the decline

in drug-related violence based on the notion of "conduct norms" (Johnson,

Golub, and Dunlap, 2000), arguing that New York street drug markets

have passed through three phases. (They vacillate between "period" and

"cohort" versions of the story.) The "heroin injection era" peaked during

1960-73; the "cocaine/crack era" peaked during 1984-89; and the

"marijuana/blunts era" started around 1990. Associated with each era are

distinct birth cohorts with distinctive behavioral patterns. "HeroinGen"

drug users (born 1945-54) were active in drug sales and property crime,

but gun use was relatively rare. "CrackGen" drug users (born 1955-69)

frequently participated in robbery and used guns for protection and 

reputation. Finally, "BluntGen" drug users (born 1970-79) are less likely

than early cohorts to engage in violence. 



Drawing on their rich ethnographic database, Johnson and colleagues

(2000) argue that these behavior changes reflect two successive

transformations of conduct norms for appropriate behavior in the

drug-using community. For example, in CrackGen's "Subculture of

Assault," a shared norm counseled: "Be aggressive and threatening to

avoid robbery. . . . Carry weapons for protection. . . . Threaten or assault

those who attempt to sell crack in your territory. Maintain your reputation

as dangerous, tough, and 'crazy,' regardless of the physical harm inflicted

or suffered" (p. 181). But for the BluntGen, the norm states: "Don't use

crack. Crackheads are s---! . . . Addicts are the scum of the earth. Stay

safe, stay alive. Don't mix cocaine or heroin with my marijuana. Shun and

exclude heroin and crack users from peer groups" (p. 185).



This norm account is fascinating and quite plausible. From a policy

perspective, it would be tremendously useful to find a way to preserve and

promote the BluntGen's more pacifist stance (though not, of course, their

consumption of blunts). Still, the evidence is causally ambiguous. Are

these conduct norms actually causes of the decline in violence during the

1990s, are they descriptions of it, or are they consequences of it?



There is little doubt that conduct norms exist and are important in shaping

deviant behavior. Cialdini, Kallgren, and Reno (1991) make a useful

distinction between injunctive norms (what others think I should do) and

descriptive norms (what others are actually doing). There is ample

evidence that purely descriptive norms--changes in the local prevalence of

a behavior--can have a self-reinforcing action. But attitudes and norms

are shaped by behavior as well as shaping it; research on cognitive

dissonance theory and self-perception theory suggest that such

conformity-based behavioral changes will tend to produce corresponding

(but retrospective) changes in relevant attitudes (see Eagly and Chaiken,

1993). Controlled social psychology experiments show that norm diffusion

effects occur and that they can be strong, but these experiments also

show that apparent norm effects are sometimes spurious (e.g., Kerr et al.,

1987).



Clearly, research on drug-using norms cannot move to the

laboratory--although one can imagine informative scenario-based

experiments embedded in field interviews. But it would be enormously

useful to make additional use of the NDRI data (and related data sources,

such as the Office of National Drug Control Policy's Pulse Check),

linking the timing of the ethnographic material more precisely to

month-to-month quantitative archival data on drug selling (or its proxies)

and violent crimes. Furthermore, archives of ethnographic data collected in

different cities during the past decade might be reanalyzed to search for

cross-city norm differences that might correlate with cross-city differences

in violent crime. Ideally, one might develop methods for identifying

"leading indicators" of emerging trends in drug using, drug selling, and

drug-related violence.



Blumstein and Cork's drug-gun diffusion account. In a series of articles

(see Blumstein, 2000a; Blumstein and Cork, 1996; Cork, 1999), Alfred

Blumstein and his collaborator Daniel Cork hypothesize a causal chain

linking the late 1980s crack epidemic to rising violence nationwide.

According to Blumstein, the 1980s growth in illicit drug markets, together

with stringent enforcement crackdowns, led to the recruitment of juvenile

drug sellers. The intense market competition together with the recruitment

and rewarding of particularly aggressive youths created a need for sellers

(as well as nonseller youths in market neighborhoods) to be armed. This

increased demand fueled an expansion in the illicit gun market and a

diffusion of guns. The linkage between drug selling and gun possession is

well established (see Decker, Pennell, and Caldwell 1997; Sheley, 1994;

Tardiff et al., 1994).[17] Cork (1999) found support for the temporal

sequence of the Blumstein account using a sophisticated diffusion

modeling analysis of time-series data from multiple cities.



The Blumstein model is a compelling account of the rise of violent crime,

but more work is needed to establish its explanatory power as an account

of the subsequent decline in violence. The model is not inconsistent with

that decline--a decline in the crack market should have reduced the need to

be armed--but future research will have to assess whether declines in the

prevalence of drug selling (as opposed to changes in other features of the

markets) have produced reductions in the likelihood of gun possession and

gun violence. 



The maturation of addicts and of illicit drug markets. Because of reduced

initiation rates, it appears that the hardcore cocaine-using population

consists mostly of an aging cohort who started using in the late 1980s, in

much the same way that heroin addicts disproportionately belong to

cohorts who initiated use in the 1970s. If this is correct, drug-related

criminality should continue to decline, absent new waves of initiation, as

addicts "mature out" of violent crime or die from drug-related illnesses or

natural causes.



Many observers were struck by the violence of 1980s crack markets

relative to earlier heroin and marijuana markets. Many have speculated

that such markets "mature" over time as (a) dealer territories are firmly

established, (b) casual users drop out of the market, and (c) hardcore users

establish reliable dealer connections. All these factors suggest a shift from

open-air public markets toward more clandestine arrangements that seem

less prone to violence.[18] But at present, this is largely speculative; there

is anecdotal and ethnographic evidence for such changes but little

systematic longitudinal research that establishes a clear trajectory over

time.



The Consequences of Prohibition and Its Enforcement



Drug Involvement As Crime



The convention in articles on drugs-crime linkages is to state that for the

purposes of the essay, the fact that drug use (and sometimes drug selling)

per se is a crime is not relevant to the analysis. But the illicit status of

street drugs is vitally important to the analysis in several ways. First, drug

prohibition is arguably necessary for Goldstein's category of systemic

(market-related) violence (MacCoun and Reuter, 2001).[19] We simply do

not observe routine violence among alcohol or tobacco vendors. Second,

Goldstein's economic-compulsive violence, although not caused by

prohibition, is surely exacerbated by it because drug prohibition almost

certainly raises the price of heroin or cocaine far above what would be

their retail market prices (MacCoun and Reuter, 2001). Finally, there are

reasons to believe that the illicit status of drugs might have subtle

criminogenic effects through several different mechanisms, including

forbidden fruit effects, labeling or stigmatization effects, and "stigma

swamping."[20] Here we highlight two such mechanisms.



Incapacitation and replacement effects. Several authors (e.g., Blumstein,

2000b; Freeman, 1996; Kleiman, 1997) have suggested that the

incarceration of drug sellers is likely to produce a weaker incapacitation

effect than would occur for other offense categories, such as property and

sex offenses. Indeed, some have speculated that a replacement process

might even produce a net increase in the prevalence of drug selling. In a

highly competitive illicit market, the incarceration of a drug seller creates

lucrative drug-selling opportunities (customers and sales territory) for

others. According to Blumstein (2000b):



The pathological rapist's crimes almost certainly are not replaced on the

street, and so one can expect his full array of crimes to be 

incapacitated. . . . A burglar's crimes may be replaced if he is serving a

fence, who would recruit a replacement; alternatively, if he is simply

operating on his own, the crimes are not likely to be replaced. And the

participant in organized vice activity such as drug dealing would be likely

to have his transactions replaced by whatever organizational structure is

serving the market demand. That replacement could be achieved by some

combination of recruiting new sellers or by increasing the rate of activity

of sellers already active in the market.



Freeman (1996) offers a formal economic model that interprets this

replacement effect in terms of the elasticity of supply of dealers with

respect to drug market wages. The supply of dealers should reflect this

sensitivity to wages as well as changes in earnings opportunities in the licit

market (i.e., shift in the supply curve) and the demand for drugs (i.e., shifts

in the demand curve). 



At present, there is surprisingly little evidence either for or against the

replacement hypothesis. One indirect argument for its plausibility is that

the explosive growth in the incarceration of drug sellers during the past

decade was not accompanied by increases in street cocaine prices, as one

might expect if the supply of street dealers was tightening (Blumstein,

2000b; see also DiNardo, 1993). Indeed, street prices have dropped

substantially (Rhodes et al., 2000). Another indirect argument is the sheer

prevalence of drug market participation in some communities during the

late 1980s, when drug sellers were being incarcerated at record levels. For

example, Saner et al. (1995) estimated that in Washington, D.C., during

1985-91, nearly one-third of African-American male residents from the

1964-67 birth cohorts were charged with drug selling.



Statistical analyses of archival data might test the replacement hypothesis

by looking for evidence of increases in the initiation to drug selling as a

function of the arrest and incarceration of dealers. Ethnographic studies

might examine whether recruitment activities increase following police

crackdowns and whether existing street dealers increase their activity. But

isolating replacement effects will be tricky; note that general deterrence

and replacement effects, if they exist, will offset each other, which may

make it hard to find any effect of sanctions on subsequent dealing.



Can enforcement amplify violence? Several authors(Eck and McGuire,

2000; MacCoun and Reuter, 2001, chapter 6; Reuter, 1989; Riley, 1998)

have argued that under certain conditions, aggressive drug enforcement

might actually increase drug-related violence. Rasmussen, Benson, and

their associates have examined whether more intense drug enforcement

increases violent crime; much of this work is summarized in Rasmussen

and Benson (1994). The mechanisms involved are quite varied. For

example, enforcement might lead to more violence in competition. Benson

and colleagues (1992) found that the violent crime rate in a community

increased with more drug arrests in a neighboring community. This, they

argue, is a displacement effect; dealers move from the targeted community

to the neighboring one and struggle over the establishment of territories.

Another mechanism works through the limited capacity of the correctional

system; increased prison space for drug offenders reduces the penalties for

other crimes, including violent crimes, and thus induces higher

victimization. Benson and Rasmussen (1991) argue that, even assuming

that prison is effective only through incapacitation and not deterrence, the

observed rise in the resources devoted to drug enforcement in Florida in

the 1980s might have increased other crime by 10 percent. 



Supply reduction versus violence reduction. An important dialogue with

respect to drug users involves the prospects and tensions of integrating use

reduction strategies with harm reduction strategies (MacCoun, 1998;

MacCoun and Reuter, 2001). We see an analogous issue with respect to

the policing of drug markets (MacCoun and Reuter, 1994). Police tactics

designed to reduce the supply of drugs (and of drug suppliers) may or may

not be the most effective means of reducing the total social harm caused

by street drug selling. Some tactics might directly reduce drug-related

violence.



One example involves efforts to drive dealers indoors (see Kennedy's 1993

analysis of Tampa's QUAD program). Of course, crack houses are not

without their harms. In an ethnographic study of the crack market in

Detroit, Mieczkowski (1990, p. 90) concludes that "tavern-style crack

houses may encourage and make possible hypersexuality among

participants and thus increase the STD and HIV rates. The use of barter as

a supplement to a cash economy in the crack trade represents further

complications in creating social policies in reaction to this behavior." Still,

indoor markets are likely to be less violent. But the effects are multiple

and hard to balance. On one hand, indoor markets are less susceptible to

police surveillance or sting operations. On the other hand, driving dealers

indoors might increase users' search costs (Moore, 1990) and thus reduce

demand. Consumers in export markets would bear a disproportionate share

of these search costs because the locals often know the local dealers and

could easily locate them. This might lead to new local markets in the areas

from which the export consumers are coming and the associated

neighborhood violence that Benson et al. (1992) examined. Further

research on these issues is needed.



Heroin maintenance. If the drugs-crime link is mediated by the high price

and conditions of sale of a drug, and if a relatively small number of

frequent users are responsible for much of the crime, then perhaps

allowing access to that drug legally for those least able to quit might

reduce associated crime. There is increasing information and interest in

exploring just this possibility for heroin (see MacCoun and Reuter, 2001).



In January 1994, Swiss authorities opened a number of

government-administered heroin maintenance clinics.[21] Registered

addicts can inject heroin at a government clinic under the care of a nurse

up to three times a day, 7 days a week. Patients have to be over 18, have

injected heroin for 2 years, and have failed at least two treatment episodes.

By the end of the initial research trials of this program, more than 800

patients had received heroin on a regular basis without any leakage into the

illicit market. No overdoses were reported among participants while they

stayed in the program. A large majority of participants had maintained the

regime of daily attendance at the clinic; 69 percent were in treatment 18

months after admission. This was a high rate relative to those found in

methadone programs. About half of the "dropouts" switched to other forms

of treatment; some chose methadone and others chose abstinence-based

therapies. The crime rate among all patients dropped during the course of

treatment, use of nonprescribed heroin dipped sharply, and unemployment

fell from 44 to 20 percent.



Critics, such as an independent review panel of the World Health

Organization, reasonably asked whether the claimed success was a result

of the heroin or the many additional services provided to trial participants.

And the evaluation relied primarily on the patients' own reports, with few

objective measures. Nevertheless, despite the methodological weaknesses,

the results of the Swiss trials provide evidence of the feasibility and

potential effectiveness of this approach. In late 1997, the Swiss

government approved a large-scale expansion of the program. A similar

program is under development in the Netherlands and in Hamburg,

Germany. 



The proposal to study heroin maintenance on a trial basis in the United

States is politically controversial and would be logistically difficult.

Moreover, the normative and moral issues are clearly complex (MacCoun

and Reuter, 2001, chapter 15). But we should not reflexively dismiss,

without serious analysis, an intervention that could in theory (and with

some fragmentary evidence) help reduce the criminality of existing heroin

users and perhaps shrink the heroin street market, thereby creating new

barriers to heroin initiation. If nothing else, serious discussion of such a

program, and perhaps even formal modeling of alternative hypotheses

about its likely effects, might significantly advance our thinking about

drug market dynamics and the possibilities for effective intervention.



Summing Up: Directions for Future Research



Here we summarize our suggestions for profitable future research, in the

order in which we discussed them:



o Methodological attention to the measurement of Goldstein's taxonomy

of drugs-violence links and to the validation of self-reports of victim and

offender causal attributions for the role of drugs in criminal offenses.



o Greater attention to the role of drug use in criminal victimization.



o Retrospective historical analysis of long-term trends in drug use, drug

arrests, and drug-related crime, including recoding of ethnographic

databases, application of the Goldstein coding scheme to homicide case

files, age/period/cohort analyses, and econometric time-series analyses.



o Determination of the causal relationships underlying comorbid drug

abuse and mental illness conditions.



o Extension and replication of the rich experimental literature on

situational moderators of alcohol-related aggression, as applied to other

drugs.



o Econometric analysis of the effects of drug price changes on

drug-related criminality.



o Assessment of the effects of the availability of licit work and licit wage

levels on criminality.



o Additional multicity analyses (and cross-neighborhood analyses within

cities) with an emphasis on understanding heterogeneity in drugs-crime

relationships: Spatial analyses, analyses of variation in the demand for

different drugs, gang versus nongang involvement, ethnic and other

demographic groupings, indoor versus outdoor markets, import versus

export versus local versus public markets, etc.



o Estimation of incapacitation versus replacement effects resulting from

the incarceration of drug sellers.



o Simulation modeling and eventual pilot tests of the efficacy of heroin 

maintenance.



One other topic that was not even hinted at in our analysis and has been

almost totally neglected in the empirical research literature also should

receive attention: the likelihood of causal linkages between illicit drug use

and such white-collar crimes as corruption, fraud, and embezzlement.[22]



This is a long list of topics. That in itself is a reminder of how little has

been done to implement and build on Goldstein's insightful taxonomy.

Advances will require an acceptance of the fact that drugs may differ

widely in the extent and form of their criminogenic effects. That

substantially complicates an already difficult enterprise but is likely to be

the source of considerable policy insight.



Notes



1. Except where noted, these statistics were reported in Drug-Related

Crime (Office of National Drug Control Policy, 2000). 



2. In fact, Goldstein et al.'s (1989) findings might not fully represent New

York City since they did not look at the entire population or a random

sample of homicides. Rather, they chose one zone in each of four different

boroughs, with the goal of sampling precincts that represented a

cross-section of New York City.



3. These findings challenge the recent generalization by White and

Gorman (2000, p. 189) that "the economic motivation explanation has not

been supported among adolescents."



4. Our understanding is that the new NIBRS (National Incident-Based

Reporting System) database perpetuates this. Officers only have to

report the circumstances of the offense (which includes drug dealing) for

aggravated assaults/homicide (considered one category in the victim-level

file).



5. Approaches might include confirmatory factor analysis, cluster analysis,

Q-sort, or Ragin's (2000) fuzzy-set approach. We are less interested in

defending a particular method than in pointing out the surprising lack of

attention to these measurement and conceptualization issues in the field.



6. Fagan (1990, p. 255) and White and Gorman (2000, p. 185) argue that,

if anything, marijuana and opiates serve to suppress aggression. Actually,

Bushman's (1990) meta-analysis found more aggression among marijuana

smokers than placebo controls in laboratory experiments. But this effect is

partly due to the fact that the placebo controls showed significantly less

aggression than nondrug controls, indicating that participants also believed

marijuana would induce passivity. 



7. Note that other psychiatric disorders are less common among substance

abusers than substance abuse is among the mentally ill (Miller, 1993).



8. Beau Kilmer's work on this section was supported by NIDA grant

R01DA12724.



9. The assumption that decriminalization (as opposed to legalization) is an

indicator of lower price is questionable. In theory, it might increase

demand by reducing the nonmoney costs, which should increase price.

However, evaluations of decriminalization in 11 U.S. States, South

Australia, the Australian Capital Territory, and the Netherlands fail to

show any effects on demand (MacCoun and Reuter, 2001).



10. The authors report the statistically significant variables, not the entire

model. The entire model is listed in Fisher, Cullen, and Turner

(forthcoming) and includes a variable for "Frequency of smoking pot or

hashish." Because the significant predictors for stalking are the same in the

published and unpublished pieces, we assume the same model was used.

Because this is likely to be the model used to predict sexual victimization

in the published piece, we report that marijuana use does not predict

sexual victimization. 



11. Even if true, high returns from crack selling do not lessen the

criminogenic consequences of the market; the issue is what share of

revenues are generated by legitimate earnings or welfare and other transfer

payments received by buyers. 



12. The question yields four binary variables about whether the arrestee

was in need of drugs/alcohol (NEEDNO), alcohol (NEEDALC), cocaine

(NEEDCOCR), and marijuana (NEEDMAR) during the crime and one

text variable (NEEDOTHR) where the coder is asked to specify if the

arrestee mentioned another drug. Curiously, the 1995 (part 2) and 1999

ADAM codebooks do not report any binary variable for heroin--widely

believed to be the major source of economic-compulsive crime. Of the

44,000 ADAM arrestees in 1999, we estimate (using the open-ended field

responses) that about 1,100 reported they needed heroin, 1,800 needed

alcohol, 2,150 needed cocaine/crack, and 700 needed marijuana. Of those

reporting that they needed heroin, about 35 percent committed

income-generating crimes.



13. ONDCP reports, based on Rhodes et al. (1995, 2000), that the

prevalence of frequent use fell by one-third between 1988 and 1993 and

then returned to its 1988 level by 1998. It is difficult to identify supporting

evidence for such a dramatic fluctuation in the figures.



14. The bookmaking business has certainly generated written records; but

that is more central to the business itself, which involves the extension of

credit and usually numerous near-simultaneous transactions between any

one buyer and seller.



15. In the District of Columbia in the mid- to late 1990s, it was reported

that some street gangs were in violent disputes over the marijuana market

(Pierre, 1996; Lattimore et al., 1997).



16. Smith and Varese (2001) model the use of coercive violence in

markets for Mafia extortion; the model can be applied to

intraorganizational violence as well.



17. Decker, Pennell, and Caldwell (1997) did not find that drug users

(rather than sellers) were more likely to be carrying a gun than other

arrestees.



18. Alfred Blumstein appeared to endorse this account in his public

comments at the 2000 Annual Meeting of the American Society of

Criminology.



19. Necessary, but not sufficient; see Zimring and Hawkins, 1997; Ousey

and Lee, 2000.



20. The term "stigma swamping" was suggested to us by Jon Caulkins as

an apt label for a phenomenon about which many have speculated (e.g.,

Jacobsen and Hanneman, 1992; McGraw, 1985; Petersilia, 1990)--the

notion that the stigma associated with arrest and even incarceration is

reduced by the sheer prevalence of those sanctions. The term "stigma

swamping" is an informal control counterpart to Kleiman's (1993) formal

control version, "enforcement swamping."



21. The earlier British experience with prescription heroin is more

notorious but less informative; see MacCoun and Reuter, 2001, chapter 12.



22. We thank Terence Dunworth for making this observation.



References



Anderson, E. (1994). Code of the streets: How the inner city environment

fosters a need for respect and a self-image based on violence. Atlantic

Monthly, 273, 80-94.



Arrestee Drug Abuse Monitoring Program (1999). 1998 annual report on

drug use among adult and juvenile arrestees (Research Report, NCJ

175656). Washington, DC: U.S. Department of Justice, National Institute

of Justice.



Baumeister, R.F., Smart, L., and Boden, J.M. (1996). Relation of

threatened egotism to violence and aggression: The dark side of high

self-esteem. Psychological Review, 103, 5-33.



Baumer, E., Lauritsen, J. L., Rosenfeld, R., and Wright, R. (1998). The

influence of crack cocaine on robbery, burglary, and homicide rates: A

cross-city longitudinal analysis. Journal of Research on Crime and

Delinquency, 35, 316-340.



Benson, B.L., Kim, I., Rasmussen, D.W., and Zuehlke, T.W. (1992). Is

property crime caused by drug use or by drug enforcement policy? Applied

Economics, 24, 679-692.



Benson, B.L., and Rasmussen, D.W. (1991). The relationship between

illicit drug enforcement and property crimes. Contemporary Policy Issues,

9, 106-115.



Block, A. (1980) East side-west side. Cardiff: University of Wales Press.



Block, C., Block, R., and Illinois Criminal Justice Information Authority.

(1998). Homicides in Chicago, 1965-1995. Ann Arbor: Intra-university

Consortium for Political and Science Research.



Blumstein, A. (2000a). Disaggregating the violence trends. In A.

Blumstein and J. Wallman (eds.), The crime drop in America (pp. 13-44).

New York: Cambridge University Press.



Blumstein, A. (2000b). The replacement of drug offenders to diminish the

effects of incarceration. Unpublished manuscript, Carnegie Mellon

University.



Blumstein, A., and Cork, D. (1996). Linking gun availability to youth gun

violence. Law and Contemporary Problems, 59, 5-24.



Blumstein, A., and Wallman, J. (eds.) (2000). The crime drop in America.

New York: Cambridge University Press.



Bourgois, P. (1996). In search of masculinity: Violence, respect and

sexuality among Puerto Rican crack dealers in East Harlem. British

Journal of Criminology, 36, 412-427.



Brownstein, H., Baxi, H., Goldstein, P., and Ryan, P. (1992). The

relationship of drugs, drug trafficking, and drug traffickers to homicide.

Journal of Crime and Justice, 15, 25-44.



Bureau of Justice Statistics (1997a). Survey of inmates in Federal

correctional facilities. Washington, DC: U.S. Department of Justice,

Bureau of Justice Statistics and Federal Bureau of Prisons. 



Bureau of Justice Statistics (1997b). Survey of inmates in State

correctional facilities. Washington, DC: U.S. Department of Justice,

Bureau of Justice Statistics.



Bureau of Justice Statistics (1991a). Survey of inmates in Federal

correctional facilities. Washington, DC: U.S. Department of Justice,

Bureau of Justice Statistics and Federal Bureau of Prisons.



Bureau of Justice Statistics (1991b). Survey of inmates in State

correctional facilities. Washington, DC: U.S. Department of Justice,

Bureau of Justice Statistics. 



Bushman, B.J. (1990). Human aggression while under the influence of

alcohol and other drugs: An integrative research review. Psychological

Science, 2, 148-152.



Bushman, B.J. (1996). Individual differences in the extent and

development of aggressive cognitive-associative networks. Personality and

Social Psychology Bulletin, 22, 811-819.



Bushman, B.J. (1997). Effects of alcohol on human aggression: Validity of

proposed explanations. In M. Galanter (ed.), Recent developments in

alcoholism, Vol. 13: Alcohol and violence: Epidemiology, neurobiology,

psychology, family issues (pp. 227-243). New York: Plenum Press.



Caulkins, J., and Reuter, P. (1996). Editorial: The meaning and utility of

drug prices. Addiction, 91, 1261-1264.



Caulkins, J., Rydell, C., Schwabe, W., and Chiesa, J. (1997). Mandatory

minimum drug sentences: Throwing away the key or the taxpayers'

money? Santa Monica: RAND Drug Policy Research Center.



Chin, K.L., and Fagan, J. (1992). The impact of crack on criminal careers:

Crime and drug involvement following initiation into cocaine smoking.

Unpublished manuscript on file with Rutgers University School of

Criminal Justice. 



Cialdini, R.B., Kallgren, C.A., and Reno, R.R. (1991). A focus theory of

normative conduct: A theoretical refinement and reevaluation of the role

of norms in human behavior. In M. Zanna (ed.), Advances in experimental

social psychology (vol. 24, pp. 201-234). New York: Academic Press.



Cohen, D., Nisbett, R.E., Bowdle, B.F., and Schwarz, N. (1996). Insult,

aggression, and the Southern culture of honor: An "experimental

ethnography." Journal of Personality and Social Psychology, 70, 945-960.



Compton, W.M., Cottler, L.B., Phelps, D.L., Abdallah, A.B., and

Spitznagel, E.L. (2000). Psychiatric disorders among drug dependent

subjects: Are they primary or secondary? American Journal on Addictions,

9, 126-134.



Cork, D. (1999). Examining space-time interaction in city-level homicide

data: Crack markets and the diffusion of guns among youth. Journal of

Quantitative Criminology, 15, 379-406.



Cottler, L.B., Compton, W.M., Mager, D., Spitznagel, E., and Janca, A.

(1992). Posttraumatic stress disorder among substance users from the

general population. American Journal of Psychiatry, 149, 664-670. 



Critchlow, B. (1986). The powers of John Barleycorn: Beliefs about the

effects of alcohol on social behavior. American Psychologist, 41, 751-764.



Decker, S.H., Pennell, S., and Caldwell, A. (1997). Illegal firearms:

Access and use by arrestees (Research in Brief, NCJ 163496).

Washington, DC: U.S. Department of Justice, National Institute of Justice.



DiNardo, H. (1993). Law enforcement, the price of cocaine, and cocaine

use. Mathematical and Computer Modeling, 17, 53-64.



Eagly, A.H., and Chaiken, S. (1993). The psychology of attitudes. Fort

Worth, TX: Harcourt Brace Jovanovich.



Eck, J., and McGuire, E. (2000). Have changes in policing reduced violent

crime? An assessment of the evidence. In A. Blumstein and J. Wallman

(eds.), The crime drop in America (pp. 207-265). New York: Cambridge

University Press.



Fagan, J. (1990). Intoxication and aggression. In M. Tonry and J.Q.

Wilson (eds.), Drugs and crime, vol. 13 of Crime and justice: A review of

research (pp. 241-320). Chicago: University of Chicago Press.



Fagan, J., and Chin, K. (1990). Violence as regulation and social control in

the distribution of crack. In M. De La Rosa, E. Lambert, and B. Gropper

(eds.), Drugs and violence: Causes, correlates, and consequences

(Research Monograph 103, pp. 8-43). Rockville, MD: U.S. Department of

Health and Human Services, National Institute on Drug Abuse.



Fishbein, D. (2000). Neuropsychological function, drug abuse, and

violence: A conceptual framework. Criminal Justice & Behavior, 27,

139-159.



Fisher, B., Cullen, F., and Turner, M. (2000). The sexual victimization of

college women (Research Report, NCJ 182369). Washington, DC: U.S.

Department of Justice, National Institute of Justice.



Fisher, B., Cullen, F. and Turner, M. (forthcoming). Being pursued:

Stalking victimization in a national study of college women. Criminology

and Public Policy. 



Fisher, B., Sloan, J., and Chunmeng, L. (1998). Crime in the ivory tower:

The level and sources of student victimization. Criminology, 36, 671-710.



Freeman, R.B. (1996). Why do so many young American men commit

crimes and what might we do about it? Journal of Economic Perspectives,

10, 25-42.



Gandossy, R., Williams, J., Cohen, J., and Harwood, H. (1980). Drugs and

crime: A survey and analysis of the literature (NCJ 159074). Washington,

DC: U.S. Department of Justice, National Institute of Justice.



Giancola, P.R. (2000). Executive functioning: A conceptual framework for

alcohol-related aggression. Experimental & Clinical Psychopharmacology,

8, 576-597.



Goldstein, P. (1985). The drug/violence nexus: A tripartite conceptual

framework. Journal of Drug Issues, 14, 493-506.



Goldstein, P., Brownstein, H.H., and Ryan, P.J.(1992). Drug-related

homicide in New York: 1984 and 1988. Crime & Delinquency, 38,

459-476.



Goldstein, P., Brownstein, H.H., Ryan, P.J., and Bellucci, P.A. (1989).

Crack and homicide in New York City, 1988: A conceptually based event

analysis. Contemporary Drug Problems, 16, 651-687.



Golub, A., and Johnson, B.D. (1997). Crack's decline: Some surprises

across U.S. cities (Research in Brief, NCJ 165707). Washington, DC:

U.S. Department of Justice, National Institute of Justice.



Harwood, H., Fountain, D., and Livermore, G. (1998). The economic costs

of alcohol and drug abuse in the United States--1992. Bethesda, MD: U.S.

Department of Health and Human Services, National Institute on Drug

Abuse.



Hindmarch, I., and Brinkmann, R. (1999). Trends in the use of alcohol and

other drugs in cases of sexual assault. Human Psychopharmacology:

Clinical and Experimental, 14, 225-231.



Inciardi, J.A. (1990). The crack-violence connection within a population

of hard core adolescent offenders. In M. De La Rosa, E. Lambert, and B.

Gropper (eds.), Drugs and violence: Causes, correlates, and consequences

(Research Monograph 103, pp. 92-111). Rockville, MD: U.S. Department

of Health and Human Services, National Institute on Drug Abuse.



Inciardi, J.A., and Pottieger, A.E. (1994). Crack-cocaine use and street

crime. Journal of Drug Issues, 24, 273-292. 



Inciardi, J.A., and Pottieger, A.E. (1991). Kids, crack, and crime. Journal

of Drug Issues, 21, 257-270.



Ito, T.A., Miller, N., and Pollock, V.E. (1996). Alcohol and aggression: A

meta-analysis on the moderating effects of inhibitory cues, triggering

events, and self-focused attention. Psychological Bulletin, 120, 60-82. 



Jacobsen, C., and Hanneman, R.A. (1992). Illegal drugs: Past, present and

possible futures. Journal of Drug Issues, 22, 105-120.



Johnson, B., Golub, A., and Dunlap, E. (2000). The rise and decline of

hard drugs, drug markets, and violence in inner-city New York. In A.

Blumstein and J. Wall-man (eds.), The crime drop in America (pp. 

164-206). New York: Cambridge University Press.



Kennedy, D. (1993, March). Closing the market: Controlling the drug

trade in Tampa, Florida (Program Focus, NCJ 139963). Washington, DC:

U.S. Department of Justice, National Institute of Justice.



Kerr, N.L., MacCoun, R.J., Hansen, C.H., and Hymes, J.A. (1987).

Gaining and losing social support: Momentum in decision making groups.

Journal of Experimental Social Psychology, 23, 119-145.



Kessler, R.C., Nelson, C.B., McGonagle, K.A., Edlund, M.J., Frank, R.G.,

and Leaf, P.J. (1996). The epidemiology of co-occurring addictive and

mental disorders: Implications for prevention and service utilization.

American Journal of Orthopsychiatry, 66, 17-31.



Kleiman, M.A.R. (1993). Enforcement swamping: A positive-feedback

mechanism in rates of illicit activity. Mathematical and Computer

Modeling, 17, 65-75.



Kleiman, M.A.R. (1997). The problem of replacement and the logic of

drug law enforcement. Drug Policy Analysis Bulletin, 3, 8-10 [Online].

Available: www.fas.org/ drugs/issue3.htm.



Klein, M., Maxson, C., and Cunningham, L. (1991). "'Crack,' street gangs,

and violence. Criminology, 29, 623-650. 



Lattimore, P.K., Trudeau, J., Riley, K.J., Leiter, J. and Edwards, S. (1997).

Homicide in eight U.S. cities: Trends, context, and policy implications

(Research Report, NCJ 167262). Washington, DC: U.S. Department of

Justice, National Institute of Justice.



Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science,

278, 45-47.



Lipton, D., and Johnson, D. (1998). Smack, crack, and score: Two decades

of NIDA-funded drugs and crime research at NDRI 1974-1994. Substance

Use & Misuse, 33, 1779-1815.



MacCoun, R. (1998). Toward a psychology of harm reduction. American

Psychologist, 53, 1199-1208.



MacCoun, R.J., and Reuter, P. (1994, March 7). Harm reduction as a

response to drug selling and violence. Paper presented at the Fifth

International Conference on the Reduction of Drug-Related Harm,

Toronto.



MacCoun, R.J., and Reuter, P. (2001). Drug war heresies: Learning from

other vices, times, and places. New York: Cambridge University Press.



Markowitz, S. (2000). An economic analysis of alcohol, drugs, and crime

in the National Criminal Victimization Survey. National Bureau of

Economic Research Working Paper 7982. Cambridge, MA: National

Bureau of Economic Research.



McElrath, K., Chitwood, D., and Comerford, M. (1997). Crime

victimization among injection drug users. Journal of Drug Issues, 27,

771-783.



McGraw, K.M. (1985). Subjective probabilities and moral judgments.

Journal of Experimental Social Psychology, 21, 501-518.



Meehl, P.E. (1995). Bootstraps taxometrics: Solving the classification

problem in psychopathology. American Psychologist, 50, 266-275.



Mieczkowski, T. (1990). The operational styles of crack houses in Detroit.

In M. De La Rosa, E. Lambert, and B. Gropper (eds.), Drugs and violence:

Causes, correlates, and consequences (pp. 60-91).



Rockville, MD: U.S. Department of Health and Human Services, National

Institute on Drug Abuse. 



Miller, N.S. (1993). Comorbidity of psychiatric and alcoholic/drug

disorders: Interactions and independent status. Journal of Addictive

Diseases, 12, 5-16.



Moore, M.H. (1990). Supply reduction and drug law enforcement. In M.

Tonry and J.Q. Wilson (eds.), Drugs and crime, vol. 13 of Crime and

justice: A review of research (pp. 109-157). Chicago: University of

Chicago Press.



Mueser, K.T., Yarnold, P.R., Rosenberg, S.D., Swett, C., Miles, K.M., and

Hill, D. (2000). Substance use disorder in hospitalized severely mentally

ill psychiatric patients: Prevalence, correlates, and subgroups.

Schizophrenia Bulletin, 26, 179-192.



National Crime Victimization Survey (2000). Criminal victimization in

the United States, 1998 statistical tables (NCJ 181585). Washington, DC:

U.S. Department of Justice, Bureau of Justice Statistics.



Nisbett, R., and Ross, L. (1980). Human inference: Strategies and

shortcomings of social judgment. Englewood Cliffs, NJ: Prentice-Hall.



Office of National Drug Control Policy. (1995). National drug control

strategy. Washington, DC: The White House.



Office of National Drug Control Policy. (2000, March). Drug-related

crime: ONDCP drug policy information Fact Sheet (NCJ 181056).

Washington, DC: The White House.



Ousey, G.C., and Lee, M.R. (2000). Examining the conditional nature of

the illicit drug market-homicide relationship: A partial test of the theory of

contingent causation. Paper presented at the 2000 meeting of the American

Society of Criminology, San Francisco.



Parker, R., and Auerhahn, K. (1998). Alcohol, drugs, and violence. Annual

Review of Sociology, 24, 291-311. 



Petersilia, J. (1990). When probation becomes more dreaded than prison.

Federal Probation, 54, 23-27.



Pierre, R.E. (1996, September 24). Marijuana's violent side. Washington

Post, A1. 



Ragin, C.C. (2000). Fuzzy-set social science. Chicago: University of

Chicago Press.



Rasmussen, D.W., and Benson, B.L. (1994). The economic anatomy of a

drug war: Criminal justice in the commons. Lanham, MD: Rowman and

Littlefield.



Rettig, R.A., and Yarmolinsky, A. (eds.) (1995). Federal regulation of

methadone treatment. Committee on Federal Regulation of Methadone

Treatment, Institute of Medicine, National Academy of Sciences.

Washington, DC: National Academy Press.



Reuter, P. (1989, March 26). An economist looks at the carnage.

Washington Post.



Reuter, P., and MacCoun, R. (1992). Street drug markets in inner-city

neighborhoods: Matching policy to reality. In J.B. Steinberg, D.W. Lyon,

and M.E. Vaiana (eds.), Urban America: Policy choices for Los Angeles

and the Nation (pp. 227-251). Santa Monica, CA: RAND.



Reuter, P., MacCoun, R., and Murphy, P. (1990). Money from crime.

Santa Monica, CA: RAND.



Rhodes, W., Scheiman, P., Pittayathikhun, T., Collins, L., and Tsarfaty, V.

(1995). What America's users spend on illegal drugs, 1988-1993. 

Washington, DC: Office of National Drug Control Policy.



Rhodes, W., Layne, M., Johnston, P., and Hozik, L. (2000). What

America's users spend on illegal drugs: 1988-1998. Washington, DC:

Office of National Drug Control Policy.



Riley, K.J. (1998). Homicide and drugs: A tale of six cities. Homicide

Studies, 2, 176-205.



Saner, H., MacCoun, R., and Reuter, P. (1995). On the ubiquity of drug

selling among youthful offenders in Washington, DC, 1985-1991: Age,

period, or cohort effect? Journal of Quantitative Criminology, 11,

337-362.



Science (1997, October 3). Frontiers in Neuroscience: The science of

substance abuse. 278, special section. 



Sheley, J.F. (1994). Drug activity and firearms possession and use by

juveniles. Journal of Drug Issues, 24, 363-382. 



Smith, A., and Varese, F. (2001, August 1). Payment, protection and

punishment: The role of information and reputation in the Mafia.

Rationality and Society, 13, 349-393.



Spunt, B., Goldstein, P., Bellucci, P., and Miller, T. (1990). Race/ethnicity

and gender differences in the drugs-violence relationship. Journal of

Psychoactive Drugs, 22, 293-303.



Spunt, B., Brownstein, H., Goldstein, P., Fendrich, M., and Liberty, H.

(1995). Drug use by homicide offenders. Journal of Psychoactive Drugs,

27, 125-134.



Stacy, A.W., Widaman, K.F., and Marlatt, G.A. (1990). Expectancy

models of alcohol use. Journal of Personality and Social Psychology, 58,

918-928.



Steadman, H.J., Mulvey, E.P., Monahan, J., Robbins, P.C., Appelbaum,

P.S., Grisso, T., Roth, L.H., and Silver, E. (1998). Violence by people

discharged from acute psychiatric inpatient facilities and by others in the

same neighborhoods. Archives of General Psychiatry, 55, 393-401.



Steele, C.M., and Josephs, R.A. (1990). Alcohol myopia: Its prized and

dangerous effects. American Psychologist, 45, 921-933.



Substance Abuse and Mental Health Services Administration. (2001).

Mid-year 2000 preliminary emergency department data from the Drug

Abuse Warning Network. Rockville, MD: U.S. Department of Health and

Human Services, Substance Abuse and Mental Health Services

Administration.



Tardiff, K., Marzuk, P.M., Leon, A.C., Hirsch, C.S., Stajic, M., Portera,

L., and Hartwell, N. (1994). Homicide in New York City: Cocaine use and

firearms. JAMA: The Journal of the American Medical Association, 272,

43-46. 



Taylor, L., Caulkins, J., and Reuter, P. (2000). Some simple economics of

illegal drug markets and violence. Unpublished manuscript, Carnegie

Mellon University.



Teplin, L.A., Abram, K.M., and McClelland, G.M. (1994). Does

psychiatric disorder predict violent crime among released jail detainees? A

six-year longitudinal study. American Psychologist, 49, 335-342.



U.S. Sentencing Commission (1995, February). Cocaine and Federal

sentencing policy. Special Report to Congress. Washington, DC: U.S.

Sentencing Commission [Online]. Available: www.ussc.gov/crack/

EXEC.HTM



White, H.R., and Gorman, D.M. (2000). Dynamics of the drug-crime

relationship. In Crime and justice 2000, vol. 1: The nature of crime:

Continuity and change (NCJ 182408, pp. 151-218). Washington, DC:

U.S. Department of Justice, National Institute of Justice.



Wolfgang, M.E. (1958). Patterns in criminal homicide. Philadelphia:

University of Pennsylvania Press. 



Zimring, F.E., and Hawkins, G. (1997). Crime is not the problem: Lethal

violence in America. New York: Oxford University Press.



Appendix A. Other Applications of Goldstein's Framework



New York



Even excluding the works of Goldstein and his colleagues, much of the

work using the tripartite framework focuses on New York during the mid-

to late 1980s. The U.S. Sentencing Commission (1995) used Goldstein's

framework to compare the incidence of violence related to the use of

powder cocaine and crack. Using expert testimony and existing literature,

and largely focusing on the studies done in New York,[1] the Commission

concluded that crack was a greater source of systemic violence than

powder cocaine, that economic-compulsive violence was relatively rare

among cocaine users, and that "neither powder nor crack cocaine excite or

agitate users to commit criminal acts and that the stereotype of a 

drug-crazed addict committing heinous crimes is not true for either form

of cocaine" (p. x).



Miami



Inciardi's (1990) survey of 611 serious juvenile delinquents in Miami and

Dade County assessed offender self-reports of drug-related systemic,

economic-compulsive, and psychopharmacological crime. In the 12

months prior to the interviews, which occurred from 1985 to 1989, about 5

percent of the sample reported being a psychopharmacological victim, 59

percent reported having committed robberies ("the majority of which were

committed to purchase drugs," p. 100), and 8 percent reported being the

perpetrators of systemic crimes. Inciardi also administered a

supplementary crack survey to 254 of these delinquents from October

1986 to November 1987. This survey and other data analyses by Inciardi

led him to conclude that the Miami crack market was much less violent

and less (juvenile) gang-related than portrayed in the media and may be

"kindler and gentler" than other large cities. He also reported that the worst

years for murders in Miami were during its cocaine wars in the early

1980s. Inciardi found that "those more proximal of the crack distribution

market were more involved in violent crime" (p. 104). This study has at

least two advantages over Goldstein et al. (1989): Crimes other than

homicide were considered, and respondents were asked about drug-related

victimization. But the drug associated with these crimes was not listed as it

was in the Goldstein et al. study.



Chicago



One source that was developed to assess homicide fluctuations and

motivations is the Chicago Homicide Dataset (CHD). Detailed information

on every homicide in the records of the Chicago Police Department is

available for 1965-95 (Block, Block, and Illinois Criminal Justice

Information Authority, 1998). CHD does not include data on specific

drugs, but its motive classification fits nicely with the tripartite framework.

The four types of drug-related motives for homicide are selling or drug

business (this includes any homicides during or because of a

transaction);[2] an argument over possession, use, quality, or cost of

drugs; getting money for drugs or acquiring drugs for personal use; and

other drug involvement (e.g., baby dies of malnutrition because the parents

were high; offender was drug crazed).[3]



The per capita drug-related homicide rate remained fairly stable from 1973

to 1984 (around 0.4 homicides per 100,000 Cook County residents), with

"arguments" at a slightly higher rate from 1974 to 1977. Homicide rates

related to all of the motives fluctuated from 1984 to 1995, but it is

interesting that the aggregate rate for every motive except

"business/transaction" was virtually the same for 1984-85 and 1995 (still

close to 0.4). The advent of crack likely explains why homicide rates

related to all of the motives increased from 1985 to 1989, but it is of

special interest that the "business/transaction" motive skyrocketed during

those years. Clearly, more might be learned by examining the specific

drugs associated with "business/transaction" homicides in Chicago over

this time period. 



Eight-City Study



To learn why city homicide rates did not change uniformly in the early

1990s, Lattimore and colleagues (1997) comprehensively examined

homicide in eight cities for 1985-94: Atlanta, Detroit, Indianapolis, Miami,

New Orleans, Richmond, Tampa, and Washington, D.C. In addition to

comparing ADAM results with UCR data for these cities, Lattimore et al.

interviewed key policymakers, law enforcement and criminal justice

officials, and community leaders in the cities. These interviews revealed

that crack was most likely associated with community violence and

homicide, while the market violence associated with marijuana was a

growing concern in Washington, D.C., and Richmond.

Methamphetamines, LSD, PCP, and heroin were not associated with

homicide rates and were rarely mentioned by local authorities. It is

important to note that Lattimore et al. found that in many cases the 

perceptions about local drug trends differed substantially from drug trends

as measured by DUF/ADAM.



Lattimore and colleagues question the relationship between crack and

market violence because the crack markets were described as highly

competitive in cities in which the homicide rate was declining, increasing,

or remaining the same (1997, p. 89). But it is not clear that the same

conclusions could be drawn if disaggregated homicide rates (by

circumstance) were considered. The authors not only looked at how

competitive the market was, they also considered the stability of prices,

transactions, and participants. Their argument that links between drugs and

homicide "appear to fall mainly on the use side" (p. 92) relies on their

findings about participants: 



The general structure of participation in crack markets and the nature,

duration, and consequences of the "crack high" may account for the

relationship between the cocaine prevalence rates among arrestees and

homicide rates. Crack users reported the large number of "buys," extensive

networks of potential suppliers, and less reliance on a primary supplier,

suggesting that transactions were likely to occur in an opportunistic

manner. The high from crack lasts as little as 10 minutes; thus, when the

high wears off, the crack user may still be in the market and motivated to

buy more of the drug--and to commit a crime to obtain the money to do so.

(p. 141)



This is essentially an argument about economic-compulsive violence,

which other crack-specific studies have dismissed (see U.S. Sentencing

Commission, 1995). While this difference may be geographic (the other

studies were primarily done in New York City), it may also be the artifact

of a bivariate analysis of two datasets (UCR and ADAM) that did not

always cover the same populations. 



National Estimates



Others used nationwide data to learn more about the drugs-crime nexus.

Caulkins and colleagues (1997) used the tripartite framework to assess the

impact that mandatory minimum sentences have on cocaine consumption

and subsequent crime. Relying on estimates from Goldstein and his

colleagues (Goldstein, Brownstein, and Ryan, 1992; Spunt et al. 1990;

Spunt et al., 1995), the National Criminal Victimization Survey, inmate

surveys, and murder data for large urban counties, Caulkins et al.

determined the number of systemic, economic-compulsive, and

psychopharmacological crimes that were drug related. Their next step was

to determine how much of this crime was related to cocaine. Based on

information from Rhodes et al. (1995), the ADAM Program, the Office of

National Drug Control Policy (1995), and Goldstein (Goldstein et al., 

1989; Goldstein, Brownstein, and Ryan, 1992; Spunt et al. 1995), Caulkins

et al. (1997) suggest that cocaine accounts for about 75 percent of

drug-related economic-compulsive crime, 50 percent of illicit

psychopharmacological homicides, and 75 percent of systemic homicides. 



Notes 



1. New York: Chin and Fagan, 1992; Fagan and Chin, 1990; Goldstein et

al., 1989. Miami: Inciardi, 1990; Inciardi and Pottieger, 1991; Inciardi and

Pottieger, 1994. Los Angeles: Klein et al., 1991. Detroit: Mieczkowski,

1990.) The Commission also cited an unpublished DEA report and a

review article by Fagan (1990). The former found "that seven crack-related

homicides were 'multi-dimensional,' with systemic being one of the

dimensions," but it is not clear where these homicides occurred and what

the other dimensions were. 



2. The codebook reads: "Use code 1 when BUSINESS is the motive for

the incident (e.g., both victim and offender involved in dealing, victim

killed as a bystander of a drug business hit, victim killed because he

interfered with the business, victim killed during a drug transaction or

because of a drug transaction)."



3. Cases where there was no positive evidence or no information are not

included. Of the 23,817 homicides occurring between 1964 and 1995,

22,282 either had no information about drug motive or were not drug

related. Unfortunately, the non-drug-related homicides cannot be separated

from the no-information group.



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



THE DRUGS-CRIME WARS: PAST, PRESENT, AND FUTURE

DIRECTIONS IN THEORY, POLICY, AND PROGRAM

INTERVENTIONS



Duane C. McBride, Curtis J. VanderWaal, and Yvonne M. Terry-McElrath



About the Authors



Duane C. McBride is a professor with and chair of the Department of

Behavioral Sciences and director of the Institute for Prevention of

Addictions at Andrews University; Curtis J. VanderWaal is a professor

with the Department of Social Work and associate director of the Institute

for the Prevention of Addictions at Andrews University; Yvonne M. 

Terry-McElrath is a research associate with the Institute for Social

Research at the University of Michigan.



Introduction



The relationship between drug use and criminal behavior has generated a

substantial body of literature in peer-reviewed journals, government

publications, and the public press. The very extent of such research--as

well as the breadth of policy positions based on or ignoring such

research--argues for the importance of a review that can summarize theory,

policy, and programmatic approaches to the issue. In this paper, we do not

attempt to provide a comprehensive review of the issues or literature.

Instead, we seek to provide a sufficient review of the most pertinent

knowledge about the drugs-crime relationship to stimulate further

discussion among researchers regarding the most important research

questions that still need attention. This discussion holds great promise for

the development of new approaches to the drugs-crime relationship.

As Brownstein has argued, "those who do the research are in the best

position to interpret their findings and offer advice based on their

conclusions" (1991, p. 132). This paper approaches the above task by

focusing on the following issues: (a) documenting the existence of the

drugs-crime relationship, (b) addressing the nature and complexity of that

relationship, (c) summarizing philosophical and theoretical contributions

that may best address the relationship, (d) reviewing both State- and 

Federal-level policy approaches to breaking the relationship, including

integrated program approaches, and (e) proposing key areas for future

research.



The Existence of the Drugs-Crime Relationship



The purpose of this section is to briefly discuss what is known about the

drugs-crime relationship. This discussion will focus on the historical

policy context; the empirical nature of the relationship overall; and

specific drugs, crimes, and populations.



Which Drugs and What Crime?



Before proceeding further, we wish to clarify what we mean by "drugs"

and provide a more complete picture of what is involved in "crime" related

to drug use. These clarifications are made in the hope that readers will

recognize that the crime aspect of the drugs-crime relationship is

multifaceted and that the current exclusion of alcohol from most

discussions of the drugs-crime relationship may be detrimental.



Substance Inclusion Decisions



The term "drugs" as used throughout this paper refers to currently illicit

substances in the United States based on Federal drug schedules. Alcohol,

prescription drugs, and other substances are excluded. Although it is

beyond the scope of the current project, it is important to at least mention

the alcohol-crime relationship. Greenfeld (1998) reminds us that an

estimated 36 percent of convicted offenders were drinking at the time they

committed their crimes and that a high correlation has been observed

between public order crimes and alcohol use. Alcohol is also strongly

related to violent crime (Coker et al., 2000; Dawkins, 1997; Ernst et al.,

1997; Parker and Auerhahn, 1998; Pihl and Peterson, 1995). Ironically,

this relationship often remains outside sentencing decisions and

monitoring procedures because alcohol is legal and therefore not subject to

the same arrest, seizure, and prosecution laws as are illicit drugs. Drug

treatment interventions, however, often include both alcohol and other

drugs. Comprehensive efforts to address crime and substance use should

include alcohol treatment in programmatic considerations.



The History of Drug Policy and the Definition of Crime



Crimes associated with drug use range from violent (such as murder and

aggravated assault) to acquisitive (burglary, forgery, fraud, and deception)

to specific drug-law violations. In addition, crimes such as bribery and

corruption are related to drug use as a result of drug policy prohibitions.

Traditionally, discussions of the drugs-crime relationship have focused

primarily on violent crime; however, it is important to recognize the

complexity of criminal acts associated with drug use. When considering

the drugs-crime relationship, this paper recommends that researchers and

policymakers include both violent and nonviolent crimes as well as drug

law violations and corruption associated with drug policy to grasp more

fully the resulting harms and societal costs (for example, see French and

Martin, 1996).



Efforts to address the drugs-crime relationship must incorporate a

realization of how the development of policy and law has contributed to

the relationship itself. Policy approaches to drug use in the United States

have historically ranged between legal markets in the 19th century to

decriminalization, harm reduction, medicalization, and strict prohibition

(as the dominant policy) in the 20th. Over time, policy has moved to

various points along this continuum, and it often resides at different points

at the same time in different locations and for different substances. Each

time policy shifts, the act of drug use takes on a slightly different character

in relation to crime. Thus, it is important to present a brief history of drug

policy in the United States, together with current possible positions in the

drug policy discussion, as each position has a unique implication for

fighting drug-related crime.



An understanding of American drug policy begins with three early

American cultural traditions that still strongly affect drug policy

discussions: (a) libertarianism, (b) the emergence of a relatively open legal 

market resulting from the libertarian perspective, and (c) Puritan

moralism. Libertarianism argues that government must have an extremely

compelling motive for interfering in the personal lives of citizens. Such

interference legitimately occurs only if a citizen's behavior is a significant,

actual risk to others (Mill, 1979). Consistent with this libertarian tradition,

early America had an open-market orientation that emphasized limited

government interference in the production and distribution of desired

goods and services.[1] Nineteenth-century national drug policy was

consistent with both libertarianism and the open market. While the Federal

Government regulated the importation of such drugs as opium and

cocaine, few regulations governed the distribution of these and other drugs

through what came to be called the patent medicine industry (Belenko,

2000; Inciardi, 2001; Musto, 1999). Patent medicines were extensively

advertised and, through them, the use of drugs such as opium and cocaine

became integrated into routine American cultural behavior patterns

(Musto, 1999).



Conflicting with both libertarianism and the market-driven approach is the

Puritan moralist perspective: Individual behaviors with the potential to

harm the community are seen as a community problem within the

legitimate purview of community action (Cherrington, 1920; Schmidt,

1995). Puritan and other religious and moral traditions present in

American history often viewed behavior such as substance use as

undermining the moral fabric of society, potentially causing the

withdrawal of God's blessing from America. The Puritan moralist

perspective dominated the early 1900s, an era of societal reform and

increasing prohibition (and thus increasing penalties for drug use). One of

the first successes of the social reform movement in the early 20th century

was the passage of the Pure Food and Drug Act of 1906, which required

the patent medicine industry to list product ingredients. The subsequent

passage of the Harrison Act of 1914 and the Marihuana Tax Act of 1937

made illegal the manufacture, sale, and possession of a variety of drugs,

including opiates and cocaine, as well as the nonmedical use of marijuana.

A strongly prohibitionist approach continued through the 1950s with the

Boggs Act of 1951 and the Narcotic Control Act of 1956, when mandatory

minimum sentences for Federal drug trafficking law violations were

strengthened and arrests without a warrant for drug charges were enabled.



The 1960s and 1970s represented a major cultural shift in the United

States. For a variety of reasons, American society experienced a "drug

revolution" during this era. There appeared to be an increase in the

proportion of individuals using drugs and in the variety of drugs used. The

evidence for this increase is seen in the number of drug-related arrests and

the increase in drug use in the general population (Musto, 1999). During

this era, drug policy initially shifted to a stronger treatment- and less

punishment-oriented stance. In 1966, the Narcotic Addict Rehabilitation

Act allowed the establishment of the civil commitment system instead of

prosecution for Federal offenders and encouraged State and local

governments to develop their own treatment programs. In 1970, the

Comprehensive Drug Abuse Prevention and Control Act consolidated and

replaced the patchwork of previous Federal drug laws. The Act created the

drug schedules in current use today and initiated the so-called "war on

drugs"; it also moved some possession or casual transfer offenses to

misdemeanors instead of felonies. This era may be considered a time when

drug use was primarily considered a medical/mental health problem to be

addressed by treatment, with lessened emphasis on criminal penalties for

possession and use.



With an apparent increase in drug use, as evidenced by an increase in drug

overdose cases and drug treatment admissions, a more prohibitionist

movement again swept the Nation. New York's Rockefeller Drug Laws

were passed in 1973, establishing mandatory prison sentences of up to 20

years for the sale of any amount of heroin or cocaine. The Anti-Drug

Abuse Acts of 1986 and 1988 continued to emphasize law enforcement

(although the 1988 Act gave more attention to treatment and prevention).

In yet another policy shift, treatment (including diversion into treatment

from the criminal justice system) and prevention received increasing

attention in the 1990s. Further, some States developed policies that

effectively decriminalized marijuana possession (removing jail/prison

penalties) and initiated policies, such as needle exchange programs, that

would reduce the dangers of injecting drugs.



Although scholars often focus on the relatively rapid development of

national drug policy, it is important to remember that many States passed

legislation prohibiting patent medicine and/or alcohol sales, as well as

marijuana use, a decade or more before similar legislation was passed by

Congress (Belenko, 2000). Because of how the United States is organized,

States often have or exercise considerable discretion regarding alcohol

and drug policies (Musto, 1999).



Essentially, the history of drug policy (and debates about where drug

policy should move in the future) can be broken down into five main

approaches: prohibition, risk reduction, medicalization,

legalization/regulation, and decriminalization (for an indepth discussion,

see McBride et al., 1999; see also Goode, 1997). Prohibition emphasizes

severe penalties for use, distribution, and production. Risk reduction uses a

public health approach to reduce the risks and harms associated with illicit

drug use and emphasizes education on risks, safer use practices,

prevention, and treatment. Medicalization calls for physician treatment of

drug addicts, viewing substance abuse primarily as a medical issue.

Legalization/regulation supports increased access to drugs through

governmental regulation of these substances, with possible distribution of

specific substances through governmentally controlled distribution

channels. Decriminalization calls for a complete end to the use of criminal

law to address individual drug use. This may imply a relatively open

market approach to drug availability and use, but that need not be the case.



Although there has been significant debate over which policy approach or

approaches might best address the drugs-crime cycle, more research is

needed that examines scientifically the effects of policy positions on both

drug use and crime. For the most part, current Federal drug law takes a

prohibitionist stance that includes a strong deterrence approach to reducing

the supply of drugs and high penalties for drug law violations. As a result,

a significant portion of the drugs-crime relationship is simply an artifact of

law and policy itself: "most directly, it is a crime to use, possess,

manufacture or distribute drugs classified as having the potential for

abuse" (Craddock, Collins, and Timrots, 1994).



The Statistical Relationship Between Drug Use and Criminal Behavior



The general conclusion of almost three decades of research on the

relationship between drug use and crime has been that there is a clearly

significant statistical relationship between the two phenomena (Austin and

Lettieri, 1976; Dorsey and Zawitz, 1999; Gandossy et al., 1980; McBride

and McCoy, 1993). Research indicates extensive drug use among arrested

populations, a high level of criminal behavior among drug users, and a

fairly high correlation between drug use and delinquency/crime in the

general population. Research also indicates significant differences in the

relationship based on drug type and type of crime. Importantly, all these

differences are further complicated by ethnic and gender issues.



The Drugs-Crime Relationship Within Various Population Groups



Drug use among arrested/incarcerated populations and crime among drug

users. From the early 1970s onward, biological and self-report data have

indicated a relatively high rate of drug use among arrested and incarcerated

populations (Arrestee Drug Abuse Monitoring Program, 2000; Austin and

Lettieri, 1976; Dorsey and Zawitz, 1999; Gandossy et al., 1980; McBride

and McCoy, 1993). In 1999, the Arrestee Drug Abuse Monitoring Program

(ADAM) collected data from more than 40,000 adults in more than 30

sites and more than 400 juveniles in 9 sites throughout the United States

(ADAM, 2000). In almost all cities where the ADAM project operates,

about two-thirds of both adult male and female felony arrestees had an

illegal drug in their bodies at the time of arrest (with higher rates among

females). Even among juveniles, the majority of arrestees were found to

have an illegal drug in their urine (with higher rates among males). The

data also suggest that, although current drug use rates among adult

arrestees are higher than those reported in the more isolated reports of the

1970s (Austin and Lettieri, 1976), these rates have remained steady for the

past 5 years (the same patterns are found among juvenile arrestees). An

argument can be made that with about two-thirds of arrestees already using

illegal drugs in the 72 hours prior to their arrest, there is not much room

for an increase.



A recent report from the Bureau of Justice Statistics (BJS) suggests that

drug use also is extensive among inmates in local jails (Wilson, 2000).

This document reports that the majority of inmates in State prisons and

local jails used drugs in the month prior to the offense that put them in

prison/jail. Interestingly, this same report also notes that about 10 percent

of jail inmates test positive for drugs while in jail.



The extent of crime among drug users also has been documented. From

the 1960s through the 1990s, surveys of drug-using populations both in

and out of treatment have consistently shown that the large majority of

users have extensive histories of criminal behavior and time served in

prison (Defleur, Ball, and Snarr, 1969; Inciardi, Horowitz, and Pottieger,

1993). This pattern applies to juveniles as well: Between 40 and 57

percent of adolescents treated for substance disorders also have committed

delinquent acts (Winters, 1998).



Drug use and crime levels among the general population. A tradition of

studies shows a correlation between drug use and delinquency in general

youth populations (Elliott and Huizinga, 1984; Elliott, Huizinga, and

Menard, 1989; Harrison and Gfroerer, 1992). Analysis from the National

Youth Survey has provided data often used to examine this relationship.

These data report a direct correlation between serious drug use and

delinquency (Johnson et al., 1991). Youths who used "hard" drugs (about

5 percent of the sample) accounted for 40 percent of all delinquencies and

60 percent of index crimes.



The Impact of Drug Type on the Drugs-Crime Relationship



The first National Institute on Drug Abuse (NIDA)-sponsored Crime and

Drugs Report (Austin and Lettieri, 1976) noted that a complex relationship

exists between type of drug use and type of crime. This relationship is

further complicated if multiple drug use exists. The 1999 ADAM report

shows that a fairly large proportion of arrestees tested positive for more

than one drug (up to 30 percent), and that reported criminal behavior

tended to include a wide variety of offenses. The ADAM data show that

while cocaine was the most likely drug found among adult arrestees in

large cities (and there is literature suggesting a significant relationship

between cocaine and violence), for many urban ADAM sites, violent

offenders were more likely to test positive for marijuana than cocaine. In

addition, property offenders were more likely to test positive for cocaine

than marijuana in most sites (ADAM, 2000).



The Impact of Crime Type on the Drugs-Crime Relationship



Drug law violations. A significant proportion of drug user arrests involve

violations of drug laws only. As noted previously, the United States

experienced wide drug policy shifts in the 20th century. Each shift has

uniquely affected crimes related to drug use and distribution. In a study of

611 juvenile cocaine users by Inciardi and colleagues in the early 1990s,

analyses showed that participants had committed more than 400,000

criminal acts in the 12 months prior to being interviewed. Of these, 60

percent were for drug law violations, mostly sales of small amounts

(Inciardi, Horowitz, and Pottieger, 1993). At the Federal level, a total of

581,000 drug arrests in 1980 nearly tripled to a record high of 1,584,000

in 1997. By 1997, 79 percent of drug arrests were for possession and 21

percent were for sales. Forty-four percent of drug arrests overall were for

marijuana offenses (Uniform Crime Reports, 1998). Drug defendants

comprised 42 percent of felony convictions (Bureau of Justice Statistics,

1999). A recent BJS Special Report (Wilson, 2000) also substantiates the

extensive percentage of drug-related crimes that result from violation of

drug laws, suggesting that about a quarter of jail inmates have a current

charge or conviction for drug law violations. Critics have argued that

since such arrests likely include many low-level users and dealers,

criminal justice processing and the stiff sentences that often are handed

down because of mandatory minimums may be inappropriate to the

offense level (McBride et al., 2001).



The violence connection. Changes in drug policy are usually driven by

concerns for public safety and the perception of a direct relationship

between drugs and violence (Brownstein, 1996, 2000). For example, the

drug policy reform movement of the early 1900s (changing from legal 

markets to strict prohibition) was accompanied by horror stories focused

on exaggerated claims of criminal behavior as a consequence of drug use.

In this literature, there was a particular emphasis on horrific violent crime

(including rape), with minority group members often portrayed as the drug

users engaged in the violent behavior. Musto (1999; see also Belenko,

2000; Hickman, 2000) documents the public concern of the time (perhaps

obsession) with Chinese opiate use, African-American cocaine use, and

the use of marijuana by Mexicans. The creation of the Narcotics Bureau

led to a type of media distribution industry focused on violence associated

with drug use, "documenting" the criminal consequences of such activity

(see Anslinger and Tompkins, 1953; Inciardi, 2001). Among the best

known of these efforts were the films "The Man with the Golden Arm"

(purporting to depict the effects of heroin use/ injection) and "Reefer

Madness" (showing the supposed behavioral consequences of marijuana

use). Although such media portrayals exaggerated the possible links

between drugs and crime, some research has connected drug use with

violence. Grogger and Willis (2000) conclude that without the

introduction of crack cocaine into urban America, 1991 crime rates would

have been about 10 percent lower. These researchers also examined the

impact of crack on specific types of violent crime and reported that the

biggest impact was on aggravated assault.



In 1985, Goldstein provided the perspective that has been most commonly

used to examine the relationship between drug use and violence.

Essentially, he argued for a tripartite scheme, where

"psychopharmacological violence" could result directly or indirectly from

the biochemical behavioral consequences of drug use;

"economic-compulsive violence" could relate to behavior/crimes engaged

in to obtain money for drugs; and "systemic violence" could emerge in the

context of drug distribution, control of markets, the process of obtaining

drugs, and/or the social ecology of drug distribution/use areas.[2] Some

researchers have concluded that there is minimal evidence regarding the

psychopharmacological impact of drugs on violence (Resignato, 2000);

however, Pihl and Peterson (1995) reviewed a wide range of studies on the

issue. They concluded that alcohol and drugs can be

psychopharmacologically related to violent acts through the release of

dopamine, which reduces inhibitory anxiety about the consequences of

aggressive behavior and increases the rewards associated with violence. In

addition, they argue that the psychopharmacological effects of drugs

interfere with the user's cognitive processing of the consequences of

potentially violent situations. It should be noted that these authors believe

that the evidence for psychopharmacological effects of alcohol use on

violence are much higher than for other drugs.



However, some indications point to the environment as being a more

powerful explanation of the drugs-violence relationship than the

psychopharmacological properties of drugs (Brownstein, 2000; Fishbein,

1998; Parker and Auerhahn, 1998). In terms of economic compulsive and

systemic violence, Collins (1990) as well as Fagan and Chin (1990) argue

that crack selling is the main contributor to the drugs-violence

relationship. Specifically, their research found that violence (mostly

robbery) emerged from the need to obtain money to purchase drugs

(predominately crack). Fagan and Chin suggest that the drugs-violence

relationship also emerges as a part of the subculture of violence.



In a 1994 study, Roth argued that drug users commit more property crime

than violent crime. A recent publication by De Li, Priu, and MacKenzie

(2000) examined the relationship between drug use and property and

violent crime in a population of probationers in Virginia. Results indicated

that drug use had a positive association with property crime, whereas drug

dealing had an association with both violent and property crime (though

the relationship was stronger for property crime). The analysis also

showed an interactive effect between drug use, drug dealing, and violent

and property crime. Among juveniles, Linnever and Shoemaker (1995)

found that arrests for both possession and selling of drugs were related to

the rate of property crime arrests. However, juvenile robbery arrest rates

were related to only drug sales arrests (not possession). A National

Institute of Justice (NIJ) Research in Brief supports this research, stating

"illegal drugs and violence are linked primarily through drug marketing"

(Roth, 1994, p. 1).



The Impact of Ethnicity and Gender



Much of the research that has been conducted on drugs and crime has not

had a sufficient focus on gender and ethnic variance. This limitation has

significant repercussions on applying findings to other population groups.

As Paniagua (1998) notes, the multicultural nature of current society must

incorporate a recognition of the complex nature of ethnicity and gender.

Specifically, individuals who share a similar ethnicity or gender will not

all be the same (i.e., recognition of language, acculturation, and

socioeconomic differences); however, it is important to recognize cultural

commonalities that may significantly affect both the extent and nature of

the drugs-crime relationship across individuals. Research that has focused

on ethnicity and gender indicates that these variables may significantly

affect various aspects of the drugs-crime relationship, including:



o Source of drugs and/or works (Taylor et al., 1994).



o Predictors of violence (Ellickson and McGuigan, 2000).



o Types of violence experienced and reactions to such violence

(Brownstein et al., 1994; Fine and Weis, 1998; Mazza and Dennerstein,

1996).



o Stress-coping factors (Vaccaro and Wills, 1998).



o Biological effects of drugs (Brady and Randall, 1999).



o Epidemiology of substance-use disorders (Brady and Randall, 1999).



o Psychiatric comorbidity (Brady and Randall, 1999).



o Social stigma issues (Brady and Randall, 1999).



o Medical consequences of drug use, including heredity issues and course

of illness (Brady and Randall, 1999).



o Assessment and treatment issues, including possible prevention settings

(Brady and Randall, 1999; Metsch et al., 1999; Paniagua, 1998).



o Differences in initiation of drug use (Doherty et al., 2000).



Summary: What We Know of the Past



The intended purpose of this section has been twofold. The first goal