Toward a Drugs and Crime Research Agenda for the 21St Century U.S

U.S. Department of Justice Office of Justice Programs ET 03 SEPT. National Institute of Justice

Special REPORT

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 SEPT. 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 ...... 1

At the Intersection of Public Health and Criminal Justice Research on Drugs and Crime ...... 11

Research on Drugs-Crime Linkages: The Next Generation ...... 65

The Drugs-Crime Wars: Past, Present, and Future Directions in Theory, Policy, and Program Interventions ...... 97

Appendix A: Summary of Proceedings...... 163

Appendix B: Forum Agenda...... 197

Appendix C: List of Participants ...... 199

iii Introduction

Henry H. Brownstein with Christine Crossland

For criminal justice practitioners who deal this knowledge not as an end in itself but with drugs and crime day in and day out, as a means to accurately define the prob- the reality of the drugs-crime nexus is lem of drugs and crime and promote indisputable. In a manual designed to help future research. The agenda for research police chiefs and sheriffs control drug was developed under NIJ and NIDA spon- abuse, the International Association of sorship at a forum held in Washington, Chiefs of Police (IACP) stated unequivocal- D.C., in April 2001. The findings of the ly its belief in “a significant though com- Drugs and Crime Research Forum are pre- plex” relationship between drug abusers sented here. and criminal offenders. Change one group, IACP proposed, and you change the other: “If there is a reduction in the number of In pursuit of the people who abuse drugs in your community, there will be a reduction in the commis- drugs-crime link sion of certain types of crime in your If we are going to make progress toward community.”1 solving the problem of drugs and crime, we need to shed light on the nature of the When IACP released its manual more than drugs-crime link by designing effective a decade ago, researchers already were responses. Developing a research agenda confirming what practitioners believed and on drugs and crime means tackling the documenting the relationship between central issue of the drugs-crime link. Is the 2 drugs and crime. Public policy and pro- link a matter of cause and effect or is it grams were and continue to be developed something far more complex? on the basis of this knowledge.3 But although researchers and practitioners There is no lack of theories. The direct alike knew the relationship existed, the cause model of the drugs-crime relation- nature of that relationship eluded them ship has attracted its share of supporters. then and continues to elude them today.4 It states simply that either drug use leads to crime or crime leads to drug use. The To shed light on the drugs-crime link simplicity is appealing. Who would not requires research, and the first step is to find it tempting to believe that reducing specify the research topics to be covered. drug use can lower the crime rate? In fact, Taking the lead, the National Institute of some policies and programs have been Justice (NIJ) and the National Institute on developed on the basis of the direct cause Drug Abuse (NIDA) brought together aca- model or the belief in a significant rela- demics and other researchers and asked tionship between drugs and crime. As them to answer three questions: What do IACP recognized, the relationship is real we know about drugs and crime, what do enough. And NIJ’s Arrestee Drug Abuse we not know, and, most important, what Monitoring program has demonstrated do we need to know? Both agencies see

1 SPECIAL REPORT / JULY 03

year after year that among people appre- Building on the past: hended and charged with a crime, a large percentage uses drugs.5 The Drugs and Crime Research Forum However, as sociologist Erich Goode has cautioned, “Even the fact that drugs and NIDA and the National Institute of Law crime are frequently found together or cor- Enforcement and Criminal Justice, NIJ’s related does not demonstrate their causal predecessor organization, were asked by connection.”6 The consensus among re- Congress in 1976 to find out what was searchers who study the issue confirms known about drugs and crime. The prod- Goode’s observation. The evidence for the uct of the agencies’ collaboration was direct cause model is just not there.7 Drugs and Crime: A Survey and Analysis of the Literature. Though not strictly a We seem more willing today to accept the research agenda, the survey was a first complexity of the drugs-crime relationship, step “to identify where the gaps in our more open to the notion that “[t]here is knowledge lie and to direct research to fill considerable uncertainty . . . about the those gaps.”10 It was intended to “set the degree to which drug use causes crime or stage for more focused future research.”11 the degree to which criminal involvement causes drug use.”8 In a recent review of In 2000, NIJ’s call for the development the literature, sociologists Helene Raskin of a research agenda was another step White and Dennis M. Gorman definitively toward meeting that need. The authors of dismissed the direct cause model. They Drugs and Crime noted at the time that concluded instead that the drugs-crime “few if any [studies] directly address the link is best explained by the common drugs-crime nexus issue.”12 This report on cause model, in which any association of the development of the research agenda drugs and crime has a cluster of causes.9 will demonstrate that although much has been learned in the intervening years Those who subscribe to the common about drugs, drug use, drug abuse, drug cause model believe that to adequately markets, and drug law enforcement, much understand the relationship of drugs to work is needed to shed light on the com- crime requires attention to many issues, plexities of the drugs-crime link. social, cultural, chemical, and biological among them. What the model means for Three papers were commissioned for the policy and practice is that any response to research forum. Each addressed the ques- drugs and crime that works in one set of tions of what we know, what we do not circumstances may not work in another. know, and what we need to know about For researchers, it means the research the drugs-crime link. Prepared by experts agenda is vast. Policy and practice can be in epidemiology, public policy, social work, informed by what we know up to this and allied disciplines, the papers served as point, but progress in responding to the the focal point and framework for discus- drugs-crime problem requires knowing sions by forum participants. (The forum more. summary, agenda, and a list of the participants are presented in appendixes A, B,

2 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

and C.) After the forum adjourned, NIDA standing of the relationship in the most created a listserv for participants to contin- productive ways? ue to exchange their thoughts. ■ What do we really know about the The discussions did not all fit the same suspected causal connection between mold. Roundtables were generated from drugs and crime? In looking at drugs one-sentence statements by participants and crime, what is the intersection at about drugs and crime. What we do not which public health and public safety know about the drugs-crime relationship meet? How can we achieve the goals was treated at length. The many strands of of greater understanding and definitive thought, lines of discussion, and themes evidence and greater mastery in design came together when Forum participants and application in policy, programs, and addressed the final question: What future techniques to prevent and reduce harm- research is most important, and what re- ful health and safety consequences of search is needed most urgently? Mindful drug use? that the next generation of researchers will be tackling the problem of drugs and crime, ■ What do we need to do to integrate Forum participants recommended topics molecular biology, genetics, and neuro- for research by their graduate students. science into discussions of drugs and crime? What do we need to do to place discussions of the drugs-crime nexus in Probing drugs and crime: the context of history? How can we clarify the question of causal inference? Three perspectives How can we use the notions of scale “At the Intersection of Public Health and and rubrics to help understand the rela- Criminal Justice Research on Drugs and tionship between drugs and crime? Crime” was commissioned by NIDA from James C. Anthony with Valerie Forman. “Research on Drugs-Crime Linkages: The Anthony asked such questions as— Next Generation” was commissioned by NIJ from Robert MacCoun, Beau Kilmer, ■ Have we made effective and adequate and Peter Reuter. Among the questions use of recent developments in science asked by MacCoun and his colleagues and technology to advance the study were— (and hence the understanding) of the ■ Are our conceptualizations of the rela- relationship between drugs and crime? tionship between drugs and crime Given the vast literature generated dur- adequate to move forward in our under- ing past decades on this subject, have standing of that nexus? How must we we adequately, appropriately, and effec- conceptualize the relationship to be able tively integrated research from both to address questions not only of con- the public safety and public health comitance and statistical correlation, but perspectives? also of social significance and causality? ■ Are the tensions between the two per- ■ To the extent that the drugs-crime re- spectives greater than our ability to lationship is causal, to what extent do overcome them? Given what we know we understand the nature of the causal now and the current tension between influences? How can we use Paul researchers in public safety and public Goldstein’s tripartite taxonomy to build health, how can we conceptualize and on work already done, and how can we organize our thinking and research to move beyond the taxonomy? How enhance our knowledge and under-

3 SPECIAL REPORT / JULY 03

can we use notions such as Bruce the dynamic tension between drug poli- Johnson’s conduct norm analysis or cy as it shifts and the drugs-crime con- Alfred Blumstein’s drugs-gun diffusion nection as policy changes? What is the hypothesis?13 value of interventions and treatment when dealing with drug-using offenders? ■ How does the considerable heterogeneity of users, substances, locations, and situations as well as differences in mar- What do we need to learn? ket dynamics affect what we need to have and to do to address the drugs- It will come as no surprise that the ques- crime nexus? How do we address the tion of what we know about drugs and question of causal influences? How will crime was eclipsed by that of what we do research in the coming decade deal with not know and what we have yet to learn. the heterogeneity of effects across The papers and accompanying discussions users, substances, cities, neighbor- yielded an abundance of ideas on research hoods, and situations? topics for the coming decades. The major themes included the following: “The Drugs-Crime Wars: Past, Present, and Future Directions in Theory, Policy, ■ Drug-related crime. and Program Interventions” was commis- ■ Drug enforcement. sioned by NIJ from Duane C. McBride, ■ Drug markets. Curtis J. VanderWaal, and Yvonne M. ■ Drug offenders. Terry-McElrath. In this paper, McBride ■ Drug policy. and his colleagues raised the following ■ Treatment and intervention. questions: ■ Drug use and abuse. ■ Ethnographic studies. ■ In the past two or three decades, what ■ Health sciences perspectives. progress has been made in our knowl- ■ Minority research. edge and understanding of the relation- ■ Research methods. ship between drugs and crime? Does ■ Victimization studies. knowledge of the statistical relationship help us understand the nature of the Categorized more broadly, the topics pro- relationship? What do we know about posed for research are the drugs-crime the nature of the nexus and what do we nexus, the social contexts of drug use and need to do now to advance the state of crime, and refining study methods and our knowledge? In the past century, designs. how have we used that knowledge to guide public policy? Could we do a bet- What explains the ter job of linking what we think and drugs-crime nexus? what we know about drugs and crime to what we do to address individual and We know that drugs and crime are related. social problems in the realms of public We also know something about the differ- health and public safety? Do we know ent ways they might be related, and per- enough about what has been tried (for haps something about the ways they may example, programs and program evalua- be related in time and space. What we tions) to know what works? have yet to learn is how they are related. In other words, we need to probe the ■ How is the idea of social capital impor- underlying dynamics of the relationship. tant to our understanding of the drugs- We do not know, for example, why so crime nexus? What is the significance of many people who commit crime also use

4 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

drugs or why some people who use drugs Give more attention to minorities. A commit crime but others who use drugs disproportionate number of the people do not commit crime. Research in this arrested, charged, and in custody for drug broad area might take several directions. and other criminal offending are from minority groups. The reason is unknown Find new ways to conceptualize the but needs to be probed. What can we drugs-crime nexus. Several years ago, learn about the involvement of various Goldstein proposed a tripartite framework ethnic and racial groups in the drugs-crime as a way to disentangle the relationship link? What can we learn about gender and between drugs and crime, specifically vio- the drugs-crime link? Some answers lent crime. Violence could be the direct might be found in comparative, multisite outcome of ingesting drugs, the result of studies of drug use and drug markets in a user’s compulsion to obtain drugs or different ethnic communities. What is the money for drugs, or a product of the disor- relationship of gender, age, race, and cul- ganization and violence inherent in the ture to drug involvement and crime? What social systems in which drugs are manu- is the effect of disparity (in income, for factured and exchanged.14 Over the years, example), prejudice, and discrimination on this framework has been useful for study- the distribution of resources used for ing drugs and violent crime but of limited treatment and prevention? How can we value for studying drugs and other types explain racial and ethnic differences in of crime. Beyond what has been learned drug use and involvement in crime? Do from this model, how can researchers people view the drugs-crime link different- conceptualize the way or ways drugs and ly because of their race, gender, or age? crime—not just violent crime—are related? From these general research areas on Combine research perspectives. Re- racial and ethnic diversity, it is possible to search on drug use illustrates how different derive many specific topics. What can we disciplines can combine forces. Social sci- find out about the relationship between ence research is beginning to merge with drugs, crime, and the increase in the num- biological research, particularly genetic re- ber of women of color who are incarcer- search. Questions include the following: ated? Have changes in the economy affected the involvement of disadvantaged ■ How can the study of genes, the social black and Hispanic/Latino males in drugs environment, and behavior help us bet- and crime? If so, how? What is the impact ter understand the link between drugs of drug-related incarceration on families and crime? and children or on prospects for education and employment in minority communities? ■ Are there physiological propensities for drug using? If so, what is the impact of the user’s environment? What do we need to know about the social context of ■ Are alcohol and marijuana complements drugs and crime? of or substitutes for other drugs? It is widely believed that drug use ad- ■ Do different drugs have different effects versely affects users. But drug use and on groups of people who are genetically crime are affected by and in turn affect different? forces operating in society at large. Drug users interact with many people: sellers ■ How can we address the ethical con- with buyers, buyers with sellers, criminal cerns of such research?

5 SPECIAL REPORT / JULY 03

offenders with their victims. There is a use and social status, and if so, how might social context of drug use. social status in turn be related to involvement in crime? Beyond using illicit drugs, Social patterns in the drug world. There to what extent are drug users and sellers have been many studies of drug users and involved in other crime? Can we realistical- some studies of drug markets. But what ly estimate how much other crime is com- do we need to learn about the social rela- mitted by drug offenders? What risk do tions and interactions of the people whose these people pose to their own health and lives are affected by drugs? safety?

The commerce of drugs and crime: drug Victims of drug users and drug use. Drug markets. Theories about and the opera- users are in some ways their own victims, tions and institutional arrangements of but are there other victims? What do we drug markets are plentiful, but not enough know about other people with whom drug research has been done to test them. users relate? How can we define for re- How stable are drug markets, and how do search and policy purposes what we mean they change over time? For example, has by “victims of drugs”? How do we define the maturation of the crack cocaine mar- victimization in this context? Are there indi- ket in some cities affected those cities’ rect victims, such as families and commu- crime rates? What is the connection be- nities, as well as direct victims? tween local market activity and fluctuations in supply and demand at the national The public’s response to drug use and level? What influences the relationship drug-related crime. Society considers between sellers and buyers? How and drug-related crime and illicit drug use as why do new markets emerge, and what affronts and responds accordingly. En- impact do they have on existing markets? forcement strategies are one example. To How are prices set in local drug markets, what extent are the responses based on a and how are wages set? real understanding of these problems? Are the responses making a difference? Patterns of use and abuse. We know something about the demographics of Enforcing drug laws. What is the effect of drug use, but what do we know about enforcement policies, programs, and prac- intergenerational patterns? How do use tices on drug use, drug dealing, and drug- patterns vary with social or biological dif- related crime? What is the relationship ferences? How do patterns of alcohol use between street-level enforcement and compare with patterns of use of other street-level drug market activity, particular- drugs? Can drug use help explain juvenile ly violent activity? What impact do drug involvement in crime or violence? Are pat- seizures, drug arrests, and asset forfei- terns of use of certain drugs, such as club ture, among other interdictions, have on drugs, designer drugs, or inhalants, differ- drug and drug-related crime? What is the ent from patterns of use of other drugs? impact of public concern about racial pro- What can we find out about how and why filing and police corruption on the ability of people start or stop using drugs? law enforcement to respond to drugs and crime? Criminal offending by drug users. There are some studies of drug offenders, but Treating drug use and abuse. With so how much do we know about how or why many different drugs and so many differ- drug offenders commit crime? Are some ent types of users, what can we say people genetically predisposed to drug about the efficacy of drug treatment in use? Is there a relationship between drug addressing drugs and crime? What is the

6 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

nexus of drug treatment and criminal jus- together? Can we find out from policy- tice? For example, what are the results of makers and practitioners what decisions treatment in correctional settings and they need to make and what questions what do evaluations reveal? Do incentives they need to answer about drugs and or disincentives help drug users to suc- crime? How can we move drug policy ceed in treatment? How do we define analysis beyond econometrics (supply and success? How important are aftercare pro- demand, for example) and begin to study grams and family interventions? What is drug use from the perspective of politics, the best way to treat drug users who are criminal justice, public health, and social dually diagnosed (for example, those who work? are also mentally ill)? What are the dropout rates for treatment, and what does it mat- Methods of studying drugs ter? What treatments work best with what and crime types of drug use? How do we distinguish users from abusers? What difference Research methods are dictated by the does that distinction make for treatment questions researchers ask. Some of the planning? questions already explored indicate that certain methodological concerns might Intervening to prevent drug use or crime. need to be addressed. Although relatively little is known about preventing drug use, the topic receives a Attention to measurement and design. great deal of attention. To what extent can What are the best measures currently media campaigns help prevent drug use? available to study drug use and involve- Is the impact of prevention programs the ment in drug markets and drug treatment? same for all social categories of users or How can they be improved? How can we irrespective of type of drug? How can we construct integrated data collection meas- educate young people about the impact ures? What is the best way to design that drugs can have on their lives? Should measures and procedures to evaluate more attention be paid to problem behav- drug control programs? What are the best ior, norm violations, and rule breaking than measures for assessing drug treatment to drug prevention? outcomes? What is the role of cost- benefit analysis in drug studies? What can Public policy. When we think about public we learn from longitudinal studies about policy on drugs, we typically do not think the long-term effects of drug use and about policy in general but rather about abuse? How can we introduce randomiza- specific aspects, such as interdiction, tion to long-term studies of drug treat- enforcement, treatment, and prevention. ment? How can statistical techniques But can we step back and think broadly developed by other sciences be adopted and measure the impact of drug policy by the social sciences? How can we over the past decade, or even the past encourage multidisciplinary teams of century? Can we learn from policy simu- researchers to work together to study lations that examine past and prospective drugs and crime? views of drug use? Can we learn from comparative studies of different coun- Ethnography. There is a long, distinguished tries? What is the impact of different tradition of ethnographic research in the directions in drug policy? What policies field of drug studies. Ethnographic stud- have worked or not worked with adult and ies, however, are almost by definition lim- juvenile drug offenders? Can research ited to a single area or a small group of examine drugs, crime, and public policy people. What might we learn by secondary analyses of ethnographic studies?

7 SPECIAL REPORT / JULY 03

What might we learn by replicating ethno- Control Strategy—2000 Annual Report, Washington, graphic studies in other communities or DC: The White House, 2000 (and earlier ONDCP among other groups of drug users? How annual reports); and Longshore, D., F. Taxman, S. Turner, A. Harrell, T. Fain, and J. Byrne, “Operation useful might it be to link ethnographic Drug TEST Evaluation,” final report submitted to the studies of community structure with National Institute of Justice, U.S. Department of studies of drug users and dealers in their Justice, 2000 (grant 97–IJ–CX–0041). communities? What can we learn from studying communities of sellers and 4. Examples of studies that explored the nature of the drugs-crime relationship in that period include users? Would it be useful to establish Brownstein, H.H., and P.J. Goldstein, “A Typology of prospective, qualitative field sites in vari- Drug Related Homicides,” in Weisheit, ed., Drugs, ous communities as a type of surveillance Crime and the Criminal Justice System: 171–192; system to monitor changing drugs and Chaiken, J.M., and M.R. Chaiken, “Drugs and drug-use patterns? Predatory Crime,” in Tonry and Wilson, eds., Drugs and Crime: 203–239; Fagan, J., “Intoxication and Aggression,” in Tonry and Wilson, eds., Drugs and Using available data and studies. How Crime: 241–320; Goldstein, P.J., H.H. Brownstein, can we make better use of available data P.J. Ryan, and P.A. Bellucci, “Crack and Homicide in to study drugs and crime? Are there obsta- New York City, 1988: A Conceptually Based Event cles to making better use of available data Analysis,” Contemporary Drug Problems 16 (1989): to learn from them what we can? What 651–687; and Johnson, B.D., T. Williams, K.A. Dei, and H. Sanabria, “Drug Abuse in the Inner City: can we learn from meta-analyses of previ- Impact on Hard-Drug Users and the Community,” ously conducted research studies of drugs in Tonry and Wilson, eds., Drugs and Crime: 9–67. and crime? 5. Arrestee Drug Abuse Monitoring, 1999 Annual Report on Drug Use Among Adult and Juvenile Arrestees, Washington, DC: U.S. Department of What is in this report? Justice, National Institute of Justice, June 2000 (NCJ 181426). Following this introduction are the three papers commissioned for the forum and 6. Goode, E., Between Politics and Reason—The appendixes containing a summary of the Drug Legalization Debate, New York: St. Martin’s forum proceedings, the agenda, and a list Press, 1997: 119. of the names and organizational affiliations 7. See, Brownstein, H.H., “What Does ‘Drug- of the participants. 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; Notes Fagan, J., “Intoxication and Aggression,” in Tonry 1. International Association of Chiefs of Police, and Wilson, eds., Drugs and Crime: 241–320; Goode, Reducing Crime by Reducing Drug Abuse: A Manual E., Between Politics and Reason—The Drug Legal- for Police Chiefs and Sheriffs, Gaithersburg, MD: ization Debate; White, H.R., and D.M. Gorman, International Association of Chiefs of Police, 1989:5. “Dynamics of the Drug-Crime Relationship,” in G. LaFree, ed., The Nature of Crime: Continuity and 2. See Tonry, M., and J.Q. Wilson, eds., Drugs and Change, vol. 1 of Criminal Justice 2000, Washington, Crime, vol. 13 of Crime and Justice: A Review of DC: U.S. Department of Justice, National Institute of Research, Chicago: University of Chicago Press, Justice, July 2000 (NCJ 182408); and Wilson, J.Q., 1990; and R. Weisheit, ed., Drugs, Crime and the “Drugs and Crime,” in Tonry and Wilson, eds., Drugs Criminal Justice System, Cincinnati: Anderson and Crime: 521–545. Publishing, 1990. 8. Office of Justice Programs, Office of Justice Pro- 3. See, for example, Forcier, M.W., “Substance grams Fiscal Year 2000 Program Plan: Resources for Abuse, Crime and Prison-Based Treatment,” the Field, Washington, DC: U.S. Department of Sociological Practice Review 2 (1991): 123–131; Justice, Office of Justice Programs, 2000: 23 (NCJ Office of National Drug Control Policy, National Drug 182238).

8 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

9. White and Gorman, “Dynamics of the Drug-Crime “life cycle” characteristics (age of onset of drug use Relationship,” in G. LaFree, ed., The Nature of and crime, for example); “economic issues” (price Crime: Continuity and Change: 193. and supply/demand, for example); treatment; and methods (sampling, for example). 10. Gandossy, R.P., J.R. Williams, J. Cohen, and H.J. Harwood, Drugs and Crime: A Survey and Analysis of 12. Ibid.: 122. the Literature, Washington, DC: U.S. Department of Justice, National Institute of Justice, 1980: xi (NCJ 13. These works are discussed in this report. 159074). 14. Goldstein, P., “The Drugs/Violence Nexus: A 11. Ibid.: 122. The survey covered five “crucial Tripartite Conceptual Framework,” Journal of Drug areas”: patterns of drug use and criminal behavior; Issues 15 (1985): 493–506.

9 At the Intersection of Public Health and Criminal Justice Research on Drugs and Crime

James C. Anthony with Valerie Forman

Introduction our planning meeting for the drugs-crime research forum: “What do we know about This paper discusses intersections of pub- the drugs-crime interrelationship?” We lic health research and criminal justice cannot provide a comprehensive answer research on the topic of drugs and crime. to this question in a relatively short paper, The drugs of interest mainly are marijuana, but we will offer a starting point for dis- heroin, and other internationally regulated cussion, focusing on suspected causal compounds of illegal origin, and such relationships between drugs and crime. internationally regulated products of legal We also present a few concluding state- origin as pharmaceutical cocaine hydro- ments that were designed to facilitate chloride, codeine, and oxycodone, which discussion at the forum on drugs-crime also may be consumed on an extraordi- research held at NIJ in April 2001. nary basis (i.e., outside the bounds of accepted medical practice). An important A burgeoning literature on a point of departure for this paper is a wide- variety of fronts ly held assumption about two goals of research on this topic. The first goal is to A scholar interested in the topic of drugs achieve greater understanding and devel- and crime has much to read. Some of the op a body of definitive evidence on drugs classics of the field include Terry and and crime. The second goal is to achieve Pellens’s The Opium Problem (1928); early greater mastery of the design and applica- papers on drug taking and sociopathy by tion of policies, programs, and techniques Kolb and Pescor, who were two of the to improve public health and public safety early clinical leaders in research at the by preventing and reducing harmful conse- facility that ultimately became the National quences of drug use. Institute on Drug Abuse’s (NIDA’s) Intra- About the Authors mural Research Program and Addiction James C. Anthony, Ph.D., The outline for this paper corresponds Research Center; and work by Dunham is a professor in with assignments delegated at a planning and Lindesmith, whose surprisingly con- the Department of meeting held at the National Institute of temporary remarks and observations start- Mental Hygiene at Johns Justice (NIJ) in January 2001. This intro- ed to systematize some of the field’s Hopkins University. Valerie ductory section provides some back- Forman is a National research questions on the social psycholo- ground notes on the literature reviewed Institute of Mental Health gy of the drugs-crime relationship. Many predoctoral fellow with the for the paper and describes an organizing of the issues that confront the drugs-crime Psychiatric Epidemiology conceptual framework that can be used to researcher today were articulated by Terry Training Program at Johns assess gaps in the current evidence. The and Pellens (1928), Kolb (1925), Pescor Hopkins University. next section identifies some tensions that (1939), Dunham (Faris and Dunham, merit discussion as we try to forge a new 1939), Lindesmith (1938), and their con- research agenda on drugs and crime. We temporaries in the first half of the 20th then address the central question posed in century.

11 SPECIAL REPORT / JULY 03

The tripartite These issues were re-articulated and a The tripartite framework clarifies three new set of themes was clarified in subse- separate types of drugs-crime relation- framework quent research, such as The Road to H ships, none of which is simple. As for clarifies three investigations led by Chein (1964), the analysis of simple solutions for these com- work of Preble and Casey as described in plex problems, a therapeutically oriented separate types “Taking Care of Business—The Heroin drug maintenance program might reduce of drugs-crime User’s Life on the Street” (1969), Cohen’s the economic-compulsive type of offend- Delinquent Boys (1955), and Robins’ De- ing without influencing the occurrence of relationships, viant Children Grown Up (1966). Two of crimes determined by poor judgment or none of which the most important emerging themes other manifestations of intoxication states. from this research offer a challenge to A successful supply-side drug eradication is simple. conventional thinking about the drugs- program might reduce both pharmacologi- crime relationship: cal and economic-compulsive types of offending, but not offending of the sys- ■ There is no single drugs-crime relation- temic variety. Imprisonment of the drug ship. Rather, there are drugs-crime rela- user within a drug-free prison environment tionships, most of which are complex might extinguish today’s crimes but might rather than simple. not influence tomorrow’s offending when the prisoner is released back to the home ■ There is no simple solution to the com- community. Even if the prisoner remains plex challenges faced when drugs-crime drug free during the immediate postre- relationships come into play. lease period, the history of incarceration and a criminal record might constrain job By way of illustration, Brownstein and opportunities and economic success to Goldstein offered and refined a tripartite the point of inducing crimes that other- conceptualization of drugs-crime relation- wise would not have been committed if ships, which serves as a useful guide to the drug user never had been incarcerated some of the surrounding issues. Within in the first place. this tripartite framework, one set of criminal offenses is described as psychophar- Illuminated in this manner, the facets of macologically induced (e.g., responses to multiple drugs-crime relationships be- intoxication states after drug taking). A come more clear; new opportunities for second set of offenses is described as research open up. As these opportunities economic-compulsive in nature (e.g., have been recognized, there has been a instrumental income-producing criminal tremendous growth in scholarship and acts as needed to stave off symptoms of research activity on the topic of drugs and withdrawal states that appear once drug crime (see exhibit 1). use has stopped). A third set of offenses is described as “systemic” and might be Scholars may benefit from an assembled understood best as a consequence of a listing or bibliography of this literature, drug user entering or living within a social now available in electronic form as a tech- context in which extraordinary drug use is nical report from the Electronic Collabora- just one of a set of often intercorrelated tory for Investigations about Drugs at criminal behaviors. That is, we do not Johns Hopkins University (Forman, 2001). need an appeal to drug intoxication, drug Readers interested in a recent compre- withdrawal states, or drug-induced com- hensive review of these publications can pulsive behavior to account for offenses turn to the Harrison and Backenheimer- observed in this third category (Goldstein, edited issue of Substance Use & Misuse 1985; Brownstein and Goldstein, 1990). on the drugs-crime nexus in the United

12 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Exhibit 1. Number of Medline citations for “drugs & crime” as of March 2001

Number of citations 16

0 6 8 1975 197 1977 197 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year

States, which was published by Stanley The rubrics. Early in their public health Einstein and Marcel Decker, Inc., in 1998. research training, we ask our predoctoral and postdoctoral fellows to master the epi- A conceptual framework for demiology of drug dependence. Here, drug drugs-crime research dependence is defined as a syndrome or “running together” of clinical features, and Confronting the accumulated body of sometimes is called drug addiction, espe- evidence and new literature, we have cially when the focus is on such clinical attempted to sort each element of evi- features as obsession-like cravings and dence in relation to a conceptual frame- compulsion-like repetitive behaviors in work originally devised for the field of which drug taking is central. The clinical psychiatric epidemiology and epidemiolog- features of the drug dependence syn- ical research in general (Anthony and Van drome include pharmacological tolerance, Etten, 1998). This conceptual framework characteristic withdrawal signs and symp- is used as we train public health scientists toms, almost obsessional thinking about for advanced research on drug depend- drugs and drug-related behavior, and other ence and related conditions. The frame- observable mental, behavioral, and social work may prove to be useful in the domain adaptational manifestations of neuroadap- of criminal justice research as well, per- tational processes that get started and haps with suitable amendments by inter- progress with repeated drug taking. ested teachers and scholars.

13 SPECIAL REPORT / JULY 03

This epidemiology of drug dependence is specially protected confidential circum- a subject matter to be mastered by the stances. To the extent that subjectively public health research fellows, just as they felt experiences such as “craving” and master the concepts, principles, and tech- obsessional thinking about drugs are cen- niques used as methodological tools in the tral clinical features for drug dependence, public health sciences. Mastery of this we cannot substitute human urine, saliva, subject matter begins with study of the or sweat samples for self-reports (Anthony, just-mentioned clinical features, the histo- Neumark, and Van Etten, 2000). ry of diagnostic criteria or case definitions used in public health research on drug Once issues of definition and measure- dependence, and what has been learned ment have been mastered, research about its neuroadaptational and genetic fellows move on to what we call the substrates. In the process, research fel- “rubrics” of epidemiology—its main sub- lows learn of patterned variation in drug headings and associated research ques- dependence syndromes, some of which tions. These main rubrics and primary can be understood in relation to the phar- associated research questions are dis- macology and pharmacokinetics of differ- played in exhibit 2. ent drugs, such as cocaine versus heroin or methamphetamine versus oxycodone. Successful research fellows learn these Research fellows also learn about different rubrics and use them to master not only measurement techniques used in labora- the state of currently available evidence tory, clinical, and field studies of the drug on each form of the drug dependence syn- dependence syndromes. For example, drome, but also the current gaps in evi- under certain conditions, an appropriate dence and the research concepts and dose of a narcotic antagonist can be used tools needed to fill the gaps in evidence. as a bioassay to check for the presence of The relationship of each rubric to an asso- dependence on heroin or other opioid ciated set of research concepts and tools drugs (e.g., via precipitated withdrawal). sometimes helps to clarify and differenti- Nonetheless, in general, the measure- ate the rubrics. Links between each rubric ments of drug dependence rely heavily on and corresponding research concepts and self-report information obtained under tools are presented in exhibit 3.

Exhibit 2. The rubrics of epidemiology

Main rubrics Primary associated research questions

Quantity How many in the population are becoming affected, have become affected, and are now affected? Location 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? Causes What accounts for some individuals becoming becoming affected whereas others are not? Mechanisms What are the underlying liked sequences of events and processes that account for the occurrence and for the persistence of the condition? Prevention and control What can be done to prevent occurrence of the condition, shorten its duration, or ameliorate its circumstances?

14 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Exhibit 3. Main concepts, research designs, and statistical tools associated with each rubric of epidemiology

Main associated Main rubrics Illustrative concepts research designs and statistical tools

Quantity Point prevalence, interval prevalence, Population census, observational ambidirectional or lifetime prevalence, and variance cross-sectional field survey, and variance estimation under complex designs Cumulative incidence and incidence Cohort and prospective study designs and multiwave density panel study design Event rate, probability distributions, Vital statistics registration methods (birth, death), and and density functions expectation rapid and continuing surveillance Location Prevalence correlate, factor, difference, Cross-sectional field studies, clinic-based and population- ratio, odds ratio, and prev. = f (incidence, based case-control and case-base studies with preva- average duration); null hypothesis; statistical lent (prevailing) cases; statistical measures of correlation precision; likelihood principle; and tests of and association; and univariate response regression significance (p-values statistical power) models for description and prediction Incidence or risk correlate, risk factor, inverse All of the above, plus clinic-based and population- risk factor, incidence difference, incidence rate based case-control and case-cohort studies with ratio, cumulative incidence ratio, and incidence incident (dynamically occurring) cases density ratio Causes Causal and preventive factors, Koch’s postu- All of the above, plus fine-grained and coarse- lates, criteria for evaluating causal significance grained matching and stratification, direct and indirect of observed associations (e.g., dose-response adjustments, modeling with statistical adjustments, relationships), counterfactuals, reciprocities, marginal and random effects models, hierarchical and effect-modification and interaction models (e.g., alternating logistic regressions), randomized trials, family history and twin studies, and instrumental variable models Mechanisms “Natural history” versus “clinical course” and All of the above, plus marginal and random effects mediation longitudinal analysis models Prevention and Efficacy versus effectiveness, preventive Randomized controlled trial, and operations and control fraction, and attributable risk systems research

Quantity. Under the rubric of quantity, Mental Health Services Administration, the main associated research question is and other Federal agencies, a substantial How many in the population are becoming fraction of the Nation’s research expendi- affected, have become affected, and are tures on drugs and crime is directed to- now affected?” In this context, “becom- ward the “report card” function of public ing affected” can refer to becoming a drug health and criminal justice research under user, developing drug dependence, initiat- the rubric of quantity. A recent National ing criminal behavior, or some combination Research Council report (Manski et al., thereof (e.g., see Gfroerer and Brodsky, 2001) tallied more than 60 Federal agen- 1992; Kosterman et al., 2000; Golub and cies with data systems designed to keep Johnson, 2001a). track of estimates on the number of drug users in households, among school- As reflected in the published scientific attending youths, among arrestees, among literature and technical reports made patients seen in emergency rooms, and in available by NIJ, the Substance Abuse and various other segments of American life.

15 SPECIAL REPORT / JULY 03

On the U.S. Department of Health and data now is regarded as impractical or too Human Services (HHS) side, we have the costly for surveys on a mass population National Household Survey on Drug Abuse scale, and there has been concern ex- (NHSDA) with growing national probability pressed that bioassays might reduce sur- samples of adolescents and adults (now vey participation rates below acceptable with a sample size of more than 70,000 values. NIDA is engaging in survey re- subjects per year); the Monitoring the search experimentation with bioassays to Future (MTF) study, which started as a complement self-report data to assess way to track drug use among graduating practical questions of this type. In the high school seniors through a national meantime, serious concerns have been probability sample each year and now expressed about the capacities of these encompasses 8th and 10th graders; and a data systems to provide evidence for poli- less intensive but more massive Centers cy evaluation (see, e.g., Manski et al., for Disease Control and Prevention (CDC) 2001). Nevertheless, evaluated from the surveillance of drug use and other health standpoint of original plans for the data, risk behaviors of teenagers in school. On these criticisms are somewhat imperti- the U.S. Department of Justice side, we nent as surveillance indicators. The have other ambitious survey operations, criticisms are asking the surveillance such as the Arrestee Drug Abuse Monitor- operations to do far more than they origi- ing (ADAM) program (formerly Drug Use nally were designed to offer. Forecasting [DUF]), which monitors drug taking among arrestees through both self- The first rubric of epidemiology also en- reports and bioassay techniques, and the compasses studies of birth cohorts that National Crime Victimization Survey. are intended to estimate risks of adversity, plot trajectories of normative develop- Evaluated as part of the public health and ment, or quantify important population criminal justice research enterprise, these characteristics such as rates of officially substantial efforts may be understood recognized offending. The concept of a best as examples of surveillance opera- cohort study is familiar to criminal justice tions. The label “surveillance” does not researchers and public health scientists trivialize the important work of the profes- alike. Prominent examples in the criminal sionals and scientists whose daily labors, justice research arena include Robins’ year in and year out, yield the hard-won classic nonconcurrent cohort study of chil- surveillance data. In fact, many of our dren seen by child guidance workers in country’s surveillance operations in this the early 20th century (1966), and the domain of inquiry truly are gems and tend work of Tracy, Wolfgang, and Figlio enti- to be regarded as the best of the best in tled Delinquency Careers in Two Birth the world. In some respects, they are the Cohorts (1990). envy of the world. Nonetheless, by definition, surveillance activities are designed The fact that the rubric of quantity is men- with timeliness and practicality in mind, tioned first does not mean that research sometimes with deliberate decisions to under this rubric is easy or a methodologi- sacrifice validity of measurement in favor cal snap. Not at all. From the standpoint of of enhanced survey response rates. data gathering, those of us who have recruited, trained, and supervised teams For example, NHSDA, MTF, and the CDC of 60 or more field worker-interviewers survey all use self-report methods to and quality control staff for data entry, measure drug-taking and crime-related documentation, and management can behavior (e.g., weapon carrying). The appreciate the operational challenges in option of bioassays to confirm self-report surveillance work. From a statistical

16 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

vantage point, the nature of the surveil- are not Avogadro’s number; rather, they are lance operations often includes interde- values expected to change over time, if not pendent observations within samples (e.g., from place to place. If we value probing sampled students within samples of quantitative criminological research exem- schools; sampled household residents plified by Cohen’s work, then we must within neighborhoods; multiple respon- ensure that the drugs-crime research agen- dents within sampled households, emer- da includes periodic repetition of surveys to gency rooms, or criminal justice facilities). yield the required estimates. These interdependencies motivate solutions that call on the calculus (e.g., in Studying the accumulated evidence on the Taylor series linearization for variance esti- drugs-crime relationship, we have been mation). In some estimation applications, able to sort much of it into the rubric of there is a need for Bayesian statistics not quantity. Quite clearly, NIJ and NIDA now yet taught widely in graduate research make a considerable investment in the training programs. basic counting tasks required to estimate and quantify such parameters as how As to the importance of these “counting” many adult arrestees and juvenile offend- operations, we may turn to a recent re- ers are taking drugs each year. Each repe- search contribution by Cohen, who incor- tition of these surveillance operations porated values from these surveillance provides evidence on variation in the esti- operations in his attempt to estimate the mates from time to time and from place to monetary value of rescuing a high-risk place. The study of this type of variation youth from a life of delinquency, crime, falls under the second rubric, which is and other socially maladaptive behavior called location. (1998). To complete this work, Cohen had to turn to an array of previous results from Location. Our second rubric is location, counting operations that ranged from the and the main associated research ques- National Institute of Mental Health Epide- tion is “Where in the population are affect- miologic Catchment Area surveys we con- ed individuals more or less likely to be ducted in Baltimore during the early 1980s found, with variation in occurrence and fre- with colleagues at four other university- quency differentiated by characteristics of based sites to work that Blumstein and his time, place, and person?” On occasion, group completed to estimate basic param- work under this rubric is guided by theory, eters of criminal justice research, e.g., an but more often the research has a more estimated 6 percent of all boys account descriptive character. for more than half of all arrests (Blumstein et al., 1986). James et al. (2002) provide an illustration of the nature of research and evidence There can be little doubt that investigators about location. The research team set in the drugs-crime arena should be interest- out to plot geographic variation in the ed in Cohen’s conclusions about varying occurrence of drug purchase opportunities programmatic investments and the mone- experienced by young adults in the United tary returns from programs to intervene States. In this figure, a “drug purchase with high-risk youths. Nevertheless, it is opportunity,” a special form of drug- somewhat startling to know that Cohen related crime opportunity, is measured had to turn back to counting evidence by a survey response to a standardized gathered in the early 1980s and before to assessment in interviews conducted for produce estimates to be used for policy NHSDA. As depicted in exhibit 4, and sub- and programmatic decisions almost two stantiated with a univariate response decades later. These quantitative estimates regression model for description, there is

17 SPECIAL REPORT / JULY 03

Exhibit 4. Prevalence of drug purchase opportunity among youths 12–24 years old, in percent

3–9% 15–19%

10–14% 20–25%

Rural females Urban females

Rural males Urban males

Source: Substance Abuse and Mental Health Services Administration, 1996 and 1997, National Household Survey on Drug Abuse, Rockville, MD: U.S. Department of Health and Human Services, Substance Abuse and Mental Health Services Administration

variation in the occurrence of these drug relation to individual-level characteristics purchase opportunities across locational (e.g., sex, age, socioeconomic status, eth- regions of the country and for young men nicity). For example, Fendrich et al. (1995) versus young women. In this context, the studied juvenile and older murderers to statistical methods are not intended to understand varying degrees of drug in- probe the causes of the observed variation volvement in murder. Locational research from place to place, nor the observed also plots temporal changes, as illustrated male-female differences. Rather, the in a recent NIJ report on the possibility of methods are used simply to help quantify new marijuana epidemics, to be described the uncertainty in the survey-based esti- below (Golub and Johnson, 2001b). mates and substantiate the presence of variation from place to place and the male- Estimates of the consistency of associa- female differences (James et al., 2002). tion between drug use and various arrest and criminal behavior types also serve to This illustration is useful because it re- illustrate analyses focused on location minds us that location refers not only to within population experience: Crime was geographic variation but also to variation in found to be more common among drug

18 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

users than among nondrug users. As as health disparities that might differential- Harrison and Gfroerer (1992) make clear ly fall on one or another racial or ethnic in their NHSDA analyses on this topic, the minority group. Here, it is a predictive as research questions they were trying to well as a descriptive purpose that can be answer concerned the number of drug achieved. However, when the task is to users, the number of individuals engaged predict and not to explain, there is no spe- in criminal behavior, and the overlap in cial calling for the methods required for these numbers. With respect to location, firm causal inferences, as depicted in their work clarified the proportion of drug exhibit 3. users who were engaged in criminal behavior and the prevalence of criminal Within the drugs-crime arena, there are behavior in relation to drug use. As is true many different examples of surveillance in the work of James et al., these investi- operations under the rubric of location, gators did not draw on the apparatus of such as we can see in recent work by causal inference, matching, or other scien- Golub and Johnson (2001b) in which they tific maneuvers to disentangle whether used DUF/ADAM data as evidence to the criminal behavior was a response to advance their claims about a new and pos- the drug use or vice versa. Nonetheless, sibly expanding epidemic of marijuana use taking a step beyond studies of officially in the United States. True to the descrip- recognized crimes, arrestees, and convict- tive character of locational research, Golub ed criminals, Harrison and Gfroerer helped and Johnson present evidence of the new confirm links between drugs and criminal and possibly expanding epidemic among behavior, but they did not seek to produce offenders in some areas (e.g., Atlanta) and definitive evidence about the causes of evidence of no epidemic in other areas drug use or criminal behavior. (e.g., Miami), but they do not seek to explain why there should be an epidemic Much of our current research enterprise at in one place but not in another. Because the interface of drugs and crime has this these data are from incarcerated individu- type of descriptive character. Substantial als, an important set of complications aris- HHS investments in the MTF study and es in their interpretation. One suspects NHSDA already have been mentioned. that the observed time trends and varia- On the NIJ side, we call on ADAM to help tion from place to place might reflect oper- clarify variation in the occurrence of drug ations of local police departments as use among arrestees across multiple juris- much or more than it reflects any underly- dictions, not only in the United States but ing change in the dynamics of marijuana also overseas. For the most part, we do epidemiology. not require these investments to yield definitive evidence that might be central This rubric of location also encompasses in causal inference. Nonetheless, the evi- studies in which the investigators may be dence from these studies helps to de- striving toward causal explanation, but scribe the location of drug taking, criminal they fall short, often demonstrated in a behavior, and the intersections of these shift toward the language of “prediction” behaviors, and sometimes to describe or and away from the language of “explana- predict the co-occurring and separate pat- tion.” Two different hypothetical conclud- terns of drug use and criminal behavior. ing statements can illustrate this point. When the research team falls short of its Analyses conducted under this rubric with- goal, the researchers may summarize their out a special push toward causal inference work by saying something like “Based on can be especially important in identifying this study’s evidence, the level of drug hot spots within geocoded areas as well use in early adolescence predicted later

19 SPECIAL REPORT / JULY 03

delinquency and criminal behavior in the causal theory and evidence, it has been young adult years.” A different verb is possible to reduce the occurrence of DS selected for the alternative, stronger form in human populations by encouraging of concluding statement: “Based on this mothers to bear their children before age study’s evidence, the levels of delinquen- 40. Hence, a strictly descriptive study pro- cy and criminal behavior in the young adult voked an effective intervention to reduce years depend on levels of drug use in the occurrence of an important genetic early adolescence.” condition.

As the focus shifts from description or It is regrettable that our studies of dispari- prediction toward explanation and causal ties affecting racially and ethnically defined inference, we move from the rubric of subgroups of the American population location to the third rubric of causes. The generally fall under the rubric of location, shift in focus calls into play a new set of as do our studies of the changing dynam- research concepts, principles, and tools, ics of household and family composition as outlined in exhibit 3. in the United States. For example, we now can say with some certainty that The yield of a Many scholars will appreciate that a single African-American males experience rates study may contribute evidence under sev- of arrest, prosecution, and incarceration study often is eral rubrics at once. For example, the peri- for drug possession offenses that cannot not clear at odic reports of NHSDA, MTF, and ADAM be explained by their rates of drug taking, routinely present evidence that falls under but we do not have good evidence on the outset or in the rubric of quantity as well as the rubric the causes of this racial disparity. Initial the stages of of location. Rarely, the authors of these inquiries suggest differential law enforce- reports seek to make causal inferences ment and judicial practices, which some- study planning. from their surveillance data. times encompass racial profiling, but rigorous scientific evidence on these The yield of a study often is not clear at practices is scarce. the outset or in the stages of study planning, and the study orientation to theory is With respect to the dynamics of house- not always a discriminating feature. Some hold and family composition, the phenom- theory-based studies have started as in- ena of youthful drug taking and related vestigations of causes but have ended up criminal offending have links back to the making contributions solely in the domains families of origin, now often characterized of prediction and description. Other atheo- by absence or infrequent appearance of retic studies end up making useful contri- the father in many of our population butions in our studies of cause. Consider groups. This is not to say that female- the first conclusive study on the topic of headed households are homogeneous or age-related risk of Down syndrome (DS) uniformly deleterious with respect to and associated mental retardation, com- socially maladaptive behavior of young pleted some 50 years ago. The investiga- people. It would be a mistake to presume tors sought to plot the risk of DS by the that the traditional mother-father house- age of the mother at the time of delivery. hold always and in all contexts is more An exponential increase in risk after age protective than a female-headed house- 40 was clear in the first graphs. We still do hold with respect to the risk of youthful not know what causes the chromosomal drug taking or delinquent behavior (e.g., trisomies that give rise to DS, nor do we see Chilcoat, 1992). Mothers often know why these trisomies and DS are mobilize family resources or draw on more common when older mothers give assets that in some measure may help birth. But even in the absence of firm compensate for absent fathers (e.g., by

20 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

involving grandparents, neighbors, church Finally, a note on ethnographic studies groups), as described by Kellam, Ensminger, should be added here. In general, the and Turner (1977), Pearson et al. (1990), sample size and “sample space” charac- and others. teristics of these studies do not make ethnography an especially fertile discipline A research agenda on race, ethnicity, and with respect to the first rubric of quantity, family or household composition can be except when the characteristic under motivated by an awareness that the study has extremely limited dispersion. drugs-crime relationships will depend to For anyone who looks to ethnographic some extent on demographic trends. studies for quantitative values, there often Against the backdrop of demographic are some unanswerable questions about trends such as these, including an in- generalizability and precision of the study creased prominence of Hispanic children estimates. In some respects, ethnography and families in the United States, it will be might be characterized as a search for the important to sustain the research agenda boundaries of no variation in a socially in the domain of locational variations of shared human characteristic. this type. Important steps in this direction have been taken in the Federal agencies This is not to say that ethnography is bar- responsible for surveillance of drug-related ren when it comes to quantitative data. behaviors, including increased attention to To the contrary, the small scale of ethno- measurement of ethnic self-identification graphic research makes it possible for (e.g., with respect to Cuban origin, Puerto ethnographers to shift directions more Rican origin, and other subgroups of the quickly than is possible in ordinary surveil- Hispanic population; with respect to lance operations. As a result, ethnographic Chinese origin, Samoan origin, and other field workers helped in the early identifica- subgroups within the Asian-Pacific Is- tion of crack cocaine, methamphetamine, lander category). Similar attention is re- and oxycodone outbreaks—years before quired in criminal justice research such as these outbreaks could be identified in ADAM and I–ADAM (International ADAM) large-sample surveillance data. and in administrative statistics compiled on operations of the criminal justice sys- Under the rubric of location, ethnographic tem in this country. field workers were among the first to note inner-city adolescents whose drug taking Whereas the human genome project is started with marijuana rather than with the challenging conventional views about more normative experiences with alcohol “race” as a scientific concept, studies on and tobacco. They also were the first to self-identified race-ethnicity will have a characterize a growing use of “blunts”— sustained importance in the NIJ-NIDA tobacco cigars hollowed out and filled with research agenda on the topic of drugs marijuana for a combined tobacco-marijuana and crime. intoxication (Golub and Johnson, 1999). In a recent round of observations, there is a This evaluation of importance can be suggestion that for some youths, the typi- grounded in an awareness of the demo- cal “gateway” drugs have been skipped— graphic trends described above, but it also an example of subgroup variation in the draws on an appreciation of what studies more typical developmental sequences of self-identified race-ethnicity may teach running through alcohol, tobacco, and mar- us about the influence of cultural contexts ijuana to drugs such as heroin, stimulants, and socially learned behaviors with respect and hallucinogens. Large-sample epidemi- to drug taking and criminal behavior. ological surveillance data now seem to

21 SPECIAL REPORT / JULY 03

confirm the initial ethnographic observa- ■ Each study is indicating ample room for tions on this topic (e.g., see Golub and gene-environment interaction or for Johnson, 1999; Golub and Johnson, influence of environmental conditions 2001a). and processes.

One of the reasons ethnographic research These results from causal research help is important under the rubric of location is substantiate a case for a future research that it can open our eyes to new concep- agenda on the genetic sources of variation tions of time, place, and personal charac- and on environmental modulation of these teristics that impinge on the drugs-crime genetic sources of variation. relationship. These ethnographic studies are especially useful in descriptions of the Randomized trials with relatively simple cultural context and socially learned be- structure also can be used to probe causal haviors described above. Their evidence hypotheses with definitive results. For can add depth and insight to otherwise example, these trials may offer our best superficially understood intersections of avenues toward definitive evidence on drug taking and criminal behavior. whether cessation of illegal drug use is followed by reductions or elimination in crim- Causes. The third rubric of epidemiology inal behavior. An alternative is to nest the pertains to the study of causes and draws study of causes within a more expanded on the research apparatus required for agenda of systems research on drugs- causal inference (exhibit 2). On occasion, crime relationships (Manski et al., 2001). To date, most this research apparatus can be quite sim- To the extent that systems research ple in concept. For example, a relatively entails a finely detailed specification of ethnographic small sample of monozygotic (MZ) twins mechanisms that link events and process- research on discordant for an important outcome is es within a system, this type of research sufficient to provide definitive evidence falls more clearly under the rubric of mech- drugs and about environment with respect to the anisms, as described below. crime has been causes of that outcome. These MZ twins are genetically matched: If they are discor- Outside of the simplicity of research on descriptive in dant for outcome, one may look for gene- discordant MZ twins and randomized con- character. environment interactions, but more often trolled trials, a complex apparatus of study one looks for differences in environmental design and statistical method is required conditions in utero (e.g., dichorionic ver- to extract definitive evidence in research sus monochorionic sacs), perinatally (e.g., on drugs-crime relationships. Given the insults at the time of delivery), or in later importance of inferences about causes in development (e.g., head trauma for one the drugs-crime relationship, it may be twin but not the other during infancy or understandable that graduate research childhood). The National Institutes of training programs have become increas- Health (NIH) investment in recent twin ingly methodological in their orientations. research to estimate heritability of different forms of drug use now generally is It may be appropriate to discuss the po- paying off in two ways: tential contribution of ethnographic research in relation to the causes of the ■ Each study is indicating at least some drugs-crime relationship. To date, most degree of heritability of drug depend- ethnographic research on drugs and crime ence, and sometimes heritability of drug has been descriptive in character. It has use, especially legal drug use (e.g., provided leads for more probing causal tobacco). investigations, but it has not produced definitive evidence on the links between

22 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

drugs and crime. In this respect, ethno- drugs-crime relationship, one might ask graphy’s contribution may be most impor- about the mechanisms of linked states tant under the rubric of location. Before and processes leading to or away from an anyone could mobilize large-sample sur- association between illegal drug use and veillance operations to study the new criminal behavior. drugs-crime phenomena connected with crack cocaine (e.g., crack and prostitution), For an illustration of these mechanisms, it was possible for ethnographers to move one may turn to the coercive process and in and make headway. To some extent, deviancy training models introduced in ethnographers have been pioneers in the work of Patterson and Dishion. Their research on methamphetamine and club Oregon Boys study has provided longitudi- drugs such as MDMA (Ecstasy), and we nal evidence on what surely must be cen- can expect more of the same in relation to tral linkages in the mechanisms underlying our first new drugs-crime outbreaks of the drugs-crime relationships (e.g., see Patter- 21st century, which involve sustained son, Dishion, and Yoerger 2000). For exam- release oxycodone. ple, studying these school-based samples of boys through ages 17–18, and using An NIJ-NIDA investment in ethnographic standardized coding of a 30-minute free research on drugs-crime relationships of discussion-interaction between best this type will continue to be important—if friends, they found substantial over-time only to help us begin to understand the correlation of deviant friendship process unusually circumscribed geographic distri- (e.g., duration of rule-breaking talk bouts butions of methamphetamine and oxy- as coded from videotape). Dishion also codone use in the United States and the has reported on a link from initial drug patterns of criminal behavior associated use to increased affiliation with deviant with use of these drugs. Ethnography can peers and onward to initiation of criminal be used to produce a catalog of causal behavior that is more consistent with the explanations for methamphetamine’s delinquency-to-drugs link that emerged in emergence as a threat to public health the longitudinal research of Johnston and and public safety in rural sectors of the colleagues based on MTF analyses pub- American Midwest and for oxycodone’s lished more than 20 years ago (Dishion emergence in small cities and towns of et al., 1996; Johnston, O’Malley, and the Appalachian mountain range, especial- Eveland, 1978), as well as on more recent ly from West Virginia southward. It is not studies (e.g., Elliott and Huizinga, 1989). clear that ethnography or any other scientific field will be capable of producing The use of multiwave longitudinal study definitive evidence about specific explana- designs to probe into suspected causal tions in this catalog of causes. Nonethe- mechanisms is well known in both public less, there is value and importance in the health and criminal justice research circles. attempt to do so, and ethnographers can The Alcohol, Drug Abuse, and Mental bring rigor and scientific discipline to this Health Administration (the precursor to process of investigating these causes. The SAMHSA), and more recently NIH and alternative seems to be to leave these OJJDP have maintained support for a investigations to the field of journalism. series of important longitudinal studies over the years (e.g., see the work of Mechanisms. Within epidemiology gener- Jessor and Jessor, Kellam and Ensminger, ally, mechanisms refer to linkages of Block and Block, McCord, Bachman, Kandel, states and processes that lead toward Robins, Elliott and Huizinga, Hawkins and expressions in clinical features of health Catalano, and many other studies of this and illness or disease. As applied to the type, as listed in compendiums such as

23 SPECIAL REPORT / JULY 03

Verdonik and Sherrod, 1984). Advantages The natural history of a disease proves to of long-term investment in these longitudi- be an important element under the rubric nal studies can be seen in the research of mechanisms. In the past, a careful de- articles from many of the research proj- scription of a disease’s natural history ects with multiwave assessments, for often has guided investigators toward example, the Pittsburgh Youth Study (e.g., underlying causal mechanisms. Loeber et al., 1998), the Denver Youth Study, and the Rochester Youth Develop- In the years before effective drug treatment Study (Loeber et al., 1999); and ments, Winick and others drew attention the research groups led by the Brooks, to the maturing out process for drug ad- Newcomb, and Bentler (e.g., see Brook et dicts, and there is a parallel literature on al., 1996, Brook et al., 2000; Newcomb maturing out with respect to criminal be- and Bentler, 1988; Newcomb 1992). havior in general (Winick, 1963). The maturing out process continues to be an One of the questions in the design of an important locus for new research on the agenda for future research on drugs and drugs-crime relationship. crime is how the evidence from large- and medium-sized samples from longitudinal Other clues about causal mechanisms are studies of this type might be linked with being produced in observational and longi- evidence from the generally much smaller tudinal studies of individual cases or fami- intensive studies of cases. Until there is lies characterized by some feature of the consensus about effective interventions to drugs-crime relationship. For example, we disrupt the drugs-crime relationship, possi- have Dunlap’s intensive studies of families bilities for a linkage exist through the con- in which one of the members is a crack cept of natural history. cocaine dealer (Dunlap and Johnson, 1996; Dunlap, Johnson, and Manwar, 1994); re- In the history of medicine and medical search such as Spunt’s study of adoles- research, the first natural historians of dis- cent offenders with a history of violent ease were clinicians and clinically oriented crime (Spunt et al., 1990), Longshore’s observers who made careful observations linkage of DUF and California Bureau of at the bedside of patients, in the absence Criminal Statistics data (Longshore 2000), of effective interventions. They watched, and the earlier related studies started by measured as best they could (e.g., body Hser, Anglin, and McGlothlin (1987); and temperature), and described change in investigations led by Inciardi, Johnson, relation to the passage of time from the and Goldstein or members of their re- first recognition of clinical features. Within search groups (e.g., see Inciardi and the realm of drugs and crime research, Russe, 1977; Inciardi 1990; Inciardi and ethnographers and social scientists gener- Pottieger, 1998; Johnson, Dunlap, and ally have taken over the responsibilities of Maher, 1998; Goldstein 1998; Spunt et al., careful clinical observers in relation to ille- 1990, 1994, 1995). gal drug use and criminal behavior. During the last 50 years, thanks to the work of Several interesting elaborations of these Robins (1966), Winick (1962), Preble and intensive case studies have developed Casey (1969), Agar (1973), Waldorf (1998), in the realm of criminal justice research. Nurco (Nurco et al., 1975, 1996; Nurco, For example, Logan (2001) has added 1998), and their successors, we have bioassays for metabolites of the neuro- learned much about the natural history of transmitter serotonin as well as testos- drug use, drug dependence, and associat- terone assays as part of his intensive ed criminal behavior through ethnographic followup studies of crack users. This and social science research. example serves to illustrate a potential

24 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

intersection of public health and criminal It is possible to make a forecast of likely justice research that should be explored in integrations of genetic research, cognitive more depth as we work through a future sciences, and the more traditional disci- agenda for research on drugs and crime. plines of behavioral and social sciences for a future agenda for NIJ and NIDA research A conceptual shuttling back and forth on drugs-crime relationships. For example, between these intensive smaller sample exhibits 5–7 represent an elaboration of studies and the generally larger sample conceptual models our research group has longitudinal studies would seem to have developed as an aid to our study of transi- advantages for investigators who work in tions from drug use to drug dependence. one or another of these arenas, and there Exhibit 5 expresses a suspected causal are a few investigators who conduct both influence of drug use on criminal behavior. types of studies (e.g., see Dishion and It also expresses a separate influence of Loeber, 1985; Dishion, Patterson, and drug dependence on criminal behavior. Reid, 1988; Dishion et al., 1996). This type These two specifications are consistent of bridgework between the microsocial with the Goldstein-Brownstein distinctions and ethnographic research traditions and between drug-related crimes that might large-scale longitudinal sample research arise from acute drug intoxication states deserves to be a deliberate focal point on versus crimes that are rooted in the eco- the future drugs-crime research agenda. nomic-compulsive behavior of an individ- This focal point is important because the ual who suffers withdrawal states as a study of causal mechanisms and process- result of sustained drug use and neuroad- es can draw attention to potentially vulner- aptation. There are many law-abiding drug able links where new interventions might dependent individuals who do not commit be directed. crimes, even when they are suffering from withdrawal pains. Hence, exhibit 5 in- In epidemiology generally, the focus of cludes a speculative causal pathway that research on causal mechanisms is shifting runs directly from withdrawal to the occur- to genes and encoded gene products, as rence of criminal behavior, over and above displayed in our most recently emerging the separately specified role of the drug subspecialties of genetic epidemiology dependence syndrome for which with- and molecular epidemiology. To some drawal serves as a manifest indicator. extent, Elliott has a head start in a potential cross-fertilization between criminal jus- We speculate that an individual’s genome tice research, genetic epidemiology, and can contribute to the drugs-crime relation- molecular epidemiology. He already has ship in different ways. Exhibit 5 concen- introduced harvesting of DNA samples in trates on a possibility that some genetic the context of his national longitudinal polymorphisms or mutations may be inter- study (Elliott, 2001). Opportunities for correlated manifestations of an underlying case-control studies and other genetically diathesis or vulnerability to make the tran- informative designs, including whole sition into drug dependence from a state genome scans nested in a case-cohort of nondependent drug taking, as reflected study design, will become possible as this in pathway 1. It also specifies a possibility research evolves. Eventually, this type of that a specific polymorphism (or mutation) work should lead us toward more defini- has an additional influence on this transi- tive evidence on causal mechanisms tion, as reflected in pathway 2. As indi- underlying the drugs-crime relationship, cated by pathway 3, we may hope for including gene-environment interactions. development of effective intervention

25 SPECIAL REPORT / JULY 03

Exhibit 5. Conceptual model of the influence of drug use and drug dependence on criminal behavior

Birth Being Drug Years Job Religion: Living cohort male using of prestige Practice in urban peers school behavior area

Onset of drug DRD2 use polymorph # 2 Aggression

Polymorph Criminal 2 # 1 behavior Diathesis Executive Polymorph # 4 dysfunction 3 Onset of ct # 3 dependence ct effel Polymorph 4 Possibleof withdrawa dire

Treatment intervention Loss of control Withdrawal

Note: Depiction of a mediational model linking a generic susceptibility trait (diathesis, path 1) with risk of making a transition from onset of drug use to onset of drug dependence and subsequent links to criminal behavior, directly and indirectly through drug-induced aggression and drug-induced disturbances in executive functions. Via path 3, treatment intervention might modify the expression of the generic diathesis (as manifest in covariation of multiple discrete polymorphisms) or might target a specific gene product or gene effect, with path 2 showing the putative gene effect and path 4 depicting the possibly specific effect of treatment intervention, over and above the intervention effect on the generic susceptibility trait.

Reproduced with permission of copyright holder James C. Anthony, 2002.

techniques that can disrupt what other- misbehavior or social maladaptation sec- wise might be an expression of the diathe- ondary to drug taking, which can occur sis. If effective, these interventions will with or without criminal behavior. Execu- slow or disrupt the natural history of drug tive dysfunction refers to impairments in dependence at a step in the process that the cognitive processes that subserve links nondependent drug taking and the human capacity to plan, direct, and control subsequent transition into drug depend- one’s future behavior within adaptational ence. This effect of intervention, by itself, boundaries and may encompass more may be sufficient to alter the drugs-crime generalized planning behavior (e.g., see relationships depicted to the right of the Tolman, Edleson, and Fendrich, 1996). exhibit. As depicted in exhibit 5, during states of The potential role of the cognitive sci- acute drug intoxication, there may be a ences is expressed in the intermediate release of aggressive behavior and a pathways that link nondependent drug tak- disruption of regulatory executive func- ing and drug dependence to later criminal tions. As levels of drug dependence in- behavior. Here, aggression may be con- crease, levels of aggressive behavior ceptualized in a generic sense as rowdy can change in an upward or downward

26 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

direction and executive dysfunctions To be sure, exhibit 5 is only a model that can occur. The complexity of interrelation- represents little more than an oversimpli- ships between aggression and executive fied representation of the complexities dysfunction is reflected in the reciprocal that link an individual’s genome with cog- causal paths between these two con- nition and behavior. Models by definition structs. Increased executive dysfunction are oversimplified representations. It is fair translates as inept decisionmaking about to ask whether the model requires addi- aggression and the subjective utility func- tional specifications, such as the possibili- tions that govern decisions about whether ty that religious convictions might tend to to commit a crime. As part of generally modulate the relationship between drug adaptive fight-flight responses and modula- taking and aggression or criminal behavior. tion of monoamine neurotransmitter In this oversimplification, exhibit 5 does signaling pathways during bouts of aggres- not convey all such possibilities. These sion, there can be a cascade of executive possibilities for elaboration of the longitu- dysfunctions: Mere rowdiness can be dinal model should help the reader under- transformed into aggravated assault. stand some of the complexities faced in

Exhibit 6. Conceptual model of the drugs-crime relationship Time = 0 Time = 1 Time = 2 Time = 3

Level Level Level Level of of of of drug drug drug drug use use use use

Level of Level of Level of Level of criminal criminal criminal criminal behavior behavior behavior behavior

Level Level Level Level of of of of dependence dependence dependence dependence

Loss of Loss of Withdrawal Loss of Withdrawal Withdrawal Loss of Withdrawal control control control control

Notes: Depiction of a simplified longitudinal mediational model that links earlier levels of drug use and dependence to later levels of criminal behavior. For example, the level of drug use might produce intoxicating states that give rise to violent criminal behavior, even when the level of drug dependence is held constant (or kept at zero levels). Here, the level of drug dependence is tapped via a measurement model with clinical features of drug dependence, such as loss of control and withdrawal as the manifest indicators for levels of dependence. In this simplification, analogous measurement models for the level of drug use and the level of criminal behavior are not drawn but may be presumed.

In this depiction of the drugs-crime relationship, there is an allowance for reciprocity between levels of drug use and levels of drug dependence, once dependence begins. That is, there is not an acyclic dose-response relationship that links drug use to drug dependence. Rather, once the drug dependence process begins after first drug taking, the dependence process becomes a determinant of later levels of drug use. Most current conceptual models do not provide for this reciprocity.

This model is one that makes no allowance for the possible effect of criminal behavior on levels of drug use or dependence, but this defect is remedied in later elaborations of this model (e.g., see exhibit 7).

Reproduced with permission of copyright holder James C. Anthony, 2002.

27 SPECIAL REPORT / JULY 03

observational studies of causal mecha- point in time influencing levels of criminal nisms that account for observed drugs- behavior at future points in time, but in crime relationships. exhibit 6, we do not specify a link from levels of criminal behavior to subsequent Exhibit 6 presents even more simplifica- drug taking or drug dependence levels. At tion to sharpen focus on the drugs-crime least in theory, and in some prior sugges- relationship specifically. The genetically tive evidence, this omission represents a based diathesis and other covariates of potentially important mis-specification of exhibit 5 are set into the background (i.e., our model for the drugs-crime relationship presumed to exist but not explicitly depict- (e.g., see Johnson et al., 1995). ed). In exhibit 6, we see a readily appreciated reciprocity between the level of drug Exhibit 7 adds a level of complexity to the taking and the level of drug dependence: model depicted in exhibit 6 and poses a (a) the more drug taking, the more we find substantive question for the agenda of increased drug dependence levels, and (b) action research: “How might an interven- the more drug dependence, the more we tion lead to change in this system of in- find increased levels of drug taking. We terrelationships?” We introduce the also see the level of criminal behavior ex- possibility that social status (e.g., status pressed as a function of levels of drug tak- in the community, socioeconomic status, ing and drug dependence, as shown in lawful income) depends on criminal behav- exhibit 5. An additional elaboration in- ior and also on the level of drug depend- volves the longitudinality of this model. ence, and that criminal behavior influences We have levels of criminal behavior at one the subsequent level of drug dependence

Exhibit 7. How an intervention might lead to a change in the drugs-crime relationship Time = 0 Time= 1 Time = 2 Time = 3

Level Level Level Level of of of of drug drug drug drug use use use use Assortative Assortative Assortative peering peering peering Level of Level of Level of Level of criminal criminal criminal criminal behavior behavior behavior behavior

Social Social Social status status status Level Level Level Level of of of of dependence dependence dependence dependence

Loss of Loss of Loss of Loss of control Withdrawal control Withdrawalcontrol Withdrawalcontrol Withdrawal

Note: Depiction of a longitudinal mediational model that links levels of drug use and levels of drug dependence with levels of criminal behavior. The model depicts criminal behavior’s influence on subsequent levels of drug use via differential association as well as a possible influence on subsequent levels of drug dependence via changes in social status. Once the drug dependence process begins, there is a reciprocity, with the level of dependence influencing the level of drug use and vice versa. As in exhibit 5, the level of drug dependence is manifest in the covariation of clinical features, such as loss of control and withdrawal signs.

Reproduced with permission of copyright holder James C. Anthony, 2002.

28 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

by its intermediate influence on social sta- similar peers (assortative peering, homo- tus. The model depicted in this exhibit phily, etc.). Fortunately, there already is also provides for a plausible link from the a cadre of criminologists and drug re- level of drug dependence to subsequent searchers who are trained in sociology criminal behavior. Namely, as drug depend- and social psychology and can readily ence increases and social status (e.g., incorporate the biomedical and clinical lawful income) falls short, criminal behav- concepts of drug dependence into their ior may increase (as in the Goldstein- research plans, if supported to do so. Brownstein tripartite model). In addition, subsequent levels of drug dependence It will be more difficult to forge a research may be influenced by the changes in agenda that integrates the genetics re- social status, either upward or downward. search and cognitive sciences constructs depicted in exhibit 5. For the most part, The model in exhibit 7 also introduces a genetics and cognitive sciences are un- conglomerate concept of “assortative known territories for most NIJ and NIDA peering,” expressing a well-known truism: investigators who have made important “birds of a feather flock together.” The contributions in past research on the occurrence of drug taking is linked to later drugs-crime relationships. For most drugs- There are formation of peer group relationships, as crime researchers, it would not be difficult mountains of is the occurrence of criminal behavior. To to integrate concepts and measurements some extent, we can say that past drug of aggressive behavior and the clinical syn- data from 20th- use and past criminal behavior influence drome of drug dependence within their century studies current peer group affiliations, and to existing research plans. Far more difficulty some extent, we can say that past peer will be encountered during the process of that have not group affiliations influence future drug use integrating genetics and the neuropsycho- yet been fully and future criminal behavior. These com- logical and neurophysiological measure- plexities are expressed by hypothesized ments of the cognitive sciences. exploited through causal paths in exhibit 7. careful analysis. We can learn a lot about the drugs-crime Conceptual models of this type are in- relationship simply by replicating and refin- complete representations of the causal ing important longitudinal research on mechanisms that lie beneath observed drugs-crime relationships that was initiat- drugs-crime relationships, yet they are ed during the second half of the 20th cen- elaborations of the Goldstein-Brownstein tury. Many of these longitudinal studies tripartite model. Nonetheless, most read- have cohorts that still are intact, and fol- ers will agree that these representations lowup studies are now underway to learn are oversimplified. If they have value, it is more as these cohorts mature into adoles- to highlight some future directions for the cence and make the transitions into young joint NIJ-NIDA research agenda on drugs- and middle adulthood. There are moun- crime relationships. tains of data from 20th-century studies that have not yet been fully exploited We do not yet have a longitudinal research through careful analysis. program to investigate the relatively simple model of interrelationships between Nonetheless, as we look forward through levels of drug use, drug dependence, and the next decades of research, the NIJ- criminal behavior as depicted in exhibit 6, NIDA agenda must go beyond what has let alone the more complex model of developed as the best 20th-century re- exhibit 7, with its sociological construct of search on the drugs-crime relationship. social status and the social-psychological Ten decades from now, if we are to construct of affiliation with behaviorally leave the 21st century with an enhanced

29 SPECIAL REPORT / JULY 03

understanding of the drugs-crime relation- suspected causal relationships. Some ship and with a greater capacity for effec- readers of this paper will know of work tive action to improve public health and that Kellam and our Johns Hopkins safety in this domain, we cannot continue research team have completed, using ran- to work within the narrow paradigms and domized field trials to probe the interrela- methodologies of the traditional scientific tionship between early aggressive and disciplines mastered by drugs-crime in- rule-breaking behavior and later drug vestigators of the 20th century. If we are involvement among boys (e.g., Kellam and successful, then in a few decades, the bio- Anthony, 1998). In essence, we decided medical, genetic, and cognitive science that more observational research on the substrates of the drugs-crime relationship link from early aggression or deviance and will no longer be a matter of mere specula- later drug use would be less important tion, as depicted in exhibits 5 and 7. There than an experimental test. Within the will be definitive evidence, solid under- framework of a randomized field trial, we standing, and effective action-plans based tried to and succeeded in reducing aggres- on what we learn from the pioneers who sive and deviant behavior of first-graders move into that now-unexplored territory. using an experimental intervention assigned at random. For the boys assigned Prevention and control. The long-term to experimental intervention, we have value of research on causal mechanisms found later reduced occurrence of drug depends on identifying potentially vulnera- involvement, and we have replicated ble linkages in the sequence of states and these results in a second cohort of first processes that lead to illegal drug use and graders (Kellam and Anthony, 1998). More criminal behavior. It may go without saying replications along these lines are needed that increasingly definitive evidence about before anyone can claim that early aggres- causes and causal mechanisms will help sion or deviance is a “cause” of later drug us achieve our goals in the domain of use, but this experimentation illustrates effective prevention and control. None- how experimental research in the domain theless, a reminder may be useful with of prevention and control can yield bene- respect to a dynamic interrelationship fits in the form of improved evidence to between etiological studies (of causes) test causal theories. This idea is not new. and the emergence of effective interven- Hawkins, Catalano, Offord, and others tions. As illustrated in the circumstance of have noted it as well (e.g., Hawkins, Von DS and maternal age, with limited evi- Cleve, and Catalano, 1991; Hawkins, dence on the underlying causal mecha- Catalano, and Miller, 1992; Jones and nisms of DS, by manipulating maternal Offord, 1989). But it is an idea that often is age we have a very effective instrument overlooked by investigators more interest- to prevent and reduce the risk of DS. As ed in theory testing and who orient them- explained in our original paper on the selves toward goal 1, described in this rubrics of epidemiology, many effective paper’s first paragraph. Under the fifth public health preventive interventions rubric, we try to orient the research to emerged before firm knowledge about serve both goal 1 and goal 2. causes and causal mechanisms became available (Anthony and Van Etten, 1998). Because elements of this rubric of prevention and control are being covered in A related concept involves the use of the companion papers that accompany randomized preventive trials to provide this working manuscript, I will close increasingly definitive evidence about this section more quickly than might be

30 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

customary. Before doing so, I would like A selective overview of to mention the contributions of operations research and systems research in this tensions faced in research domain, which often have been neglected on drugs and crime in epidemiology. Over the years, the thoughtful and quantitatively sharp work Numerous tensions are faced at the in- of Blumstein and colleagues has contin- tersection of public health and criminal ued to inspire an important line of re- justice research on the drugs-crime rela- search on prevention and control that is tionship. This section identifies and de- pertinent to the drugs-crime relationship. scribes a selection of these tensions, and Although I am not confident about all of in some instances recommendations are the data or assumptions of the underlying offered for NIJ and NIDA action to help analysis approaches, I have been especial- resolve the tensions. ly impressed by the directions taken by Blumstein colleagues Caulkins and Cohen Tensions in theoretical in this regard. perspectives, concept, and definition For example, Cohen (1998) discusses potential synergy of programs and distin- Heterogeneity at the intersections of pub- guishes the aggregate benefits of pro- lic health and criminal justice research is grams designed to reduce crime versus not limited to differences of opinions and the sum of the benefits of individual judgment about empirical observations, programs. It is possible that no single the inferences we can draw from these program would help city residents feel observations, and the uses to which we safe enough to derive lifestyle-related apply the observations (e.g., cost analyses expenditure benefits (e.g., walking a mile of alternative programs). There are some through a rough neighborhood versus tak- fundamental tensions within and across ing a taxicab because of concerns about theoretical perspectives and also approach. safety). Combinations of programs might The concept of scale. Ecologists work do so. This distinction ties into the con- with a concept of scale that may help us cept of marginal costs versus average understand some of the tensions men- costs associated with drug-using and tioned above and may serve as an axis delinquent youths or criminals, where the of orientation as we turn to future direc- marginal costs exclude fear of crime and tions for research (e.g., see Brown, 1995; private security expenditures because Wiens et al., 1986). As a concept, scale these costs are largely unaffected by any resonates with what educational researchers one criminal’s actions. and behavioral and social scientists often Caulkins and his colleagues developed a have termed multilevel or hierarchical mod- challenging line of systems research that els, as in Ennett’s and the Duncans’ research can ultimately yield new ideas and evi- with young people nested within ecological dence about policy instruments in relation niches of higher order such as classrooms, to the drugs-crime relationship. The evolu- schools, or families (e.g., see Ennett et al., tion of this work toward selection of poli- 1997; Duncan et al., 1997; Duncan, cies and programmatic instruments at Duncan, and Hops, 1998) and our own different stages of a drug-taking epidemic research group’s nesting of individual is especially important (Caulkins, Crawford, drug users and collections of drug users and Reuter, 1993; Behrens et al., 1999). in their neighborhoods of residence (e.g.,

31 SPECIAL REPORT / JULY 03

Bobashev and Anthony, 1998; Petronis drugs-crime relationship to be of one order and Anthony, 2000). Parker and Toth of magnitude when we are investigating (1990) also have appealed to related individuals who all reside in the same local macro versus micro concepts in their area (e.g., as in much of the ethnographic research on alcohol and homicide, as have research on the drugs-crime nexus), a dif- Patterson and colleagues in their research ferent order of magnitude when we work on peer groups (Patterson, Dishion, and with individuals and data from multiple Yoerger, 2000). Bronfenbrenner’s ecologi- neighborhoods, but with matching on cal systems theory for human develop- neighborhood in the analysis, and a dif- mental research slices scale into macro, ferent order of magnitude when our data meso, and micro divisions that many in- are from individuals across the Nation, vestigators have found useful (Bronfen- with no analytical attention to who lives brenner, 1979, 1986). near whom, except perhaps during the process of estimating variances for confi- Although not with any direct reference to dence limits and standard errors (e.g., see a formal ecological concept of scale, we Bobashev and Anthony, 1998). can see resonance of this concept in Markowitz and Grossman’s studies of Although not clearly within the scope of taxes and regulations on alcohol and their the original ecological concept of scale, an hypothesized effects on criminal behavior investigator may work inward from the (Markowitz and Grossman, 2000), the boundaries of the whole organism toward research of Caulkins and colleagues on subunits, ultimately leading to the signal- national drug policy and programmatic ini- ing pathways between neurons, messen- tiatives (e.g., Behrens et al., 2001), and ger systems originating from genetic Holder’s research on preventive interven- material, and the simplest proteins and tions directed toward communities in the encoding genes themselves. This the United States (Holder 1993, 2001; elaboration of the concept of scale creates Holder et al., 1999, 2000). Scale is worked yet another tension, in part because the outward from the individual organism in concepts of genetics and signaling path- the direction of larger social groups, organ- ways for neurotransmission are more izations, and geopolitical units. In public familiar in the public health research com- health and criminal justice research, we munity but are not yet in the mainstream often refer to pre-established institutional of graduate or postdoctoral research train- or geopolitical boundaries (schools, cen- ing in the criminal justice research com- sus tracts, nations) when we work at munity. To illustrate, when I have talked higher scale. In ecology, mathematical with my criminal justice research col- models and methods such as advanced leagues about Elliott’s inclusion of DNA wavelet analysis are used to allow the assays in the most recent waves of his empirical data to inform scale—as in National Youth Survey, many of them have research on landscape ecology (Anthony asked, “Why?” To be sure, some skeptical and Bradshaw, 2001). behavioral genetics colleagues also have asked, “Why?” but this is an instance in Some tensions arise in research when which the same verbal behavior has ori- investigators ignore scale in their theoreti- gins in substantially different theoretical cal perspectives or empirical research models. My point is that tension can arise reports. For example, most of us work when concepts of scale are not made within a conceptual framework that leads explicit. us to comprehend estimates of the drugs- crime relationship without reference to Some of the work at the intersections of scale. However, one should expect the public health and criminal justice research

32 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

will be to make our concepts of scale ex- The literature also shows heterogeneity plicit. In some respects, this will be more in the typologies of criminal behavior or readily accomplished as we work from social maladaptation. Notions of childhood the whole organism outward, and the task conduct disorder followed by Antisocial may be more difficult as we try to inte- Personality Disorder appear prominently grate molecular biology, genetics, and in some formulations, but are absent neuroscience into our discussions. None- elsewhere (Loeber and Schmaling, 1985; theless, this hard work will be essential as Stevens, Kaplan, and Bauer, 2001; Lang- we make a 21st-century science of drugs- behn and Cadoret, 2001). crime relationships. These definitions and constructs in our Orienting definitions and constructs. theoretical perspectives demand work at The literature also displays considerable the intersection of public health and crimi- heterogeneity in orienting definitions and nal justice research. If we cannot bridge constructs. On the public health side, there these different approaches or marry them often has been an orientation toward drug to produce adaptive offspring, they will use and drug dependence or addiction as prove to be an unending source of unre- useful constructs in their own right. One solved tension with implications for reorientation often has been called the search progress. Unresolved tensions “medical model,” but it amounts to little slow down our progress in research more than an analysis of empirical syn- that depends on a peer review process, dromes (i.e., co-occurring manifestations whether the peer review occurs at the of the neuroadaptational processes that stage of reviewing proposals or of vetting get started when drug use begins, fol- journal articles. lowed by a cascade of secondary and terti- ary adaptations, some of them occurring At NIJ and NIDA, an important part of the in the domain of social adaptational roles research agenda can be a series of meet- and responsibilities). In a later section of ings or technical workshops. The charge this paper, I will return to this syndrome to workshop participants is to bridge these concept. On the criminal justice side, drug orienting concepts and definitions across use and the drug problems associated disciplines or create an articulation be- with drug use often are treated as if they tween concepts that will accelerate are not interesting in their own right but research progress on drugs-crime rela- are something akin to interchangeable tionships rather than slow it down. observable manifestations of something else that is more fundamental, such as the Orienting conceptual frameworks and “problem behavior syndrome” construct theories. The originating biomedical first elucidated by Jessor and Jessor branches of public health research some- (1977) some 30 years ago. A more recent times take theory as a given or work with version of this concept is a general de- theory in the background when there are viance construct used by Scheier, Botvin, emergent problems of human suffering and others in empirical studies (e.g., see and disease to be solved. For example, Scheier and Botvin, 1996), and there also the important 20th-century line of investi- is a recent respecification of the Jessor gations required to identify lung cancer and Jessor model, with elaborations as an adverse consequence of tobacco that encompass the epidemiological con- smoking was guided more by implicit cepts of risk factors and protective factors concepts of carcinogenesis secondary to (Jessor, 1998). tobacco smoking. Strongly articulated,

33 SPECIAL REPORT / JULY 03

explicit theories, if any, would have been be useful in a synthesis or integration of mis-specified and incomplete in that they various theoretical perspectives that range could not possibly have incorporated the from the disciplines of molecular or behav- postsmoking DNA adducts, protein ioral genetics to those of econometrics adducts, and gene-encoded metabolizing and the other social sciences. However, enzymes now prominent in the models there clearly is diversity and tension even of carcinogenesis. In criminal justice within fields as narrow as behavior genet- research, true to its origins in the social ics, where some work is oriented toward and behavioral sciences, the theoretical developmental family processes (e.g., as underpinnings are made more explicit advanced in the recent work of Neider- (e.g., see Thornberry, 1997; Kaplan, 1995). hiser and colleagues), and other work is One might say that without explicit theory, not (e.g., see Neiderhiser et al., 1998, the research in this domain stands little 1999; Neiderhiser, 2001; Brennan, Med- chance in peer review, no matter how nick, and Jacobsen, 1996; Tehrani et al., important the empirical contribution. 1998; Kotler et al., 1999).

This is another source of tension at the These tensions surface most clearly in intersections of and sometimes within the debate and discussion of an intersection domains of public health and criminal jus- of genetics research and studies of the tice research. In Public Health Service drugs-crime relationship. Many investiga- study sections, I have seen study section tors from social science backgrounds are members be less than enthusiastic about hesitant to take part in discussions of proposed epidemiological research on genetics, gene expression, and mecha- drug use and Antisocial Personality Dis- nisms of inheritance that might account order because the applicants had not ori- for covariation of drug-taking behavior and ented themselves to the major theories of criminal offending. This hesitation can be deviance well known in criminal justice cir- traced in part back to serious and impor- cles: “inadequate conceptual model” is tant concerns about ethical issues, eugen- the phrase that comes to mind. I also have ics, and the like. Some of the hesitation observed major differences of opinion can be traced back to a gap in graduate The intersection about scientific priority among experts in education: Graduates of social science the criminal justice and social science training programs often have not mas- of the Human world, some of whom are comfortable tered the basics of human biology and Genome Project, with “psychologizing” constructs within genetics. their theories (e.g., the self-derogation gene expression, models developed by Kaplan), and others Looking from a different perspective, an and proteomics who are more focused on constructs with observer can see other sources of tension a behavior analytic origin (e.g., coercive in relation to conceptual framework and with research on process and deviancy training models theories. Graduates of human biology and drugs-crime developed by Patterson, Dishion, and their genetics programs often have not mas- research groups in Oregon). tered the basics of behavioral and social relationships sciences research. Tension in relation to theoretical models merits close has been readily apparent in this NIJ-NIDA The intersection of the Human Genome attention at NIDA collaboration, which has offered a chance Project, gene expression, and proteomics to step back and look over a broad ex- with research on drugs-crime relationships and NIJ. panse of scientific progress in public merits close attention at NIDA and NIJ. To health and criminal justice research on the some extent, this intersection can be culti- drugs-crime association. This broad per- vated in a gradual process of shaping new spective creates germs of ideas that might investigators. NIDA’s peer review of its

34 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

portfolio of research training programs and is especially useful because the environ- individual career development awards can mental conditions experimentally manipu- specify requirements for cross-discipline lated in this lab have conceptual linkages mastery. On one side, new social science back to the deviancy training, inept parent- investigators can be required to master ing, and parent-infant relationship models the basics of human biology and genetics. developed by Patterson, Dishion, Brook, On the other side, new human biology and and others (Higley et al., 1996a, 1996b; genetics investigators can be required to Higley, Suomi, and Linnoila, 1996a, 1996b; master the basics of behavioral and social Patterson, Dishion, and Yoerger, 2000; sciences. Dishion et al., 1996; Brook et al., 1996; Brook, Tseng, and Cohen, 1996). NIDA already is sponsoring a series of training workshops for new investigators to More examples of this type of bridgework expose them to the different disciplines are emerging from the work of the research that now contribute to its research mission. pioneers who try to keep pace with evolv- The initial workshops have focused on ing contributions from the NIH Human epidemiology, pharmacology, and neuro- Genome Project. The NIJ-NIDA research science and introduced participants to agenda can be enriched by a technical those fields. Future workshops are planned, report series that brings examples of this with a broad agenda that cuts across the type to the community of investigators behavioral and social sciences, including and research trainees. ethnography and behavior genetics, as well as domains of medical sciences such as Tension that involves approach proteomics, drug development, and NIDA’s or methods clinical trials network. Review of the drugs-crime literature cre- Sustained investment in research educa- ates an opportunity for developing new tion of these types will be needed at NIJ insights about the sometimes different and NIDA. Without attention to pharmacol- approaches and methods that have been ogy, neuroscience, and pharmacogenetics, developed in public health and criminal it will be difficult for future investigators justice research work groups. For exam- to develop a fundamental understanding ple, ethnography with small groups has of the pharmacological and economic- expanded to almost large-sample ethno- compulsive categories of offending in the graphic research that bears some resem- Goldstein-Brownstein tripartite conceptual blance to large-sample survey research, framework. Without grounding in the but in many ways is different. To an out- social sciences, it will be difficult for them sider, this expansion is a puzzle to be to develop a fundamental understanding solved and has not yet been grasped. In of the systemic categories. the public health research domain, the original role of an ethnographer bore some There now are investigators who can resemblance to the role of the medical bridge the gaps that appear as canyons practitioner as a student of the natural his- between disciplines. Elliott’s attempt to tory of disease. The original natural history articulate his work with the NIH human studies were intensive case studies, with genetics initiative provides one example. the doctor at the bedside of individual sick In a primate lab run by Steve Suomi at patients making careful systematic obser- NIH, research on gene-environment inter- vations about this individual case and then actions as substrates of aggressive behav- that individual case, in the days when ior, social maladaptation, and drug use there might have been symptom pallia- provides another example. This research tion (e.g., cold cloths for fever), but no

35 SPECIAL REPORT / JULY 03

effective curative interventions to change prerequisites: having a favorable and trust- the clinical course of disease. This has ing relationship between study partici- some resonance with Agar’s concepts of pants and researchers, ensuring complete the ethnographer’s attention to the details confidentiality of responses, meeting labor of behavior and verbal expression and of or equipment and programming costs writing the narrative and taking down the associated with beeping the participants stories of drug users in their own words several times per day, and addressing dif- (Agar, 1973). The link from this role of the ficulties faced when the participants are ethnographer to large-sample ethnography illiterate or challenged by technology. remains unclear. Several recent studies of delinquent and Measurement methods pioneered in antisocial behavior may help clarify the util- behavioral sciences research and intro- ity of ESM procedures in research on the duced to studies of drug taking by Larson, drugs-crime relationship. For example, Kaplan, and Schiffman have started to sur- Farnworth (2000) studied a group of young face in criminal justice research as well. Australian offenders on probation and Experience Sampling Methods (ESM), found that these respondents were en- originally developed to study the daily lives gaged in such productive activities as of high school students, have now been employment or education an estimated 10 introduced in research on drug use (e.g., percent of the time. Compared with refer- see Csikzentmihalyi and Larson, 1987, ence norms for Australian adolescents, 1992). Their ESM procedure requires offenders spent 30 percent more time on study participants to wear an electronic passive leisure activities. An estimated 42 pager device that beeps at randomly percent of the time, offenders on proba- scheduled intervals, signaling the partici- tion reported being bored, while 62 per- pant to record some predetermined cent of the time they were involved in aspects of his or her present feelings, unchallenging activities. The use of ESM activities, and/or surrounding environmen- to integrate studies of drug-taking and tal conditions. Usually, dozens of self- criminal behaviors will provide new and reports are collected over a week or more important evidence on relationships that to capture as much of participants’ daily generally have been studied via retrospec- living as possible. One advantage of this tive reconstruction of behavior over long method is the ability of the researcher to spans of developmental time. examine drug use specific to each individual, given the assessment of his or her On another measurement front, there is baseline characteristics for comparison. arelated tension that involves the use of An additional benefit is the possibility of bioassay methods to study recent and past taking into account measured social con- drug taking. Wish has been a pioneer in text of the behavior (e.g., see Farnworth, the use of these methods for research on 2000). ESM also creates new opportuni- arrestees, and recent studies by Harrison ties to investigate the determinants of and Fendrich are extending this reach into drug-taking behavior that might be unique general household population samples of to each individual and each situation (e.g., the type surveyed for the National House- see Kaplan and Lambert, 1995). hold Survey on Drug Abuse (Wish, 1988; Yacoubian, Wish, and Perez, 2001; Fendrich, These evolving ESM procedures require 2001; Harrison, 2001). In future research, a number of conditions if reliability and one may anticipate these differences in validity are to be enhanced. Kaplan and approach to sustain a tension until a gener- Lambert (1995) identified the following al consensus has evolved.

36 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

With respect to approach in the domain procedures and disclosure statements are Computational of statistical methodology, computational required. advances have contributed to an accelera- advances have tion of innovation. There is a resulting air NIJ officials can initiate a useful dialogue contributed to of optimism for what might be accom- with OHRP on this important research plished, as in the domains of longitudinal topic. Perhaps more than any other gov- an acceleration latent transition modeling, multilevel or ernment agency, NIJ can help to stimulate of innovation nested models, approaches to nonignor- a dialogue and negotiate a reorientation of able missing data, and alternative meth- current practices in a manner that fosters in the domain ods of research on directed acyclic new and creative research on the drugs- of statistical pathways with mediation versus cyclical crime relationship without a lapse in pathways with reciprocities. At the same research ethics or slippage in the protec- methodology. time, there is a tension because these tion of human subjects in this research. new statistical approaches have not become integrated in most research training Outside the Federal Government, re- programs, and there remains certain skep- searchers now face increasingly thorny ticism about heavily modeled data. challenges in the protection of their research participants and the assurance of Limitations on numeracy keep many of us confidentiality in relation to research data. from probing the assumptions of complex For example, research that includes models, whether these are models of be- assessments of tobacco smoking now havior in individual studies, econometric requires special handling as a result of models, or operations and systems re- legal action by the tobacco industry. These search models to probe alternative pro- requirements apparently extend to crimi- gram and policy decisions. Tension may be nal justice research in which tobacco inevitable in the face of such complexities. smoking is approached as a self-reported indicator of deviance. The integration of molecular biology and genetics into these Tension involving research ethics research agendas, and even the introduc- NIJ can play an important role in relation to tion of experience sampling methods or investigations that probe drugs-crime rela- bioassays for drug testing, raise new tionships. At present, a good part of the questions in the domain of research ethics, NIJ role has been ceded to HHS and its some of which have been scrutinized in new Office of Human Research Protections randomized experimental designs. These (OHRP). True to its origins in NIH, OHRP is challenges deserve the close attention of oriented toward the standards of experi- these research communities, with OHRP mental medical intervention research (e.g., and its NIJ counterpart in suitable roles. randomized trials to test safety and efficacy of new drugs). OHRP specifications for informed consent procedures and disclo- Does drug use cause crime? sure statements share this orientation. A focal point Many behavioral and social sciences re- Each author of working papers for the searchers have expressed concern that drugs-crime research forum was asked to the standards and specifications of experi- identify a circumscribed set of research mental medical research are not appro- issues and probe what we really know priate for studies of the drugs-crime about them. Mindful of other sections to relationship. For example, in ethnograph- be written, we have been able to organize ic and observational survey research, dif- these research issues in relation to a sin- ferent specifications for informed consent gle focal point, expressed in the question,

37 SPECIAL REPORT / JULY 03

“Does drug use cause crime?” One ad- Hill, 1965). Today, students of epidemi- vantage of this specific focal point is that ology learn them as criteria for judging it has a broad range and can encompass whether an association is causal or guide- many different strands of evidence devel- lines for evaluating the evidence of a oped in public health and criminal justice causal relationship, together with an research. Another advantage is that it is analysis of the relative strengths and acrucial open question for research on weaknesses of evidence from randomized crime and drugs. As characterized by trials, prospective and longitudinal studies, Harrison and Backenheimer (1998), “Re- retrospective studies, and case-control search has not been able to validate a comparisons. Exhibit 8 presents these causal link between drug use and criminal criteria and guidelines. behavior.” Before reviewing these criteria, four clarifi- When confronted with an etiological re- cations may be in order. First, the criteria search question such as “Does drug use for evaluating causal significance of ob- cause crime?,” a public health scientist served associations represent standards typically might turn to a 20th-century elab- of scientific evidence that are substantially oration of the 19th century Henle-Koch different from the standards used to judge postulates or conditions for judging causal evidence in civil and criminal pro- whether a specific disease might be ceedings. For some segments of this caused by specific bacteria. For a time, paper’s readership, the question, “Does this 20th-century elaboration was known drug use cause crime?” may sound silly: as Hill’s postulates (after Sir Austin “Of course drug use causes crime. My Bradford Hill, a medical statistician) and grandmother could tell you that.” (This also as Evans’s postulates (after A.S. was the type of reaction TV/radio personal- Evans, an epidemiologist; Evans, 1976; ity Rush Limbaugh gave to some of the

Exhibit 8. Criteria and guidelines for judging the causal significance of an observed association

Criteria/guidelines Associated questions

Temporal relationship Is the temporal sequencing consistent with the idea that A causes B, or is there an ambiguity or the possibility that B causes A? Biological or other theoretical plausibility Is the idea that A causes B supported by theory or by trustworthy common experience and wisdom? Biological or other theoretical plausibility Is the available evidence consistent with the suspected causal link between A and B, or is there considerable inconsistency across studies? Alternative explanations ruled out If we are skeptical that A causes B or that B causes A, are there other specific alternative explanations for the observed statistical relationship between A and B, such as some background factor C that accounts for a spurious association between A and B? Dose-response or gradient relationship Is there regularity in the observed plot of B as a response to A? Where we see more of A, do we see more of B? Strength of association How strong is the relationship? Is it strong enough to make other alternative explanations less plausible? Cessation effects In this extension of the dose-response criterion, do levels of B drop substantially when A no longer is present?

38 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

early work that Howard Chilcoat, Tom idea of reciprocities between drug use Dishion, and I published on the topic of and criminal behavior, do not necessarily whether inner-city mothers and fathers undermine inferences about drug use as might be able to help protect their children a cause of crime. We face a problem of against risk of early-onset drug use if they slightly different conformation in our re- maintained levels of parental vigilance search on drug dependence: The use of generally associated with good parental adrug is an absolutely necessary but not supervision and monitoring.) sufficient condition for development of clinical syndromes of drug dependence, Our response to these gentle readers is to but once the drug dependence process beg forbearance. Of course, some of what has started, the drug dependence takes our grandparents learned to be true is not on a life of its own and becomes a deter- true, and the analysis of responsibility for mining influence for subsequent drug use negligent or criminal acts in the individual (i.e., drug use causes drug dependence, case (as in a court of law) necessarily has and then drug dependence causes drug a different set of standards of evidence. use). For example, evidence beyond a reasonable doubt is not the same as the defini- Our third clarification is to ask first whether tive evidence referenced in the first it is plausible that there is no association paragraph of this paper. between drug use and crime or criminal behavior or whether there might be an Our second clarification is that we acknowl- inverse association (the more crime, the edge a possibility that delinquent or crimi- less drug use). In our review of available nal behavior might be a cause of drug use, evidence, we must acknowledge the pos- the chicken-egg problem referenced by sibility that in some subsegments of Inciardi and advanced with evidence by human experience, there well may be a others. This possibility surfaces when one negative association between drug use considers earlier sociological models of and criminal behavior (e.g., in the highly There is a deviance (e.g., Sutherland, Matza) or later disciplined and controlled environments sociopsychological developmental models of industrial espionage), just as we must generally for youthful deviance, antisocial behavior, acknowledge the fact that some 90-year- consistent overall and delinquency, such as the coercive olds have smoked a pack or more of interaction and deviancy training models tobacco cigarettes virtually each day of pattern of positive introduced by Patterson, Dishion, and col- adult life and have not developed lung can- and sometimes leagues; Coie and his colleagues at Duke; cer. We also acknowledge the high proba- and Kaplan at Texas A&M University (e.g., bility that in certain times and places or in quite strong see Patterson, Reid, and Dishion, 1992; certain subsegments of population experi- associations Patterson, Dishion, and Yoerger, 2000; ence, there is no association between Dishion et al., 1996; Coie and Lenox, drug use and criminal behavior (e.g., see between illegal 1994; Sandstrom and Coie, 1999; Bagwell Blum and Associates’ studies of clinicians drug use and et al., 2000; Hubbard et al., 2001; Kaplan, and professionals who used LSD before 1995). For example, minor rule violations it was regulated by the Food and Drug criminal behavior in early childhood, well before the years of Administration; Blum and Associates, of other types. starting drug use, might be followed by 1964). general peer rejection, differential association or affiliation with other rejected and Notwithstanding these exceptional circum- deviant peers, and subsequent group- stances, there is a generally consistent fostered delinquency and norm violations, overall pattern of positive and sometimes including illegal drug use. We note that quite strong associations between illegal this possibility, and the more advanced drug use and criminal behavior of other

39 SPECIAL REPORT / JULY 03

types. These associations are observed the onset of that criminal behavior. Judg- not only in samples of offenders in the ments about this criterion or guideline can criminal justice system (e.g., DUF and become difficult when there are potential ADAM), but also in general household reciprocities. For example, when sus- population samples. This evidence has tained medicinal use of phenacetin and some vulnerability due to constraints on acetaminophen compounds (e.g., Tylenol, methods (e.g., refusals by study partici- Datril) was being investigated as a cause pants to give informed consent for partici- of interstitial nephritis and end-stage renal pation), but recent consistent evidence disease (ESRD), one of the complications from general population surveys indicates was the possibility that the earliest clinical that the observed association extends features of ESRD include headaches. Of beyond officially recognized crimes and course, headaches can promote the sus- does not suffer the transition bias that is tained use of pain-relieving medicines, present in DUF, ADAM, and other criminal including the acetaminophen compounds. justice samples (e.g., perhaps the arrested or incarcerated offenders were caught The drugs-crime relationship presents this because of impairments from drug use, or type of temporal complexity, as was seen the drug use of an offender is a manifesta- in exhibits 6 and 7. Earlier aggression, con- tion of a more general characteristic of duct problems, and criminal behavior may carelessness that might lead more readily function as a direct cause of illegal drug to apprehension by the authorities). use (e.g., see Kellam and Anthony, 1998), and possibly as an indirect cause (e.g., by Fourth, a “cloud of confusion” sometimes promoting affiliation with other delinquent descends when people begin talking about and drug-using peers). Earlier drug use causes and causation. We will try to be also may function to promote later growth clear. Although we are asking whether of conduct problems or criminal behavior drug use causes crime, we are not saying (e.g., see Johnson et al., 1995). that there are no other causes of crime. This issue sometimes is subject to misin- Criterion/guideline 2: Biological terpretation. For this reason, it might be or other theoretical plausibility more sensible to express the question in a different way: “Under what conditions, if Carrying books of matches is associated any, does criminal behavior, as a response with the risk of developing lung cancer, variable, depend in any substantive way tends to precede rather than follow the on drug use, such that we might be able onset of lung cancer, and has at least a to shape criminal behavior by shaping moderately strong association with lung drug use?” This question is not as pithy cancer. However, except with respect to as, “Does drug use cause crime?” but it the associated characteristic of tobacco might help us escape the cloud of confu- smoking, we have no biological or other sion when we try to review available evi- theoretical plausibility to link carrying dence pertinent to this issue of causal matches per se with the etiology of lung inference. cancer. Even if the matches-cancer association were to withstand the challenges Criterion/guideline 1: posed by the other criteria for evaluating Temporal relationship causal significance of an association, we would be inclined to ask about the under- If illegal drug use is believed to be a cause lying theory and its plausibility and coher- of criminal behavior, then we require evi- ence in relation to known relationships dence that illegal drug use has preceded and facts.

40 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

The tripartite model for the drugs-crime Licata et al. (1993) administered a high The tripartite nexus represents a substantiation of plau- dose, low dose, and no dose of cocaine sible causal links from illegal drug use to and found that subjects in the high-dose model for the criminal behavior. Other related strands in group expressed significantly greater drugs-crime nexus the fabric of plausibility have been men- aggression than subjects in the control tioned (e.g., differential crime opportunity, group; the low-dose group did not differ represents a differential association). from the control group. substantiation of The plausibility of a link between drug use Notwithstanding these strands of plausibil- plausible causal and aggressive or violent crimes rests to ity, there also is a considerable amount of links from illegal some extent on neuroscience theory and inconsistency in the observed data and observed clinico-pathological associations, some complexity in relation to dose- drug use to as in contemporary thinking about cocaine’s response analyses. For example, Crowley criminal behavior. influence on limbic-hypothalamic sub- et al. (1992) found no increase in aggres- strates of aggression (Davis, 1996). In sion when cocaine was administered in addition, there is a line of preclinical and primate lab research; Darmani and col- clinical laboratory experiments that has leagues (1990) found increased aggres- helped to solidify the plausibility of a link sion among mice that were given from drug use to aggressive or violent relatively low doses of cocaine but not behavior, and possibly to the types of when the mice were given higher doses norm violations associated with nonviolent of cocaine. Moro et al. (1997) found reduc- crime. The evidence on links from the use tions in the total number and length of of psychostimulant drugs (e.g., metham- aggressive activities in mice after amphet- phetamine, cocaine) and aggression is amine administration. Cherek et al. (1989), noteworthy in this respect. Administration studying humans, examined the relation- of cocaine to hamsters during adoles- ship between d-amphetamine on aggres- cence increased the number of bites and sion using point subtractions and found attacks indicative of a surge of offensive an increase in aggression among those aggression (Harrison et al., 2000). Moore receiving 10 mg per 70 kg of body weight and Thompson (1978), studying pigeons, but a decrease in aggression when 20 mg found that high doses of cocaine elicited per 70 kg of body weight was administered. aggressive behavior. In some species, increased levels of aggression also have Police experience on the street implicates been observed with the administration of dissociative drugs such as phencyclidine amphetamine stimulant drug—not only (PCP) in relation to violent and aggressive when a large single dose (e.g., Melega et behavior and crime. We have been able to al., 1997), but also after sustained lower find some supportive experimental labora- doses (Haber, Barchas, and Barchas, 1981) tory evidence consistent with this street- are administered. These psychostimulants wise experience (e.g., Burkhalter and also may increase the risk of self-directed Balster, 1979; McCardle and Fishbein, aggression (e.g., see Peffer-Smith et. al., 1989). Nevertheless, even with PCP, there 1983). is a complex pattern of inconsistent evidence that does not ring true with the Experimental laboratory research with experience on the street and common human subjects also has produced sup- wisdom about PCP. Tyler and Miczek portive evidence along these lines, often (1982), Emley and Hutchinson (1983), with computerized point-subtraction meth- and Miczek and Haney (1994) reported ods used to evoke aggression after the no increase in aggressive behavior after drug has been administered and under experimental administration of PCP and control (no drug) conditions. For example, an erratic increase in aggression only in a

41 SPECIAL REPORT / JULY 03

subgroup of animals receiving low doses. an individual who otherwise would be in- Hence, it may be that PCP promotes volved in criminal behavior. Despite exam- aggression only in certain subgroups of ples of this type, and notwithstanding the population (e.g., see McCardle and contrary evidence, the drugs-crime re- Fishbein, 1989); and in some experiments, search literature now includes a generally animals receiving high doses of PCP were consistent replication of positive associa- more likely to be victims of aggression by tions between illegal drug use and criminal nondrugged animals (e.g., see Russell, behavior (e.g., see Harrison and Backen- Greenberg, and Segal, 1984; Tyler and heimer, 1998). Miczek, 1982). The body of laboratory experiments on In sum, there is some plausibility to the drugs and aggressive or violent behavior Given the idea that drug use might promote criminal is not as consistent as one might expect. behavior, with strands of plausibility com- As described under criterion/guideline 2, multiplicity of ing from neuroscience theory, the com- under some circumstances, laboratory drugs, types of mon wisdom and experience of criminal experiments have established a small set justice officials and drug users, and lab- of drugs as causal agents in relation to crimes, and oratory experiments. The links between aggression and violence. However, for varieties of being a drug user and becoming a crime most drugs and many circumstances, victim represent an understudied phenom- there are negative findings, and the social contexts, enon, and the inconsistent patterns of lab- evidence is not consistent with causal it may be oratory evidence provoke us to investigate links from drug taking to aggressive and the possibility that there might be sub- violent behavior. inevitable that stantial heterogeneity within the popula- the accumulated tion with respect to links from drug use to Given the multiplicity of drugs, types of aggressive behavior or to crime (e.g., see crimes, and varieties of social contexts, body of evidence Parker and Rebhun, 1995). it may be inevitable that the accumulated body of evidence on the drugs-crime rela- on the drugs- History demonstrates one of the difficul- tionship appears inconsistent. Variation in crime relationship ties with this criterion for judging causal the quality of the research also has a bear- significance of associations. Time and time ing on consistencies or inconsistencies in appears again, new evidence has contradicted the evidence. As every first-year graduate inconsistent. what appeared to be a biologically plausi- student learns, research with imprecise ble or theoretically pleasing link between measurements will tend to yield evidence a suspected cause and a suspected re- of no relationship even when a relation- sponse. Today’s biologically plausible or ship exists; research with measurements theory-driven causal inference may be of limited validity will tend to yield evi- tomorrow’s “old wives’ tale.” As is true dence of relationships where none exists. for the other criteria and guidelines, by itself this one counts for little. Although not generally introduced as a feature of studying consistency of relationships between causes and effects, a Criterion/guideline 3: developmental perspective may help to Consistency of the association lead the reader to a greater appreciation of We already have clarified the possibility inconsistencies and complexities faced in that no association or a negative associa- research on the drugs-crime relationship. tion might exist for certain subsegments That is, the timing of the onset of the drug of population experience. For example, at taking may condition the later expression some point, drug taking may incapacitate of criminal behavior and may lead to

42 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

greater heterogeneity in the population regression, we are seeking to hold con- with respect to the drugs-crime relation- stant the baseline level of conduct prob- ship. For example, we have some evi- lems and study boys who have a high dence on the possibility that earlier-onset initial level of conduct problems but who drug use is associated with later risk of start drinking alcohol early on and com- developing drug problems (e.g., see pared them with boys with an equally high Anthony and Petronis, 1995). We also initial level of conduct problems but for have evidence that prompts us to concep- whom alcohol consumption is delayed tualize earlier-onset drug use as a type of until adolescence. Approaching the con- precocious adolescent development that trast in this manner, we may discover that may disrupt normative developmental early-onset drinking dampens the growth trajectories (e.g., see Newcomb, 1992; trajectory for conduct problems; the steep- Dawes et al., 2000). There may be a ten- est trajectory for growth of conduct prob- dency to interpret these disruptions as lems may be observed for boys with high sources of increased levels of later crimi- initial levels of conduct problems but with- nal behavior, consistent with the idea that out the impairments associated with early risk of drug problems are increased for drinking. The early drinking might lead early-onset drug users; this has been the to retardation in the growth of conduct perspective our research group has taken problems for boys who otherwise would in its studies of this topic (e.g., Johnson escalate to very high levels of conduct et al., 1995; Anthony and Petronis, 1995). problems in adolescence. Nonetheless, it is possible that precocious (i.e., early onset) drug taking is followed This is a somewhat counterintuitive propo- by disproportionately greater increases in sition, and it may run counter to common frequency of drug use and in risk of drug wisdom and experience with respect to problems but that the early-onset drug use the effects of early-onset drinking or drug dampens the level of criminal behavior use and the later lifecourse of young peo- that otherwise might occur if the drug use ple. However, our intuitions and common had not started so early. experience about these circumstances tend to reflect a type of population- Our study of early-onset alcohol use and averaged summary of developmental the later developmental trajectory of con- trajectories and generally do not encom- duct problems represents a case in point. pass all varieties of human experience. In that study, cited above under criterion/ We mention this open research question guideline 1 (Johnson et al., 1995), we as an example of the complexities faced found that baseline levels of conduct prob- in developmental research on the drugs- lems were greater for boys who had start- crime relationship and as a possible expla- ed drinking alcohol before the adolescent nation for the inconsistencies observed in years without parental permission and that drugs-crime research. The timing of the growth of conduct problems was greater drug use may induce subgroup variation for these early-onset alcohol users—when in the drugs-crime relationship, which compared with boys whose drinking did then is interpreted as inconsistency in not start until later. Similar relationships and a challenge to causal significance of were observed for girls with early-onset the observed associations. alcohol use—when compared with girls whose drinking did not start until later. Fortunately, complete consistency of evi- However, a discussion of this research dence is not required. What is required is with Blumstein has prompted us to re- a focused probing of the circumstances approach this problem with a different under which the drugs-crime relationship comparison in mind. Using random effects is a causal relationship, with a deliberate

43 SPECIAL REPORT / JULY 03

effort to ferret out situations in which there The challenge for the courageous skepti- is no causal linkage between drug use and cal critic is to assert a specific background criminal behavior. Deliberate scientific pur- factor or set of background factors that suit of these circumstances and situations might account for the observed A-B rela- may require investigators to look overseas, tionship and that have not been consid- where use of such drugs as marijuana, ered explicitly or taken into account in a cocaine, and heroin are not treated as study plan or description of completed criminal behaviors. In social contexts of work. For example, observing the suspect- this type, by studying the developmental ed causal association between tobacco trajectories of criminal behavior among smoking and risk of lung cancer, the statis- young people with and without early drug- tician Sir Ronald Fisher posed a question taking experiences, we may be able to illu- of the following type: “Can’t we explain minate some of the inconsistencies now the observed association as a manifesta- observed in the drugs-crime evidence tion of an underlying predisposition or lia- available to us. For example, longitudinal bility that determines both the tobacco studies of children growing up in the smoking and the lung cancer?” Netherlands are underway. The recent effective decriminalization of marijuana use In relation to the drugs-crime relationship, in the Netherlands creates a social context the most plausible background factors for research on this drug and later criminal seem to be of the variety named by Fisher, behavior that merits attention on the NIJ- namely, unmeasured predispositions; in NIDA research agenda. this instance, the predispositions might involve who abides by the conventional rules of society, who is willing to run afoul Criterion/guideline 4: Alternative of the law by taking a drug illegally, and explanations ruled out who is willing to commit crimes other than This criterion or guideline represents the crime of drug possession for personal the Achilles heel for much of the prior use. To some extent, these predisposi- research on the possible causal links be- tions may be a manifestation of family her- tween illegal drug use and criminal behav- itage, a manifestation of early experiential ior and represents a general difficulty for conditions and processes, or a synthesis observational research in general. Observ- of both. Nevertheless, no matter what ing a possible causal relationship between their origin, until these predispositions are antecedent A and response B, the skepti- taken into account, they represent a plau- cal critic always can ask, “Isn’t there some sible form of alternative explanation when- unrecognized background factor C that ever a drugs-crime relationship is found in can account for the A-B relationship that our empirical studies. you have observed in this study?” If so, “Isn’t this a poorly developed conceptual One line of response to this criticism has model?” been to measure personality or facets of temperament in observational studies and To some extent, these are a coward’s to re-estimate the drugs-crime association questions about the drugs-crime relation- with personality or temperament held con- ship in specific and about empirical re- stant (e.g., via stratification or statistical search in general. Of course, there might adjustment in a regression model). But be some unrecognized background factor this response always is subject to the criti- in empirical research plans and in complet- cism that the wrong facets of personality ed studies; if not the hand of the mischie- or temperament were measured or that vous Norse god Loki, then something else the measurement of personality or tem- of a less celestial nature. perament was not as good as it should have been.

44 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

It is in relation to this criterion that we Longitudinal subjects-as-their-own-controls now have new opportunities for research research designs also can help rule out at the intersection of public health and alternative explanations in the sense that criminal justice research on the drugs- each individual participant is carrying for- crime relationship. Three important op- ward a within-individual set of propensities portunities at this intersection involve (1) to become engaged in illegal drug use and genetics, twin, and family research; (2) other criminal offending. In these longitu- longitudinal studies with “subjects as their dinal designs, in an otherwise law-abiding own controls” designs; and (3) controlled individual, if we were to observe that crim- experimental trials. inal offending occurs only in the immediate aftermath of a drug intoxication Future genetics, twin, and family studies experience or only in the stages of with- can help to narrow the alternative explana- drawal after drug dependence, we would tions in a useful manner. For example, in have additional evidence of a drugs-crime an earlier section we described a design association at the individual level. These that exploits the genetic matching of longitudinal designs remain vulnerable to monozygotic twins to search for environ- a possible counterclaim that there is an mental conditions that contribute to the underlying predisposition that links earlier occurrence of disease. Discordant MZ illegal drug use to later criminal offending twin designs also can be used to hold only during the context of drug intoxica- constant predispositions or liabilities linked tion or withdrawal states. That is, the to the individual genome of the twins, observed association between illegal drug while studying differences in the trajectory intoxication or withdrawal and the later of criminal behavior for the MZ twin criminal offending is a spurious artifact of whose illegal drug use starts first versus uncontrolled confounding: There is some- the MZ twin whose illegal drug use starts thing else in the background, a third vari- later (or not at all). able that explains the observed sequence.

Alternative twin and family research de- Medical and public health research is signs can be used to narrow other expla- host to a variety of examples of this type nations of the observed drugs-crime of spurious confounding. One of them relationship (e.g., studies of discordant involves the connection between chicken- siblings, studies based on the transmis- pox and shingles. For most people, chick- sion disequilibrium test when specific enpox occurs early in life and shingles polymorphisms are under investigations). occurs late in life. There sometimes is an Cadoret and colleagues have offered exceptional case of shingles occurring with recent illustrations of the power of twin no prior history of chickenpox in childhood, studies in which some twins have been but these exceptional cases might be separated at birth, but these “natural understood as instances of “clinically inap- experiments” have become scarce in the parent” infections (i.e., with mild or mini- United States and other parts of the world mal symptoms in childhood, so mild as to where twins now generally are kept to- pass without notice). A longitudinal re- gether in their new adoptive families (e.g., search design on this topic can lead to the see Cadoret et al., 1986, 1995; Cadoret, impression that chickenpox causes shin- Leve, and Devor, 1997). Tsuang et al. gles, in the sense that shingles rarely or (2001) provide a recent useful overview of never occurs unless chickenpox occurs pertinent findings from the Harvard Twin first. This observed longitudinal link be- Study. tween chickenpox and shingles satisfies

45 SPECIAL REPORT / JULY 03

the requirement described above: Criminal described earlier, we sought to test this behavior occurs only after a bout of illegal hypothesis by constructing an experimen- drug use. The fly in the ointment in our tal trial in which we disrupted the develop- chickenpox-shingles example is that we ment of early aggression and rule breaking now know that chickenpox does not cause (e.g., Kellam and Anthony, 1998). We used shingles. Rather, it is an underlying virus the power of randomization to hold con- that causes both of these clinical phenom- stant the profile of alternative explanations ena. Exposure to the chickenpox virus (her- that might account for later illegal drug use pes zoster) is the cause of chickenpox in and criminal behavior. In a current fol- childhood and is the cause of shingles in lowup study of the youths who participat- later life when the virus emerges from an ing in this trial, we will be testing whether otherwise dormant or latent state of no the primary school intervention had a sus- activity. The apparent linkage from earlier tained impact on illegal drug use and crimi- chickenpox to later shingles is due to an nal behavior. If so, we might expect a underlying third variable, the herpes zoster weakened association between illegal infection, which accounts for the appear- drug use and criminal behavior in the sub- ance of both outcomes. group of youths exposed to the active behavioral intervention arms of our study. The analogy to research on illegal drug use and later criminal offending should be A related opportunity to test the drugs- clear. Even when longitudinal research crime relationship and to use randomiza- shows us examples of participants who tion to rule out alternative explanations become engaged in criminal behavior only involves controlled trials of new therapeu- in the context of drug intoxication or with- tic interventions directed toward illegal drawal states, we cannot be confident that drug use and drug dependence of adoles- the illegal drug use is the cause of the cents. Observational studies now suggest associated criminal offending. Some un- that entry into drug treatment reduces the known underlying cause may be account- rate of criminal offending, but these stud- ing for both outcomes. ies leave open possibilities for alternative explanations (e.g., selection biases in the The third approach, involving randomized assignment of subjects to treatment, im- trials, offers a way to bring these unknown balances in the other determinants of underlying variables into check. This ap- criminal offending). Randomization in the proach already has been described in rela- setting of controlled trials of new thera- tion to our research group’s studies of an peutic interventions creates an opportuni- alternative explanation for the drugs-crime ty to constrain these selection biases and relationship. Namely, we advanced the bring into balance the alternative sources hypothesis that a predisposing characteris- of variation in criminal offending (e.g., see tic in the form of early aggression or rule Manski et al., 2001). breaking is a potentially modifiable determinant of both later illegal drug use and By adding followup measurements of criminal behavior or other sorts. This posttreatment criminal behavior to current hypothesis does not reject the possibility and newly emerging randomized con- that illegal drug use causes later criminal trolled trials of therapeutic interventions, behavior, but it introduces one alternative NIJ and NIDA can help foster new evi- explanation for the observations associa- dence on the degree to which illegal drug tion between illegal drug use and criminal use is a cause of criminal offending. Alter- behavior (i.e., the earlier aggression or ten- native explanations for the observed dency to break rules and social norms). As drugs-crime association and other determinants of the offending behavior can

46 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

either be brought into balance by random- b. What about the reverse causal pathway ization or held constant as measured and the possibility that increasing mari- covariates in statistical models of analysis. juana use might promote later increases Some examples of past research along in unconventionality? these lines are described under criterion/ guideline 7. c. As levels of marijuana use increase, are there later dose-response or gradient- like increases in unconventionality? Criterion/guideline 5: Dose-response or gradient In light of the population heterogeneity relationship mentioned above, this dose-response cri- Absence of a dose-response or gradient terion might be especially troublesome in relationship does not rule out causal as- research on the drugs-crime relationship. sociations; there are good examples of For example, consider the drug user whose threshold relationships with no clear gradi- increasing bouts of intoxication yield less ent. Nonetheless, there are examples in criminal behavior than otherwise might which the probability or rate of criminal occur and whose intoxication-associated behavior is observed to be lower with carelessness leads to apprehension and lower levels or frequencies of illegal drug detoxification and outpatient treatment use and is observed to be greater as lev- prior to a bench appearance. The detoxifi- els or frequencies increase. cation and treatment might be followed by a return to the baseline level of criminal In one recent and especially informative behavior (i.e., a higher level of criminal longitudinal cohort study, Brook et al. behavior than was observed during the (2001) studied the developmental trajecto- period of intoxication) and an impression ry of marijuana use from childhood into that treatment was ineffectual with adulthood and found that behavioral and respect to the frequency of criminal attitudinal indicators of unconventionality behavior. (e.g., attitudes tolerant of norm violations) had a gradient relationship with later in- Criterion/guideline 6: creases in marijuana involvement. The Strength of association research team also found that as levels of unconventionality increased, so did Weak associations seem especially vulner- marijuana involvement. These gradient able to sources of spuriousness and bias. relationships between unconventionality One benchmark standard for strength is and marijuana use help to substantiate a the association between tobacco smoking possible causal link between earlier un- and lung cancer: The risk of dying from conventionality and later developmental lung cancer is estimated to be 10 times trajectories of marijuana involvement. or greater for persistent tobacco smokers However, as in the circumstance of re- than for nonsmokers. Toward the other search on the drugs-crime relationship, end of the spectrum of magnitude is a this research report leaves us with unan- widely appreciated but quite modest swered questions of the following variety: strength of association between being male and illegal use of drugs: The risk of a. What about the predisposition that becoming an illegal drug user is an esti- links unconventionality to the earliest mated 1.5 to 3.0 times greater for an marijuana use? Where does the un- American male than for an American conventionality come from, and is this female (Anthony and Helzer, 1995). predisposition the same as the predisposition to smoke marijuana?

47 SPECIAL REPORT / JULY 03

When a positive Examining the range of study estimates Reprise: Does drug use cause on the drugs-crime relationship, there are crime? What do we not know? relationship is some studies with extremely large rela- observed, the tionships, but when a positive relationship This review of a specific hypothesized is observed, the strength of relationship causal relationship was intended to high- strength of tends to be quite modest. This generally light some of what we know about the relationship modest relationship may imply that alter- drugs-crime relationship. Its main purpose native explanations (e.g., predispositions) was to provoke discussion and help in a tends to be are sufficient to account for the observed process of identifying weaknesses and relationship. gaps in evidence that might be used to quite modest. guide a future research agenda.

Criterion/guideline 7: Evaluated in relation to these conventional Cessation effects criteria or guidelines for judging the causal significance of observed associations, the Cessation effects already have been men- reader may have a better appreciation for tioned in the context of our discussion of the uncertainty conveyed in a recent sum- alternative explanations under criterion/ mary statement cited above: “Research guideline 4. There are many studies of co- has not been able to validate a causal link occurring maturation processes that lead between drug use and criminal behavior” to fading of both illegal drug use and other (Harrison and Backenheimer, 1998). The criminal behavior, especially since the available evidence is ambiguous with re- work of Winick. The observational studies spect to temporal relationships. of McGlothlin, Anglin, and Hser in Cali- fornia and the work of Nurco, Lerner, and Instead, we offer a series of discussion colleagues in Baltimore also shed light on points about what we might not yet know. declines in criminal behavior during periods of abstinence or reduced illegal drug Is the evidence on a temporal relation- use. The literature includes numerous ship compelling? Illegal drug use pre- studies of what has happened to crime cedes formal criminal behavior in some of involvement after cessation of drug use, these studies, but what about the earlier based on observational studies. antecedents of both drug use and crime in the form of rule breaking, misbehavior, As noted under criterion/guideline 4, some and minor norm violations? One can imag- of the strongest evidence about cessation ine a co-occurrence process that begins effects can come from randomized experi- with expression of irritable temperament ments in which drug treatment or other or aggression in the preschool years, fol- interventions are used to disrupt illegal lowed by rule breaking or norm violations drug use, with subsequent evaluation of in the primary school years, and then later crime as an outcome of treatment. As co-occurrence of illegal drug use and delin- noted under criterion/guideline 6, for some quent or criminal offending. Our own re- segments of the drug-using population, search group and others have added some the cessation of drug use is followed by evidence on the possibility that drug tak- increases in frequency of criminal behavior ing that starts by age 11 might promote (i.e., once impairments associated with growth trajectories for later conduct prob- intoxication are reduced). lems among both boys and girls. The pattern of co-occurrence of conduct problems

48 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

and drug use is a centerpiece of Jessor’s Cadoret and his colleagues, some with problem behavior theory, and there is rea- a sharper focus on mechanisms of inheri- son to look to experiments that will help tance (e.g., assays of genetic polymor- us differentiate these problem behaviors phisms), and some with a focus on (e.g., differential response of each form personality and early social environment. of problem behavior to different interven- It is not clear that studies to date have tions, as suggested in Dishion’s early provided adequate control over these Adolescent Transitions experiment). sources of co-variation. Nonetheless, the longitudinal study of individuals over time Plausibility? Our focus has been oriented has provided evidence from subjects-as- toward the individual, but there is a per- their-own-controls designs, and the ran- spective on the drugs-crime relationship domized trials of interventions provide that is more ecological or contextual in some evidence that, despite common orientation. For example, a social environ- causes, an intervention directed toward ment characterized by illegal drug use of illegal drug use can reduce frequency of individuals might give rise to norm viola- criminal behavior. Even if there are com- tions and criminal offending of other sorts, mon causes (e.g., inherited traits), for and not necessarily in the form of offend- many observers, the longitudinal evidence ing by the drug users but rather in the coupled with experimental evidence is form of offending by others. The mugging sufficient to draw the inference that illegal of a heavily intoxicated drug user by a drug use causes criminal behavior. Rea- group of nonusing passersby serves as sonable people will disagree about this one example of aggregate effects of illegal inference from available observations, and drug use on crime that would not be the points of disagreement will lead us apparent in individual-oriented studies but to specific experiments or new studies would require multilevel studies of interre- to gather evidence that will be more lationships between individuals. compelling.

Consistency? What about the exceptions Gradient? Is it possible that some of the to a general pattern of observations? It inconsistency in observations about the seems likely that the drugs-crime associa- drugs-crime relationship can be traced to tion varies from time to time, place to (a) selective attention either to the lower place, and subgroup to subgroup. The end of drug involvement (e.g., among chil- study of variation in these patterns of asso- dren, adolescents, or high school seniors ciation will help to disclose the boundary followed through the college years; see conditions and mechanisms that give Schulenberg et al., 1994) or to the higher rise to strong, weak, and possibly inverse end (e.g., among arrestees or clients in associations. Research across borders drug treatment programs); (b) possible and in settings such as the Amsterdam thresholds in the gradient relationship, cannabis environment can help illuminate with between-sample heterogeneity with these boundary conditions. respect to the effective threshold; or (c) an uncertain metric for assessing the type or Alternative explanations? Several lines level of drug involvement? As described in of research have been started on the com- the prior section entitled “Criterion/guide- mon causes for both illegal drug use and line 2: Biological or other theoretical plausi- other criminal behavior, some of them bility,” we have noted some inconsistent originating in family history studies and pharmacological effects across dosage the clever adoption paradigm adapted by

49 SPECIAL REPORT / JULY 03

levels of the same drug and across differ- Conclusion ent drugs. If we are to appropriately sort the consistent and inconsistent findings In the final section of this working paper, of field studies on the drugs-crime rela- Iwould like to integrate the organizing tionships, it may be necessary to reach for conceptual framework presented in the greater specificity with respect to dosage section on the rubrics with the ecological levels or intensity of drug use and also concept of scale described previously. with respect to the pharmacological differ- Here, there is an adaptation of the formal ences observed in laboratory experiments. ecological concept of scale that includes It no longer is enough to sort drugs into the microcosm and an extension of the the non-scientific colloquial “soft” and concept that reaches to the macrocosm of “hard” categories, nor to lump all “illegal the international regulatory environment. drug use” as if there were no hetero- geneities of effect across the various The integration of the five rubrics and the forms of internationally regulated drugs. concept of scale is depicted in exhibit 9. The best field studies of the 21st century The result is a two-dimensional grid with will abandon these relatively crude clas- the rubrics on one axis and scale on the sifications and will not carry forward an other axis and showing the conceptual obsolete tradition from the earlier ground- domain where research on drugs and re- breaking days of drugs-crime research. search on criminal offending intersect. Each rubric-scale intersection or subunit Strength of association? Due to uncer- in the grid can be populated by past and tainties about reciprocal and dynamic inter- current examples of research on the relationships between drugs and crime, it drugs-crime relationship. In some sub- would be advantageous to look closely at units, density of past and current research studies with fine-grained temporal analysis is quite high; work in these domains may of the drugs-crime relationship and to esti- require strengthening, or perhaps these mate strength of association prospective- investigators should be left alone to do ly. This should be done in a manner that their work. In other subunits, we have allows change in the level of criminal be- done little or no past research activity; havior to be gauged in relation to change these subunits might warrant attention in in the level of drug use and vice versa, or a new agenda for drugs-crime research. with an expression of the relative risk of criminal acts with and without antecedent Starting in the upper left-hand corner of illegal drug use. this framework, we have the intersection of quantity research with the microcosm Cessation? Our recent National Research represented by the genes we inherit from Council committee expressed concern our forebears. We may expect one day to that selection effects, transition biases, have an investigation that produces quan- or other artifacts might lead to a spurious titative estimates of the frequency of inference that criminal behavior declines homozygotes and heterozygotes with or stops when illicit drug use is ended, respect to genes that are implicated in the either with or without intervention. The drugs-crime relationship, just as we now evidence on this criterion might require have these estimates for the frequency of special scrutiny in light of concerns such alleles mapped to apolipoprotein E4 and as these. other genes or polymorphisms implicated in the risk of developing Alzheimer’s disease.

50 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Exhibit 9. A conceptual framework for research on the drugs-crime relationship

Scale from microcosm to macrocosm

Genes & simple Nations & gene Individual Social global The products organisms groups regions main rubrics ABCDE F G H

1. Quantity

2. Location

3. Causes

4. Mechanism

5. Prevention & control

Working our way to the far upper right- investigacao/espad99/indice.htm); and hand corner, we stop at the level of cross-national studies supported by NIDA Nations. To the best of my knowledge, in Latin America (e.g., Brook et al., 2001), we have a limited set of quantitative esti- including our own PACARDO Project mates for rates of drug-taking behaviors (Anthony, 2000). and criminal justice statistics at the level of Nations; but definitive evidence on vari- In the middle range, between the micro- ation across regions of the globe is lacking cosm of the gene and the macrocosm of and represents a current gap in knowl- global regions, we have collections of esti- edge. To some extent, this gap can be mates for various social and geopolitical filled by cross-national studies now under- groups. In aggregate, these estimates can way, such as the World Health Organiza- help us to draw generalizations about the tion’s (WHO’s) recent World Mental Health relative magnitude of problems associated 2000 research initiative being led by Ron with drug taking of one sort or another Kessler at Harvard and T. Bedirhan Ustun (e.g., marijuana use versus cocaine use) at WHO, with collaborators in more than or with criminal behavior of one sort of 20 countries around the globe (Kessler, another (e.g., aggravated assault versus 1999); the European School Survey Pro- shoplifting or vandalism). ject on Alcohol and Drugs (www.ipdt.pt/

51 SPECIAL REPORT / JULY 03

A few Even within the rubric of quantity, there In a middle position in this framework, are many gaps. For example, our quantita- somewhat overlapping the different investigators tive estimates often are based strictly on segments, the important line of research have started officially recognized offending and do not being conducted by Higley, Suomi, and encompass unrecognized offending. With their colleagues in relation to gene- to integrate respect to drug taking, there is a plethora environment interactions merits attention. genetic variation of evidence on the prevalence of drug use This research, already mentioned in one and drug dependence but not much evi- of the preceding sections, touches on in their studies dence on the incidence or risk of becom- aggressive behavior, social conditions of on such topics as ing drug dependent. Here, also, we have child rearing, and drug use. Using a pri- big gaps in the evidence that warrant mate model, this research group has been drugs and crime. some attention as we design an agenda able to extend the line of research on for future research. infant-mother relations that Harlow initiated. The group is engaged in experimental The intersections of the location, causes, manipulation of the early conditions of and mechanisms rubrics with the scale infant rearing, crossed with genetic predis- dimension brings us closer to evidence on positions that in the wild have been found variation from place to place, time to time, to be related to aggressive behavior and or in relation to personal characteristics. A excess mortality. The evidence from this few investigators have started to integrate research serves as an important example genetic variation in their studies on such of how the effects of an apparently nox- topics as drugs and crime, and soon we ious inherited predisposition might be may have more definitive evidence on the modulated by a change in child-rearing relative frequency of different polymor- environments. Does this animal model of phisms or gene-encoded protein products gene-environment interaction also hold for for different subgroups of the population aggressive children, with later implications or in different geopolitical zones. We for their drug-taking behavior? Questions can expect ecological analyses of the such as this one merit discussion in rela- between-subgroup and between-zone tion to the proposed drugs-crime research rates, with new evidence on location. agenda, if only to choose not to pursue these lines of research. Similarly, working outward from the simplest gene products to more complex Turning to the last row of the framework, products of gene-environment interaction, Ioffer some speculations about gaps in the sex hormones research of separate research on prevention and control. At the research groups led by Logan and by level of scale that reaches from micro- Angold, Costello, and others should pro- cosm to the whole organism, I see a gap vide us with more evidence on rates of in research on underlying brain structures antisocial behavior, drug use, and offend- that subserve neuropsychological func- ing in relation to levels of testosterone and tioning of clear importance in the choice other hormones before and after drug use behavior of drug users and offenders. To (Logan, 2001; Federman et al., 1997). The the extent that drug users and offenders initial evidence is not expected to allow are making choices about various ele- causal inference, but the understanding of ments in their behavioral repertoires, we locational variation will allow us to sharpen may be able to understand variations in our causal theories and to integrate new response to prevention and control inter- biological perspectives on the drugs-crime ventions as a function of neuropsycholo- relationship. gical performance (e.g., with respect to direction, control, and planning). Our

52 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

comprehension of this variation can more than one level of scale. These increase through a program of research attempts deserve encouragement. on fMRI brain imaging and neuropsychological testing under experimentally con- Before closing, we must turn to the empty trolled paradigms (e.g., aggression evoked spaces created in the circle but not includ- through computerized point subtraction or ed as part of the two-dimensional grid. other procedures). In time, we should be Within these spaces, we have important able to evaluate the degree to which re- drugs-crime research that does not fall sponse to these interventions depends on neatly into the two-dimensional conceptu- brain structure and function as manifest in al framework. I am thinking of the recent neuropsychological tests as well as in re- ethnographic studies of the gangs in- sponse to genetic predispositions of the volved in drug sales (e.g., Levitt and type now being characterized in Suomi’s Venkatesh, 1998), and some of the other primate laboratory and elsewhere. recent innovative qualitative research on drug trafficking (e.g., Natarajan, Clarke, Working our way to the bottom right-hand and Belanger, 1996; Natarajan and Be- side of this matrix, we find the intersec- langer, 1998), which shed new light on the tions with social groups and contexts of structure and organization of the criminal increasingly larger scale, not only the peer organizations that sustain drug supply and group and family of origin or procreation, influence drug-related criminal offending but also the larger neighborhood, employ- around the world. There also are good ment context, the community at large, and recent examples of operations research across national boundaries. As we plot focused on the organization and adminis- examples of intervention research in this tration of criminal justice agencies and the two-dimensional framework, it is easier to deployment of law enforcement, prosecu- find examples of individual investigations tion, and judicial resources (e.g., see with narrow breadth of scale. For example, Maltz, 1996). To the extent that these One of the we can find an intervention focused on the investigations guide us toward useful evi- challenges for community but without elements of inter- dence about prevention and control, and vention directed toward specific individuals to the extent that they focus on individuals future research in the community. We can find many in- or small groups of individuals (e.g., in a on drugs and terventions directed toward individual city or State), they may be placed in the arrestees but not toward the social groups space on the left-hand side of the figure, crime will be of which the arrestees are members. between the grid and the surrounding to encourage circle. To the extent that these control- One of the challenges for those who seek oriented investigations are directed to- broadband to shape the future research agenda on ward international drug trafficking (e.g., research that drugs and crime will be to encourage see Montagne, 1990), they may be placed broadband research that cuts across multi- in the space on the right-hand side of the cuts across ple levels of scale. This is not to say that figure. multiple levels we should eliminate narrow-band research because it often is necessary to solve a There are other research programs and of scale. research problem through focus, and initiatives that do not fall neatly within the focus is one of the defining characteristics two-dimensional grid presented in exhibit of narrow-band research. Nonetheless, as 9. Methodological research constitutes we look over some of the more exciting one set of examples (e.g., Wish et al., research projects now underway, we can 2000; Harrison, 1997, 2001; Fendrich, see that the excitement is coming from 2001). Proposed new research on drug the investigators’ attempts to encompass prices and a consumer product index for

53 SPECIAL REPORT / JULY 03

illegal drugs represents another set (e.g., Anthony, J.A., and Bradshaw, G.A. (2001). see Manski et al., 2001). Wavelet analysis as an approach to investigate the reciprocal relationship between As we move toward a new drugs-crime ecological pattern and process. The prem- research agenda for NIJ and NIDA at the ises and problems with spatial analysis intersection of public health and criminal in landscape ecology. Paper presented justice studies, it is important to remem- at Pattern, Process, Scale, & Hierarchy: ber the two major themes mentioned in Interactions in Human-Dominated and the introduction to this paper: Natural Landscapes, the 16th Annual Symposium of the U.S. Chapter of In- ■ There is no single drugs-crime relation- ternational Association of Landscape ship. Rather, there are drugs-crime rela- Ecology, Memorial Union, Arizona State tionships, most of which are complex University, Tempe, Arizona, April 25–29. rather than simple. Anthony, J.C. (2000). Cross-national stud- ■ There is no simple solution to the com- ies on clusters of drug use. NIH grant plex challenges faced when drugs-crime award 5 R01 DA10502–02. relationships come into play. Anthony, J.C., and Helzer, J.E. (1995). Epi- The two-dimensional grid encircled in demiology of drug dependence. In M.T. exhibit 9 offers no simple solutions to the Tsuang, M. Tohen, and G.E.P. Zahner (eds.), complex challenges faced when drugs- Psychiatric epidemiology (pp. 361–406). crime relationships come into play. That New York: Wiley & Sons. grid is only a tool that may help us identify important gaps in the research evidence, Anthony, J.C., Neumark, Y., and Van Etten, gaps that must be filled as we work to- M.L. (2000). Do I do what I say? A per- ward a more complete understanding of spective on self-report methods in drug the drugs-crime relationship and more dependence epidemiology. In A.A. Stone, effective action plans that apply new J.S. Turkkan, C.A. Bachrach, J.B. Jobe, understanding in the service of public H.S. Kurtzman, and V.S. Cain (eds.), The health and safety. In an important sense, science of self-report: Implications for the empty spaces encircling the two- research and practice. Mahwah, NJ: dimensional grid also can be useful tools Lawrence C. Erlbaum Publishers. as we try to identify and fill the gaps in evidence. These empty spaces can serve to Anthony, J.C., and Petronis, K.R. (1995). remind us that no conceptual framework Early-onset drug use and risk of later drug is all encompassing. We must “think out- problems. Drug and Alcohol Dependence, side the box” in this regard. If we organize 40(1), 9–5. our scientific resources simply to continue our current lines of research, we will not Anthony, J.C., and Van Etten, M.L. (1998). achieve lasting reductions in illegal drug Epidemiology and its rubrics. In A. Bellack use and drug-related crimes, and we will and M. Hersen (eds.), Comprehensive clin- never lay claim to great victories in the ical psychology (vol. 1, pp. 355–390). service of public health and safety. Oxford, UK: Elsevier. Bagwell, C.L., Coie, J.D., Terry, R.A., and Lochman, J.E. (2000). Peer clique partici- References pation and social status in preadoles- Agar, M. (1973). Ripping and running. New cence. Merrill-Palmer Quarterly, 46(2), York: Seminar Press. 280–305.

54 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Behrens, D.A., Caulkins, J.P., Tragler, G., Brook, J.S., Brook, D.W., Arencibia- and Feichtinger, G. (2000). Optimal control Mireles, O., Richter, L., and Whiteman, of drug epidemics: Prevent and treat— M. (2001). Risk factors for adolescent But not at the same time. Management marijuana use across cultures and across Science, 46(3), 333–347. time. Journal of Genetic Psychology, 162(3), 357–374. Behrens, D.A., Caulkins, J.P., Tragler, G., Haunschmied, J.L., and Feichtinger, G. Brook, J.S., Tseng, L.J., and Cohen, P. (1999). A dynamic model of drug initiation: (1996). Toddler adjustment: Impact of Implications for treatment and drug con- parents’ drug use, personality, and parent- trol. Mathematical Biosciences, 159(1), child relations. Journal of Genetic Psycho- 1–20. logy, 157(3), 281–295.

Blum, R., and Associates (1964). Utopi- Brook, J.S., Whiteman. M., Finch, S.J., ates: The use and users of LSD 25. New and Cohen, P. (1996). Young adult drug York: Atherton Press. use and delinquency: Childhood antecedents and adolescent mediators. Journal Blumstein, A., Cohen J., Roth, J.A., and of the American Academy of Child Ado- Visher, C.A. (eds.), and the National Re- lescent Psychiatry, 35(12), 1584–1592. search Council (1986). Criminal careers and career criminals. Washington, DC: Brook, J.S., Whiteman, M., Finch, S.J., National Academy Press. Morojele, N.K., and Cohen, P. (2000). In- dividual latent growth curves in the devel- Bobashev, G., and Anthony, J.C. (1998). opment of marijuana use from childhood Clusters of marijuana use in the United to young adulthood. Journal of Behavioral States. American Journal of Epidemiology, Medicine, 23(5), 451–464. 148(12), 1168–1174. Brown, J.H. (1995). Macroecology. Bobashev, G., and Anthony, J.C. (2000). Chicago: University of Chicago Press. Use of alternating logistic regression in studies of drug use clustering. Substance Brownstein, H.H., and Goldstein, P.J. Use & Misuse, 35(6–8), 1051–1073. (1990). Research and the development of public policy: the case of drugs and violent Brennan, P.A., Mednick, S.A., and Jacob- crime. Journal of Applied Sociology, 7, sen, B. (1996). Assessing the role of genet- 77–92. ics in crime using adoption cohorts. CIBA Foundation Symposium, 194, 115–123. Burkhalter, J.E., and Balster, R. (1979). Effects of phencyclidine on isolation- Bronfenbrenner, U. (1979). The ecology induced aggression in mice. Psychological of human development: Experiments by Reports, 45(2), 571–576. nature and design. Cambridge, MA: Harvard University Press. Cadoret, R.J., Leve, L.D., and Devor, E. (1997). Genetics of aggressive and violent Bronfenbrenner, U. (1986). Ecology of behavior. Psychiatric Clinics of North the family as a context for human develop- America, 20(2), 301–322. ment: Research perspectives. Develop- mental Psychology, 22, 723–742. 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.

55 SPECIAL REPORT / JULY 03

Cadoret, R.J., Yates, W.R., Troughton, E., Csikzentmihalyi, M., and Larson, R. (1992). Woodworth, G., and Stewart, M.A. (1995). Validity and reliability of the experience Adoption study demonstrating two genet- sampling method. In M.W. De Vries (ed.), ic pathways to drug abuse. Archives of The experience of psychopathology: In- General Psychiatry, 52(1), 42–52. vestigating mental disorders in their natural settings. New York: Cambridge Caulkins, J., Crawford, G., and Reuter, P. University Press. (1993). Simulation of adaptive response: A model of drug interdiction. Mathematical Darmani, N.A., Hadfield, M.G., Carter, and Computer Modeling, 17, 37–52. W.H., Jr., and Martin, B.R. (1990). Acute and chronic effects of cocaine on isolation- Chein, I., Gerard, D.L., Lee, R.S., and induced aggression in mice. Psychophar- Rosenfeld, E. (1964). The road to H: macology (Berlin), 102(1), 37–40. Narcotics, delinquency, and social policy. New York: Basic Books. Davis, W.M. (1996). Psychopharmacologic violence associated with cocaine abuse: Cherek, D.R., Steinberg, J.L., Kelly, T.H., Kindling of a limbic dyscontrol syndrome? and Sebastian, C.S. (1989). Effects of Progress in Neuropsychopharmacology d-amphetamine on human aggressive and Biological Psychiatry. 20(8), 1273–1300. responding maintained by avoidance of provocation. Pharmacology Biochemistry Dawes, M.A., Antelman, S.M., Vanyukov, and Behavior, 34(1), 65–71. M.M., Giancola, P., Tarter, R.E., Susman, E.J., Mezzich, A., and Clark, D.B. (2000). Chilcoat, H.D. (1992). Parent monitoring Developmental sources of variation in lia- and initiation of drug use in elementary bility to adolescent substance use disor- school children. Doctoral thesis. The ders. Drug and Alcohol Dependence, Johns Hopkins University, Baltimore, MD. 61(1), 3–14.

Cohen, A.K. (1955). Delinquent boys. Dishion, T.J., and Loeber, R. (1985). Ado- Glencoe, IL: Free Press. lescent marijuana and alcohol use: The role of parents and peers revisited. Ameri- Cohen, M.A. (1998). The monetary value can Journal of Drug and Alcohol Abuse, of saving a high-risk youth. Journal of 11, 11–25. Quantitative Criminology, May, 19–33. Dishion, T.J., Patterson, G.R., and Reid, J. Coie, J.D., and Lenox, K.F. (1994). The (1988). Parent and peer factors associated development of antisocial individuals. with drug sampling in early adolescence: Progress in Experimental Personality Implications for treatment. In E.R. Rahdert and Psychopathology Research, 45–72. and J. Grabowski (eds.), Adolescent drug abuse: Analyses of treatment research Crowley, T.J., Mikulich, S.K., Williams, (NIDA Research Monograph 77, pp. 69–93). E.A., Zerbe, G.O., and Ingersoll, N.C. Washington, DC: U.S. Department of (1992). Cocaine, behavior, and alcohol- Health and Human Services, National In- solution drinking in monkeys. Drug and stitute on Drug Abuse. Alcohol Dependence, 29(3), 205–223. Dishion, T.J., Spracklen, K.M., Andrews, Csikzentmihalyi, M., and Larson, R. (1987). D.W., and Patterson, G.R. (1996). Devi- Validity and reliability of the experience ancy training in male adolescent friend- sampling method. Journal of Nervous and ships. Behavior Therapy, 27, 373–390. Mental Diseases, 175, 526–536.

56 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Duncan, T.E., Duncan, S.C., and Hops, Faris, D.E.L., and Dunham, H.W. (1939). H. (1998). Latent variable modeling of Mental disorder in urban areas. Chicago: longitudinal and multilevel alcohol use Hafner/Chicago University Press. data. Journal of Studies on Alcohol, 59(4), 399–408. Farnworth, L. (2000). Time use and leisure occupations of young offenders. American Duncan, T.E., Duncan, S.C., Hops, H., and Journal of Occupational Therapy, 54(3), Alpert, A. (1997). Multi-level covariance 315–325. structure analysis of intra-familial substance use. Drug and Alcohol Depend- Federman, E.B., Costello, E.J., Angold, A., ence, 46(3), 167–180. Farmer, E.M., and Erkanli, A. (1997). De- velopment of substance use and psychi- Dunlap, E., and Johnson, B.D. (1996). atric comorbidity in an epidemiologic study Family and human resources in the devel- of white and American Indian young ado- opment of a female crack-seller career: lescents: The Great Smoky Mountains Case study of a hidden population. Journal Study. Drug and Alcohol Dependence, of Drug Issues, 26(1), 175–198. 44(2–3), 69–78.

Dunlap, E., Johnson, B.D., and Manwar, A. Fendrich, M. (2001). Feasibility of biologi- (1994). A successful female crack dealer: cal measurement in drug surveys. NIH Case study of a deviant career. Deviant grant award 3 R01 DA12425–01A1S1. Behavior, 15(1), 1–25. Fendrich, M., Macksey-Amiti, M.E., Elliott, D.S. (2001). NYS family study: Goldstein, P., Spunt, B., and Brownstein, Problem alcohol use & problem behavior. H.B. (1995). Substance involvement NIH grant award 1 R01 AA11949–01A2. among juvenile murderers: Comparisons with older offenders based on interviews Elliott, D.S., and Huizinga, D. (1989). The with prison inmates. International Journal relationship between delinquent behavior of the Addictions, 30(11), 1363–1382. and ADM problems. In C. Hampton (ed.), Juvenile offenders with serious drug, alco- Forman, V. (2001). ELCID Technical Report hol and mental health problems. Washing- 2001.1. Baltimore: Johns Hopkins Univer- ton, DC: U.S. Government Printing Office. sity [Online]. Available: www.jhu.edu/ ~janthony/elcid/techrep.0801.pdf. Emley, G.S., and Hutchinson, R.R. (1983). Effects of phencyclidine on aggressive Gfroerer, J., and Brodsky, M. (1992). The behavior in squirrel monkeys. Pharma- incidence of illicit drug use in the United cology Biochemistry and Behavior, 18(2), States, 1962–1989. British Journal of 163–166. Addiction, 87(9), 1345–1351.

Ennett, S.T., Flewelling, R.L., Lindrooth, Goldstein, P.J. (1985). The drugs/violence R.C., and Norton, E.C. (1997). School and nexus: A tripartite conceptual framework. neighborhood characteristics associated Journal of Drug Issues, 15(4), 493–506. with school rates of alcohol, cigarette, and marijuana use. Journal of Health and Goldstein, P.J. (1998). Drugs, violence, and Social Behavior, 38(1), 55–71. Federal funding: A research odyssey. Sub- stance Use & Misuse, 33(9), 1915–1936. Evans, A.S. (1976). Causation and disease: The Henle-Koch postulates revisited. Yale Golub, A.L., and Johnson, B.D. (1999). Journal of Biology and Medicine, 49, Cohort changes in illegal drug use among 175–195. arrestees in Manhattan: From the Heroin

57 SPECIAL REPORT / JULY 03

Injection Generation to the Blunts Gener- Harrison, R.J., Connor, D.F., Nowak, C., ation. Substance Use & Misuse, 34(13), and Melloni, R.H. (2000). Chronic low- 1733–1763. dose cocaine treatment during adolescence facilitates aggression in hamsters. Golub A., and Johnson, B.D. (2001a). Physiology and Behavior, 69(4–5), 555–562. Variation in youthful risks of progression from alcohol and tobacco to marijuana and Hawkins, J.D., Catalano, R.F., and Miller, to hard drugs across generations. Ameri- J.Y. (1992). Risk and protective factors for can Journal of Public Health, 91(2), alcohol and other drug problems in adoles- 225–232. cence and early adulthood: Implications for substance abuse prevention. Psycho- Golub, A.L., and Johnson, B.D. (2001b). logical Bulletin, 112(1), 64–105. The rise of marijuana as the drug of choice among youthful adult arrestees (Research Hawkins, J.D., Von Cleve, E., and Catalano, in Brief, June, NCJ 187490). Washington R.F. (1991). Reducing early childhood DC: U.S. Department of Justice, National aggression: Results of a primary prevention Institute of Justice. program. Journal of the American Academy of Child and Adolescent Psychiatry, 30(2), Haber, S., Barchas, P.R., and Barchas, J.D. 208–217. (1981). A primate analogue of amphetamine-induced behaviors in humans. Higley, J.D., Mehlman, P.T., Higley, S.B., Biological Psychiatry, 16(2), 181–196. Fernald, B., Vickers, J., Lindell, S.G., Taub, D.M., Suomi, S.J., and Linnoila, M. (1996a). Harrison, L. (1997). The validity of self- Excessive mortality in young free-ranging reported drug use in survey research: An male nonhuman primates with low cere- overview and critique of research meth- brospinal fluid 5-hydroxyindoleacetic acid ods. In L. Harrison and A. Hughes (eds.), concentrations. Archives of General Psy- The Validity of Self-Reported Drug Use: chiatry, 53(6), 537–543. Improving the Accuracy of Survey Esti- mates (NIDA Research Monograph 167, Higley, J.D., Mehlman, P.T., Poland, R.E., pp. 17–36). Rockville, MD: U.S. Depart- Taub, D.M., Vickers, J., Suomi, S.J., and ment of Health and Human Services, Linnoila, M. (1996b). CSF testosterone National Institute on Drug Abuse. and 5-HIAA correlate with different types of aggressive behaviors. Biological Psy- Harrison, L. (2001). Validity of self-reported chiatry, 40(11), 1067–1082. drug use in population surveys. NIH grant award 5 R01 DA10930–03. Higley, J.D., Suomi, S.J., and Linnoila, M. (1996a). A nonhuman primate model Harrison, L., and Gfroerer, J. (1992). The of type II excessive alcohol consump- intersection of drug use and criminal be- tion? Part 1. Low cerebrospinal fluid 5- havior: Results from the National House- hydroxyindoleacetic acid concentrations hold Survey on Drug Abuse. Crime and and diminished social competence corre- Delinquency, 38, 422–443. late with excessive alcohol consumption. Alcoholism: Clinical & Experimental Re- Harrison, L.D., and Backenheimer, M. search, 20(4), 629–642. (1998). Editors’ introduction: Evolving insights into the drug-crime nexus. Sub- Higley, J.D., Suomi, S.J., and Linnoila, M. stance Use & Misuse, 33(9), 1763–1777. (1996b). A nonhuman primate model of type II alcoholism? Part 2. Diminished

58 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

social competence and excessive aggres- Rosa, E.Y. Lambert, and B. Gropper (eds.), sion correlates with low cerebrospinal fluid Drugs and violence: Causes, correlates, 5-hydroxyindoleacetic acid concentrations. and consequences (Research Monograph Alcoholism: Clinical & Experimental Re- 103, pp. 92–111). Rockville, MD: U.S. search, 20(4), 643–650. Department of Health and Human Serv- ices, National Institute on Drug Abuse. Hill, A.B. (1965).The environment and disease: association or causation? Pro- Inciardi, J.A., and Pottieger, A.E. (1998). ceedings of the Royal Society of Medicine, Drug use and street crime in Miami: An 58, 295–300. (almost) twenty-year retrospective. Sub- stance Use & Misuse, 33(9), 1839–1870. Holder, H.D. (1993). Prevention of alcohol- related accidents in the community. Ad- Inciardi, J.A., and Russe, B.R. (1977). diction, 88(7), 1003–1012. Professional thieves and drugs. Inter- national Journal of the Addictions, 12(8), Holder, H.D. (2001). Prevention of alcohol 1087–1095. problems in the 21st century: Challenges and opportunities. American Journal on James, K.E., Wagner, F.A., and Anthony, Addictions, 10(1), 1–15. J.C. (2002). Regional variation in drug purchase opportunity among youths in the Holder, H., Flay, B., Howard, J., Boyd, G., United States, 1996–1997. Journal of Voas, R., and Grossman, M. (1999). Urban Health, 79(1), 104–112. Phases of alcohol problem prevention research. Alcoholism: Clinical & Experi- Jessor, R. (1998). New perspectives on mental Research, 23(1), 183–194. adolescent risk behavior. In R. Jessor (ed.), New perspectives on adolescent risk be- Holder, H.D., Gruenewald, P.J., Ponicki, havior (pp. 1–12). Cambridge, UK: Cam- W.R., Treno, A.J., Grube, J.W., Saltz, R.F., bridge University Press. Voas, R.B., Reynolds, R., Davis, J., Sanchez, L., Gaumont, G., and Roeper, P. (2000). Jessor, R., and Jessor, S.L. (1977). Prob- Effect of community-based interventions lem behavior and psychosocial develop- on high-risk drinking and alcohol-related ment: A longitudinal study of youth. New injuries. Journal of the American Medical York: Academic Press. Association, 284(18), 2341–2347. Johnson, B.D., Dunlap, E., and Maher, L. Hser, Y.I., Anglin, M.D., and McGlothlin, (1998). Nurturing for careers in drug use W. (1987). Sex differences in addict and crime: Conduct norms for children and careers. 1. Initiation of use. American juveniles in crack-using households. Sub- Journal of Drug and Alcohol Abuse, stance Use & Misuse, 33(7), 1511–1546. 13(1–2), 33–57. Johnson, E.O., Arria, A.M., Borges, G., Hubbard, J.A., Dodge, K.A., Cillessen, Ialongo, N., and Anthony, J.C. (1995). The A.H., Coie, J.D., and Schwartz, D. (2001). growth of conduct problem behaviors The dyadic nature of social information from middle childhood to early adoles- processing in boys’ reactive and proactive cence: Sex differences and the suspected aggression. Journal of Personality and influence of early alcohol use. Journal of Social Psychology, 80(2), 268–280. Studies on Alcohol, 56(6), 661–671.

Inciardi, J.A. (1990). The crack-violence Johnston, L.D., O’Malley, P.M., and Eve- connection within a population of hard- land, L.K. (1978). Drugs and delinquency: core adolescent offenders. In M. De La A search for causal connections. In D.B.

59 SPECIAL REPORT / JULY 03

Kandel (ed.), Longitudinal research on drug initiation: Patterns and predictors of first use (pp.137–156). Washington, DC: use in adolescence. American Journal of Hemisphere Publishing Co. Public Health, 90(3), 360–366.

Jones, M.B., and Offord, D.R. (1989). Kotler, M., Barak, P., Cohen, H., Averbuch, Reduction of antisocial behavior in poor I.E., Grinshpoon, A., Gritsenko, I., Nemanov, children by nonschool skill-development. L., and Ebstein, R.P. (1999). Homicidal Journal of Child Psychology and Psychiatry behavior in schizophrenia associated with and Allied Disciplines, 30(5), 737–750. a genetic polymorphism determining low catechol O-methyltransferase (COMT) Kaplan, C.D., and Lambert, E.Y. (1995). activity. American Journal of Medical The daily life of heroin-addicted persons: Genetics, 88(6), 628–633. The biography of specific methodology. In E.Y. Lambert, R.S. Ashery, and R.H. Langbehn, D.R., and Cadoret, R.J. (2001). Needle (eds.), Qualitative methods in drug The adult antisocial syndrome with and abuse and HIV research (NIDA Research without antecedent conduct disorder: Monograph 157, pp. 100–116). Rockville, Comparisons from an adoption study. MD: U.S. Department of Health and Comprehensive Psychiatry, 42(4), Human Services, National Institute on 272–282. Drug Abuse. Levitt, S.D., and Venkatesh, S.A. (1998). Kaplan, H.B. (ed.) (1995). Drugs, crime, An economic analysis of a drug-selling and other deviant adaptations. New York: gang’s finances. NBER Working Paper No. Plenum Press. 6592. Cambridge, MA: National Bureau of Economic Research. Kellam, S.G., and Anthony, J.C. (1998). Targeting antecedents to prevent tobacco Licata, A., Taylor, S., Berman, M., and smoking: Findings from an epidemiologi- Cranston, J. (1993). Effects of cocaine on cally based randomized field trial. Ameri- human aggression. Pharmacology Bio- can Journal of Public Health, 88(10), chemistry and Behavior, 45(3), 549–552. 1490–1495. Lindesmith, A.R. (1938). A sociological Kellam, S.G., Ensminger, M.E., and Turner, theory of drug addiction. American Journal R.J. (1977). Family structure and the men- of Sociology, 43, 593–613. tal health of children. Concurrent and longitudinal community-wide studies. Loeber, R., Farrington, D.P., Stouthamer- Archives of General Psychiatry, 34(9), Loeber, M., Moffitt, T., and Caspi, A. (1998). 1012–1022. The development of male offending: Key findings from the first decade of the Kessler, R.C. (1999). The World Health Pittsburgh Youth Study. Studies on Crime Organization International Consortium in and Crime Prevention, 7(2), 141–172. Psychiatric Epidemiology (ICPE): Initial work and future directions. Acta Psychi- Loeber, R., and Schmaling, K.B. (1985). atrica Scandinavica, 99(1), 2–9. The utility of differentiating between mixed and pure forms of antisocial child Kolb, L. (1925). Drug addiction in its rela- behavior. Journal of Abnormal Child tion to crime. Mental Hygiene, 9, 77. Psychology, 13(2), 315–335.

Kosterman, R., Hawkins, J.D., Guo, J., Loeber, R., Wei, E., Stouthamer-Loeber, Catalano, R.F., and Abbott, R.D. (2000). M., Huizinga, D., and Thornberry, T.P. The dynamics of alcohol and marijuana (1999). Behavioral antecedents to serious

60 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

and violent offending: Joint analyses from Comparison of source of supply and distri- the Denver Youth Survey, Pittsburgh Youth bution networks. International Journal of Study and the Rochester Youth Develop- the Addictions, 25(5), 557–577. ment Study. Studies on Crime and Crime Prevention, 8(2), 245–264. Moore, M.S., and Thompson, D.M. (1978). Acute and chronic effects of cocaine on Logan, T.K. (2001). HIV risk behavior and extinction-induced aggression. Journal of violence among crack users. NIH grant the Experimental Analysis of Behavior, award 5 R29 DA11578–03. 29(2), 309–318.

Longshore, D.Y. (2000). Linking DUF and Moro, M., Salvador, A., and Simon, V.M. criminal history data. NIH grant award 5 (1997). Effects of repeated administration R03 DA11009–02. of d-amphetamine on agonistic behaviour of isolated male mice. Behavioural Phar- Maltz, M.D. (1996). From Poisson to the macology, 8(4), 309–318. present: Applying operations research to problems of crime and justice. Journal of Natarajan, M., and Belanger, M. (1998). Quantitative Criminology, 12(1), 3–61. Varieties of drug trafficking organizations: A typology of cases prosecuted in New Manski, C.F., Pepper, J.V., and Petrie, C.V. York City. Journal of Drug Issues, 28(4), (eds.), and the Committee on Data and Re- 1005–1025. search for Policy on Illegal Drugs (2001). Informing America’s policy on illegal drugs: Natarajan, M., Clarke, R.V., and Belanger, What we don’t know keeps hurting us. M. (1996). Drug dealing and pay phones: Washington, DC: National Academy Press. The scope for intervention. Security Journal, 7(4), 245–251. Markowitz, S., and Grossman, M. (2000). The effects of beer taxes on physical child Neiderhiser, J.M. (2001). Understanding abuse. Journal of Health Economics, the roles of genome and envirome: 19(2), 271–282. Methods in genetic epidemiology. British Journal of Psychiatry Supplement, 40, McCardle, L., and Fishbein, D.H. (1989). S12–7. The self-reported effects of PCP on human aggression. Addictive Behaviors, Neiderhiser, J.M., Pike, A., Hetherington, 14(4), 465–472. E.M., and Reiss, D. (1998). Adolescent perceptions as mediators of parenting: Melega, W.P., Raleigh, M.J., Stout, D.B., Genetic and environmental contributions. Huang, S.C., and Phelps, M.E. (1997). Developmental Psychology, 34(6), Ethological and 6-[18F]fluoro-L-DOPA-PET 1459–1469. profiles of long-term vulnerability to chronic amphetamine. Behavioural Brain Re- Neiderhiser, J.M., Reiss, D., Hetherington, search, 84(1–2), 259–268. E.M., and Plomin, R. (1999). Relationships between parenting and adolescent adjust- Miczek, K.A., and Haney, M. (1994). Psycho- ment over time: Genetic and environmen- motor stimulant effects of d-amphetamine, tal conditions. Developmental Psychology, MDMA and PCP: Aggressive and schedule- 35, 680–692. controlled behavior in mice. Psychophar- macology (Berlin), 115(3), 358–365. Newcomb, M.D. (1992). Understanding the multidimensional nature of drug use Montagne, M. (1990). The social epidemi- and abuse: The role of consumption, risk ology of international drug trafficking: factors and protective factors. In M. Glantz

61 SPECIAL REPORT / JULY 03

and R. Pickens (eds.), Vulnerability to drug parenting involvement in the Woodlawn abuse. Washington, DC: American Psycho- community. Child Development, 61(2), logical Association. 434–442.

Newcomb, M.D., and Bentler, P.M. (1988). Peffer-Smith, P.G., Smith, E.O., and Byrd, Consequences of adolescent drug use: L.D. (1983). Effects of d-amphetamine on Impact on the lives of young adults. self-aggression and posturing in stumptail Newbury Park, CA: Sage Publications. macaques. Journal of the Experimental Analysis of Behavior, 40(3), 313–320. Nurco, D. (1998). A long-term program of research on drug use and crime. Sub- Pescor, M.J. (1939). The Kolb classification stance Use & Misuse, 33(9), 1817–1839. of drug addicts. Public Health Reports, Supp. 155. Nurco, D.N., Bonito, A.J., Lerner, M., and Balter, M.B. (1975). Studying addicts over Petronis, K.R., and Anthony, J.C. (2000). time: Methodology and preliminary find- Perceived risk of cocaine use and experi- ings. American Journal of Drug and ence: Do they cluster within U.S. neigh- Alcohol Abuse, 2(2), 183–196. borhoods and cities? Drug and Alcohol Dependence, 57(3), 183–192. Nurco, D.N., Kinlock, T., O’Grady, K., Lerner, M., and Hanlon, T.E. (1996). Per- Preble, E., and Casey, J.J. (1969). Taking ceptions of social pathology in the neigh- care of business—the heroin user’s life on borhood and the etiology of narcotic the street. International Journal of the addiction: A retrospective study. Journal Addictions, 4, 1–24. of Nervous and Mental Disease, 184(1), 35–42. Robins, L.N. (1966). Deviant children grown up: A sociological and psychiatric Parker, R.N., and Rebhun, L.A. (1995). study of sociopathic personality. Balti- Alcohol and homicide: A deadly combina- more, MD: Williams and Wilkins. tion of two American traditions. Albany: State University of New York Press. Russell, J.W., Greenberg, B.D., and Segal, D.S. (1984). The effects of phencyclidine Parker, R.N., and Toth, A.M. (1990). on spontaneous aggressive behavior in Family, intimacy, and homicide: A macro- the rat. Biological Psychiatry, 19(2), social approach. Violence and Victims, 195–202. 5(3), 195–210. Sandstrom, M.J., and Coie, J.D. (1999). A Patterson, G.R., Dishion, T.J., and Yoerger, developmental perspective on peer rejec- K. (2000). Adolescent growth in new tion: Mechanisms of stability and change. forms of problem behavior: Macro- and Child Development, 70(4), 955–966. micro-peer dynamics. Prevention Science, 1(1), 3–13. Scheier, L.M., and Botvin, G.J. (1996). Purpose in life, cognitive efficacy, and gen- Patterson, G.R., Reid, J.R., and Dishion, eral deviance as determinants of drug T.J. (1992). Antisocial boys. Eugene, OR: abuse in urban black youth. Journal of Castalia Publishing Co. Child and Adolescent Substance Abuse, 5(1), 1–26. Pearson, J.L., Hunter, A.G., Ensminger, M.E., and Kellam, S.G. (1990). Black Schulenberg, J., Bachman, J.G., O’Malley, grandmothers in multigenerational house- P.M., and Johnston, L.D. (1994). High holds: Diversity in family structure and

62 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

school educational success and subse- Tracy, P., Wolfgang, M.E., and Figlio, R.M. quent substance use: A panel analysis fol- (1990). Delinquency careers in two birth lowing adolescents into young adulthood. cohorts. New York: Plenum Press. Journal of Health and Social Behavior, 35(1), 45–62. Tsuang, M.T., Bar, J.L., Harley, R.M., and Lyons, M.J. (2001). The Harvard twin Spunt, B., Brownstein, H., Goldstein, P., study of substance abuse: What we have Fendrich, M., and Liberty, H.J. (1995). learned. Harvard Review of Psychiatry, Drug use by homicide offenders. Journal 9(6), 267–279. of Psychoactive Drugs, 27(2), 125–134. Tyler, C.B., and Miczek, K.A. (1982). Spunt, B.J., Goldstein, P.J., Bellucci, P.A., Effects of phencyclidine on aggressive and Miller, T. (1990). Race/ethnicity and behavior in mice. Pharmacology Bio- gender differences in the drugs-violence chemistry and Behavior, 17(3), 503–510. relationship. Journal of Psychoactive Drugs, 22(3), 293–303. Verdonik, F., and Sherrod, L.R. (eds.) (1984). An inventory of longitudinal re- Spunt, B., Goldstein, P., Brownstein, H., search on childhood and adolescence. and Fendrich, M. (1994). The role of mari- New York: Social Science Research juana in homicide. International Journal of Council. the Addictions, 29(2), 195–213. Waldorf, D. (1998). Misadventures in the Stevens, M.C., Kaplan, R.F., and Bauer, drug trade. Substance Use & Misuse, L.O. (2001). Relationship of cognitive 33(9), 1957–1991. ability to the developmental course of antisocial behavior in substance-dependent Wiens, J.A., Addicott, J.F., Case, T.J., and patients. Progress in Neuropsychophar- Diamond, J. (1986). Overview: The impor- macology and Biological Psychiatry, 25(8), tance of spatial and temporal scale in eco- 1523–1536. logical investigations. In J. Diamond and T.J. Case (eds.), Community Ecology (pp. Tehrani, J.A., Brennan, P.A., Hodgins, S., 145–153). New York: Harper and Row. and Mednick, S.A. (1998). Mental illness and criminal violence. Social Psychiatry Winick, C. (1962). Maturing out of narcotic and Psychiatric Epidemiology, 33 Supp. 1, addiction. U.N. Bulletin on Narcotics, S81–85. 14(1), 1–7.

Terry, C.E., and Pellens, M. (1928). The Wish, E.D. (1988). Identifying drug-abusing opium problem. New York: Committee on criminals. In C.G. Leukefeld and F.M. Tims Drug Addictions and Bureau of Social (eds.), Compulsory treatment of drug abuse: Hygiene. Research and clinical practice (NIDA Research Monograph 86, pp. 139–159). Thornberry, T.P. (ed.) (1997). Developmental Washington, DC: U.S. Government theories of crime and delinquency. New Printing Office. Brunswick, NJ: Transaction Publishers. Wish, E.D., Gray, T., Sushinsky, J., Yacou- Tolman, R.M., Edleson, J.L., and Fendrich, bian, G.S., and Fitzgerald, N. (2000). An M. (1996). The applicability of the theory experiment to enhance the reporting of of planned behavior to abusive men’s ces- drug use by arrestees. Journal of Drug sation of violent behavior. Violence and Issues, 30(1), 55–76. Victims, 11(4), 341–354.

63 SPECIAL REPORT / JULY 03

Yacoubian, G.S., Wish, E.D., and Perez, or methods” and “Criterion/guideline 2: D.M. (2001). A comparison of saliva test- Biological or other theoretical plausibility.” ing to urinalysis in an arrestee population. Correspondence concerning this paper Journal of Psychoactive Drugs, 33(3), should be addressed to James. C. Anthony, 289–294. Electronic Collaboratory for Investigations About Drugs, JHU School of Hygiene and Public Health, 624 North Broadway, Author’s note Room 893, Baltimore, MD 21205; e-mail [email protected]. Valerie Forman coauthored the following sections: “Tension that involves approach

64 Research on Drugs-Crime Linkages: The Next Generation

Robert MacCoun, Beau Kilmer, and Peter Reuter

Introduction believed the offender was under the influence of drugs or alcohol (National The association between drugs and crime Crime Victimization Survey, 2000). in the public mind is so strong that a recent psychology experiment showed ■ A recent estimate of the economic the word “drug” tightly linked to such costs of drug abuse reported that 60 words as “choke,” “knife,” “fight,” and percent were associated with crime and “wound” in participants’ associative mem- criminal justice (Harwood, Fountain, and ory networks (Bushman, 1996). Although Livermore, 1998). it is routine in academia to deride public ignorance of all things criminological, in Considerable complexities and nuances this case the public is hardly deluded. underlie these associations. Although Consider the following facts:1 many of these subtleties were anticipated by astute observers in the 1970s (see ■ Across 35 cities in 1998, between 40 Gandossy et al., 1980), the past decade and 80 percent of male arrestees in has seen a solid scholarly consensus form the Arrestee Drug Abuse Monitoring around the following principles (see (ADAM) Program tested positive for at Fagan, 1990; Parker and Auerhahn, 1998; least one drug at arrest (Arrestee Drug White and Gorman, 2000): Abuse Monitoring Program, 1999). 1. Many different data sources establish About the Authors ■ Nearly one-quarter (22 percent) of a raw correlation between drug use Robert MacCoun is with the Federal prison inmates and one-third and other criminal offenses. But corre- Goldman School of Public (33 percent) of State prison inmates— lation does not equal causation: In Policy and Jurisprudence nearly 40 percent of State inmates con- principle, drug use might cause (pro- and Social Policy Program victed of robbery, burglary, or motor mote, encourage) crime; criminality at Boalt Hall School of Law, vehicle theft—reported being under the might cause (promote, encourage) University of California influence of drugs at the time of their drug use; and/or both might be caused at Berkeley. Please email (promoted, encouraged) by some set any comments to offense (Bureau of Justice Statistics, maccoun@socrates. 1997a, 1997b). of “third variables”—environmental, berkeley.edu. Beau Kilmer situational, dispositional, and/or biolog- ■ is with the Drug Policy Among State and Federal prison ical. In fact, all three pathways have Research Center at RAND. inmates, 27 percent of those serving empirical support in at least some set- Peter Reuter is with the sentences for robbery and 30–32 per- tings and populations. School of Public Affairs cent of those serving sentences for bur- and Department of glary said they committed their offense 2. These causal influences are probabilis- Criminology Criminology at to buy drugs (Bureau of Justice Statistics, tic, not deterministic. Most drug users the University of Maryland. 1991a, 1991b). are not otherwise criminally active, and the vast majority of drug-using ■ In the 70 percent of cases in which the incidents neither cause nor accom- victim formed an opinion, 31 percent pany other forms of criminality.

65 SPECIAL REPORT / JULY 03

Nevertheless, drugs clearly play an Drugs-crime linkages: important causal role in violent and property crime. Expanding the Goldstein taxonomy 3. These causal influences are contingent, not unconditional. There is little Goldstein’s framework evidence that drug use per se directly causes people to become aggressive Paul Goldstein’s (1985) conceptual essay in some direct and unconditional man- offered a tripartite classification of drugs- ner or that criminality per se causes violence connections: someone to use drugs. The drugs- ■ crime link varies across individuals, Psychopharmacological: Violence due over time within an individual’s devel- to the direct acute effects of a psy- opment, across situations, and possi- choactive drug on the user. bly over time periods (as a function of ■ Economic-compulsive: Violence com- the dynamics of drug epidemics and, mitted instrumentally to generate possibly, drug control policies). money to purchase expensive drugs. 4. That drug use can causally influence ■ Systemic: Violence associated with the criminality does not necessarily impli- marketing of illicit drugs, such as turf cate the psychopharmacological prop- battles, contract disputes, and so on. erties of the drug. Intoxication, the need or desire to raise money to buy Goldstein and his colleagues (Brownstein et drugs, and the nature of illicit markets al., 1992; Goldstein et al., 1989; Goldstein, are distinct mechanisms by which Brownstein, and Ryan, 1992) applied this drugs can cause crime. Thus, drug scheme empirically to homicides in New prohibition cannot be only a response York State (1984) and New York City to drug-related crime, but it may also (1988). They found that drugs and alcohol be a causal antecedent to some drug- were important causes for a large share of related crime. all homicides in both samples. For 1988, near the height of the crack epidemic, 5. Alcohol is a drug, and it stimulates or they classified 53 percent of 414 homi- augments a great deal of criminal cides as drug or alcohol related; there was behavior, almost certainly more than also a substantial percentage whose drug- the street drugs combined. relatedness could not be determined. Of We expect that understanding the consid- those homicides that could be determined erable heterogeneity of effects across to be drug or alcohol related, 14 percent users, substances, cities, neighborhoods, were psychopharmacological (68 percent and situations—and the interactions alcohol, 16 percent crack), 4 percent were among these factors—will be the central economic-compulsive, and 74 percent focus of drugs-crime research during the were systemic (61 percent crack, 27 per- remainder of this decade. This paper cent powder cocaine). By contrast, in reviews the existing literature, focusing 1984, before the crack surge, only 42 per- particular attention on Goldstein’s (1985) cent of homicides were drug or alcohol taxonomy, the temporal dynamics of drug related; 59 percent of those were psy- markets, and the consequences of prohibi- chopharmacological (79 percent alcohol), tion. These highlight some of the ques- 3 percent were economic-compulsive, and tions that should drive this research. 21 percent were systemic. The difference

66 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

between the findings of the two years Chicago Homicide Dataset; Block, Many studies have might reflect differences in geography to Block, and Illinois Criminal Justice some extent (New York State versus New Information Authority, 1998). Despite tried to determine York City), but it also reminds us that the fluctuations, the 1995 homicide whether crimes these numbers are not eternal verities; rate was strikingly similar to the 1985 they result from complex and historically rate for all drug-related motives except were drug related, dependent market dynamics. for homicides that resulted from a but few have drug transaction; the latter increased assessed whether Subsequent applications tenfold from 1985 to 1995. the offender’s drug The generalizability of Goldstein et al.’s 4. Results from Lattimore et al.’s (1997) (1989) original findings was limited by homicide study of eight cities, which need, drug use, their location (New York) and timing (the included surveys of local officials and or role in the height of the crack explosion; see U.S. ADAM/UCR (Uniform Crime Reports) Sentencing Commission, 1995, 106).2 analyses for 1985–94, suggest that drug market Many studies have tried to determine drugs other than cocaine and crack was directly whether crimes were drug related, but were not associated with homicide few have assessed whether the offend- trends “in any discernible way.” They responsible for er’s drug need, drug use, or role in the also found that the drug market struc- the crime. drug market was directly responsible for ture was less associated with violence the crime. Although most of the studies than was expected. that used this framework were conducted by Goldstein and his colleagues in New The Lattimore et al. study questioned the York (Parker and Auerhahn, 1998), there role of crack and systemic crime because are others worthy of attention, especially the crack markets were described as high- given their unique approaches. General ly competitive in cities where the homi- findings include the following: cide rate was declining, increasing, or remaining the same (1997, p. 89). It is not 1. Non-NDRI (National Development and clear, however, that the same conclusions Research Institutes, Inc.) studies of could be drawn if disaggregated homicide New York City in the mid- to late rates (by circumstance) were considered. 1980s found that crack sellers are (Additional discussion and methodological more violent than other drug sellers descriptions of these studies are reported and that their violence is not con- in appendix A.) fined to the drug-selling context (U.S. Sentencing Commission, 1995, citing Limitations of existing research Fagan and Chin, 1990). on the Goldstein framework 2. Studies of juvenile delinquents in The Goldstein tripartite framework has Miami in the mid- to late 1980s found been a boon to drug research reviewers— that they were much more likely to it is invaluable as an organizing scheme— commit a drug-related economic- but still, we are struck by the relative rarity compulsive crime than a psychophar- of actual empirical applications. Existing macological or systemic crime applications overrepresent New York, and (Inciardi, 1990).3 they overrepresent the crack epidemic at its height relative to earlier and later peri- 3. The per capita drug-related homicide ods. In fairness, the taxonomy was not rate remained fairly stable in Chicago proposed until 1985, but it could be ap- from 1973 to 1984 and fluctuated plied retrospectively to earlier homicide from 1985 to 1995 (data are from the case files. In our view, such comparisons

67 SPECIAL REPORT / JULY 03

The prevailing would be invaluable. There has been coded separately, there would be no effort little consistency in the methods used to to force events into a single classification. view about implement the scheme (e.g., Goldstein’s Psychometric analysis could be used to psychophar- trained coders versus Inciardi’s survey test the hypothesized latent structure.5 approach). Little has been learned from Such analyses pose enormous logistical macological (as that methodological diversity because, to difficulties, but the payoffs for advancing opposed to our knowledge, no two methods have our understanding of drug violence would ever been applied to the same sample of surely justify the effort. economic- cases for comparative purposes. Indeed, if compulsive or one imagines a three-dimensional matrix In the remainder of this section, we will of major cities by time periods by meth- examine other ways in which Goldstein’s systemic) violence ods, almost every cell is empty and there taxonomic scheme might be expanded is that it is rare are almost no vectors with more than one and refined. cell occupied. This spotty record makes it and attributable hard to identify either temporal trends or Psychopharmacological violence mostly to alcohol the influence of local variations on drug popularity, drug market structures, or poli- The prevailing view about psychophar- rather than cies and enforcement practices. Finally, macological (as opposed to economic- illicit drugs. the scheme has been applied mostly to compulsive or systemic) violence is that it homicide and less often to other, more is rare and attributable mostly to alcohol prevalent violent crimes.4 rather than illicit drugs (Fagan, 1990; Parker and Auerhahn, 1998; White and Parker and Auerhahn (1998) complain that Gorman, 2000). According to Fagan (1990, Goldstein’s categories are not mutually p. 243): exclusive. This critique presumes a classi- cal set-theoretic approach that, in our opin- [I]ntoxication does not consistently ion, is neither feasible nor scientifically lead to aggressive behavior . . . only useful for drugs-violence research. Mu- limited evidence that consumption of tually exclusive categories are not neces- alcohol, cocaine, heroin, or other sub- sary for scientific classification (Meehl, stances is a direct, pharmacologically 1995), and they are usually impossible to based cause of crime. achieve using sparse and noisy archival data (Ragin, 2000). But we agree with According to Parker and Auerhahn (1998, Parker and Auerhahn’s (1998) contention p. 306): that “the Goldstein tripartite framework . . . Our review of the literature finds a is not treated as a set of testable proposi- great deal of evidence that the social tions but rather as a set of assumptions environment is a much more power- about the nature of drug- and alcohol- ful contributor to the outcome of related violence.” violent behavior than are pharmacological factors associated with any of In our view, an understanding of the taxo- 6 metric properties of drug-related violence the substances reviewed here. ought to emerge inductively from more The Goldstein et al. (1989) analysis pro- fine-grained coding of the underlying fea- vides some support for these claims; only tures of these events—whether various 14 percent of the drug-related homicides drugs were found as evidence, the results appeared to be psychopharmacological, of toxicology on the offender and the vic- and these largely involved alcohol either tim, various features of witness reports, alone or in combination with other drugs. prior record information, and so on. Be- But one in seven is hardly a trivial fraction, cause each property or attribute would be

68 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

and those results reflect the peak of the attention to situational cues (Steele crack market wars, when systemic homi- and Josephs, 1990), and reduced self- cides were occurring in unprecedented attention (Ito, Miller, and Pollock, 1996). numbers, inflating the denominator. ■ Social threats to self-identity or self- Moderators. Examining the literature esteem (Baumeister, Smart, and Boden, cited in many recent review essays, it is 1996) that seem particularly relevant in difficult to avoid the suspicion that some “cultures of honor” (see Anderson, authors hold neuropharmacological factors 1994; Bourgois, 1996; Cohen et al., to a stricter standard of proof than the 1996). sociological factors under study. If the psychopharmacological claim is that marijua- Moreover, the absence of evidence does The absence of na, heroin, or cocaine ingestion directly not equal evidence of absence; the labora- promotes violent behavior absent any situ- tory literature on drugs and aggression is evidence does not ational provocation or stressors, then that simply too spotty at present to permit any equal evidence claim is probably false. But evidence for firm conclusions. Almost the entire experi- Drug x Situation and Drug x Psychology mental literature on moderators of the of absence; the interaction effects hardly exonerates drug drugs-aggression relationship has exam- laboratory use as a causal factor. It may be that no ined alcohol rather than illicit drugs. drug is sufficient to produce aggression in literature on drugs isolation from psychological and situational Comorbidity: Drugs in association with and aggression is moderators. But it seems clear that some mental illness or alcoholism. A second drugs—certainly alcohol—can amplify the potential class of moderators of the drugs- simply too spotty psychological and situational facilitators of aggression link involve comorbid condi- at present to aggression. Relevant moderators (see tions—substance abuse in tandem with Bushman, 1997; Fagan, 1990; Ito, Miller, schizophrenia or other psychoses, person- permit any firm and Pollock, 1996) include: ality disorders, or alcoholism. Numerous studies have identified a high prevalence conclusions. ■ Situational stressors and frustrators (see of illicit substance abuse among individu- Ito, Miller, and Pollock, 1996). als diagnosed with psychiatric disorders (e.g., Compton et al., 2000; Kessler et al., ■ Expectancy effects: personal and cultur- 1996; Mueser et al., 2000).7 The causal al beliefs about the effect of the drug on nexus of these comorbid conditions is behavior, and local norms about tolera- unclear. The MacArthur Violence Risk ble versus unacceptable conduct when Assessment Study (Steadman et al., under the influence (e.g., Critchlow, 1998), a prospective followup study of 1986; Stacy, Widaman, and Marlatt, clients admitted to acute psychiatric inpa- 1990). tient facilities, found that substance abuse increased the probability of violent behav- ■ Disinhibition (e.g., Parker and Auerhahn, ior, but this was true for both psychiatric 1998; but see Fagan, 1990). patients and matched community controls. Neither drug dependence nor psychi- ■ Impaired cognitive functioning, includ- atric illness predicted subsequent violent ing reduced executive functioning crime in a 6-year followup of released jail (self-control and decisionmaking ability; detainees (Teplin, Abram, and McClelland, Fishbein, 2000; Giancola, 2000), reduced 1994).

69 SPECIAL REPORT / JULY 03

Drug use and victimization8 difficult it is to assess the relationship. First, as Goldstein (1985) argues, it is diffi- Increased victimization provides another cult to obtain this information because vic- mechanism by which drugs can become tims do not want to talk to the police while linked with violence. Although this catego- intoxicated and often do not remember ry can be subsumed under Goldstein’s the details of the offense; thus, it may psychopharmacological category, treating go unreported. Second, the victimization it as a fourth category might have merit surveys that ask about substance use usu- because the causal mechanisms differ ally include it as a predictor but do not ask and it has been largely neglected by re- whether it contributed to a specific event. searchers. There are a number of reasons Third, many of these surveys only ask to expect that drug users ought to be par- about (or report) general drug use, not ticularly vulnerable to criminal victimiza- about specific drugs or the circumstances tion, especially when intoxicated. First, of their use. Finally, the label “victim” is intoxicated people often appear (and often problematic when the participants sometimes are) more vulnerable than are codisputants; indeed, the “victim” other targets for such offenses as robbery, may have initiated the provocation. In our rape, or hate crimes. Second, intoxicated view, these concerns are valid, but they people are often obnoxious, annoying, do not undermine the importance of vic- and/or offensive in their appearance, timization as a research topic. conduct, and speech. Third, intoxication makes people’s conduct unpredictable and The ubiquity of alcohol has made it the ambiguous—intoxication impairs the per- subject of victimization work for 50 years, ception of signals, but it also impairs the and there is general agreement about its transmission of clear signals to others. role in victimizations, especially sexual Finally, in an active illicit drug market, drug assaults. The research on drugs is not as sellers are sometimes both intoxicated robust, but there are some important find- and flush with cash. ings that should be addressed in future works on drugs and crime. The following Fagan (1990) notes that the vulnerability of sections provide insight about this rela- drug users to victimization has been long tionship by examining existing victimiza- recognized. For example, Wolfgang (1958) tion studies of the general population, studied “victim-precipitated homicides” by women, and hard drug users. assessing the incidence of intoxication among victims. And Fagan (1990) reviews The general population. The Nation’s evidence from animal studies showing largest victimization survey, the National that “substances that induce changes in Criminal Victimization Survey (NCVS), does an opponent’s behavior might result in not ask about victim drug use, but it is increased aggression by a drug-free used in conjunction with other data to pro- attacker . . .” (p. 251). vide insight about drugs and crime. Using NCVS, Markowitz’s (2000) multivariate Although Goldstein (1985) acknowledged analysis of almost 450,000 observations that the victimization of drug users consti- found that marijuana decriminalization (a tuted a distinct drugs-violence linkage, he proxy for lower marijuana prices) will did not include it as a separate category in result in a higher incidence of robbery and his classification scheme. Since then, the assault while higher cocaine prices will victimization of drug users has received lit- decrease these crimes.9 Neither measure tle attention in the drugs and crime litera- was significantly related to rape or sexual ture. This is not surprising given how assault. When victims’ perceptions of

70 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

offender drug and alcohol use during Beyond using prices and self-reports, assaults were used as the dependent some researchers have drug-tested variable, the significance of marijuana rape victims to assess their drug use. decriminalization and cocaine prices was Hindmarch and Brinkmann (1999) found ambiguous (significance depends on that 41 percent of the 1,033 participants model specification). For perceived use tested negative for alcohol and other during robberies, neither was significant. drugs, 37 percent tested positive for alco- Although Markowitz suggests the percep- hol, 19 percent tested positive for cannabi- tion variable is questionable because of noids, and 0.6 percent tested positive for underreporting, these findings raise ques- flunitrazepam (Rohypnol); however, the tions about the causal relationship and lack of information about participant char- the role of drug use by victims, especially acteristics and site locations would pre- marijuana. vent researchers from creating the necessary control groups. Based on an instrument similar to NCVS, Fisher et al.’s (1998) survey of 3,472 ran- Drug users. Tardiff et al. (1994) found domly selected college students found that 31 percent of one sample of homi- that regularly taking recreational drugs pre- cide victims tested positive for cocaine dicted an increased likelihood of a violent metabolites. This rate did not vary for victimization but not of a theft victimiza- firearm deaths versus nonfirearm deaths. tion. For the general population, Cottler et McElrath, Chitwood, and Comerford al.’s (1992) survey of a probability sample (1997) surveyed 308 intravenous drug of 2,663 household residents found that users who were receiving methadone those who had used cocaine or heroin and/or inpatient drug treatment about their more than five times in their lives were victimization and drug use in the previous more than three times as likely to have 6 months. Those reporting heroin use experienced a physical attack than non- were significantly less likely to be victims users. Those who used marijuana more of violent and property crimes. McElrath than five times (no use of other drugs) and et al. argue that heroin users sometimes those who used pills or hallucinogens have “running partners” who may also more than five times were no more likely look out for each other, thus decreasing to have experienced a physical attack than victimization. Crack cocaine users were nonusers. This is one of the few studies four times as likely to be victims of proper- that presents its results by drug and raises ty crime than nonusers, leading the questions about the situations in which authors to suggest, “it is possible that the hard drug users put themselves. drug-seeking behavior associated with crack-cocaine places users in contact with Women. Much of the victimization a larger pool of motivated offenders.” research focuses on women because many of the studies are about sexual Drug-user-on-drug-user crime was also assault. Fisher, Cullen, and Turner (2000) addressed in Inciardi’s delinquency study randomly selected 4,446 college women (1990). Respondents were asked about to participate in their National College not only drug-related offenses they com- Women Victimization Study. That study mitted but also drug-related victimizations; did not find that marijuana use was a sig- 4.6 percent reported being victims of nificant predictor of sexual victimization psychopharmacological-related crimes, and stalking.10 These findings are consis- 39.9 percent reported being victims of tent with Markowitz’s claim that the price drug robberies, and 9.0 percent reported of cocaine and marijuana are not signifi- being victims of systemic violence. cant predictors of sexual victimization.

71 SPECIAL REPORT / JULY 03

Although every youth in the survey used sales by experienced users who could not at least one drug daily, it is not clear maintain legitimate jobs were less than whether the victimizations occurred while minimum wage.11 the victim was under the influence. But the argument that drug selling has Crime victim surveys and offender surveys replaced other income-generating crime require respondents to make attributions might reflect limitations of recent work. about the causes of offenders’ behavior. First, as we have noted, most studies Such causal attributions are susceptible to applying the Goldstein framework were numerous well-documented biases (e.g., conducted at the peak of the crack epi- Nisbett and Ross, 1980), but to date there demic, when the sheer prevalence of has been little methodological work vali- street drug sales was probably at an all- dating these survey responses. time high (see Saner, MacCoun, and Reuter, 1995). Second, most studies have Economic-compulsive violence largely examined crimes with violent outcomes rather than robberies or burglaries Arrested and incarcerated offenders report in which no homicide occurred. One that they committed their offenses to exception is the Caulkins et al. (1997) raise money to purchase drugs. Of course, study, which attributed a substantial this might be a convenient rationalization fraction of robberies and burglaries to or excuse for antisocial behavior. Should economic-compulsive crime, and a size- we believe them? able fraction of those economic-compulsive crimes to cocaine. At least for heroin addiction, the answer is probably yes. Studies of heroin “careers” The ADAM Program provides some oppor- show that the frequency of criminal activi- tunities for studying these issues (e.g., ty tends to covary with periods of intense Arrestee Drug Abuse Monitoring Program, use (see Fagan, 1990, for review), and 1999). The ADAM/DUF (Drug Use Fore- addicts significantly reduce their criminal casting) instrument was modified in 1995 involvement during periods of methadone to include a question asking whether the maintenance (see review in Rettig and arrestee needed drugs or alcohol at the Yarmolinsky, 1995). But in studies applying time of the offense.12 Appendix B summa- the Goldstein taxonomy (see above), rizes data for the period 1995 to 1999 for economic-compulsive criminality has been this survey item. As one would expect, relatively rare. White and Gorman (2000) these attributions are more common for argue, “[B]ecause there is more money in income-generating offenses (14 percent) crack distribution than in previous illegal than for non-income-generating offenses drug markets, drug dealing may have obvi- (10 percent)—a reliable but quite modest ated the need to commit property crimes difference. and income-generating violent crimes” (p. 189). Indeed, in our survey of drug sell- Our understanding and interpretation of ers in Washington, D.C., in the late 1980s economic-compulsive crime ought to (Reuter, MacCoun, and Murphy, 1990), evolve as the scientific understanding more than 40 percent reported keeping of drug dependence evolves. Recent some drugs for their own consumption— decades have seen great progress in the 39 percent of crack sellers and 69 percent understanding of such phenomena as of heroin sellers. However, the claim tolerance, withdrawal, reinforcement, and about the high returns for crack selling is drug craving (see Science, 1997). Leshner probably no longer correct. Bourgois (1997, pp. 45–46) notes that many assume (1996) reports that proceeds from crack the following:

72 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

[T]he more dramatic the physical A brief history of the markets. There withdrawal symptoms, the more seri- was an epidemic of initiation into heroin ous or dangerous the drug must be. use in the 1970s; after that, heroin initia- This thinking is outdated . . . many of tion rates remained low until the late the most addicting and dangerous 1990s. The number of heroin addicts (a drugs do not produce severe physical function of the number of initiates and the symptoms upon withdrawal. . . . length of their addiction careers) remained What does matter tremendously is fairly stable at about 750,000 from 1981 to whether or not a drug causes what 1997.13 During that period, most heroin we now know to be the essence of purchases were made by an aging cohort addiction: compulsive drug seeking of experienced users. and use, even in the face of negative health and social consequences. Powder cocaine and crack had a similar dynamic, only with different parameters. There are also intriguing new findings from Powder cocaine initiation rates were high behavioral economics research on the from about 1975 to 1988; the number of price elasticity of demand for cocaine and dependent users has been quite stable opiates—the percentage decline in de- since about 1988. The crack epidemic mand for a 1-percent increase in price. The came later, from about 1982 to 1990 conventional wisdom is that addicts are (depending on the city; see Blumstein and relatively insensitive to price, at least in the Cork, 1996). Estimates of the number of short run, because they are enslaved to dependent users of either crack or powder their drug and must find ways to obtain it cocaine range from 600,000 to 3,600,000 to avoid withdrawal symptoms. If addicts (see Rhodes et al., 2000). were relatively insensitive to price, one would expect price increases to produce Many retailers are now also frequent increased economic-compulsive crime. But users (Arrestee Drug Abuse Monitoring recent studies (reviewed in Caulkins and Program, 1999). Selling seems to be Reuter, 1996) suggest considerable price opportunistic for many users; sudden sensitivity, with elasticities for cocaine access to an unusually large source of ranging from –0.7 to –2.0. A possible cash may lead a regular buyer to become explanation for the high elasticity among a seller for a day. Thus, at the low end of heavy users is that they spend most of the market, it may be difficult to distin- their earnings on the drug and may re- guish systemic from psychopharmacologi- spond to the increased difficulty of main- cal violence. taining desired consumption levels (i.e., avoiding withdrawal) by seeking treatment. Enforcement against these markets, as measured by years of jail time per ton of drugs, probably declined through the early Systemic violence 1980s but then intensified from 1985 The third of Goldstein’s categories is sys- onward. In 1990, the Colombian govern- temic violence. This has been narrowly ment aggressively attacked the principal interpreted as referring to struggles for exporters of cocaine from Colombia. competitive advantage. We suggest here There are a number of indications that that drug markets generate violence in a this led to a temporary tightening of the variety of ways and that market violence cocaine market; otherwise, prices have varies systematically over time and place. declined throughout the period, while consumption has been declining modestly since 1988.

73 SPECIAL REPORT / JULY 03

Conceptual issues. The markets for illegal 2. The value of the drugs themselves. drugs operate without the usual protec- The cocaine that fills a plastic sand- tions against fraud and violence offered by wich bag is worth thousands of dol- the civil tort system. The state, instead of lars. The return to sudden, situational attempting to facilitate transactions, aims violence could be very high. to disrupt them. Contracts cannot be enforced through written documents and 3. The intensity of law enforcement. the legal system; agreements are made Transactions are conducted under hurriedly, sometimes in ambiguous code, considerable uncertainty as a conse- and orally.14 Territories cannot be allocated quence of increased law enforcement. through bidding for desirable locations Intensified enforcement increases the because there is no enforceable owner- incentives for violence by raising the ship of property for these purposes. adverse consequences of identifying someone as a potential informant. Yet the illegality itself is insufficient to generate high levels of violence in the market. 4. The indirect consequence of drug use. Prostitution, although frequently unsightly Users are more violent and aggres- and sometimes a nuisance, does not gen- sive, and this encourages dealers to erate much by way of additional violence. prefer selling out of doors or in highly Bookmaking, notwithstanding the drama protected settings. It also promotes of the film “The Sting,” was also a gener- unreliable behavior among user/ ally peaceful affair; bookies were more dealers and thus more retaliation by likely to die in bed than on the job. Even their suppliers. for some drugs, the markets generate little violence; marijuana in general does not It is probably the combination of these fac- spark much injury as the result of compe- tors, rather than any one of them, that titive or transactional disputes.15 accounts for the extraordinary violence associated with crack markets in the late Some drug markets, however, are clearly 1980s. That violence seems to have fallen violent; many participants are at risk of substantially in the late 1990s, perhaps being killed or seriously wounded by oth- reflecting the aging of participants in crack ers in the same business, either as buyers markets (Golub and Johnson, 1997), al- or sellers, and there are unintended shoot- though violence itself, as well as enforce- ings of innocent bystanders. The crack ment, may also have selected out the market is thought to be particularly prone most violent participants; Taylor, Caulkins, to market-related violence. and Reuter (2000) present a model in which violence declines with more intense Why are these drug markets, particularly enforcement as a consequence of selec- for crack, so violent? We suggest that four tive incarceration. factors contribute: Competitive and internal violence. 1. The youth of participants. Rates for Attention has been given to violence gen- violent crime peak early, at about ages erated by competition between sellers. 18–22. The young are particularly likely Less attention has been given to violence to lack foresight and thus engage in within selling organizations, although the violence to settle disputes. The crack older literature on organized crime and ille- market was the first mass drug mar- gal markets reported a great deal on this ket in which most of the sellers were (e.g., Block, 1980). young.

74 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Criminal organizations are hindered inter- decline in market-related violence may nally by lack of access to the civil courts. reflect changes in organizational structure. Employment contracts cannot be enforced except privately. Managerial succession is Other market characteristics and vio- complicated by the specificity of reputa- lence. Exhibit 1 presents a simple classifi- tion within the organization; a promising cation of markets according to whether midlevel manager cannot readily provide buyers and sellers come from the neigh- evidence of performance to another borhood or elsewhere. We believe that potential employer so higher level man- this taxonomy, originally identified for pur- agers get weaker market signals and may poses of analyzing vulnerability to enforce- withhold deserved promotions or merit ment (Reuter and MacCoun, 1992), may increases. This gives incentives to lower also be useful in the study of violence. level agents to use violence for upward Markets characterized by mostly resident mobility. dealers and customers are labeled local markets. Export markets are ones in Symmetrical with successional violence is which residents of the neighborhood sell disciplinary violence. Managers have rea- drugs to nonresidents. Markets in which son to fear subordinates who can provide mostly nonresident dealers sell to local evidence against them; the longer lasting residents are characterized here as import The violence in the relationship, the greater the potential markets. Finally, markets in which both atomistic markets for harm from informing. Thus, managers sellers and customers are mostly nonresi- may use violence as a tool to reduce risks dents are labeled here as public markets has different of informing. They have more incentive because they tend to occur at such large sources than for doing so than do high-level dealers in public locations as parks, train or bus sta- transactions with low-level dealers be- tions, or schoolyards. that in markets cause the information about these acts serviced by will spread more rapidly and extensively.16 Each class of market differs in the poten- There are numerous stories of this kind tial for violence. Local markets, precisely larger selling of violence in Colombian drug-dealing because they involve buyers and sellers organizations. organizations. who know each other, do not lend themselves to territorial competition. At the Thus, the violence in atomistic markets other extreme are public markets, in has different sources than that in markets which buyers and sellers cannot readily serviced by larger selling organizations. find each other except at specific loca- Which generates greater violence from a tions; the incentives for territoriality are given set of participants cannot be deter- consequently greater. mined theoretically, but some of the

Exhibit 1. Types of illicit drug markets

Customers Dealers Mostly residents Mostly outsiders

Mostly residents Local market Export market Mostly outsiders Import market Public market

75 SPECIAL REPORT / JULY 03

The case for Transactional violence may also vary in factor in the steep rise in American vio- these dimensions. Local markets discour- lence from 1985 to 1990, and the even crack’s role age cheating of buyers as a consequence steeper drop since 1993 (see Blumstein in the crime of the ongoing connections between buy- and Wallman, 2000). In our view, the case ers and sellers; a local customer is more for crack’s role in the crime rise is quite rise is quite likely to spread information effectively to compelling; its role in the post-1993 compelling; its other potential customers than one who decrease is more subtle and by no means has little connection to other buyers. It is an open-and-shut case. role in the not clear whether much of the transac- post-1993 tional violence comes from buyers, as Many discussions of the crime drop fail opposed to associates and rival sellers. to distinguish between a decline in the decrease is crack market and a decline in the linkage more subtle. If this is correct, then the maturation of between crack and crime—but a decline in cocaine and heroin markets will tend to the crack-crime link is part of the crime reduce market-related violence by reduc- drop outcome to be explained. It is true ing the size of all but local markets. More- that DUF (and now ADAM) data show over, as a result of the dissemination of declines in positive cocaine tests among beepers and cell phones, an increasing arrestees in many cities (e.g., Arrestee share of cocaine transactions may be Drug Abuse Monitoring Program, 1999). occurring in locations (apartments, restau- And the reduced violence attributable to rants, offices) that are agreed on by the crack selling has made crack markets less buyer and seller for their mutual conven- visible. But nationwide, hardcore cocaine ience. Johnson, Golub, and Dunlap (2000, use remained surprisingly stable during p. 191, table 6.1) report that in New York the 1990s (Rhodes et al., 2000). Indeed, City in the 1990s, the “seller style” includ- from 1990 to 1998, there were rising ed phone and delivery services as well as cocaine mentions in emergency rooms freelancers. Poor and socially isolated (Substance Abuse and Mental Health cocaine users still frequently conduct Services Administration, 2001) and rising transactions in exposed locations, chosen cocaine seizures. Nevertheless, recent precisely because they facilitate the com- multicity comparisons (Baumer et al., ing together of buyers and sellers. So 1998; Lattimore et al., 1997) indicate reli- probably do many heroin addicts, given able positive correlations between various their generally impoverished state. The indices of crack use and homicide and ability to choose locations on the basis of other offense rates. specific situational need not only reduces territorially motivated violence but also Various experts have suggested that the reduces the vulnerability of buyers to changing dynamics of drug markets may robbery and other victimization because matter as much or more as any decline in fewer of them need to congregate at spe- total market activity (e.g., Ousey and Lee, cific locations, which thus become less 2000). Below, we consider a few more attractive to predators. complex accounts of the link between crack market dynamics and violence.

NDRI’S conduct norm account. Johnson The temporal dynamics of and his colleagues at NDRI (Lipton and drug markets Johnson, 1998) have produced a valuable In the past several years, numerous interdisciplinary, multimethod program of authors have examined the emergence research on street drug markets in New and decline of crack markets as a key York, spanning several decades. They

76 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

recently offered an account of the decline not, of course, their consumption of in drug-related violence based on the blunts). Still, the evidence is causally notion of “conduct norms” (Johnson, ambiguous. Are these conduct norms Golub, and Dunlap, 2000), arguing that actually causes of the decline in violence New York street drug markets have during the 1990s, are they descriptions of passed through three phases. (They it, or are they consequences of it? vacillate between “period” and “cohort” versions of the story.) The “heroin injec- There is little doubt that conduct norms tion era” peaked during 1960–73; the exist and are important in shaping deviant “cocaine/crack era” peaked during behavior. Cialdini, Kallgren, and Reno 1984–89; and the “marijuana/blunts era” (1991) make a useful distinction between started around 1990. Associated with injunctive norms (what others think I each era are distinct birth cohorts with dis- should do) and descriptive norms (what tinctive behavioral patterns. “HeroinGen” others are actually doing). There is ample drug users (born 1945–54) were active in evidence that purely descriptive norms— drug sales and property crime, but gun changes in the local prevalence of a use was relatively rare. “CrackGen” drug behavior—can have a self-reinforcing users (born 1955–69) frequently participat- action. But attitudes and norms are ed in robbery and used guns for protection shaped by behavior as well as shaping it; and reputation. Finally, “BluntGen” drug research on cognitive dissonance theory users (born 1970–79) are less likely than and self-perception theory suggest that early cohorts to engage in violence. such conformity-based behavioral changes will tend to produce corresponding (but Drawing on their rich ethnographic data- retrospective) changes in relevant atti- base, Johnson and colleagues (2000) tudes (see Eagly and Chaiken, 1993). argue that these behavior changes reflect Controlled social psychology experiments two successive transformations of con- show that norm diffusion effects occur duct norms for appropriate behavior in the and that they can be strong, but these drug-using community. For example, in experiments also show that apparent CrackGen’s “Subculture of Assault,” a norm effects are sometimes spurious shared norm counseled: “Be aggressive (e.g., Kerr et al., 1987). and threatening to avoid robbery. . . . Carry weapons for protection. . . . Threaten or Clearly, research on drug-using norms can- assault those who attempt to sell crack in not move to the laboratory—although one your territory. Maintain your reputation as can imagine informative scenario-based dangerous, tough, and ‘crazy,’ regardless experiments embedded in field inter- of the physical harm inflicted or suffered” views. But it would be enormously useful (p. 181). But for the BluntGen, the norm to make additional use of the NDRI data states: “Don’t use crack. Crackheads are (and related data sources, such as the s---! . . . Addicts are the scum of the earth. Office of National Drug Control Policy’s Stay safe, stay alive. Don’t mix cocaine Pulse Check), linking the timing of the or heroin with my marijuana. Shun and ethnographic material more precisely to exclude heroin and crack users from peer month-to-month quantitative archival data groups” (p. 185). on drug selling (or its proxies) and violent crimes. Furthermore, archives of ethno- This norm account is fascinating and quite graphic data collected in different cities plausible. From a policy perspective, it during the past decade might be reana- would be tremendously useful to find lyzed to search for cross-city norm differ- away to preserve and promote the ences that might correlate with cross-city BluntGen’s more pacifist stance (though differences in violent crime. Ideally, one

77 SPECIAL REPORT / JULY 03

The linkage might develop methods for identifying cocaine-using population consists mostly “leading indicators” of emerging trends in of an aging cohort who started using in between drug drug using, drug selling, and drug-related the late 1980s, in much the same way selling and gun violence. that heroin addicts disproportionately belong to cohorts who initiated use in the possession is Blumstein and Cork’s drug-gun diffu- 1970s. If this is correct, drug-related crimi- well established. sion account. In a series of articles (see nality should continue to decline, absent Blumstein, 2000a; Blumstein and Cork, new waves of initiation, as addicts “mature 1996; Cork, 1999), Alfred Blumstein and out” of violent crime or die from drug- his collaborator Daniel Cork hypothesize a related illnesses or natural causes. causal chain linking the late 1980s crack epidemic to rising violence nationwide. Many observers were struck by the vio- According to Blumstein, the 1980s growth lence of 1980s crack markets relative to in illicit drug markets, together with strin- earlier heroin and marijuana markets. gent enforcement crackdowns, led to the Many have speculated that such markets recruitment of juvenile drug sellers. The “mature” over time as (a) dealer territories intense market competition together with are firmly established, (b) casual users the recruitment and rewarding of particu- drop out of the market, and (c) hardcore larly aggressive youths created a need for users establish reliable dealer connec- sellers (as well as nonseller youths in mar- tions. All these factors suggest a shift ket neighborhoods) to be armed. This from open-air public markets toward more increased demand fueled an expansion in clandestine arrangements that seem less the illicit gun market and a diffusion of prone to violence.18 But at present, this is guns. The linkage between drug selling largely speculative; there is anecdotal and and gun possession is well established ethnographic evidence for such changes (see Decker, Pennell, and Caldwell 1997; but little systematic longitudinal research Sheley, 1994; Tardiff et al., 1994).17 Cork that establishes a clear trajectory over (1999) found support for the temporal time. sequence of the Blumstein account using a sophisticated diffusion modeling analysis of time-series data from multiple cities. The consequences of The Blumstein model is a compelling prohibition and its account of the rise of violent crime, but enforcement more work is needed to establish its explanatory power as an account of the Drug involvement as crime subsequent decline in violence. The model is not inconsistent with that decline—a The convention in articles on drugs-crime decline in the crack market should have linkages is to state that for the purposes reduced the need to be armed—but future of the essay, the fact that drug use (and research will have to assess whether sometimes drug selling) per se is a crime declines in the prevalence of drug selling is not relevant to the analysis. But the illicit (as opposed to changes in other features status of street drugs is vitally important of the markets) have produced reductions to the analysis in several ways. First, in the likelihood of gun possession and drug prohibition is arguably necessary for gun violence. Goldstein’s category of systemic (market- related) violence (MacCoun and Reuter, The maturation of addicts and of illicit 2001).19 We simply do not observe routine drug markets. Because of reduced initia- violence among alcohol or tobacco vention rates, it appears that the hardcore dors. Second, Goldstein’s economic- compulsive violence, although not caused

78 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

by prohibition, is surely exacerbated by it Freeman (1996) offers a formal economic because drug prohibition almost certainly model that interprets this replacement raises the price of heroin or cocaine far effect in terms of the elasticity of supply above what would be their retail market of dealers with respect to drug market prices (MacCoun and Reuter, 2001). wages. The supply of dealers should Finally, there are reasons to believe that reflect this sensitivity to wages as well as the illicit status of drugs might have subtle changes in earnings opportunities in the criminogenic effects through several dif- licit market (i.e., shift in the supply curve) ferent mechanisms, including forbidden and the demand for drugs (i.e., shifts in fruit effects, labeling or stigmatization the demand curve). effects, and “stigma swamping.”20 Here we highlight two such mechanisms. At present, there is surprisingly little evidence either for or against the replace- Incapacitation and replacement effects. ment hypothesis. One indirect argument Several authors (e.g., Blumstein, 2000b; for its plausibility is that the explosive Freeman, 1996; Kleiman, 1997) have growth in the incarceration of drug sellers suggested that the incarceration of drug during the past decade was not accompa- sellers is likely to produce a weaker inca- nied by increases in street cocaine prices, pacitation effect than would occur for as one might expect if the supply of street other offense categories, such as property dealers was tightening (Blumstein, 2000b; and sex offenses. Indeed, some have see also DiNardo, 1993). Indeed, street speculated that a replacement process prices have dropped substantially (Rhodes might even produce a net increase in the et al., 2000). Another indirect argument is prevalence of drug selling. In a highly com- the sheer prevalence of drug market par- petitive illicit market, the incarceration of a ticipation in some communities during the drug seller creates lucrative drug-selling late 1980s, when drug sellers were being opportunities (customers and sales territo- incarcerated at record levels. For example, ry) for others. According to Blumstein Saner et al. (1995) estimated that in Wash- (2000b): ington, D.C., during 1985–91, nearly one- third of African-American male residents The pathological rapist’s crimes from the 1964–67 birth cohorts were almost certainly are not replaced on charged with drug selling. the street, and so one can expect his full array of crimes to be incapacitat- Statistical analyses of archival data might ed. . . . A burglar’s crimes may be test the replacement hypothesis by look- replaced if he is serving a fence, who ing for evidence of increases in the initia- would recruit a replacement; alterna- tion to drug selling as a function of the tively, if he is simply operating on his arrest and incarceration of dealers. Ethno- own, the crimes are not likely to be graphic studies might examine whether replaced. And the participant in recruitment activities increase following organized vice activity such as drug police crackdowns and whether existing dealing would be likely to have his street dealers increase their activity. But transactions replaced by whatever isolating replacement effects will be tricky; organizational structure is serving the note that general deterrence and replace- market demand. That replacement ment effects, if they exist, will offset each could be achieved by some combina- other, which may make it hard to find any tion of recruiting new sellers or by effect of sanctions on subsequent dealing. increasing the rate of activity of sellers already active in the market.

79 SPECIAL REPORT / JULY 03

Police tactics Can enforcement amplify violence? reducing the total social harm caused by Several authors (Eck and McGuire, 2000; street drug selling. Some tactics might designed to reduce MacCoun and Reuter, 2001, chapter 6; directly reduce drug-related violence. the supply of drugs Reuter, 1989; Riley, 1998) have argued that under certain conditions, aggressive One example involves efforts to drive (and of drug drug enforcement might actually increase dealers indoors (see Kennedy’s 1993 suppliers) may or drug-related violence. Rasmussen, Benson, analysis of Tampa’s QUAD program). Of and their associates have examined course, crack houses are not without their may not be the whether more intense drug enforcement harms. In an ethnographic study of the most effective increases violent crime; much of this work crack market in Detroit, Mieczkowski is summarized in Rasmussen and Benson (1990, p. 90) concludes that “tavern-style means of (1994). The mechanisms involved are crack houses may encourage and make reducing the total quite varied. For example, enforcement possible hypersexuality among partici- might lead to more violence in competi- pants and thus increase the STD and HIV social harm tion. Benson and colleagues (1992) found rates. The use of barter as a supplement caused by street that the violent crime rate in a community to a cash economy in the crack trade rep- increased with more drug arrests in a resents further complications in creating drug selling. neighboring community. This, they argue, social policies in reaction to this behavior.” is a displacement effect; dealers move Still, indoor markets are likely to be less from the targeted community to the violent. But the effects are multiple and neighboring one and struggle over the hard to balance. On one hand, indoor mar- establishment of territories. Another kets are less susceptible to police surveil- mechanism works through the limited lance or sting operations. On the other capacity of the correctional system; hand, driving dealers indoors might in- increased prison space for drug offenders crease users’ search costs (Moore, 1990) reduces the penalties for other crimes, and thus reduce demand. Consumers in including violent crimes, and thus in- export markets would bear a dispropor- duces higher victimization. Benson and tionate share of these search costs be- Rasmussen (1991) argue that, even cause the locals often know the local assuming that prison is effective only dealers and could easily locate them. This through incapacitation and not deterrence, might lead to new local markets in the the observed rise in the resources devot- areas from which the export consumers ed to drug enforcement in Florida in the are coming and the associated neighbor- 1980s might have increased other crime hood violence that Benson et al. (1992) by 10 percent. examined. Further research on these issues is needed. Supply reduction versus violence reduction. An important dialogue with Heroin maintenance. If the drugs-crime respect to drug users involves the pros- link is mediated by the high price and con- pects and tensions of integrating use ditions of sale of a drug, and if a relatively reduction strategies with harm reduction small number of frequent users are re- strategies (MacCoun, 1998; MacCoun and sponsible for much of the crime, then per- Reuter, 2001). We see an analogous issue haps allowing access to that drug legally with respect to the policing of drug mar- for those least able to quit might reduce kets (MacCoun and Reuter, 1994). Police associated crime. There is increasing infor- tactics designed to reduce the supply of mation and interest in exploring just this drugs (and of drug suppliers) may or may possibility for heroin (see MacCoun and not be the most effective means of Reuter, 2001).

80 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

In January 1994, Swiss authorities opened and Reuter, 2001, chapter 15). But we a number of government-administered should not reflexively dismiss, without heroin maintenance clinics.21 Registered serious analysis, an intervention that could addicts can inject heroin at a government in theory (and with some fragmentary evi- clinic under the care of a nurse up to three dence) help reduce the criminality of exist- times a day, 7 days a week. Patients have ing heroin users and perhaps shrink the to be over 18, have injected heroin for 2 heroin street market, thereby creating new years, and have failed at least two treat- barriers to heroin initiation. If nothing else, ment episodes. By the end of the initial serious discussion of such a program, and research trials of this program, more than perhaps even formal modeling of alterna- 800 patients had received heroin on a tive hypotheses about its likely effects, regular basis without any leakage into the might significantly advance our thinking illicit market. No overdoses were reported about drug market dynamics and the possi- among participants while they stayed in bilities for effective intervention. the program. A large majority of participants had maintained the regime of daily attendance at the clinic; 69 percent were Summing up: Directions for in treatment 18 months after admission. This was a high rate relative to those future research found in methadone programs. About half Here we summarize our suggestions for of the “dropouts” switched to other forms profitable future research, in the order in of treatment; some chose methadone and which we discussed them: others chose abstinence-based therapies. The crime rate among all patients dropped ■ Methodological attention to the meas- during the course of treatment, use of urement of Goldstein’s taxonomy of nonprescribed heroin dipped sharply, and drugs-violence links and to the validation unemployment fell from 44 to 20 percent. of self-reports of victim and offender causal attributions for the role of drugs Critics, such as an independent review in criminal offenses. panel of the World Health Organization, reasonably asked whether the claimed ■ Greater attention to the role of drug use success was a result of the heroin or the in criminal victimization. many additional services provided to trial participants. And the evaluation relied pri- ■ Retrospective historical analysis of long- marily on the patients’ own reports, with term trends in drug use, drug arrests, few objective measures. Nevertheless, and drug-related crime, including recod- despite the methodological weaknesses, ing of ethnographic databases, applica- the results of the Swiss trials provide evi- tion of the Goldstein coding scheme to dence of the feasibility and potential effec- homicide case files, age/period/cohort tiveness of this approach. In late 1997, the analyses, and econometric time-series Swiss government approved a large-scale analyses. expansion of the program. A similar pro- ■ Determination of the causal relation- gram is under development in the Nether- ships underlying comorbid drug abuse lands and in Hamburg, Germany. and mental illness conditions. The proposal to study heroin maintenance ■ Extension and replication of the rich on a trial basis in the United States is politi- experimental literature on situational cally controversial and would be logistically moderators of alcohol-related aggres- difficult. Moreover, the normative and sion, as applied to other drugs. moral issues are clearly complex (MacCoun

81 SPECIAL REPORT / JULY 03

■ Econometric analysis of the effects of Notes drug price changes on drug-related criminality. 1. Except where noted, these statistics were reported in Drug-Related Crime (Office of National Drug ■ Assessment of the effects of the avail- Control Policy, 2000). ability of licit work and licit wage levels 2. In fact, Goldstein et al.’s (1989) findings might not on criminality. fully represent New York City since they did not look at the entire population or a random sample of homi- ■ Additional multicity analyses (and cross- cides. Rather, they chose one zone in each of four neighborhood analyses within cities) different boroughs, with the goal of sampling pre- with an emphasis on understanding het- cincts that represented a cross-section of New York City. erogeneity in drugs-crime relationships: Spatial analyses, analyses of variation in 3. These findings challenge the recent generalization the demand for different drugs, gang by White and Gorman (2000, p. 189) that “the eco- versus nongang involvement, ethnic and nomic motivation explanation has not been support- other demographic groupings, indoor ed among adolescents.” versus outdoor markets, import versus 4. Our understanding is that the new NIBRS (National export versus local versus public mar- Incident-Based Reporting System) database perpetu- kets, etc. ates this. Officers only have to report the circumstances of the offense (which includes drug dealing) ■ Estimation of incapacitation versus for aggravated assaults/homicide (considered one replacement effects resulting from the category in the victim-level file). incarceration of drug sellers. 5. Approaches might include confirmatory factor analysis, cluster analysis, Q-sort, or Ragin’s (2000) ■ Simulation modeling and eventual fuzzy-set approach. We are less interested in defend- pilot tests of the efficacy of heroin ing a particular method than in pointing out the sur- maintenance. prising lack of attention to these measurement and conceptualization issues in the field. One other topic that was not even hinted 6. Fagan (1990, p. 255) and White and Gorman at in our analysis and has been almost (2000, p. 185) argue that, if anything, marijuana and totally neglected in the empirical research opiates serve to suppress aggression. Actually, literature also should receive attention: the Bushman’s (1990) meta-analysis found more aggres- likelihood of causal linkages between illicit sion among marijuana smokers than placebo controls drug use and such white-collar crimes as in laboratory experiments. But this effect is partly due to the fact that the placebo controls showed 22 corruption, fraud, and embezzlement. significantly less aggression than nondrug controls, indicating that participants also believed marijuana This is a long list of topics. That in itself is would induce passivity. a reminder of how little has been done to implement and build on Goldstein’s in- 7. Note that other psychiatric disorders are less sightful taxonomy. Advances will require common among substance abusers than substance abuse is among the mentally ill (Miller, 1993). an acceptance of the fact that drugs may differ widely in the extent and form of 8. Beau Kilmer’s work on this section was supported their criminogenic effects. That substan- by NIDA grant R01DA12724. tially complicates an already difficult enterprise but is likely to be the source of 9. The assumption that decriminalization (as opposed to legalization) is an indicator of lower price is ques- considerable policy insight. tionable. In theory, it might increase demand by reducing the nonmoney costs, which should increase

82 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

price. However, evaluations of decriminalization in 16. Smith and Varese (2001) model the use of coer- 11 U.S. States, South Australia, the Australian Capital cive violence in markets for Mafia extortion; the Territory, and the Netherlands fail to show any ef- model can be applied to intraorganizational violence fects on demand (MacCoun and Reuter, 2001). as well.

10. The authors report the statistically significant vari- 17. Decker, Pennell, and Caldwell (1997) did not find ables, not the entire model. The entire model is listed that drug users (rather than sellers) were more likely in Fisher, Cullen, and Turner (forthcoming) and includes to be carrying a gun than other arrestees. a variable for “Frequency of smoking pot or hashish.” Because the significant predictors for stalking are the 18. Alfred Blumstein appeared to endorse this same in the published and unpublished pieces, we account in his public comments at the 2000 Annual assume the same model was used. Because this is Meeting of the American Society of Criminology. likely to be the model used to predict sexual victimization in the published piece, we report that marijuana 19. Necessary, but not sufficient; see Zimring and use does not predict sexual victimization. Hawkins, 1997; Ousey and Lee, 2000.

11. Even if true, high returns from crack selling do 20. The term “stigma swamping” was suggested to not lessen the criminogenic consequences of the us by Jon Caulkins as an apt label for a phenomenon market; the issue is what share of revenues are gen- about which many have speculated (e.g., Jacobsen erated by legitimate earnings or welfare and other and Hanneman, 1992; McGraw, 1985; Petersilia, transfer payments received by buyers. 1990)—the notion that the stigma associated with arrest and even incarceration is reduced by the sheer 12. The question yields four binary variables about prevalence of those sanctions. The term “stigma whether the arrestee was in need of drugs/alcohol swamping” is an informal control counterpart to (NEEDNO), alcohol (NEEDALC), cocaine (NEED- Kleiman’s (1993) formal control version, “enforce- COCR), and marijuana (NEEDMAR) during the crime ment swamping.” and one text variable (NEEDOTHR) where the coder is asked to specify if the arrestee mentioned another 21. The earlier British experience with prescription drug. Curiously, the 1995 (part 2) and 1999 ADAM heroin is more notorious but less informative; see codebooks do not report any binary variable for MacCoun and Reuter, 2001, chapter 12. heroin—widely believed to be the major source of economic-compulsive crime. Of the 44,000 ADAM 22. We thank Terence Dunworth for making this arrestees in 1999, we estimate (using the open- observation. 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 References 35 percent committed income-generating crimes. Anderson, E. (1994). Code of the streets: 13. ONDCP reports, based on Rhodes et al. (1995, How the inner city environment fosters a 2000), that the prevalence of frequent use fell by need for respect and a self-image based one-third between 1988 and 1993 and then returned on violence. Atlantic Monthly, 273, 80–94. to its 1988 level by 1998. It is difficult to identify supporting evidence for such a dramatic fluctuation in Arrestee Drug Abuse Monitoring Program the figures. (1999). 1998 annual report on drug use 14. The bookmaking business has certainly generat- among adult and juvenile arrestees (Re- ed written records; but that is more central to the search Report, NCJ 175656). Washington, business itself, which involves the extension of cred- DC: U.S. Department of Justice, National it and usually numerous near-simultaneous transac- Institute of Justice. tions between any one buyer and seller.

15. In the District of Columbia in the mid- to late Baumeister, R.F., Smart, L., and Boden, 1990s, it was reported that some street gangs were J.M. (1996). Relation of threatened ego- in violent disputes over the marijuana market (Pierre, tism to violence and aggression: The dark 1996; Lattimore et al., 1997). side of high self-esteem. Psychological Review, 103, 5–33.

83 SPECIAL REPORT / JULY 03

Baumer, E., Lauritsen, J. L., Rosenfeld, R., Brownstein, H., Baxi, H., Goldstein, P., and and Wright, R. (1998). The influence of Ryan, P. (1992). The relationship of drugs, crack cocaine on robbery, burglary, and drug trafficking, and drug traffickers to homicide rates: A cross-city longitudinal homicide. Journal of Crime and Justice, analysis. Journal of Research on Crime 15, 25–44. and Delinquency, 35, 316–340. Bureau of Justice Statistics (1997a). Benson, B.L., Kim, I., Rasmussen, D.W., Survey of inmates in Federal correctional and Zuehlke, T.W. (1992). Is property facilities. Washington, DC: U.S. Depart- crime caused by drug use or by drug ment of Justice, Bureau of Justice enforcement policy? Applied Economics, Statistics and Federal Bureau of Prisons. 24, 679–692. Bureau of Justice Statistics (1997b). Benson, B.L., and Rasmussen, D.W. Survey of inmates in State correctional (1991). The relationship between illicit facilities. Washington, DC: U.S. Depart- drug enforcement and property crimes. ment of Justice, Bureau of Justice Contemporary Policy Issues, 9, 106–115. Statistics.

Block, A. (1980) East side-west side. Bureau of Justice Statistics (1991a). Cardiff: University of Wales Press. Survey of inmates in Federal correctional facilities. Washington, DC: U.S. Depart- Block, C., Block, R., and Illinois Criminal ment of Justice, Bureau of Justice Justice Information Authority. (1998). Statistics and Federal Bureau of Prisons. Homicides in Chicago, 1965–1995. Ann Arbor: Intra-university Consortium for Bureau of Justice Statistics (1991b). Political and Science Research. Survey of inmates in State correctional facilities. Washington, DC: U.S. Depart- Blumstein, A. (2000a). Disaggregating the ment of Justice, Bureau of Justice violence trends. In A. Blumstein and J. Statistics. Wallman (eds.), The crime drop in America (pp. 13–44). New York: Cambridge Univer- Bushman, B.J. (1990). Human aggression sity Press. while under the influence of alcohol and other drugs: An integrative research re- Blumstein, A. (2000b). The replacement of view. Psychological Science, 2, 148–152. drug offenders to diminish the effects of incarceration. Unpublished manuscript, Bushman, B.J. (1996). Individual differ- Carnegie Mellon University. ences in the extent and development of aggressive cognitive-associative networks. Blumstein, A., and Cork, D. (1996). Linking Personality and Social Psychology Bulletin, gun availability to youth gun violence. Law 22, 811–819. and Contemporary Problems, 59, 5–24. Bushman, B.J. (1997). Effects of alcohol Blumstein, A., and Wallman, J. (eds.) on human aggression: Validity of proposed (2000). The crime drop in America. New explanations. In M. Galanter (ed.), Recent York: Cambridge University Press. developments in alcoholism, Vol. 13: Alcohol and violence: Epidemiology, neu- Bourgois, P. (1996). In search of masculini- robiology, psychology, family issues ty: Violence, respect and sexuality among (pp. 227–243). New York: Plenum Press. Puerto Rican crack dealers in East Harlem. British Journal of Criminology, 36, 412–427.

84 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Caulkins, J., and Reuter, P. (1996). Editor- American Journal of Psychiatry, 149, ial: The meaning and utility of drug prices. 664–670. Addiction, 91, 1261–1264. Critchlow, B. (1986). The powers of John Caulkins, J., Rydell, C., Schwabe, W., and Barleycorn: Beliefs about the effects of Chiesa, J. (1997). Mandatory minimum alcohol on social behavior. American drug sentences: Throwing away the key Psychologist, 41, 751–764. or the taxpayers’ money? Santa Monica: RAND Drug Policy Research Center. Decker, S.H., Pennell, S., and Caldwell, A. (1997). Illegal firearms: Access and use by Chin, K.L., and Fagan, J. (1992). The im- arrestees (Research in Brief, NCJ 163496). pact of crack on criminal careers: Crime Washington, DC: U.S. Department of and drug involvement following initiation Justice, National Institute of Justice. into cocaine smoking. Unpublished manuscript on file with Rutgers University DiNardo, H. (1993). Law enforcement, School of Criminal Justice. the price of cocaine, and cocaine use. Mathematical and Computer Modeling, Cialdini, R.B., Kallgren, C.A., and Reno, 17, 53–64. R.R. (1991). A focus theory of normative conduct: A theoretical refinement and Eagly, A.H., and Chaiken, S. (1993). The reevaluation of the role of norms in human psychology of attitudes. Fort Worth, TX: behavior. In M. Zanna (ed.), Advances in Harcourt Brace Jovanovich. experimental social psychology (vol. 24, pp. 201–234). New York: Academic Press. Eck, J., and McGuire, E. (2000). Have changes in policing reduced violent crime? Cohen, D., Nisbett, R.E., Bowdle, B.F., An assessment of the evidence. In A. and Schwarz, N. (1996). Insult, aggression, Blumstein and J. Wallman (eds.), The and the Southern culture of honor: An crime drop in America (pp. 207–265). New “experimental ethnography.” Journal of York: Cambridge University Press. Personality and Social Psychology, 70, 945–960. Fagan, J. (1990). Intoxication and aggression. In M. Tonry and J.Q. Wilson (eds.), Compton, W.M., Cottler, L.B., Phelps, Drugs and crime, vol. 13 of Crime and jus- D.L., Abdallah, A.B., and Spitznagel, E.L. tice: A review of research (pp. 241–320). (2000). Psychiatric disorders among drug Chicago: University of Chicago Press. dependent subjects: Are they primary or secondary? American Journal on Addic- Fagan, J., and Chin, K. (1990). Violence tions, 9, 126–134. as regulation and social control in the distribution of crack. In M. De La Rosa, E. Cork, D. (1999). Examining space-time Lambert, and B. Gropper (eds.), Drugs and interaction in city-level homicide data: violence: Causes, correlates, and conse- Crack markets and the diffusion of guns quences (Research Monograph 103, pp. among youth. Journal of Quantitative 8–43). Rockville, MD: U.S. Department of Criminology, 15, 379–406. Health and Human Services, National Institute on Drug Abuse. Cottler, L.B., Compton, W.M., Mager, D., Spitznagel, E., and Janca, A. (1992). Post- Fishbein, D. (2000). Neuropsychological traumatic stress disorder among sub- function, drug abuse, and violence: A con- stance users from the general population. ceptual framework. Criminal Justice & Behavior, 27, 139–159.

85 SPECIAL REPORT / JULY 03

Fisher, B., Cullen, F., and Turner, M. Golub, A., and Johnson, B.D. (1997). (2000). The sexual victimization of college Crack’s decline: Some surprises across women (Research Report, NCJ 182369). U.S. cities (Research in Brief, NCJ Washington, DC: U.S. Department of 165707). Washington, DC: U.S. Depart- Justice, National Institute of Justice. ment of Justice, National Institute of Justice. Fisher, B., Cullen, F. and Turner, M. (forthcoming). Being pursued: Stalking victim- Harwood, H., Fountain, D., and Livermore, ization in a national study of college G. (1998). The economic costs of alcohol women. Criminology and Public Policy. and drug abuse in the United States— 1992. Bethesda, MD: U.S. Department of Fisher, B., Sloan, J., and Chunmeng, L. Health and Human Services, National (1998). Crime in the ivory tower: The level Institute on Drug Abuse. and sources of student victimization. Criminology, 36, 671–710. Hindmarch, I., and Brinkmann, R. (1999). Trends in the use of alcohol and other Freeman, R.B. (1996). Why do so many drugs in cases of sexual assault. Human young American men commit crimes and Psychopharmacology: Clinical and Ex- what might we do about it? Journal of perimental, 14, 225–231. Economic Perspectives, 10, 25–42. Inciardi, J.A. (1990). The crack-violence Gandossy, R., Williams, J., Cohen, J., and connection within a population of hard Harwood, H. (1980). Drugs and crime: A core adolescent offenders. In M. De La survey and analysis of the literature (NCJ Rosa, E. Lambert, and B. Gropper (eds.), 159074). Washington, DC: U.S. Depart- Drugs and violence: Causes, correlates, ment of Justice, National Institute of and consequences (Research Monograph Justice. 103, pp. 92–111). Rockville, MD: U.S. Department of Health and Human Giancola, P.R. (2000). Executive function- Services, National Institute on Drug ing: A conceptual framework for alcohol- Abuse. related aggression. Experimental & Clinical Psychopharmacology, 8, 576–597. Inciardi, J.A., and Pottieger, A.E. (1994). Crack-cocaine use and street crime. Goldstein, P. (1985). The drug/violence Journal of Drug Issues, 24, 273–292. nexus: A tripartite conceptual framework. Journal of Drug Issues, 14, 493–506. Inciardi, J.A., and Pottieger, A.E. (1991). Kids, crack, and crime. Journal of Drug Goldstein, P., Brownstein, H.H., and Ryan, Issues, 21, 257–270. P. J. (1992). Drug-related homicide in New York: 1984 and 1988. Crime & Delinquency, Ito, T.A., Miller, N., and Pollock, V.E. 38, 459–476. (1996). Alcohol and aggression: A meta- analysis on the moderating effects of Goldstein, P., Brownstein, H.H., Ryan, P.J., inhibitory cues, triggering events, and self- and Bellucci, P.A. (1989). Crack and homi- focused attention. Psychological Bulletin, cide in New York City, 1988: A conceptual- 120, 60–82. ly based event analysis. Contemporary Drug Problems, 16, 651–687. Jacobsen, C., and Hanneman, R.A. (1992). Illegal drugs: Past, present and possible futures. Journal of Drug Issues, 22, 105–120.

86 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Johnson, B., Golub, A., and Dunlap, E. Leshner, A.I. (1997). Addiction is a brain (2000). The rise and decline of hard drugs, disease, and it matters. Science, 278, drug markets, and violence in inner-city 45–47. New York. In A. Blumstein and J. Wall- man (eds.), The crime drop in America Lipton, D., and Johnson, D. (1998). (pp. 164–206). New York: Cambridge Smack, crack, and score: Two decades University Press. of NIDA-funded drugs and crime research at NDRI 1974–1994. Substance Use & Kennedy, D. (1993, March). Closing the Misuse, 33, 1779–1815. market: Controlling the drug trade in Tampa, Florida (Program Focus, NCJ MacCoun, R. (1998). Toward a psychology 139963). Washington, DC: U.S. Depart- of harm reduction. American Psychologist, ment of Justice, National Institute of 53, 1199–1208. Justice. MacCoun, R.J., and Reuter, P. (1994, Kerr, N.L., MacCoun, R.J., Hansen, C.H., March 7). Harm reduction as a response and Hymes, J.A. (1987). Gaining and los- to drug selling and violence. Paper pre- ing social support: Momentum in decision sented at the Fifth International Confer- making groups. Journal of Experimental ence on the Reduction of Drug-Related Social Psychology, 23, 119–145. Harm, Toronto.

Kessler, R.C., Nelson, C.B., McGonagle, MacCoun, R.J., and Reuter, P. (2001). K.A., Edlund, M.J., Frank, R.G., and Leaf, Drug war heresies: Learning from other P.J. (1996). The epidemiology of co- vices, times, and places. New York: occurring addictive and mental disorders: Cambridge University Press. Implications for prevention and service utilization. American Journal of Ortho- Markowitz, S. (2000). An economic analy- psychiatry, 66, 17–31. sis of alcohol, drugs, and crime in the National Criminal Victimization Survey. Kleiman, M.A.R. (1993). Enforcement National Bureau of Economic Research swamping: A positive-feedback mecha- Working Paper 7982. Cambridge, MA: nism in rates of illicit activity. Mathe- National Bureau of Economic Research. matical and Computer Modeling, 17, 65–75. McElrath, K., Chitwood, D., and Comerford, M. (1997). Crime victimization among Kleiman, M.A.R. (1997). The problem of injection drug users. Journal of Drug replacement and the logic of drug law Issues, 27, 771–783. enforcement. Drug Policy Analysis Bulletin, 3, 8–10 [Online]. Available: McGraw, K.M. (1985). Subjective proba- www.fas.org/ drugs/issue3.htm. bilities and moral judgments. Journal of Experimental Social Psychology, 21, Klein, M., Maxson, C., and Cunningham, 501–518. L. (1991). “‘Crack,’ street gangs, and violence. Criminology, 29, 623–650. Meehl, P.E. (1995). Bootstraps taxomet- rics: Solving the classification problem in Lattimore, P.K., Trudeau, J., Riley, K.J., psychopathology. American Psychologist, Leiter, J. and Edwards, S. (1997). Homi- 50, 266–275. cide in eight U.S. cities: Trends, context, and policy implications (Research Report, Mieczkowski, T. (1990). The operational NCJ 167262). Washington, DC: U.S. De- styles of crack houses in Detroit. In M. De partment of Justice, National Institute of La Rosa, E. Lambert, and B. Gropper Justice. (eds.), Drugs and violence: Causes, correlates, and consequences (pp. 60–91). 87 SPECIAL REPORT / JULY 03

Rockville, MD: U.S. Department of Health meeting of the American Society of and Human Services, National Institute on Criminology, San Francisco. Drug Abuse. Parker, R., and Auerhahn, K. (1998). Miller, N.S. (1993). Comorbidity of psychi- Alcohol, drugs, and violence. Annual atric and alcoholic/drug disorders: Inter- Review of Sociology, 24, 291–311. actions and independent status. Journal of Addictive Diseases, 12, 5–16. Petersilia, J. (1990). When probation becomes more dreaded than prison. Moore, M.H. (1990). Supply reduction and Federal Probation, 54, 23–27. drug law enforcement. In M. Tonry and J.Q. Wilson (eds.), Drugs and crime, vol. Pierre, R.E. (1996, September 24). 13 of Crime and justice: A review of re- Marijuana’s violent side. Washington search (pp. 109–157). Chicago: University Post, A1. of Chicago Press. Ragin, C.C. (2000). Fuzzy-set social sci- Mueser, K.T., Yarnold, P.R., Rosenberg, ence. Chicago: University of Chicago S.D., Swett, C., Miles, K.M., and Hill, D. Press. (2000). Substance use disorder in hospi- talized severely mentally ill psychiatric Rasmussen, D.W., and Benson, B.L. patients: Prevalence, correlates, and sub- (1994). The economic anatomy of a drug groups. Schizophrenia Bulletin, 26, 179–192. war: Criminal justice in the commons. Lanham, MD: Rowman and Littlefield. National Crime Victimization Survey (2000). Criminal victimization in the Rettig, R.A., and Yarmolinsky, A. (eds.) United States, 1998 statistical tables (1995). Federal regulation of methadone (NCJ 181585). Washington, DC: U.S. treatment. Committee on Federal Regula- Department of Justice, Bureau of Justice tion of Methadone Treatment, Institute of Statistics. Medicine, National Academy of Sciences. Washington, DC: National Academy Press. Nisbett, R., and Ross, L. (1980). Human inference: Strategies and shortcomings of Reuter, P. (1989, March 26). An economist social judgment. Englewood Cliffs, NJ: looks at the carnage. Washington Post. Prentice-Hall. Reuter, P., and MacCoun, R. (1992). Street Office of National Drug Control Policy. drug markets in inner-city neighborhoods: (1995). National drug control strategy. Matching policy to reality. In J.B. Steinberg, Washington, DC: The White House. D.W. Lyon, and M.E. Vaiana (eds.), Urban America: Policy choices for Los Angeles Office of National Drug Control Policy. and the Nation (pp. 227–251). Santa (2000, March). Drug-related crime: Monica, CA: RAND. ONDCP drug policy information Fact Sheet (NCJ 181056). Washington, DC: Reuter, P., MacCoun, R., and Murphy, P. The White House. (1990). Money from crime. Santa Monica, CA: RAND. Ousey, G.C., and Lee, M.R. (2000). Examining the conditional nature of the Rhodes, W., Scheiman, P., Pittayathikhun, illicit drug market-homicide relationship: T., Collins, L., and Tsarfaty, V. (1995). What A partial test of the theory of contingent America’s users spend on illegal drugs, causation. Paper presented at the 2000 1988–1993. Washington, DC: Office of National Drug Control Policy.

88 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Rhodes, W., Layne, M., Johnston, P., and psychiatric inpatient facilities and by oth- Hozik, L. (2000). What America’s users ers in the same neighborhoods. Archives spend on illegal drugs: 1988–1998. of General Psychiatry, 55, 393–401. Washington, DC: Office of National Drug Control Policy. Steele, C.M., and Josephs, R.A. (1990). Alcohol myopia: Its prized and danger- Riley, K.J. (1998). Homicide and drugs: ous effects. American Psychologist, 45, A tale of six cities. Homicide Studies, 2, 921–933. 176–205. Substance Abuse and Mental Health Saner, H., MacCoun, R., and Reuter, P. Services Administration. (2001). Mid-year (1995). On the ubiquity of drug selling 2000 preliminary emergency department among youthful offenders in Washington, data from the Drug Abuse Warning Net- DC, 1985–1991: Age, period, or cohort work. Rockville, MD: U.S. Department of effect? Journal of Quantitative Crimin- Health and Human Services, Substance ology, 11, 337–362. Abuse and Mental Health Services Administration. Science (1997, October 3). Frontiers in Neuroscience: The science of substance Tardiff, K., Marzuk, P.M., Leon, A.C., abuse. 278, special section. Hirsch, C.S., Stajic, M., Portera, L., and Hartwell, N. (1994). Homicide in New York Sheley, J.F. (1994). Drug activity and City: Cocaine use and firearms. JAMA: firearms possession and use by juveniles. The Journal of the American Medical Journal of Drug Issues, 24, 363–382. Association, 272, 43–46.

Smith, A., and Varese, F. (2001, August 1). Taylor, L., Caulkins, J., and Reuter, P. Payment, protection and punishment: The (2000). Some simple economics of illegal role of information and reputation in the drug markets and violence. Unpublished Mafia. Rationality and Society, 13, 349–393. manuscript, Carnegie Mellon University.

Spunt, B., Goldstein, P., Bellucci, P., and Teplin, L.A., Abram, K.M., and McClelland, Miller, T. (1990). Race/ethnicity and gender G.M. (1994). Does psychiatric disorder differences in the drugs-violence relation- predict violent crime among released jail ship. Journal of Psychoactive Drugs, 22, detainees? A six-year longitudinal study. 293–303. American Psychologist, 49, 335–342.

Spunt, B., Brownstein, H., Goldstein, P., U.S. Sentencing Commission (1995, Fendrich, M., and Liberty, H. (1995). Drug February). Cocaine and Federal sentencing use by homicide offenders. Journal of policy. Special Report to Congress. Wash- Psychoactive Drugs, 27, 125–134. ington, DC: U.S. Sentencing Commission [Online]. Available: www.ussc.gov/crack/ Stacy, A.W., Widaman, K.F., and Marlatt, EXEC.HTM G.A. (1990). Expectancy models of alcohol use. Journal of Personality and Social White, H.R., and Gorman, D.M. (2000). Psychology, 58, 918–928. Dynamics of the drug-crime relationship. In Crime and justice 2000, vol. 1: The Steadman, H.J., Mulvey, E.P., Monahan, nature of crime: Continuity and change J., Robbins, P.C., Appelbaum, P.S., Grisso, (NCJ 182408, pp. 151–218). Washington, T., Roth, L.H., and Silver, E. (1998). Vio- DC: U.S. Department of Justice, National lence by people discharged from acute Institute of Justice.

89 SPECIAL REPORT / JULY 03

Wolfgang, M.E. (1958). Patterns in crimi- Zimring, F.E., and Hawkins, G. (1997). nal homicide. Philadelphia: University of Crime is not the problem: Lethal violence Pennsylvania Press. in America. New York: Oxford University Press.

90 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Appendix A. Other (juvenile) gang-related than portrayed in the media and may be “kindler and gen- applications of Goldstein’s tler” than other large cities. He also report- framework ed that the worst years for murders in Miami were during its cocaine wars in the New York early 1980s. Inciardi found that “those more proximal of the crack distribution Even excluding the works of Goldstein market were more involved in violent and his colleagues, much of the work crime” (p. 104). This study has at least using the tripartite framework focuses on two advantages over Goldstein et al. New York during the mid- to late 1980s. (1989): Crimes other than homicide were The U.S. Sentencing Commission (1995) considered, and respondents were asked used Goldstein’s framework to compare about drug-related victimization. But the the incidence of violence related to the drug associated with these crimes was use of powder cocaine and crack. Using not listed as it was in the Goldstein et al. expert testimony and existing literature, study. and largely focusing on the studies done in New York,1 the Commission concluded that crack was a greater source of sys- Chicago temic violence than powder cocaine, that One source that was developed to assess economic-compulsive violence was rela- homicide fluctuations and motivations is tively rare among cocaine users, and that the Chicago Homicide Dataset (CHD). De- “neither powder nor crack cocaine excite tailed information on every homicide in the or agitate users to commit criminal acts records of the Chicago Police Department and that the stereotype of a drug-crazed is available for 1965–95 (Block, Block, and addict committing heinous crimes is not Illinois Criminal Justice Information Au- true for either form of cocaine” (p. x). thority, 1998). CHD does not include data on specific drugs, but its motive classifica- Miami tion fits nicely with the tripartite framework. The four types of drug-related Inciardi’s (1990) survey of 611 serious motives for homicide are selling or drug juvenile delinquents in Miami and Dade business (this includes any homicides dur- County assessed offender self-reports ing or because of a transaction);2 an argu- of drug-related systemic, economic- ment over possession, use, quality, or compulsive, and psychopharmacological cost of drugs; getting money for drugs or crime. In the 12 months prior to the inter- acquiring drugs for personal use; and views, which occurred from 1985 to 1989, other drug involvement (e.g., baby dies about 5 percent of the sample reported of malnutrition because the parents were being a psychopharmacological victim, high; offender was drug crazed).3 59 percent reported having committed robberies (“the majority of which were The per capita drug-related homicide rate committed to purchase drugs,” p. 100), remained fairly stable from 1973 to 1984 and 8 percent reported being the perpetra- (around 0.4 homicides per 100,000 Cook tors of systemic crimes. Inciardi also County residents), with “arguments” at administered a supplementary crack sur- a slightly higher rate from 1974 to 1977. vey to 254 of these delinquents from Homicide rates related to all of the mo- October 1986 to November 1987. This tives fluctuated from 1984 to 1995, but it survey and other data analyses by Inciardi is interesting that the aggregate rate for led him to conclude that the Miami crack every motive except “business/transac- market was much less violent and less tion” was virtually the same for 1984–85

91 SPECIAL REPORT / JULY 03

and 1995 (still close to 0.4). The advent of how competitive the market was, they crack likely explains why homicide rates also considered the stability of prices, related to all of the motives increased transactions, and participants. Their argu- from 1985 to 1989, but it is of special ment that links between drugs and homi- interest that the “business/transaction” cide “appear to fall mainly on the use motive skyrocketed during those years. side” (p. 92) relies on their findings about Clearly, more might be learned by examin- participants: ing the specific drugs associated with “business/transaction” homicides in The general structure of participation Chicago over this time period. in crack markets and the nature, duration, and consequences of the “crack high” may account for the relation- Eight-city study ship between the cocaine prevalence To learn why city homicide rates did not rates among arrestees and homicide change uniformly in the early 1990s, rates. Crack users reported the large Lattimore and colleagues (1997) com- number of “buys,” extensive net- prehensively examined homicide in works of potential suppliers, and less eight cities for 1985–94: Atlanta, Detroit, reliance on a primary supplier, sug- Indianapolis, Miami, New Orleans, Rich- gesting that transactions were likely mond, Tampa, and Washington, D.C. In to occur in an opportunistic manner. addition to comparing ADAM results with The high from crack lasts as little as UCR data for these cities, Lattimore et al. 10 minutes; thus, when the high interviewed key policymakers, law en- wears off, the crack user may still be forcement and criminal justice officials, in the market and motivated to buy and community leaders in the cities. more of the drug—and to commit a These interviews revealed that crack was crime to obtain the money to do so. most likely associated with community (p. 141) violence and homicide, while the market violence associated with marijuana was a This is essentially an argument about growing concern in Washington, D.C., and economic-compulsive violence, which Richmond. Methamphetamines, LSD, other crack-specific studies have dis- PCP, and heroin were not associated with missed (see U.S. Sentencing Commission, homicide rates and were rarely mentioned 1995). While this difference may be geo- by local authorities. It is important to note graphic (the other studies were primarily that Lattimore et al. found that in many done in New York City), it may also be the cases the perceptions about local drug artifact of a bivariate analysis of two trends differed substantially from drug datasets (UCR and ADAM) that did not trends as measured by DUF/ADAM. always cover the same populations.

Lattimore and colleagues question the National estimates relationship between crack and market violence because the crack markets were Others used nationwide data to learn more described as highly competitive in cities in about the drugs-crime nexus. Caulkins and which the homicide rate was declining, colleagues (1997) used the tripartite frame- increasing, or remaining the same (1997, work to assess the impact that mandatory p. 89). But it is not clear that the same minimum sentences have on cocaine con- conclusions could be drawn if disaggregat- sumption and subsequent crime. Relying ed homicide rates (by circumstance) were on estimates from Goldstein and his col- considered. The authors not only looked at leagues (Goldstein, Brownstein, and Ryan,

92 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

1992; Spunt et al. 1990; Spunt et al., Notes 1995), the National Criminal Victimization Survey, inmate surveys, and murder data 1. New York: Chin and Fagan, 1992; Fagan and Chin, for large urban counties, Caulkins et al. 1990; Goldstein et al., 1989. Miami: Inciardi, 1990; Inciardi and Pottieger, 1991; Inciardi and Pottieger, determined the number of systemic, 1994. Los Angeles: Klein et al., 1991. Detroit: economic-compulsive, and psychopharma- Mieczkowski, 1990.) The Commission also cited an cological crimes that were drug related. unpublished DEA report and a review article by Fagan Their next step was to determine how (1990). The former found “that seven crack-related much of this crime was related to cocaine. homicides were ‘multi-dimensional,’ with systemic being one of the dimensions,” but it is not clear Based on information from Rhodes et al. where these homicides occurred and what the other (1995), the ADAM Program, the Office of dimensions were. National Drug Control Policy (1995), and Goldstein (Goldstein et al., 1989; Goldstein, 2. The codebook reads: “Use code 1 when BUSI- Brownstein, and Ryan, 1992; Spunt et al. NESS is the motive for the incident (e.g., both victim and offender involved in dealing, victim killed as a 1995), Caulkins et al. (1997) suggest that bystander of a drug business hit, victim killed be- cocaine accounts for about 75 percent of cause he interfered with the business, victim killed drug-related economic-compulsive crime, during a drug transaction or because of a drug trans- 50 percent of illicit psychopharmacological action).” homicides, and 75 percent of systemic 3. Cases where there was no positive evidence or no homicides. 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.

93 SPECIAL REPORT / JULY 03

Appendix B. Arrestees Needing Drugs and/or Alcohol at the Time of the Offense, 1995–99

Appendix B. Arrestees Needing Drugs and/or Alcohol at the Time of the Offense, 1995–99 Income-generating offenses Non-income-generating offenses City Total % needing drugs and/or alcohol Total % needing drugs and/or alcohol

Albuquerque 249 40 1,308 19 Anchorage 105 16 723 9 Atlanta 1,526 17 2,833 9 Birmingham 1,216 17 3,646 12 Chicago 1,825 26 4,183 17 Cleveland 1,569 16 4,191 12 Dallas 1,934 12 3,432 8 Denver 1,195 8 5,842 6 Des Moines 182 20 744 10 Detroit 903 9 2,876 8 Ft. Lauderdale 1,209 19 4,032 13 Houston 1,257 6 4,252 5 Indianapolis 2,447 15 5,248 8 Laredo 185 13 531 5 Las Vegas 355 26 1,638 14 Los Angeles 4,022 10 6,951 7 Miami 1,395 15 2,182 11 Minneapolis 179 21 953 10

94 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Appendix B. Arrestees Needing Drugs and/or Alcohol at the Time of the Offense, 1995–99 (continued)

Income-generating offenses Non-income-generating offenses City Total % needing drugs and/or alcohol Total % needing drugs and/or alcohol

New Orleans 2,072 16 4,020 10 New York 3,162 16 6,247 16 Oklahoma City 394 14 1,298 9 Omaha 678 13 3,249 5 Philadelphia 2,201 21 1,645 17 Phoenix 1,828 15 5,929 7 Portland 1,550 11 5,032 10 Sacramento 389 14 1,307 9 Salt Lake City 333 17 1,044 13 San Antonio 2,060 8 5,570 4 San Diego 2,407 8 3,982 7 San Jose 1,549 8 4,441 6 Seattle 301 21 1,090 13 Spokane 261 20 1,063 12 St. Louis 1,160 17 2,592 12 Tucson 308 14 1,965 9 Washington, DC 1,529 10 3,200 8 Total 43,935 14% (n=6,141) 109,239 10% (n=10,431)

Notes: Percentages rounded to nearest whole percentage point. Observations with missing data for any of these variables were deleted. Sixty-four observations from 1998 and 374 observations from 1999 were not considered because of a unique charge-coding strategy. Income-generating offenses include burglary, burglary tools, prostitution, embezzlement, larceny/theft, pickpocketing/jostling, robbery, stolen property, stolen vehicle, and drug sales. Source: 1995–99 data from the Arrestee Drug Abuse Monitoring Program

95 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

Introduction The existence of the The relationship between drug use and drugs-crime relationship criminal behavior has generated a substan- The purpose of this section is to briefly dis- tial body of literature in peer-reviewed cuss what is known about the drugs-crime journals, government publications, and the relationship. This discussion will focus on public press. The very extent of such the historical policy context; the empirical research—as well as the breadth of policy nature of the relationship overall; and spe- positions based on or ignoring such cific drugs, crimes, and populations. research—argues for the importance of a review that can summarize theory, policy, and programmatic approaches to the Which drugs and what crime? issue. In this paper, we do not attempt to Before proceeding further, we wish to provide a comprehensive review of the clarify what we mean by “drugs” and pro- issues or literature. Instead, we seek to vide a more complete picture of what is About the Authors provide a sufficient review of the most involved in “crime” related to drug use. pertinent knowledge about the drugs- Duane C. McBride is a These clarifications are made in the hope professor with and chair of crime relationship to stimulate further dis- that readers will recognize that the crime the Department of Behavioral cussion among researchers regarding the aspect of the drugs-crime relationship is Sciences and director of the most important research questions that multifaceted and that the current exclu- Institute for Prevention still need attention. This discussion holds sion of alcohol from most discussions of of Addictions at Andrews great promise for the development of new the drugs-crime relationship may be University; Curtis J. VanderWaal approaches to the drugs-crime relation- detrimental. is a professor with the ship. As Brownstein has argued, “those Department of Social Work who do the research are in the best posi- and associate director of tion to interpret their findings and offer Substance inclusion decisions the Institute for the Prevention advice based on their conclusions” (1991, of Addictions at Andrews The term “drugs” as used throughout this p. 132). This paper approaches the above University; Yvonne M. paper refers to currently illicit substances Terry-McElrath is a research task by focusing on the following issues: in the United States based on Federal associate with the Institute for (a) documenting the existence of the drug schedules. Alcohol, prescription Social Research at the drugs-crime relationship, (b) addressing drugs, and other substances are excluded. University of Michigan. the nature and complexity of that relation- Although it is beyond the scope of the ship, (c) summarizing philosophical and current project, it is important to at least theoretical contributions that may best mention the alcohol-crime relationship. address the relationship, (d) reviewing Greenfeld (1998) reminds us that an esti- both State- and Federal-level policy mated 36 percent of convicted offenders approaches to breaking the relationship, were drinking at the time they committed including integrated program approaches, their crimes and that a high correlation has and (e) proposing key areas for future been observed between public order research.

97 SPECIAL REPORT / JULY 03

crimes and alcohol use. Alcohol is also policy has moved to various points along strongly related to violent crime (Coker et this continuum, and it often resides at dif- al., 2000; Dawkins, 1997; Ernst et al., ferent points at the same time in different 1997; Parker and Auerhahn, 1998; Pihl and locations and for different substances. Peterson, 1995). Ironically, this relation- Each time policy shifts, the act of drug use ship often remains outside sentencing takes on a slightly different character in decisions and monitoring procedures relation to crime. Thus, it is important to because alcohol is legal and therefore not present a brief history of drug policy in the subject to the same arrest, seizure, and United States, together with current possi- prosecution laws as are illicit drugs. Drug ble positions in the drug policy discussion, treatment interventions, however, often as each position has a unique implication include both alcohol and other drugs. for fighting drug-related crime. Comprehensive efforts to address crime and substance use should include alcohol An understanding of American drug policy treatment in programmatic considerations. begins with three early American cultural traditions that still strongly affect drug policy discussions: (a) libertarianism, (b) the The history of drug policy and emergence of a relatively open legal the definition of crime market resulting from the libertarian per- Crimes associated with drug use range spective, and (c) Puritan moralism. Liber- from violent (such as murder and aggra- tarianism argues that government must vated assault) to acquisitive (burglary, for- have an extremely compelling motive for gery, fraud, and deception) to specific interfering in the personal lives of citizens. drug-law violations. In addition, crimes Such interference legitimately occurs only such as bribery and corruption are related if a citizen’s behavior is a significant, actual to drug use as a result of drug policy prohi- risk to others (Mill, 1979). Consistent with bitions. Traditionally, discussions of the this libertarian tradition, early America had drugs-crime relationship have focused pri- an open-market orientation that empha- marily on violent crime; however, it is sized limited government interference in important to recognize the complexity of the production and distribution of desired criminal acts associated with drug use. goods and services.1 Nineteenth-century When considering the drugs-crime rela- national drug policy was consistent with tionship, this paper recommends that both libertarianism and the open market. researchers and policymakers include both While the Federal Government regulated violent and nonviolent crimes as well as the importation of such drugs as opium drug law violations and corruption associ- and cocaine, few regulations governed the ated with drug policy to grasp more fully distribution of these and other drugs the resulting harms and societal costs (for through what came to be called the patent example, see French and Martin, 1996). medicine industry (Belenko, 2000; Inciardi, 2001; Musto, 1999). Patent medicines Efforts to address the drugs-crime rela- were extensively advertised and, through tionship must incorporate a realization of them, the use of drugs such as opium and how the development of policy and law cocaine became integrated into routine has contributed to the relationship itself. American cultural behavior patterns Policy approaches to drug use in the (Musto, 1999). United States have historically ranged between legal markets in the 19th century Conflicting with both libertarianism and to decriminalization, harm reduction, med- the market-driven approach is the Puritan icalization, and strict prohibition (as the moralist perspective: Individual behaviors dominant policy) in the 20th. Over time, with the potential to harm the community

98 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

are seen as a community problem within governments to develop their own treat- the legitimate purview of community ment programs. In 1970, the Comprehen- action (Cherrington, 1920; Schmidt, 1995). sive Drug Abuse Prevention and Control Puritan and other religious and moral tradi- Act consolidated and replaced the patch- tions present in American history often work of previous Federal drug laws. The viewed behavior such as substance use as Act created the drug schedules in current undermining the moral fabric of society, use today and initiated the so-called “war potentially causing the withdrawal of on drugs”; it also moved some posses- God’s blessing from America. The Puritan sion or casual transfer offenses to misde- moralist perspective dominated the early meanors instead of felonies. This era may 1900s, an era of societal reform and in- be considered a time when drug use was creasing prohibition (and thus increasing primarily considered a medical/mental penalties for drug use). One of the first health problem to be addressed by treat- successes of the social reform movement ment, with lessened emphasis on criminal in the early 20th century was the passage penalties for possession and use. of the Pure Food and Drug Act of 1906, which required the patent medicine indus- With an apparent increase in drug use, as try to list product ingredients. The subse- evidenced by an increase in drug overdose quent passage of the Harrison Act of 1914 cases and drug treatment admissions, a and the Marihuana Tax Act of 1937 made more prohibitionist movement again illegal the manufacture, sale, and posses- swept the Nation. New York’s Rockefeller sion of a variety of drugs, including opiates Drug Laws were passed in 1973, estab- and cocaine, as well as the nonmedical lishing mandatory prison sentences of up use of marijuana. A strongly prohibitionist to 20 years for the sale of any amount of approach continued through the 1950s heroin or cocaine. The Anti-Drug Abuse with the Boggs Act of 1951 and the Nar- Acts of 1986 and 1988 continued to em- cotic Control Act of 1956, when mandato- phasize law enforcement (although the ry minimum sentences for Federal drug 1988 Act gave more attention to treat- trafficking law violations were strength- ment and prevention). In yet another poli- ened and arrests without a warrant for cy shift, treatment (including diversion into drug charges were enabled. treatment from the criminal justice system) and prevention received increasing The 1960s and 1970s represented a major attention in the 1990s. Further, some cultural shift in the United States. For a States developed policies that effectively variety of reasons, American society expe- decriminalized marijuana possession rienced a “drug revolution” during this (removing jail/prison penalties) and initiat- era. There appeared to be an increase in ed policies, such as needle exchange pro- the proportion of individuals using drugs grams, that would reduce the dangers of and in the variety of drugs used. The evi- injecting drugs. dence for this increase is seen in the number of drug-related arrests and the Although scholars often focus on the rela- increase in drug use in the general popu- tively rapid development of national drug lation (Musto, 1999). During this era, policy, it is important to remember that drug policy initially shifted to a stronger many States passed legislation prohibiting treatment- and less punishment-oriented patent medicine and/or alcohol sales, as stance. In 1966, the Narcotic Addict well as marijuana use, a decade or more Rehabilitation Act allowed the establish- before similar legislation was passed by ment of the civil commitment system Congress (Belenko, 2000). Because of instead of prosecution for Federal offend- how the United States is organized, States ers and encouraged State and local often have or exercise considerable

99 SPECIAL REPORT / JULY 03

The history of discretion regarding alcohol and drug poli- The statistical relationship cies (Musto, 1999). between drug use and criminal drug policy can behavior be broken down Essentially, the history of drug policy (and debates about where drug policy should The general conclusion of almost three into five main move in the future) can be broken down decades of research on the relationship approaches: into five main approaches: prohibition, risk between drug use and crime has been reduction, medicalization, legalization/ that there is a clearly significant statistical prohibition, risk regulation, and decriminalization (for an relationship between the two phenomena reduction, indepth discussion, see McBride et al., (Austin and Lettieri, 1976; Dorsey and 1999; see also Goode, 1997). Prohibition Zawitz, 1999; Gandossy et al., 1980; medicalization, emphasizes severe penalties for use, dis- McBride and McCoy, 1993). Research indi- legalization/ tribution, and production. Risk reduction cates extensive drug use among arrested uses a public health approach to reduce populations, a high level of criminal be- regulation, and the risks and harms associated with illicit havior among drug users, and a fairly high decriminalization. drug use and emphasizes education on correlation between drug use and delin- risks, safer use practices, prevention, and quency/crime in the general population. treatment. Medicalization calls for physi- Research also indicates significant differ- cian treatment of drug addicts, viewing ences in the relationship based on drug substance abuse primarily as a medical type and type of crime. Importantly, all issue. Legalization/regulation supports these differences are further complicated increased access to drugs through govern- by ethnic and gender issues. mental regulation of these substances, with possible distribution of specific sub- The drugs-crime relationship stances through governmentally controlled within various population distribution channels. Decriminalization groups calls for a complete end to the use of criminal law to address individual drug Drug use among arrested/incarcerated use. This may imply a relatively open- populations and crime among drug market approach to drug availability and users. From the early 1970s onward, bio- use, but that need not be the case. logical and self-report data have indicated a relatively high rate of drug use among Although there has been significant debate arrested and incarcerated populations over which policy approach or approaches (Arrestee Drug Abuse Monitoring Pro- might best address the drugs-crime cycle, gram, 2000; Austin and Lettieri, 1976; more research is needed that examines Dorsey and Zawitz, 1999; Gandossy et scientifically the effects of policy positions al., 1980; McBride and McCoy, 1993). In on both drug use and crime. For the most 1999, the Arrestee Drug Abuse Moni- part, current Federal drug law takes a pro- toring Program (ADAM) collected data hibitionist stance that includes a strong from more than 40,000 adults in more deterrence approach to reducing the sup- than 30 sites and more than 400 juveniles ply of drugs and high penalties for drug in 9 sites throughout the United States law violations. As a result, a significant (ADAM, 2000). In almost all cities where portion of the drugs-crime relationship is the ADAM project operates, about two- simply an artifact of law and policy itself: thirds of both adult male and female “most directly, it is a crime to use, pos- felony arrestees had an illegal drug in their sess, manufacture or distribute drugs clas- bodies at the time of arrest (with higher sified as having the potential for abuse” rates among females). Even among juve- (Craddock, Collins, and Timrots, 1994). niles, the majority of arrestees were found to have an illegal drug in their urine (with

100 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

higher rates among males). The data also 1991). Youths who used “hard” drugs suggest that, although current drug use (about 5 percent of the sample) accounted rates among adult arrestees are higher for 40 percent of all delinquencies and 60 than those reported in the more isolated percent of index crimes. reports of the 1970s (Austin and Lettieri, 1976), these rates have remained steady The impact of drug type on the for the past 5 years (the same patterns are drugs-crime relationship found among juvenile arrestees). An argument can be made that with about two- The first National Institute on Drug Abuse thirds of arrestees already using illegal (NIDA)-sponsored Crime and Drugs Report drugs in the 72 hours prior to their arrest, (Austin and Lettieri, 1976) noted that a there is not much room for an increase. complex relationship exists between type of drug use and type of crime. This rela- A recent report from the Bureau of Justice tionship is further complicated if multiple Statistics (BJS) suggests that drug use drug use exists. The 1999 ADAM report also is extensive among inmates in local shows that a fairly large proportion of jails (Wilson, 2000). This document reports arrestees tested positive for more than that the majority of inmates in State pris- one drug (up to 30 percent), and that ons and local jails used drugs in the month reported criminal behavior tended to prior to the offense that put them in prison/ include a wide variety of offenses. The jail. Interestingly, this same report also ADAM data show that while cocaine was notes that about 10 percent of jail inmates the most likely drug found among adult test positive for drugs while in jail. arrestees in large cities (and there is literature suggesting a significant relationship The extent of crime among drug users between cocaine and violence), for many also has been documented. From the urban ADAM sites, violent offenders were 1960s through the 1990s, surveys of drug- more likely to test positive for marijuana using populations both in and out of treat- than cocaine. In addition, property offend- ment have consistently shown that the ers were more likely to test positive for large majority of users have extensive his- cocaine than marijuana in most sites tories of criminal behavior and time served (ADAM, 2000). in prison (Defleur, Ball, and Snarr, 1969; Inciardi, Horowitz, and Pottieger, 1993). This pattern applies to juveniles as well: The impact of crime type on the Between 40 and 57 percent of adoles- drugs-crime relationship cents treated for substance disorders also Drug law violations. A significant propor- have committed delinquent acts (Winters, tion of drug user arrests involve violations 1998). of drug laws only. As noted previously, the United States experienced wide drug poli- Drug use and crime levels among the cy shifts in the 20th century. Each shift general population. A tradition of studies has uniquely affected crimes related to shows a correlation between drug use and drug use and distribution. In a study of delinquency in general youth populations 611 juvenile cocaine users by Inciardi and (Elliott and Huizinga, 1984; Elliott, Huizinga, colleagues in the early 1990s, analyses and Menard, 1989; Harrison and Gfroerer, showed that participants had committed 1992). Analysis from the National Youth more than 400,000 criminal acts in the Survey has provided data often used to 12 months prior to being interviewed. Of examine this relationship. These data these, 60 percent were for drug law viola- report a direct correlation between serious tions, mostly sales of small amounts drug use and delinquency (Johnson et al.,

101 SPECIAL REPORT / JULY 03

(Inciardi, Horowitz, and Pottieger, 1993). violence associated with drug use, ”docu- At the Federal level, a total of 581,000 menting” the criminal consequences of drug arrests in 1980 nearly tripled to a such activity (see Anslinger and Tompkins, record high of 1,584,000 in 1997. By 1997, 1953; Inciardi, 2001). Among the best 79 percent of drug arrests were for pos- known of these efforts were the films session and 21 percent were for sales. “The Man with the Golden Arm” (purport- Forty-four percent of drug arrests overall ing to depict the effects of heroin use/ were for marijuana offenses (Uniform injection) and “Reefer Madness” (show- Crime Reports, 1998). Drug defendants ing the supposed behavioral conse- comprised 42 percent of felony convic- quences of marijuana use). Although such tions (Bureau of Justice Statistics, 1999). media portrayals exaggerated the possible A recent BJS Special Report (Wilson, links between drugs and crime, some 2000) also substantiates the extensive research has connected drug use with vio- percentage of drug-related crimes that lence. Grogger and Willis (2000) conclude result from violation of drug laws, suggest- that without the introduction of crack ing that about a quarter of jail inmates cocaine into urban America, 1991 crime have a current charge or conviction for rates would have been about 10 percent drug law violations. Critics have argued lower. These researchers also examined that since such arrests likely include many the impact of crack on specific types of low-level users and dealers, criminal jus- violent crime and reported that the biggest tice processing and the stiff sentences impact was on aggravated assault. that often are handed down because of mandatory minimums may be inappropri- In 1985, Goldstein provided the perspec- ate to the offense level (McBride et al., tive that has been most commonly used 2001). to examine the relationship between drug use and violence. Essentially, he argued Changes in drug The violence connection. Changes in for a tripartite scheme, where “psycho- drug policy are usually driven by concerns pharmacological violence” could result policy are usually for public safety and the perception of a directly or indirectly from the biochemical driven by concerns direct relationship between drugs and vio- behavioral consequences of drug use; lence (Brownstein, 1996, 2000). For exam- “economic-compulsive violence” could for public safety ple, the drug policy reform movement of relate to behavior/crimes engaged in to and the perception the early 1900s (changing from legal obtain money for drugs; and “systemic markets to strict prohibition) was accom- violence” could emerge in the context of of a direct panied by horror stories focused on exag- drug distribution, control of markets, the relationship gerated claims of criminal behavior as a process of obtaining drugs, and/or the consequence of drug use. In this litera- social ecology of drug distribution/use between drugs ture, there was a particular emphasis on areas.2 Some researchers have concluded and violence. horrific violent crime (including rape), with that there is minimal evidence regarding minority group members often portrayed the psychopharmacological impact of as the drug users engaged in the violent drugs on violence (Resignato, 2000); how- behavior. Musto (1999; see also Belenko, ever, Pihl and Peterson (1995) reviewed a 2000; Hickman, 2000) documents the wide range of studies on the issue. They public concern of the time (perhaps obses- concluded that alcohol and drugs can be sion) with Chinese opiate use, African- psychopharmacologically related to violent American cocaine use, and the use of acts through the release of dopamine, marijuana by Mexicans. The creation of which reduces inhibitory anxiety about the the Narcotics Bureau led to a type of consequences of aggressive behavior and media distribution industry focused on increases the rewards associated with

102 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

violence. In addition, they argue that the Institute of Justice (NIJ) Research in Brief psychopharmacological effects of drugs supports this research, stating “illegal interfere with the user’s cognitive process- drugs and violence are linked primarily ing of the consequences of potentially vio- through drug marketing” (Roth, 1994, p. 1). lent situations. It should be noted that these authors believe that the evidence The impact of ethnicity for psychopharmacological effects of alco- and gender hol use on violence are much higher than for other drugs. Much of the research that has been conducted on drugs and crime has not had a However, some indications point to the sufficient focus on gender and ethnic vari- environment as being a more powerful ance. This limitation has significant reper- explanation of the drugs-violence relation- cussions on applying findings to other ship than the psychopharmacological population groups. As Paniagua (1998) properties of drugs (Brownstein, 2000; notes, the multicultural nature of current Fishbein, 1998; Parker and Auerhahn, society must incorporate a recognition of 1998). In terms of economic compulsive the complex nature of ethnicity and gen- and systemic violence, Collins (1990) as der. Specifically, individuals who share a well as Fagan and Chin (1990) argue that similar ethnicity or gender will not all be crack selling is the main contributor to the the same (i.e., recognition of language, drugs-violence relationship. Specifically, acculturation, and socioeconomic differ- their research found that violence (mostly ences); however, it is important to recog- robbery) emerged from the need to obtain nize cultural commonalities that may money to purchase drugs (predominately significantly affect both the extent and crack). Fagan and Chin suggest that the nature of the drugs-crime relationship drugs-violence relationship also emerges across individuals. Research that has as a part of the subculture of violence. focused on ethnicity and gender indicates that these variables may significantly In a 1994 study, Roth argued that drug affect various aspects of the drugs-crime users commit more property crime than relationship, including: violent crime. A recent publication by De Li, Priu, and MacKenzie (2000) examined ■ Source of drugs and/or works (Taylor the relationship between drug use and et al., 1994). property and violent crime in a population of probationers in Virginia. Results indicat- ■ Predictors of violence (Ellickson and ed that drug use had a positive association McGuigan, 2000). with property crime, whereas drug dealing had an association with both violent and ■ Types of violence experienced and reac- property crime (though the relationship tions to such violence (Brownstein et al., was stronger for property crime). The 1994; Fine and Weis, 1998; Mazza and analysis also showed an interactive effect Dennerstein, 1996). between drug use, drug dealing, and vio- ■ Stress-coping factors (Vaccaro and Wills, lent and property crime. Among juveniles, 1998). Linnever and Shoemaker (1995) found that arrests for both possession and selling of ■ Biological effects of drugs (Brady and drugs were related to the rate of property Randall, 1999). crime arrests. However, juvenile robbery arrest rates were related to only drug ■ Epidemiology of substance-use disor- sales arrests (not possession). A National ders (Brady and Randall, 1999).

103 SPECIAL REPORT / JULY 03

■ Psychiatric comorbidity (Brady and development of drug policy at all lev- Randall, 1999). els of government.

■ Social stigma issues (Brady and Randall, 4. Hyperbole, demagoguery, demoniza- 1999). tion, and perhaps even naivete have historically characterized the drugs- ■ Medical consequences of drug use, crime debate (and may still). However, including heredity issues and course of there is a clear statistical relationship illness (Brady and Randall, 1999). between drug use and crime. The majority of drug users have extensive ■ Assessment and treatment issues, histories of involvement with crimes including possible prevention settings and the criminal justice system; most (Brady and Randall, 1999; Metsch et al., arrestees are current drug users; and 1999; Paniagua, 1998). there is a correlation between drug use and delinquency/crime in general ■ Differences in initiation of drug use populations. A large proportion of this (Doherty et al., 2000). criminal activity is a result of drug law violations. Summary: What we know of the past 5. Although there is some evidence that drug costs may be related to property The intended purpose of this section has crimes and robberies, and that distri- been twofold. The first goal has been to bution and subcultural elements sur- review the history of American drug policy rounding drug use may be related to (as well as possible drug policy positions) violence, there is debate about the within the framework of the relationship evidence for a strong and continuous among policy, drug use, and crime. The connection between drug use and vio- second purpose has been to summarize lence. This relationship is also compli- the statistical documentation of the drugs- cated by the type of drug use, the crime relationship. Hopefully, this review category of crime, and ethnicity and has served to remind readers of the fol- gender. lowing issues:

1. American drug policy originated in the antithetical cultural traditions of rela- The nature and complexity of tively open-market/libertarian values the drugs-crime relationship and Puritan moralist social reform. These traditions still affect current As White and Gorman (2000) note, three debates about the drugs-crime rela- main explanatory models exist for grap- tionship, as well as the various policy pling with the drugs-crime relationship: positions between these two end- ■ Drug use causes or leads to crime. points on the policy continuum. ■ Crime causes or leads to drug use. 2. States have a history of experimenting with drug policies in advance of, and ■ The relationship is purely coincidental or sometimes in opposition to, Federal is based in a common etiology. action on the same issues. Based on their evaluations of the research 3. Public safety concerns have been supporting and/or refuting each of the the underlying rationale for the three main models above, they conclude

104 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

that “one single model cannot account for On the other hand, some researchers the drug-crime relationship. Rather, the argue that a level of general delinquency drug-using, crime-committing population often precedes drug use (Elliott, Huizinga, is heterogeneous, and there are multiple and Menard, 1989). The subcultural expla- paths that lead to drug use and crime” nation is used here as well: Involvement in (White and Gorman, 2000, p. 151). Ten criminal activity and/or subcultures pro- years earlier, Collins (1990) also rejected vides “the context, the reference group, simple explanatory models for the com- and the definitions of a situation that are plex relationship. The debates over both conducive to subsequent involvement the direction of a drugs-crime relationship with drugs” (White and Gorman, 2000, as well as the etiological variables that p. 174; see also White, 1990). Individuals may be involved in the common occur- with deviant lifestyles and/or personalities rence of both drugs and crime have signifi- may also use substances for the purposes cant implications for attempts to intervene of self-medication (Khantzian, 1985; White in the drugs-crime cycle. and Gorman, 2000) or to provide a “reason” for deviant acts (Collins, 1993; White The direction of the relationship: and Gorman, 2000). Although Apospori Searching for a cause and associates (1995) concluded that the relationship between early delinquency At the popular and sometimes at the gov- and subsequent drug use was relatively ernmental level, the drugs-crime relation- weak, Bui, Ellickson, and Bell (2000) found ship is often clearly causal: Drug use what they called a modest relationship causes crime. Models such as Goldstein’s between delinquency in grade 10 and tripartite scheme (1985) have been used greater drug use in grade 12. Importantly, to illustrate this approach, specifying they found no significant differences by psychopharmacological, economic- ethnicity for this relationship. Hser, Anglin, compulsive, and systemic causes of vio- and Powers (1993) found that addicts who lence. As noted previously, arguments ceased narcotic use were less likely to that focus on the psychopharmacological engage in criminal behavior over a 24-year properties of various drugs cite research followup period. that indicates that stimulants may increase aggressiveness and paranoia and that Although there is some evidence of direc- many drugs have a strong disinhibiting tionality in the drugs-crime relationship, effect that could seriously interfere with researchers who have attempted to ad- judgment (Pihl and Peterson, 1995). dress this issue generally have concluded Economic arguments posit that the cost of that the relationship is extremely complex drugs, coupled with high unemployment and defies attempts to sort out directional- among drug users, results in the commis- ity. Work by Nurco and colleagues on crim- sion of property crimes to support drug inal careers initially found that increases in use (16 percent of jail inmates committed narcotic drug use were often followed by their current offense to get money for increases in criminal activity; conversely, drugs; BJS, 1999). Those who argue for a periods with no drug use were associated systemic approach maintain that drug use with less criminal activity of all types simply has a subcultural relationship with (these results applied for white, African- criminal behavior: Because it is illegal, American and Hispanic narcotics addicts; drug use involves the user in criminal sub- Nurco, Cisin, and Balter, 1981; Hanlon et cultures that often lead to future deviance al., 1990). However, in a subsequent 1993 (Fagan and Chin, 1990). article, Nurco, Kinlock, and Balter found that narcotic drug users had very early

105 SPECIAL REPORT / JULY 03

involvement in what these researchers call development of street identity for survival “precocious criminal activity.” This activity (Collison, 1996), social isolation that pre- pattern occurred prior to the onset of vents access to the social and economic addiction, and therefore simply could not systems of society (Harrell and Peterson, be attributed to addiction itself. A recent 1992; Stephens, 1991), and lack of what is article by Maxwell and Maxwell (2000) now referred to as human and social capi- provides another example of the confus- tal (described later in this chapter). Dembo ing directionality, suggesting that drug use and his colleagues have studied the drugs- has a very complex relationship with types crime relationship among high-risk youths of deviant behavior for women. Their find- entering the juvenile justice system through- ings suggest that frequent use of crack, out the last decade. In an important 1994 combined with early onset of crack use, is article, Dembo and colleagues found that related to prostitution. Drug selling, how- both delinquency and drug use emerge ever, was found to relate to decreased within the context of family problems and prostitution as it provided another opportu- peer deviant behavior. These researchers nity for income to purchase drugs. On a found that for both males and females, as broader level, Curtis (1999) found that well as African-Americans and whites, drug use rates did not decrease in either family alcohol and drug use, emotional Any simple the general or at-risk populations during problems, arrest history, and peer deviant the 1990s; however, there was a wide- behavior were related to continuing drug attempt to only spread decrease in urban crime during the use. Based on these models, any simple deter drug use same time period. He argues that market attempt to only deter drug use through and cultural forces were behind the ob- severe punishment or treatment will not through severe served changes in substance use patterns result in less crime or substance use, as punishment or and consequences: street drug dealers such approaches do not address the com- exerted higher control on both the drug plex cause of both behaviors (Harrell and treatment will not use of those who worked for them as well Peterson, 1992). result in less crime as the violence often associated with street drug dealing. or substance Summary use, as such A common origin Research on understanding the nature of the drugs-crime relationship illustrates that approaches do not One of the traditions of research on the no simple causal model can explain the address the drugs-crime relationship has emphasized phenomena. Rather, the statistical relation- that drug use and crime may not have a ship between the two activities may be a complex cause of direct causal relationship (White and result of their common etiological origin. both behaviors. Gorman, 2000), but may emerge in the As the purpose of this paper is to present same contextual milieu and have the a background for discussion of possible same antecedent variables such as poor research agendas to expand and reform social support systems, difficulty in research on the drugs-crime relationship, school, and membership in a deviant peer it is important to ground such a system- group (Hamid, 1998; Inciardi, Horowitz, wide effort in theoretical frameworks and Pottieger, 1993; Lurigio and Swartz, that allow for the complex nature of the 2000). These variables have been suggest- relationship. Such frameworks can be ed to include such issues as neighborhood then used to help shape possible future context (McBride and McCoy, 1982), the research.

106 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Philosophical and behavior. The theory calls attention to inherent personal characteristics that theoretical contributions affect individual behavior, including com- to addressing the drugs- petence, self-esteem, and self-direction crime relationship (Germain and Gitterman, 1995). This section will provide a theoretical Definitive theories of behavior that have framework for reviewing current program- been used to explain crime and deviance matic approaches to breaking the drugs- have varied. Since the 1960s, the follow- crime cycle. The theoretical approaches ing theories have been predominant: to be presented include both overarching anomie, social disorganization, differential behavioral theories and philosophies spe- association, social control, deterrence, cific to justice system programming. labeling, and conflict (Liska, Krohn, and Messner, 1989). Recently, however, atten- Overarching theoretical tion has been directed to new approaches approaches with the hope that theoretical and research advances will better support pre- While recognizing the existence of a wide vention and treatment: “integrated theory, range of theories on human behavior, this general theory, lifecourse transitions, and paper uses ecosystems theory as an over- social capital appear to offer promise for all framework for examining the drugs- the future” (Bartollas, 2000, p. 564). We crime relationship. Within this framework, will focus specifically on social capital the concept of social capital has emerged since it is a relatively new theory with recently as a promising approach to break- the potential to explain many complex ing the drugs-crime cycle. relationships.

Ecosystems theory. Human behavior, Social capital. The social sciences have including participation in drug use or crimi- always had an interest in the relationship nal activities, takes place within the broad- among community organization, social er social environment: circumstances, interaction, and individual behavior. Today, social norms, cultural conditions, and inter- the concept of social capital increasingly is actions with others (Kirst-Ashman, 2000). used to understand the extent of commu- Ecosystems theory acts as an organizing nity interaction and its effects. Social capi- framework (as opposed to a definitive the- tal was originally defined by Coleman ory of behavior or development) that calls (1988) as the quality and depth of relation- for an active awareness that the interac- ships between people in a family and tion of biology; interpersonal relationships; community. Putnam (1993) developed the culture; and legal, economic, organization- concept to include “the networks, norms al, and political forces affects an individ- and trust that facilitate coordination and ual’s behavior (Beckett and Johnson, cooperation for mutual benefit” (p. 2). The 1995; Kirst-Ashman, 2000). It should be World Bank Group (2002) modified the noted that the relative influences of each definition to include “the institutions, rela- force are likely to change throughout the tionships, and norms that shape the quali- lifecourse of each person. Essentially, ty and quantity of a society’s interactions” ecosystems theory helps provide the per- (p. 1). Finally, Rose (2000) emphasized the spective needed to understand the utility of social capital by defining it as “the breadth of systems (micro, mezzo, and stock of networks [relationships between macro) involved in any discussion of individuals] that are used to produce human behavior, as well as specific theo- goods and services in society” (p. 1422). ries that might be useful in addressing Increasing evidence shows that social

107 SPECIAL REPORT / JULY 03

Increasing capital and the social cohesion and norma- capital (informal friendship and family re- tive environment enabling its development lationships versus formal institutional evidence shows are critical for community and individual arrangements) and the quantity versus that social capital quality of life. The productive utility of in- quality of the social networks involved. formal face-to-face associations and for- and the social mal organizational networks has been The concept of social capital can be cohesion and noted, for example, in the areas of eco- applied to breaking the drugs-crime rela- nomic development (World Bank Group, tionship in several ways. First, high levels normative 2000), political participation (Putnam, of social capital in communities may play environment 2000; Putnam and Campbell, 2000), health arole in preventing drug use and other promotion (Baum, 1997, 2000; Kawachi deviant behavior through the presence of enabling its et al., 1997; Kawachi, Kennedy, and Glass, stronger formal and informal social bonds development 1999; Veenstra, 2000), and general quality and networks. The presence of anti-drug- of life at the individual and community use norms within more informal structures are critical for levels (Billings, 2000; Caspi et al., 1998; (such as family networks, communities of community and Lerner, 2000; Parcel and Menaghan, 1993; faith, and neighborhoods) may contribute Popay, 2000). to lower drug use rates. Conversely, lower individual levels of community social capital may be quality of life. Recent studies based on social control associated with greater access to drugs and social bonding theories have devel- and more lenient social norms and low- oped highly innovative solutions to crime ered social controls regarding the use of prevention, linking the levels of collective drugs or association with drug users. efficacy (Sampson and Raudenbush, 1999; Second, drug users who have recently Sampson, Raudenbush, and Earls, 1997; entered the criminal justice system may Fagan, 1987), community cohesion and/or find that the presence of high levels of integration (Hirschfield and Bowers, 1997; social capital in a community result in a Jobes, 1999; Kawachi, Kennedy, and stronger network of diversion options. Wilkinson, 1999; Kennedy et al., 1998; This could be due, in part, to formal and Lee, 2000; Mullen and Donnermeyer, informal network interest in restorative 1985; Walklate, 1998), local informal net- justice (described later in this chapter) works (Bursik, 1999; Savelsberg, 1999), versus punishment approaches to crime and youth family dynamics (Brannigan, intervention. Third, once a drug offender 1997; Hagan, 1995, 1997; Macmillan, is incarcerated, high levels of social capital 1995; Sampson and Laub, 1990) to crime within the offender’s home community rates in a given neighborhood. might better preserve networks of support for reintegration upon the offender’s re- Despite the extent of recent studies apply- lease. Offenders might more easily obtain ing the concept of social capital, very little jobs, receive support for continued sobri- research has been conducted to measure ety, and receive reinforcement for socially the relationship between social capital and appropriate behaviors. Finally, communi- drug use. The only related (and very limit- ties with high levels of social capital might ed) evidence points to the role of social have strong formal (vertical) social net- capital in preventing youth behavior prob- works in the form of coalitions or collabo- lems (Parcel and Menaghan, 1993). Put- ratives working to reduce substance use. nam (2000) found that this was especially Such agency connections may help focus true for those at higher risk for parental the community on policy development abuse. As effective intervention programs related to drug prevention and treatment are developed, it is essential to differenti- systems in homes, schools, and busi- ate between the various forms of social nesses. Such strong, integrated social

108 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

networks may offer a larger range of serv- to incorporate factors that best support ices and may develop more formal hori- the inherent personal characteristics that zontal relationships with other service affect individual behavior and they address providers, thereby improving the coordi- the broader context of the social environ- nated delivery of services and care to ment. These concepts have significant those with drug or alcohol problems. implications for how programmatic interventions may occur within the criminal jus- One example of the impact that social- tice system. capital-based concepts are currently having on the drugs-crime relationship in the Retributive justice. The traditional crimi- United States is the recent establishment nal justice perspective of retributive justice of the Office of Faith-Based and Commu- generally sees drug abuse as a willful nity Initiatives in the White House. This choice made by an offender capable of action has focused the Nation’s attention choosing between right and wrong and on the role of faith-based institutions in acting on that choice. The approach the provision of drug treatment, aftercare, emphasizes deterrence through strict and other services. Such interventions penalties, including increasing arrests, may be particularly important in poor and developing tougher sentencing laws, and minority communities with large numbers building new prisons to hold and punish of high-risk individuals, where there are offenders (McBride et al., 2001). Imple- few (if any) traditional drug treatment pro- mentation of this perspective does tem- grams. However, these same communi- porarily reduce the number of criminals on ties are often served by churches and the streets as well as interrupt an offend- other faith-based organizations that care er’s drug use. However, drug-using offend- deeply about the members of their com- ers do not appear to alter their behavior in munity and are well established in serv- the face of punishment alone (Goldkamp, ice provision. While concerns about 1994). Thus, it is highly likely that offend- church-state separation, attempts at ers will recidivate, and the cycle of drug proselytization, and teachings of bigotry use and crime will continue (Hora, Schma, and prejudice have prompted some to de- and Rosenthal, 1999). mand a clear ban on the use of public funds to support faith-based institutions, Therapeutic jurisprudence and restora- others have begun to carefully examine tive justice. Therapeutic jurisprudence the potential of these organizations to has been defined as “the use of social sci- improve the lives of their clients. At presence to study the extent to which a legal ent, there has not been sufficient research rule or practice promotes the psychologi- to determine the effectiveness of treat- cal and physical well-being of the people it ment in faith-based settings. affects” (Slobogin, 1995, p. 196). Within this framework, key players from the justice system (including judges, prosecu- Criminal justice philosophies tors, and defense attorneys) move from An examination of recent approaches to adversarial roles to problem solvers as part intervention in the drugs-crime cycle re- of a collaborative team while still perform- quires a brief review of major criminal jus- ing their traditional roles of guardians of tice philosophies and recent conceptual community protection, administrators of developments. Philosophies with the the law, and protectors of due process greatest promise for success acknowl- (Spangenberg and Beeman, 1998). Thera- edge the complex relationship between peutic jurisprudence specifically addresses drugs and crime. In addition, they attempt the needs and problems of drug offenders from a medical, therapeutic perspective.

109 SPECIAL REPORT / JULY 03

Drug addiction is viewed as a problem community reparative boards, family group with deeply rooted biological, psychologi- conferencing, and circle sentencing (see cal, and social influences, and substance Bazemore and Umbreit, 2001). The shared abusers are seen as having a condition features of these approaches include: that requires treatment. From this perspective, the criminal justice system ■ Promoting citizen and community own- offers the best opportunity some offenders ership of the criminal justice system. will ever have to confront and overcome ■ their drug use and its consequences. Pro- Providing an opportunity for the victim grammatic approaches that often employ and other community members to con- therapeutic justice principles include drug front the offender about his or her courts, restorative conferencing, cross- behavior. systems case management, coerced and ■ Providing opportunities for the offender voluntary drug treatment programs, day to learn about the impact of the crime reporting centers, and intensive monitor- and to take responsibility and be held ing approaches. Each of these approaches accountable for the offense. will be reviewed in greater detail later in this paper. ■ Creating meaningful consequences developed by the victim, the communi- Within the past decade, a justice philoso- ty, and sometimes by the offender and phy associated with the principles under- his or her support system. lying therapeutic jurisprudence has emerged: restorative justice. Used primarily for non- Although concerns and implementation violent adult and juvenile offenders, the issues exist regarding restorative justice restorative justice approach (also termed (such as some resistance by the victims’ restorative conferencing) attempts to bal- rights movement, the need for collabora- ance the needs of victims, the community, tive relations with the community at large, and offenders. Unlike retributive justice, and potential clashes with current sen- which is concerned primarily with punish- tencing and corrections law), the philo- ing the offender, restorative justice seeks sophical approach shows promise as a to repair the damage inflicted by the future direction in addressing drugs and crime. This approach makes the criminal crime (Smith, 2001). process less formal by involving the victim and community members in the planning and implementation of the sentencing. Summary Rather than asking what should be done Human behavior is an extremely complex to punish the offender, restorative justice phenomenon, and theories imply that pro- asks the following questions (Zehr, 1990): grams that acknowledge the multiple systems and factors that affect behavior will ■ What is the nature of the harm resulting have the greatest chance for realistically from the crime? assisting in behavior change—in this case, reducing both drug use and crime. ■ What needs to be done to repair the harm? Although programmatic interventions focusing on punishment and deterrence ■ Who is responsible for the repair? alone can temporarily reduce drug and crime rates, long-term solutions seem to Restorative justice has been implemented favor interventions based on principles in a number of programmatic methods, similar to those of therapeutic jurispru- including victim-offender mediation, dence as well as restorative justice.

110 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

State- and Federal-level with cancer treatments and to increase appetite in patients with AIDS. Many policy approaches to State medical marijuana laws were allow- breaking the drugs-crime ed to expire or were repealed following relationship Marinol’s approval (Dogwill, 1998). As noted previously, American drug policy Current efforts at marijuana medicalization is undergoing continual modification. Thus, began in the mid-1990s as a result of the observed relationship between crime media pressure and general dissatisfaction levels associated with drug use and drug with Marinol and other antiemetic drugs policy is constantly changing. There are (Dogwill, 1998). As of the end of the 2000 currently a broad array of drug policy legislative year, 28 States had statutes movements that may directly affect the providing for the medicinal use of marijua- drugs-crime relationship. The most wide- na (Pacula et al., 2001). The type of laws spread and potentially influential of these enacted by States varies, and States may policy changes include marijuana medical- have more than one law type. The list ization and/or decriminalization, lessening below shows the number of States with of the powder and crack cocaine sentenc- currently operating laws and a brief ing disparity, current activity surrounding description of the laws and related pro- club drugs, revisiting the concept of tections (Pacula et al., 2001): mandatory minimum sentencing, treatment versus prison, and model State drug 1. Therapeutic research programs laws. Each of these movements will be (TRPs): 14 (only 6 of which are cur- briefly described below, with a focus on rently operational). TRPs are adminis- how the proposed policy changes may tered by State health departments or affect the drugs-crime relationship. pharmacy boards and must be approved by the Food and Drug Administra- tion and adhere to specific Federal Marijuana medicalization regulations. Protection is provided Movement toward the medicalization of only to approved and participating marijuana has been ongoing since the patients, physicians, and pharmacies, 1970s (see Belenko, 2000; Goode, 1997). and for specified ailments not respond- The two actions that preceded the move- ing to other available treatments. ment were the National Commission on Marihuana and Drug Abuse report in 1972 2. Physician prescription laws: 13. These that called for reduced penalties for pos- laws are of three types: One allows session, and the unpublished 1975 trial of physicians to discuss the medical ben- United States v. Randal, which allowed efits of marijuana with patients; the the use of a medical necessity defense second allows physicians to prescribe for marijuana possession when a glauco- marijuana for medical purposes; and ma patient was arrested for growing his the third provides an affirmative own plants (Belenko, 2000). By the end defense for physician discussion or of 1982, 31 States and the District of prescription of marijuana. These laws Columbia had enacted medical marijuana protect physicians only, not patients. provisions (Markoff, 1997). However, 3. Medical necessity laws: 10. These in 1986, the Food and Drug Administra- actions provide a defense from pros- tion approved the use of the brand- ecution to patients and/or caregivers name drug Marinol (dronabinol, delta-9- for possessing marijuana for medical tetrahydrocannabinol, or THC) to prevent purposes if obtained via physician the nausea and vomiting often occurring

111 SPECIAL REPORT / JULY 03

recommendation, certification, or jail/prison penalties for subsequent pos- authorization. session offenses. MacCoun and Reuter (1997) have suggested that a better term 4. Rescheduling laws: 3. These laws re- might be depenalization. While the exact schedule marijuana to categories that definition of decriminalization is debated, recognize an acceptable use for mari- complex, and inconsistently applied, a re- juana and/or claim a lower potential for view of State statutes shows significant abuse. variation regarding possible penalties for simple marijuana possession ranging from Of the four types of laws noted above, no monetary penalties and no incarcera- only TRPs are federally sanctioned. Al- tion to fines in the five figures and multiple though the other three types of laws have years in prison (ImpacTeen Illicit Drug been or are being challenged in court, no Team, 2002). In addition, anecdotal reports firm ruling has been given that would suggest that some local police depart- clearly identify the final outcome of med- ments simply do not enforce existing ical marijuana initiatives. Although the out- marijuana possession laws. All of this sug- come of the medical marijuana debate is gests that States (and communities) show unknown, the policies in question have significant variance in marijuana policy, several ramifications for the drugs-crime and the impact of this variance should be relationship (Pacula et al., 2001). These examined to determine the possible rami- include potential decreases in marijuana- fications for arrests, black-market prices, related arrests due to a supportable de- use rates, and associated harms. fense for medical use, significant changes in black-market marijuana prices between States with varying medicalization policies, Lessening of the powder and changes in the ability or willingness to crack cocaine sentencing prosecute recreational marijuana users, disparity changes in possession penalties, and dif- There has been considerable public and ferences in use rates for both adults and research focus on the current sentencing adolescents. differences between the possession or sale of powder versus crack cocaine. Marijuana decriminalization Sentencing disparities emerged in the 1980s in the context of large increases in The decriminalization of marijuana posses- crack cocaine use, together with the con- sion in law or in enforcement policy has clusion that crack cocaine caused signifi- been evolving for many years. In the cantly more harm than powder cocaine to early 1970s, the National Commission on the individual and the community through Marihuana and Drug Abuse called for the increased violence (McBride et al., 2001). decriminalization of simple marijuana pos- Congress eventually enacted legislation session. This would mean the removal of mandating 5-year prison terms for the all criminal penalties; possession would be possession or sale of 5 grams of crack neither a felony nor a misdemeanor. In cocaine. This same legislation mandated practice, the application of such a simple the same penalty (5 years) for the posses- definition is complex. Although 11 States sion of 500 grams of powder cocaine indicate that they have decriminalized mar- (Sentencing Project, 1998). Thus, the ijuana, an examination of those statutes Federal Government defined the mandato- indicates that, operationally, decriminaliza- ry minimum sentencing disparity of crack tion means the removal of incarceration to powder cocaine at 100:1. The ramifica- for first or second marijuana possession tions of this policy became apparent fairly offenses but may include fines and/or

112 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

early in its application: There were signifi- department mentions of Ecstasy in 1994. cant increases in the prison population, in In 1999, the number had risen to 2,850 the number of drug users in prison, and (DAWN, 2000). Results of these increases specifically in the number of African- have been felt in both research and policy. Americans in prison (Beck and Mumola, Research focus on the psychopharmaco- 1999; Mumola, 1999). Currently, 86 per- logical effects of Ecstasy is growing (for cent of all Federal crack cocaine defen- example, see Boot, McGregor, and Hall, dants are African-American (Sentencing 2000), as are attempts to provide valid Project, 1998). In 1995, the U.S. Sentencing information about the effects and dangers Commission recommended the elimination of its use (Larkin, 2000). At the Federal of the sentencing disparity between crack policy level, the Ecstasy Anti-Proliferation and powder forms of cocaine, arguing that Act was enacted in October 2000. The Act the policy had not accomplished its goal of directs the U.S. Sentencing Commission reducing crack use but had resulted in sig- to increase penalties for Ecstasy traffick- nificant unintended consequences. The ing as part of an increased deterrence recommendation was not acted upon. In approach to use. State laws also are 1997, the same group recommended mov- changing, with substantial numbers of ing to a 5:1 sentencing ratio, the Clinton States moving to schedule Ecstasy and/or administration recommended a 10:1 ratio, to increase penalties for sales (ImpacTeen and an additional bill was introduced in the Illicit Drug Team, 2002). Senate specifying a 20:1 ratio. No action was ever taken, however, and the initial Reconsidering mandatory sentencing disparity remains at the original minimum sentencing Federal level of 100:1. It is important to note that at the State level, sentencing dis- Mandatory minimum sentencing plays a parity is not universally mandated (but may significant role in the drugs-crime relation- be specified in State sentencing guide- ship and has been a major component of lines). Some States, such as Michigan, the war on drugs. Initially, it was thought have begun to modify the disparity in their that high mandatory penalties for drug law laws (Sentencing Project, 1998). violations (such as serving at least 85 percent of an assigned sentence) would have The growing club drug reaction a deterrent effect on drug use, related criminal behavior, and associated costs The general term “club drugs” refers to a (see McBride et al., 2001). However, the “number of illicit drugs, primarily synthet- primary results of mandatory minimum ic, that are most commonly encountered sentencing likely have been to increase at nightclubs and ‘raves’” (Drug Enforce- dramatically the number of drug-related ment Administration Intelligence Division, arrests and the proportion of prisoners 2000, p. 1). Examples of club drugs in- who are drug users (Harlow, 1998; Mumola, clude Ecstasy, Ketamine, Rohypnol, and 1999). Mandatory minimums for drug GHB (gamma-hydroxybutyrate). Both use charges may play a significant role in the rates and emergency department men- shifting of power from judges to prosecu- tions for these substances (especially tors, prison overcrowding, and a break- Ecstasy) have recently increased. Johns- down in truth-in-sentencing laws because ton, O’Malley, and Bachman (2001) report of early release due to prison overcrowd- that use of Ecstasy in the past 12 months ing. In reality, prison overcrowding often among 12th graders increased from 6 per- makes mandatory minimum sentencing cent in 1999 to 8 percent in 2000. Accord- laws all but impossible to enforce (see ing to the Drug Abuse Warning Network McBride et al., 2001). (DAWN), there were only 25 emergency

113 SPECIAL REPORT / JULY 03

Those who question the appropriateness Farabee, Prendergast, and Anglin (1998) of mandatory minimum sentences have determined that findings generally sup- been supported by studies suggesting that ported the use of coercive measures to this approach to addressing the drugs- increase the likelihood that an offender crime relationship is not effective and is will enter and remain in treatment. Speci- more costly than treatment (for example, fically, they concluded that compulsory see Caulkins et al., 1998). Significant activi- substance abuse treatment is “an effec- ty at the State and Federal level is focusing tive source of treatment referral, as well on mandatory minimum sentencing revi- as a means for enhancing retention and sion. Along with seeking to reduce the compliance” (p. 7). Since researchers gen- crack/powder sentencing discrepancy, the erally agree that length of time in treat- U.S. Sentencing Commission has been ment is strongly related to treatment actively supporting efforts to reevaluate success, coercing offenders into treat- mandatory minimum sentencing (Sen- ment and then applying graduated sanc- tencing Project, 1998). New York (the State tions to motivate continued participation is that played a major role in the introduction a potentially successful strategy. It can of mandatory minimum sentencing for certainly be stated that coerced treatment drug offenders via the Rockefeller Drug plays a major role in treatment referrals. Laws) is seriously considering significant Recent studies indicate that the criminal modification of its policies. The proposed justice system is responsible for 40 to 50 New York modifications focus on an expan- percent of community-based treatment sion of treatment services, a reduction in program referrals (Farabee, Prendergast, the range of mandatory minimum sen- and Anglin, 1998). Rates of referral vary tences, and an expansion of judicial discre- widely by substance, with marijuana and tion (Sengupta, 2001). If and when these methamphetamine referrals occurring sig- changes take place (at the national level nificantly more often than referrals for and/or in specific States), it will be impor- other substances (Drug Abuse Warning tant to examine their impact on the drugs- Network, 2000). crime relationship. However, Taxman (2000) argues that merely mandating an offender to treat- Treatment versus prison ment does little to increase motivation Coerced treatment (also referred to as or success. Simpson and colleagues compulsory, mandated, or involuntary (Simpson et al., 1997; Simpson, Joe, and treatment) is a heavily debated issue. Brown, 1997) have found that failure to Some oppose the practice on philosophi- address motivation and readiness for treat- cal or constitutional grounds, while many ment reduces treatment effectiveness. In treatment clinicians maintain that treat- addition, Farabee et al. (1999) maintain ment can be successful only if a person that the application of mandated treatment is truly motivated to change. Other re- varies widely, ranging from simple referral searchers (Anglin and Maugh, 1992; to treatment to strict graduated sanctions Salmon and Salmon, 1983) and policymak- with heavy monitoring and clear penalties ers have argued that few chronic addicts for failure. More research is needed to will voluntarily agree to enter and remain determine which offender types may ex- in treatment without external coercion. perience the greatest benefits of coerced In a review of research examining the treatment, and with which levels of treat- relationship between various levels of ment structures and settings (e.g. residen- legal pressure and treatment outcomes, tial versus intensive outpatient with heavy monitoring).

114 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Reports on the promise of coerced offenders into 18-month residential treat- treatment have prompted some State ment programs in lieu of prison time legislatures to adopt various forms of (Gallagher, 2001). These plans resulted pri- corrections-initiated drug treatment for marily from the recommendations of an nonviolent drug-using offenders. The fol- independent commission charged to study lowing is a review of these State initia- the impact of drug cases on New York tives, as well as a Federal measure under State courts. The principal recommenda- current consideration. tion of this commission was to “launch a systematic, statewide approach to the California. State voters recently passed delivery of ‘coerced’ drug treatment to the Substance Abuse and Crime Preven- nonviolent addicts in every jurisdiction” tion Act of 2000, which targets $128 mil- (New York State Commission on Drugs lion per year to help counties develop the and the Courts, 2000, p. 7). capacity to provide drug treatment, literacy training, family counseling, and vocational Massachusetts. The Department of training services for an expected 36,000 Public Health’s Bureau of Substance new treatment clients per year (San Abuse Services recently reported that Francisco Examiner, 2000). integrating such services across the State resulted in significant improvements in a Arizona. The Arizona Drug Medicalization, number of categories, including reductions Prevention and Control Act of 1996 in crime involvement, psychological prob- requires mandatory treatment and pro- lems, and use of health services, as well hibits incarceration of first- and second- as improvements in employment levels time drug offenders. A 1998 Arizona and abstinence rates (Massachusetts De- Supreme Court report concluded that the partment of Public Health, 2000). Based in State saved $2.5 million in its first year by part on these successes, ballot initiative sending users into treatment rather than Proposition P was introduced in the 2000 prison (Arizona Supreme Court, 1999). general election to divert drug forfeiture Although critics claim it is too early to money from police and district attorneys argue for program effectiveness due to to treatment centers. The measure failed, selection bias and lack of long-term recidi- possibly due to claims that the initiative vism rates, the study found that 77 per- was a cover for efforts to decriminalize cent of offenders tested drug free at the dangerous drugs (Boston Globe, 2000). end of their outpatient treatment programs. In addition, probationers who National. The U.S. Senate is currently received treatment were twice as likely to considering the recently introduced Drug be employed (90 versus 41 percent), to Abuse Education, Prevention, and Treat- finish community service requirements ment Act of 2001 (S. 304, 2001). The (85 versus 40 percent), and to complete measure would, among other things, probation successfully (85 versus 22 per- authorize new funding grants to States for cent) when compared with those who did the purpose of providing drug treatment not complete treatment. services to inmates and residential treatment facilities. New York. Governor Pataki recently unveiled a plan to reform the State’s Rockefeller Drug Laws by cutting mini- Model State drug laws mum sentences from 15 to 8 1/3 years for In 1992, the President’s Commission on some offenses, giving judges increased Model State Drug Laws was charged with discretion in sentencing, and giving prose- the task of creating a compilation of model cutors the power to divert repeat drug State laws that would effectively address

115 SPECIAL REPORT / JULY 03

drug and alcohol use (President’s Com- and members (National Alliance for Model mission on Model State Drug Laws, State Drug Laws, 2001). Several States 1993). After a series of public hearings, have passed legislation using the model drug treatment program site visits, and laws as a framework for laws specifically meetings with various individuals, agen- tailored to their needs, including Arizona, cies and groups, a total of 44 model laws Arkansas, Georgia, Iowa, Kansas, Louisi- and policies were developed. In its report, ana, Mississippi, New Jersey, North the Commission noted that Carolina, Pennsylvania, and Utah (National Alliance for Model State Drug Laws, [T]he legislative remedies offered 2001). However, no known evaluations of within do not rely exclusively on pun- the impact of these laws currently exist. ishment and deterrence to “solve” Additional efforts by the Alliance to assist drug problems. Instead, the goal of with drug policy revision include providing this report is to establish a compre- national and Federal agencies with assis- hensive continuum of responses and tance on State and local laws and policies. services, encompassing prevention, education, detection, treatment, rehabilitation, and law enforcement Summary to allow individuals and communities Trends in State- and Federal-level policies to fully address alcohol and other aimed at the drugs-crime relationship can drug problems. Tough sanctions are (and indeed do) move in different direc- used to punish those individuals who tions for different substances. Although refuse to abide by the law. More there has been considerable movement to importantly, the recommended sanc- modify marijuana laws at the State level, tions are designed to be construc- no comparable action has been seen at tive, attempting to leverage alcohol the Federal level. The movement toward The President’s and other drug abusers into treat- reducing the sentencing disparity between Commission ment, rehabilitation, and ultimately, crack and powder cocaine (as well as re- recovery. (pp. 1–2) duce overall penalties) is co-occurring with on Model State State and Federal trends to increase the Drug Laws’ five The five main policy areas are as follows scheduling and penalties for club drugs (see appendix A for a listing of specific such as Ecstasy. A further concern raised main policy areas model laws and policies within these by this section is that although research are economic areas): economic remedies, community may indicate the legitimacy and wisdom mobilization, crimes code enforcement, of revising current policy (such as moving remedies, treatment, and drug-free families/schools/ to coerced treatment instead of incarcera- community workplaces (President’s Commission on tion), there is often significant resistance Model State Drug Laws, 1993). to such actions based on the fear of fur- mobilization, ther escalations of the drugs-crime con- Following the compilation of the model crimes code nection or negative voter reaction. The laws and policies, The National Alliance for nature of public policy is complex and re- enforcement, Model State Drug Laws (Alliance) was ciprocal: The public elects policymakers organized as a nonprofit group that would treatment, who support the majority view. This tends serve as an ongoing resource for States to make legislators cautious about sup- and drug-free considering implementation of legislation porting changes in drug policy. Therefore, based on the model laws. The Alliance families/schools/ the development of possible public policy has held several conferences across the that might contradict traditional viewpoints workplaces. United States to work with elected and can be highly problematic (Tonry, 1996). appointed officials, substance abuse pro- However, the breadth and scope of poten- fessionals, and other community leaders tial legislative actions is impressive. With

116 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

an increasing number of States developing receive drug treatment services in a sig- innovative laws based on examples such nificantly restrictive prison-based thera- as the Model State Drug Laws, there is peutic community. Nonviolent drug-using need for researchers to examine the pos- offenders might receive sentencing and sible effects of such policy changes. ongoing supervision from a drug court and participate in minimally restrictive victim- This paper has reviewed a wide variety of offender mediation, along with mandated data describing the drugs-crime relation- attendance in intensive outpatient drug ship and its complex nature, conceptual treatment services. frameworks that may help interpret the relationship, and the implications of policy Many jurisdictions struggle to integrate for the relationship. An important part of substance abuse treatment into their crim- society’s reaction to the relationship has inal justice systems, which often view been to develop programs to intervene such efforts as adjunct services rather with or break the drugs-crime cycle. Al- than primary, integrated components. though such intervention attempts have Taxman (2000) notes six threats that im- occurred for more than a century, they pede the implementation of treatment have become increasingly sophisticated as services: policy makers and clinicians have come to further understand and apply research ■ Lack of clear crime control goals for findings and relevant conceptual models. treatment services. The next section of this paper examines ■ many of the intervention programs that Lack of clear assessment and eligibility have been used and assesses key pro- requirements. gram elements that have shown some ■ Insufficient treatment duration to effect success at intervening in the drugs-crime behavioral change. relationship. ■ Lack of supervision and sanctions/ rewards to reinforce treatment goals. Integrated programmatic approaches to breaking the ■ Lack of objective drug testing to monitor drugs-crime cycle treatment progress. ■ In developing programmatic interventions Insufficient case management services. designed to break the drugs-crime cycle Many researchers and practitioners have among offenders, it is essential to ensure argued that to address these threats, a that neither community safety nor offend- comprehensive and integrated approach er accountability be compromised in any should be used to maximize treatment way, particularly for violent and chronic success and minimize future harm to offenders. However, as noted previously, the community (Anglin and Hser, 1990; drug-related crimes exist along a continu- Inciardi et al., 1997; Taxman, 1998, um of severity ranging from index crimes— Farabee et al., 1999; Martin et al., 1999; such as murder and armed robbery—to Taxman, 1998). Taxman (2000) argues for more minor offenses such as nonviolent a systems approach in which “correctional drug possession. Interventions such as and treatment agencies build a delivery drug treatment should be provided along a system that cuts across and integrates the continuum as well. Drug-involved offend- systems, reduces duplication in efforts to ers who commit serious crimes might create and recreate processes for unique

117 SPECIAL REPORT / JULY 03

programs, and emphasizes empirically manner, are destined to produce faulty driven programmatic components” (pp. and inadequate recommendations and 5–6). decisions. Careful assessment mechanisms not only will help identify those The following review will discuss interven- services that are most needed by offend- tions designed to break the drugs-crime ers, but also will prevent system duplica- cycle among offenders using an integrated tion leading to inefficient and poorly approach that can be applied throughout coordinated service delivery. By properly the range of sentencing alternatives.3 This assessing and coordinating services at approach, which integrates restorative intake, the justice system can more effec- justice with an ecosystems framework, tively work towards preventing increasing includes the following components: imme- levels of future recidivism and drug use. diate and comprehensive assessment; judicial processing, including the use of Offender evaluation generally occurs in drug courts; supervision and monitor- two phases: initial screening, followed by ing, including graduated sanctions and more comprehensive assessment. The cross-systems case management; cross- primary purpose of initial screening is to systems collaboration; the drug treatment determine if the need for a more compre- service continuum; and aftercare. hensive assessment exists. Thus, it is Common inappropriate to use screening instruments to formulate a diagnosis or decide screening Comprehensive assessment and treatment planning treatment needs. Screening instruments instruments also filter out individuals with medical, Appropriate client selection, assessment, psychological, or legal problems that include the CAGE and placement have been identified as need to be addressed prior to placement. Questionnaire, critical components of the treatment con- Common screening instruments include tinuum (Simpson and Curry, 1997–98; the CAGE Questionnaire, the Michigan the Michigan Taxman, 1998; Farabee et al., 1999). Sub- Alcoholism Screening Test, and the Alcoholism stance abuse problems are usually en- Offender Profile Index (for more detailed meshed within a wide variety of other descriptions of these tools, see Inciardi, Screening Test, issues. Thus, comprehensive assessment 1994). and the Offender is necessary to successfully address alcohol and other drug problems. If the screening instrument indicates an Profile Index. alcohol or other drug problem, a more Assessment. Assessment usually occurs comprehensive assessment is needed. At at the point of intake into the criminal jus- minimum, a comprehensive assessment tice system (often at either centralized should include: intake centers or police stations). Intake recommendations can heavily affect judi- ■ An indepth examination of the severity cial decisions; it is imperative that intake and nature of the alcohol and other drug personnel be thoroughly trained in the use abuse identified by the screening of comprehensive assessment tools. Such process. training should include incorporation of culture and ethnicity issues in comprehen- ■ A more thorough assessment of addi- sive evaluations, as well as dealing with tional problems flagged during screening the complexities of clients with multiple and further inquiry into problems that diagnoses. A poorly conducted assess- may not have been identified up to that ment, using techniques and measurement point. instruments that do not consider the ■ A strong effort to use multiple methods offender’s entire life situation in a holistic and sources.

118 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Components of a comprehensive assess- among incarcerated drug-using popula- ment instrument include: tions. Since the early 1970s, researchers have called attention to the special needs ■ History and current patterns of alcohol of jail inmates with mental illness (Gibbens, and other drug use. 1979; Gold, 1973; Verma, 1979). Although indepth studies on the prevalence of ■ Past and current involvement in the mental illness in prisons are very limited, criminal justice system, including any researchers estimate that around 7 to 9 history of violent behavior and manifes- percent of jail inmates are mentally ill tations of antisocial personality and psy- (BJS, 1999, as cited in Lurigio and Swartz, chopathology. 2000, p. 67). Rates of mental illness among those who are alcohol or drug ■ Family and social support systems. dependent are believed to be much high- ■ Medical history and current health sta- er. Peters and colleagues (1992) found tus, including HIV/AIDS screening. that, of jail inmates who were receiving substance abuse treatment, more than ■ Mental health history and current status, half self-reported a history of depression, including screening for any history of 45 percent reported serious anxiety or abuse, anxiety, or depression. tension, and 19 percent had a history of suicidal thoughts. Among juveniles, the ■ Educational and vocational history and Northwestern Juvenile Project has esti- needs. mated that two-thirds of juvenile detainees have one or more alcohol, drug, or Two commonly used assessment instrumental disorders (Teplin, 2001). Because ments are the Addiction Severity Index depression is also a consistent predictor (ASI) and the Wisconsin Uniform Sub- of therapeutic noncompliance, it is impor- stance Abuse Screening Battery (adapted tant to make sure that an alcohol or other from the well-known Minnesota Multiphasic drug-diagnosed arrestee is properly as- Personality Inventory). The Wisconsin sessed and treated for depression or other instrument is composed of four separate mental disorders (Markou, Kosten, and sub-instruments: the Alcohol Dependence Koob, 1998). Scale, the Offender Drug Use History, the Client Management Classification inter- The conditions and care received by the view, and the Megargee Offender Typ- detained mentally ill have been found to ology. Important supplemental tests to be grossly inadequate (Alemagno, 2001; these comprehensive assessment instru- Birmingham et al., 2000; Lurigio and ments include the AIDS Initial Assess- Lewis, 1987). Outcome studies suggest ment Jail/Prison Supplement and various that to serve this population better, the biological tests to determine recent drug most effective approach includes ade- or alcohol use, including urinalysis, breath- quately training jail and prison personnel alyzer tests, blood tests, hair analysis, and to meet emergency situations, perform sweat tests (for more detailed descrip- basic assessments, and make appropriate tions of all of these tools, see Inciardi, referrals to community-based mental 1994). health services where safety concerns can be adequately monitored. Such an Comorbidity issues. Researchers report approach would have the added benefit of high rates of depression in street drug- also avoiding community-based service using populations (McBride et al., 2000). duplication (Cox, Landsberg, and Paravati, Additionally, a wide variety of data suggest 1989; Lurigio, 2000). that there is a high rate of comorbidity

119 SPECIAL REPORT / JULY 03

Treatment planning. The treatment plan and often uses an intensive supervision should be based on the client’s needs, and treatment program based on graduat- problems, strengths, and resources as ed sanctions (described below). Drug identified in the assessment process, and courts are partnerships between justice it should seek to use assessment informa- system personnel (prosecutors, defense tion to match the client with the best attorneys, and judges); treatment special- treatment modality and level of risk ists; and other social service personnel (Inciardi, 1994; McLellan et al., 1997; see (National Association of Drug Court Pro- also Taxman, 2000). Although clients fessionals, 2000). Drug courts allow should participate in the planning process judges to take a more active role than was to improve buy-in and treatment compli- provided by such previous options as man- ance, they cannot dictate treatment goals. dated lengthy sentences and to partner Treatment planning goals and objectives with community resources and agencies. should be specific, measurable, and attain- Judges draw on a variety of professionals able. They should also be flexible enough in assessing needs and recommending to adapt to emerging client needs as they services. They are then actively involved move through the criminal justice and in the decisionmaking process regarding treatment systems. Goals must conform what services are to be received. Judges to the limitations imposed by the court, also monitor compliance and apply sanc- parole or probation department, or other tions when a lack of compliance is evi- criminal justice agency that has jurisdiction dent. Some of the most unique and over the client. Good treatment plans also essential principles of drug courts include are designed to address issues related to immediate and upfront intervention; coor- treatment attrition, noncompliance, and dinated, comprehensive supervision; inadequate progress (Inciardi, 1994). access to a wide variety of treatment services including long-term treatment and At the conclusion of intake and assess- aftercare; and graduated sanctions and ment, intake officers generally have the incentive programs (Tauber, 1994; for option of dismissing the case with no fur- more indepth information on suggested ther action, placing the offender in a diver- organizational factors, see Berman and sion program, or referral to further justice Anderson, 1999; Cooper, 1997; McBride system processing. et al., 1999; National Association of Drug Court Professionals, 2000; Peyton and Judicial processing Gossweiler, 2001). If a decision is made to formally refer an Evaluations of drug courts have been offender to court for further processing, mixed. Concern has been expressed over judges will generally use the assessment evaluation research methodology, wide and arrest report as well as other facts to variations in populations served, and lack determine disposition and, if necessary, of consistent standards for assessment sentencing. In most jurisdictions, fact- and referral (Inciardi, McBride, and Rivers, finding and adjudication take place in 1996; U.S. General Accounting Office, conventional court systems. However, in 1997). More recent reviews by Belenko an attempt to play a more active role in (1998) and Covington (2001) have conclud- breaking the linkage between substance ed that drug courts have not been subject- use and crime, the judicial system devel- ed to consistent or methodologically oped the drug court. strong evaluations that define terms clearly (from program elements to definitions Specifically, a drug court takes responsibili- of success), examine the long-term impact ty for less serious drug-using offenders,

120 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

of drug courts using appropriate compari- Graduated sanctions. Judicial processing son groups, or identify what program ele- within systems such as drug courts often ments contribute to successful outcomes. relies on graduated sanctions for supervi- Peyton and Gossweiler (2001) suggest the sion purposes. This approach helps ensure need for more comprehensive policies and offender rights and deters noncompliance. protocols consistently applied in all drug Graduated sanctions are based on the the- courts. This would contribute significantly oretical foundation of procedural justice, to methodologically strong evaluations. which posits that compliance is enhanced by procedures that are perceived as fair With the above concerns noted, evidence (Taxman, Soule, and Gelb, 1999). Lack of still points to a positive impact for drug compliance is a significant problem across courts: high treatment retention, increased the justice system. Studies indicate that sobriety, and reductions in recidivism have as many as 61 percent of probationers fail been noted in many drug court locations; to comply with release conditions (Langan in addition, savings in jail costs can be and Cunniff, 1992), and that 30 to 80 per- substantial (Drug Strategies, 1997; cent of new prison intakes each year are Cooper, 1997; Harrell, Cavanagh, and probation and parole violators (Burke, Roman, 2000). A recent evaluation of a 1997; Rhine, 1993). Some critics have midwestern drug court by Spohn and col- expressed concerns that graduated sanc- leagues (2001), which used a comparison tions are a form of “net widening,” in group design and controlled for a variety which probationers are given technical vio- of social and behavioral characteristics, lations for positive urinalysis tests. Such concluded that drug court participants had positive tests have become the equivalent significantly lower rates of recidivism than of crimes, although they are described by those who received standard court pro- the drug treatment system as relapses. cessing. To be successful, drug courts require a long-term outlook, significant Taxman, Soule, and Gelb (1999) state that initial resource allocation, and available the efficacy of graduated sanctions results treatment slots (Platt, 2001). Additional from the use of structured, incremental research is needed to address the signifi- responses to noncompliant behavior and cant issues critics have raised regarding from an emphasis on swift response to the scientific support for drug court noncompliant acts through a series of spe- enthusiasm. cific sanctions that vary based on such factors as the nature and number of violations. The concept of graduated sanctions Supervision and monitoring applies to the following: As stated in the introduction to this sec- ■ The type of initial treatment intervention tion, interventions for drug-using offend- (outpatient, residential, or types of col- ers must ensure community safety as well laborative services). as offender accountability. Programmatic approaches designed to help accomplish ■ The service delivery sentencing context safety and accountability goals include (from community diversion to incarcera- supervision via a system of graduated tion with coerced drug treatment in a sanctions, use of drug monitoring and State training school). testing to substantiate accountability, and system oversight and coordination ■ Overall intervention/treatment program through cross-systems case management. outcome goals. ■ Progress within the program (McBride et al., 1999).

121 SPECIAL REPORT / JULY 03

Taxman, Soule, and Gelb (1999) state that critics regard as net widening. Regular to be effective, graduated sanctions must drug testing is often part of an overall include three specific elements: strategy in which both treatment and criminal justice systems use graduated sanc- ■ Inform offenders about infractional tions to monitor compliance. Advocates of behavior and the potential conse- such strategies recommend that testing quences for such behavior. must be conducted frequently and randomly. Researchers (Marlowe, 2001; ■ Ensure that all members of the graduat- Taxman, Soule, and Gelb, 1999) have rec- ed sanctions judicial team adhere to the ommended several compliance-gaining agreed-on sanctions model. strategies, including clarification of negative and positive behaviors as well as ■ Strive to uphold offender dignity. swift, certain, and progressive responses. Use of a behavioral contract informing the It is important to use a team approach in offender of the graduated sanctions menu which treatment providers and criminal should be developed at intake or at the justice personnel share information about time of court-ordered probation. Such a progress or relapse issues. It is also im- sanctions menu should reflect certainty, portant to ensure that offenders are tested consistency, parsimony, proportionality, as long as they are under criminal justice and progressiveness (Taxman, Soule, and system supervision. Gelb, 1999), and it should provide for A wide variety of testing methods exists equivalent responses that allow for tailor- for illicit drugs, with variation in reliability ing sanctions to specific cases. and validity among testing procedures. Research specifically evaluating graduated The most widely practiced technique is sanctions approaches is very limited. How- urinalysis. Urinalysis offers a number of ever, the use of this approach is quite advantages compared with other testing common within drug courts. In addition, methods, including ease in obtaining a initial studies indicate that offenders in a sample, ability for sample retest, and low pretrial intervention program that used cost (Jacobs, DuPont, and Gold, 2000). graduated sanctions had lower rearrest However, subjects can easily tamper with rates for both short- and long-term (1-year) samples, and testing only reflects drug followup (Harrell, 1998). In addition, the use within the last few days. The window cost-benefits of graduated sanctions indi- of detection is also small for blood sam- Although a cate promise (Greenwood and Turner, pling, although results are highly reliable. 1993; Rivers and Trotti, 1995). In contrast, hair analysis allows for detec- combination of tion of long-term use (within the last 90 modalities is likely Drug monitoring and testing. In recent days), but provides unreliable data for years, drug testing programs have be- studying variables other than simple drug to offer the most come increasingly widespread in criminal presence. The least invasive testing tech- accurate results, justice settings (Jacobs, DuPont, and niques include sweat patch, saliva testing, Gold, 2000). In 1998, 71 percent of jails and nail testing, but the wider utility of privacy and reported having a policy to test inmates these approaches remains to be studied. feasibility issues for drug use; however, only 8 percent Although a combination of modalities is imposed mandatory treatment in response likely to offer the most accurate results, usually determine to positive test results. Instead, the most privacy and feasibility issues usually deter- which methods are common responses to positive testing mine which methods are used in practice involved punitive sanctions ranging from (Jacobs, DuPont, and Gold, 2000). Com- used in practice. loss of privileges to adding time to the prehensive outcome studies are needed sentence (Wilson, 2000), a practice that to evaluate the linkages between drug

122 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

testing and expected (negative) conse- services, and the workforce. In addition, quences for positive results. CMs may intervene in crisis situations or assist offenders with relapse prevention Cross-systems case management, strategies such as developing non-drug- including TASC. Case management pro- related leisure activities. Intensive case vides one way for criminal justice systems management services are most critical to coordinate the comprehensive needs during the vulnerable 2-month period fol- of offenders. Case management has lowing discharge from primary treatment. emerged as a strategy to connect clients They provide continuity of care while to needed resources throughout the serv- simultaneously working to move the client ice continuum, at intake, during treatment, toward independence. and after treatment. Case management results in more rapid service access (Bo- Although a CM can help an offender navi- kos et al., 1992), higher levels of goal gate through the interconnected array of attainment (Godley et al., 1994; Rapp, treatment services, it is also clear that 1997), longer lengths of stay in treatment such services must be provided in the (Rapp et al., 1998), reductions in drug use context of the justice system. Drug courts, (Rapp, 1997), improved employment func- probation offices, and other criminal jus- tioning (Siegal et al., 1996) and improved tice system components must work with connection to needed resources over time CMs to coordinate an offender’s move- (Dennis, Karuntzos, and Rachal, 1992; ment through the justice system via the Godley et al., 1994; Schlenger, Kroutil, and use of graduated sanctions. The graduated Roland, 1992) when compared with stan- sanctions process allows the judge or pro- dard treatment services. Research sug- bation officer to maintain an appropriate gests that case management may be balance between community protection effective as an adjunct to substance abuse and offender rehabilitation. However, treatment for two reasons: Retention in judges generally have neither the time nor treatment is generally associated with bet- the training to ensure that offenders re- ter outcomes, and one of case manage- ceive a continuum of services. According ment’s primary goals is to keep the client to a recent NIJ examination of case engaged in the treatment process (Kolden management within the criminal justice et al., 1997; Siegal et al., 1995, 1996, system (Healey, 1999), optimum case 1997); and treatment is more likely to suc- management models currently combine ceed when a client’s non-substance-abuse two broad approaches: strengths-based problems are also being addressed (e.g. case management—focusing on a client’s financial problems, family problems, etc.; self-identified strengths and talents when see Siegal, 1998). developing a service plan, and assuming a client’s ability to use these strengths to Case managers (CMs), who are often move toward “socially acceptable choic- mental health or social workers, support es” (Clark, 1997; Enos and Southern, and reinforce treatment goals throughout 1996; Rapp et al., 1998; Siegal et al., the treatment continuum by providing the 1997); and assertive case management— following three functions: assessment requiring active involvement of the CM in (Babor et al., 1991); treatment planning seeking out and delivering services to and goal setting, linking, monitoring and clients as opposed to passive service pro- advocating (Ballew and Mink, 1996), in- vision (Healey, 1999; Inciardi, McBride, cluding navigating the often-confusing and Rivers, 1996). Within the criminal jus- social service system (Spear and Skala, tice setting, CMs combine support and 1995); and assisting in offender reintegra- positive regard for a client’s strengths with tion with home or other placement, social clear disapproval of the behaviors that led

123 SPECIAL REPORT / JULY 03

the client to become involved with the processes and 10 critical elements justice system. (Bureau of Justice Assistance, 1995). The four processes are: Healey (1999) notes that criminal justice case management often involves a con- ■ Identification of appropriate drug- scious blurring of roles between CMs, involved offenders. mental health providers, substance abuse counselors, domestic violence program ■ Assessment of treatment needs. counselors, and other social service ■ providers. Taxman and Sherman (1998) Referral to appropriate services and have suggested that much of the role con- placement. fusion can be reduced through a systemic ■ Continuous case management at all approach to case management, including points along the criminal justice process- agreed-on role clarifications and resource ing continuum (Anglin, Longshore, and allocation. Significant cross-training is Turner, 1999). often necessary to allow such blurring to take place without confusion of appropri- The 10 critical elements involve: ate role responsibility or misunderstand- ings regarding philosophical differences ■ Broad-based support within both the (Healey, 1999). criminal justice and treatment systems with formal communication systems. Effective use of assessment data within a case management framework requires ■ Independence as a unit with designated a complex information system that can administrator. ensure the availability of relevant information to those involved in service provision ■ Appropriate staff training on TASC poli- (Taxman and Sherman, 1998). If services cies and procedures. are to be integrated effectively, it is crucial that intake, assessment, and progress ■ An established data collection system. information be shared and not be need- ■ lessly duplicated. Such information can Explicit and agreed-on eligibility criteria. play a major role in increased service deliv- ■ Documented assessment/referral ery efficiency and improve the outcome of screening procedures. provided services (for further discussion of this area, see Mahoney et al., 1998). ■ Documented policies and procedures for drug testing. Perhaps the best example of a programmatic approach incorporating cross-systems ■ Offender monitoring procedures, includ- case management is TASC: Treatment ing reporting procedures (Bureau of Alternatives for Safe Communities (also Justice Assistance, 1995). known as Treatment Alternatives to Street Programs can be Crime, or Treatment Accountability for The usual position of a TASC program is Safer Communities). TASC is recognized that of a neutral party. Most program sites perceived as using as an offender management model do not provide treatment services of their nonbiased referral (Anglin, Longshore, and Turner, 1999) that own, nor are they an official member of links criminal justice system legal sanc- the criminal justice system. Thus, the judgments and tions with drug treatment program thera- programs can be perceived as using non- case management peutic interventions (Sigmon et al., 1999; biased referral judgments and case man- see also Inciardi and McBride, 1991). agement decisions. decisions. The TASC approach consists of 4 distinct

124 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Evaluations of TASC programs have been Cross-systems collaboration mixed, based on whether the evaluation is examining operational/procedural issues or By definition, the drugs-crime relationship outcome issues. Operational/procedural crosses currently accepted jurisdictional evaluation results (see Anglin, Longshore, responsibilities and requires system part- and Turner, 1999) have been consistently nerships. The promising components positive, citing strong screening and iden- described so far in this paper demand the tification of drug-using offenders (Toborg successful integration of a wide variety of et al., 1976); effective linkages with the services and jurisdictions, including crimi- criminal justice system; increased ethnic nal justice, drug treatment, social services, diversity in treatment; and increased and public health. Effective use of immedi- treatment participation (Collins et al., ate and comprehensive assessment, drug 1982); improvements in treatment reten- courts, communication necessary for suc- tion (Hubbard et al., 1989; Inciardi and cessful use of graduated sanctions, cross- McBride, 1991), and considerable cost- systems case management in the form of benefit ratios when compared with any agencies such as TASC, and post-criminal- form of incarceration (System Sciences, justice transition services to reintegrate 1979). Outcome evaluations have been drug users back into the community—all mixed. Studies focusing on recidivism of these approaches are based on an inte- generally show that TASC clients either grated care system. Yet, as Sigmon et al. have higher recidivism rates or no signifi- (1999) note, the adjudication process is cant differences in recidivism compared historically an adversarial system, and cre- with control groups (Anglin, Longshore, ating successful partnerships that involve and Turner, 1999; Owens et al., 1997). a variety of individual agencies is often However, as TASC uses higher monitoring difficult. levels, results on recidivism may simply To build the infrastructure required to sup- indicate “net widening”; those who are port cross-systems interactions, collabora- watched more are caught more. This may tive efforts are becoming widespread. indicate a possibility of higher public safety Eisenburg and Fabelo (1996) argue that in TASC communities, rather than program failure to develop an integrated infrastruc- failure. Anglin, Longshore, and Turner’s ture not only negatively affects the out- (1999) review of five TASC programs cho- comes of individual programs, but also sen to reflect similar programmatic and hastens treatment decay. Such infrastruc- population characteristics (including adher- tures have a variety of names but one ence to the 10 critical elements) indicated essential goal: to have representatives favorable outcomes for service delivery, from key agencies and services join drug-use days, drug crimes, and sexual together to identify the problems their activity while high on drugs. However, community is seeking to target, develop these results were either modest or were effective goals and strategies to address confined to high-risk offenders. Anglin, those problems, and then oversee the Longshore, and Turner conclude that more implementation of those goals and strate- problematic offenders may receive the gies (Sigmon et al., 1999). The types of highest benefit from program participation. problems such collaborative efforts ad- Covington (2001) reminds program address should not be narrowly construed. ministrators and researchers that TASC Sigmon and colleagues (1999) refer to programs have generally not received con- adjudication partnerships as an “umbrella sistent methodologically strong long-term concept under which many interagency outcome evaluations. Future research efforts can be classified” (p. 2). should focus on these issues.

125 SPECIAL REPORT / JULY 03

While collaborative formation usually coddled (Lipton, 1998). However, research results from grassroots efforts of local involving numerous large-scale studies leaders (Sigmon et al., 1999), the recent consistently demonstrates that treatment emergence of State- and county-level has beneficial outcomes. These federally managed-care models often require funded and independently evaluated stud- provider subcontracts and collaboration ies—including the Drug Abuse Treatment (McBride et al., 1999). Key agency mem- Outcome Study (DATOS), the National bers for collaboratives addressing drugs Treatment Improvement Evaluation Study and crime would include justice system (NTIES), the Treatment Outcome Prospec- agencies (offices of the prosecution, the tive Study (TOPS), and the Drug Abuse defense, and the court), as well as other Reporting Program (DARP)—have all con- groups such as law enforcement, welfare, firmed drug abuse treatment efficacy State and local corrections, managed through 1-year followup. These findings behavioral health care, community treat- remained valid when controlling for type ment, the health department, and State of service received (residential long-term, and local managed-care initiatives (Mull, outpatient drug-free, or outpatient metha- 1998; Sigmon et al., 1999). Such a mem- done maintenance) as well as drug and bership list would allow two essential client type (U.S. General Accounting types of individuals: “1) those who under- Office, 1998). However, the National stand and have an interest in the broad Research Council (2001) has questioned and specific problems of community wel- the strength of these studies’ conclusions, fare, justice, alcohol and other drug abuse, arguing that because the studies lacked and health and social services, and 2) randomized assignment, researchers community leaders who can ensure that “could not provide rigorous evidence productive change occurs” (McPhail and on the relative effectiveness or efficacy Wiest, 1995, p. 28). of particular drug-by-treatment combinations, or estimate the absolute effect size, Although each collaborative will be uniquely cost-effectiveness, or benefit-cost ratio of tailored to the community it serves, re- treatment” (p. 230). views of collaborative efforts have identified several critical elements for success Cost savings for treatment relative to (Sigmon et al., 1999, pp. 2–4; see also incarceration, interdiction, and health care Bureau of Justice Assistance, 1995; Mc- expenditures have been estimated by two Bride et al., 1999). These include leadership recent studies. The first, the California designation, membership integration, goal Drug and Alcohol Treatment Assessment setting, development of a team approach, (CALDATA), examined the effectiveness, emphasis on a long-term view, research costs, and benefits of providing alcohol and evaluation, efforts to develop broad- and drug treatment in California (Gerstein based community support, and sustainable et al., 1997). Economic savings to the funding (see appendix B for a more thor- California taxpayer both during and after ough discussion of these elements). treatment were estimated to be worth $10,000 per client, yielding a 1:7 cost- Continuum of drug treatment benefit ratio (the greatest share of the services benefits was found in crime reductions, with smaller savings in health care and Many policymakers, particularly legislators, welfare costs). The study also reported a oppose funding for drug treatment in cor- 68-percent reduction in drug selling and a rectional facilities, believing that the public 60-percent reduction in arrests resulting wants offenders punished rather than from drug treatment. In the second study,

126 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

RAND researchers developed an econom- In a critique of the original 1994 RAND ic model to estimate the relative cost- model, the Office of National Drug Control effectiveness of four cocaine-control Policy (ONDCP)-funded National Research programs: three “supply control” pro- Council reviewers argued that RAND’s grams (source-country control, interdic- conclusions were “based on problematic tion, and domestic enforcement) and a estimates of treatment effectiveness “demand control” program treating heavy drawn from uncontrolled observational users (Rydell and Everingham, 1994). studies” (National Research Council, Results indicated that for every dollar 2001, p. 225), and that the assumptions spent on drug treatment, $7 would have and economic modeling procedures used to be spent on incarceration and $25 on by RAND researchers were flawed in interdiction to achieve the same degree of other ways and therefore not useful for reduction in cocaine use (cost savings policymaking (Manski, Pepper, and would vary depending on factors such as Thomas, 1999). Caulkins, Chiesa, and treatment setting, length of time in treat- Everingham (2000) offered an extensive ment, and degree of treatment structure). response to the latter set of criticisms, Further, they argued that even when only showing that modifying the model to in- looking at modest in-treatment effects corporate the reviewers’ suggested (assuming 0-percent post-treatment effec- changes did not in fact materially alter tiveness through abstinence), cost savings the conclusions. As for the concern that for treatment exceeded those that would RAND’s characterization of treatment was be achieved through incarceration and overly optimistic, the evidence is ambigu- interdiction. This study was later updated ous. Indeed, some have criticized their to distinguish among a variety of types of model for being overly pessimistic (Caul- Although billions domestic enforcement and used a more kins, Chiesa, and Everingham, 2000). optimistic assumption concerning how Clearly, future research in this area is of dollars are responsive consumption is to enforcement- needed to clarify and tighten assumptions, spent each year induced price increases. Caulkins and his improve methodologies, and incorporate colleagues (Caulkins et al., 1997) conclud- more carefully controlled data from drug to support drug ed that treatment outcome studies (for more com- abuse treatment, prehensive information on the economics treatment is more cost-effective than of drug treatment services, see Cart- the large majority either enforcement approach [con- wright, 2001). of offenders do ventional or federal] at reducing both cocaine consumption and cocaine Inmate participation in treatment. Al- not receive drug spending. Treatment is solidly but not though billions of dollars are spent each treatment services exceptionally more cost-effective year to support drug abuse treatment, the than the federal-level enforcement large majority of offenders do not receive of any kind. programs at reducing consumption; drug treatment services of any kind. it has a 1.6:1 edge over conventional ONDCP spent approximately 20 percent enforcement and close to a 3:1 ad- of its $18.4 billion budget on drug treat- vantage over mandatory minimums. ment in fiscal year 2000 (ONDCP, 2000). (p. 51) More than half of such Federal funding was allocated to support State block They also found treatment to be “enor- grants. In addition to these amounts, mously more cost-effective (on the order State, county, and local governments (as of 70 times more cost-effective) at reduc- well as private funding sources) con- ing spending on cocaine” (p. 51) than tributed significant funds to drug treat- enforcement strategies that shrink con- ment efforts (U.S. General Accounting sumption primarily by driving up prices.

127 SPECIAL REPORT / JULY 03

Office, 1998). However, it is unclear what that drug treatment for offenders is proportion of the total available funds have effective. been targeted toward treatment of drug- using offenders. Regarding offender treat- Effectiveness of drug treatment for ment services, 83 percent of State and 73 offenders. Drug treatment for offenders percent of Federal prisoners reported past is being taken seriously by even the drug use in 1997, with 57 percent of State strongest advocates of incarceration for and 45 percent of Federal prisoners re- drug possession and use. Flooded court porting use in the month prior to their dockets, overcrowded prisons, and high offense (Mumola, 1999). However, report- recidivism rates of drug-using offenders ed participation in drug treatment in have convinced even those most skeptical Federal and State prisons is minimal in of treatment that it is impossible to incar- most cases. The 1997 Survey of Inmates cerate all the illegal drug users in the in State and Federal Correctional Facilities Nation. Scientific research on the brain is (Mumola, 1999) reported decreases in the offering clues into the nature of drug percentage of both State and Federal independence, leading most to agree with mates undergoing drug treatment. It is the conclusions of NIDA: “Prolonged use important to note that these trends are dif- of these drugs eventually changes the ficult to interpret without knowing more brain in fundamental and long-lasting about the increases in actual drug treat- ways, explaining why people cannot just ment capacity within State and Federal quit on their own, why treatment is essen- systems relative to inmate population tial” (Leshner, 2001). This view has also increases. been adopted by ONDCP, which states that “chronic, hardcore drug use is a dis- Local jails have fared about the same as ease, and anyone suffering from a disease Federal and State facilities. According to needs treatment” (ONDCP, 2001, p. 1). BJS’s 1998 Annual Survey of Jails (Wilson, Recognizing both the public safety bene- 2000), 66 percent of jail inmates were fits from breaking the cycle of drug use actively involved with drugs prior to their and crime as well as the potential safety current incarceration, and 74 percent re- risks of allowing drug-addicted criminals ported past drug involvement. Almost on the streets (Taxman, 2000), ONDCP’s three-quarters of local jails (90 percent in National Drug Control Strategy advocates larger jurisdictions) state that they provide a two-pronged approach to the problem: substance abuse treatment or programs punish criminals for their behaviors while for their inmates. However, 64 percent of mandating sanctions-based drug treat- that total are self-help programs; only 12 ment. However, questions remain as to percent of jail jurisdictions (mostly large which treatment programs are effective, jurisdictions) provided detoxification, coun- and for which drug users. seling, and education in addition to self- help programs. There is a substantial Three major cautions must be noted when difference between what jails say they reviewing the mostly quasi-experimental provide and what inmates report. The per- drug treatment outcome studies. First, centage of inmates who actually reported many studies rely on client self-reports, that they participated in substance abuse which are least valid for higher penalty treatment or programs since their admis- drugs, recent use, and those involved with sion to jail was estimated at 10 percent the criminal justice system (for further lim- (19 percent for those who had used drugs itations on the validity of self-report drug at the time of the offense). Despite these use, see Hser, 1997). A second and relat- low rates of participation in treatment, a ed problem is selection bias. Both the broad range of studies continues to show selection of those who elect to enter

128 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

treatment (and are thus perhaps viewed conformity such as employment or mar- as being more motivated to remain in riage (i.e., if an individual is employed, he treatment) and program terminations may or she has a greater incentive to adhere to leave only those participants who are treatment in order to not get fired; or, if most ready and capable of succeeding married, an individual may have a greater when released into the community. Such incentive to do well to prevent a spouse “weeding out” of participants who may from leaving); use of credible outcome be more likely to fail than succeed could measures; identifying appropriate followup lead researchers to incorrectly conclude periods; linking retention to outcomes; greater treatment effects than would be and identifying treatment components seen through more careful attention to that promote recovery. treatment design with randomized assignment to treatment groups (U.S. General Treatment settings. Overall, the size and Overall, the size Accounting Office, 1998; Pelissier et al., consistency of treatment effects across 2000). Third, making a generalization many reasonably good studies tend to and consistency of based on the issues just noted, a recent lend credibility to consistent claims of treatment effects National Research Council report (2001) treatment effectiveness. The following notes that very few randomized controlled section reviews a sample of recent out- across many research studies have been conducted come evaluations for offenders in a variety reasonably good on drug treatment outcomes, thereby of treatment settings, moving from more casting some doubt on the cause of some restrictive to less restrictive settings. Out- studies tend to outcomes. come measures that are typically used to lend credibility to gauge drug treatment effectiveness in Despite these challenges, however, some such studies include reduced frequency consistent claims researchers are paying more attention to or amount of drug used; relapse time or of treatment improving the scientific rigor of these eval- length of abstinence period; crime, arrest, uations to achieve the greatest accuracy and conviction rates; and maintenance of effectiveness. possible. The National Research Council parole or probation status. report summarized five recent treatment evaluation studies that were, in the com- Prison-based therapeutic communities. mittee’s view, “the methodological state Therapeutic communities (TCs) are gener- of the art in drug treatment research” ally intensive, long-term, self-help-based, (2001, p. 227). The studies, none of which highly structured residential treatment included drug-using offenders, were noted programs for chronic, hardcore drug for their random treatment assignment, users. Although still rooted in a self-help treatment fidelity, measurement reliability approach, prison-based TCs are more and validity, and continuous rather than likely than community-based TCs to have dichotomous outcome measurements. professionally trained staff, with inmates The committee also discussed in some being given a reasonable level of power detail the ways in which drug treatment and rewards without too much program outcome studies could be strengthened control (Wexler, 1995; see also ONDCP, through improved methodological and sta- 1996). Three TC approaches will be re- tistical rigor. In a separate review (in the viewed below. same volume) of drug treatment in the criminal justice system, Covington (2001) Wexler and colleagues have reported on suggested guidelines for evaluating crimi- the effectiveness of the Stay ‘N Out TC nal justice system-based drug treatment. program used by the Department of These guidelines included controlling for Corrections in New York State (Wexler, self-selection bias; controlling for stake in Falkin, and Lipton, 1990; Wexler et al., 1992). TC inmates were compared with

129 SPECIAL REPORT / JULY 03

inmates assigned to milieu therapy, coun- 29-percent reincarceration rate compared seling, or a no-treatment group (composed with 74 percent for program dropouts. of those who volunteered for TC treat- Similarly, although 54 percent of program ment but were placed on a waiting list). graduates were not convicted of any Comparing male post-treatment arrest crime (including minor offenses), only 25 rates, the groups receiving counseling and percent of the comparison group and 15 no treatment were equally likely to be percent of program dropouts were not arrested (40 and 41 percent, respectively), convicted of a crime. Again, these findings while those receiving milieu therapy had should be viewed with some caution an arrest rate of 35 percent, and those given that participants who remained in receiving TC group treatment had an treatment were acknowledged to have arrest rate of 27 percent. One significant been more highly motivated to succeed flaw in this finding is the researchers’ fail- than program dropouts. It is also impossi- ure to account for other background vari- ble to separate out the effects of the 6 ables, causing some to question the months of community followup treatment strength of the treatment effect (Pelissier (Pelissier et al., 2000). The second study et al., 2000). However, time-in-treatment (Field, 1989) found that approximately 75 effects were also noted that showed a percent of program completers were not strong positive relationship between the reincarcerated, compared with 37 percent number of months in the TC program in the comparison group. In contrast, only and the percentage of inmates who were 15 percent of participants who dropped successfully discharged from parole. out of treatment after less than 2 months Specifically, the percentage of male TC in the program were not reincarcerated inmates who had successful parole dis- during the 3-year followup. charges grew from 49 percent for those in treatment for less than 3 months to 58 A major concern of this and similar studies percent for those in treatment for 3 to 6 is the high dropout rates from voluntary months. Positive rates further increased to drug treatment programs. For example, 62 percent when inmates participated in a Field (1992) highlighted that, of 220 volun- TC from 6 to 9 months and to 77 percent teer inmates who had been admitted to for those in a TC from 9 to 12 months. Cornerstone over a 2-year period, 65 with- Those who eventually failed on parole drew after spending 1 to 2 days in the were still able to stay drug and crime free program, 58 withdrew after spending for significantly longer periods than the between 2 to 6 months in the program, comparison groups. 43 withdrew after spending at least 6 months in the program, and 43 graduated. Field (1985, 1989) conducted two evalua- Simpson and colleagues (1997) have esti- tions of the Cornerstone Program, a TC for mated that, on average, only 50 percent of alcohol- and drug-dependent inmates in all addicts who voluntarily enter treatment Oregon’s correctional system that also actually complete the recommended treat- required at least 6 months of followup ment course. High dropout rates tend to treatment in the community. Participants confuse conclusions about treatment out- had to be granted minimum security sta- comes because those who remain in tus by the prison superintendent. Treat- treatment could be arguably more motivat- ment clients had, on average, about 12 ed to remain drug and crime free than prior arrests, 6 prior convictions, and 6 those who drop out. As has been noted years of adult incarceration. In the first earlier, however, offenders who are given 3-year followup study (1985), program graduated sanctions as a form of coerced graduates were found to have had a treatment generally stay in treatment

130 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

longer, complete treatment programs, and the comparison group. However, those report less drug use while in treatment who completed Crest did much better, programs than those in voluntary treat- and those who completed Crest plus ment (Simpson et al., 1997; Hubbard et aftercare were the least likely to have a al., 1989). new arrest. Specifically, less than one-third of clients with aftercare had a new arrest, The Key-Crest program is a corrections- compared with more than two-thirds of based, three-stage treatment model pro- the comparison group (Martin et al., 1999). gram that operates within Delaware’s correctional system. The first stage, the Long-term residential treatment. Prison- Key, is modeled on the Stay ‘N Out pro- based long-term residential treatment is gram and includes a 12-month intensive generally considered to last between 6 to residential TC that is based in the institu- 12 months. Participants often live together tion but segregated from the rest of the in units separated from the regular inmate Prison-based inmates. The second stage, the Crest population. These units are specifically residential Outreach Center, is a transitional TC in designed to focus on drug treatment. The which inmates work during the day and degree of structure can vary, but generally treatment is return to a community-based, more tradi- a professional drug treatment staff coordi- generally more tional TC environment during their non- nates all programs and services. Com- working hours. In the third or aftercare pared with TCs, prison-based residential likely than stage, clients have completed work re- treatment is generally more likely to therapeutic lease and are now on parole or other include professional therapeutic interven- supervision. Intervention at this stage usu- tions using standard treatment approach- communities to ally involves group or individual counseling es. For example, the Bureau of Prisons include as well as the opportunity to return to the includes programming on criminal life- work-release TC for booster sessions. style confrontation, cognitive and interper- professional While earlier studies (Martin, Butzin, and sonal skill building, and relapse prevention therapeutic Inciardi, 1995; Inciardi et al., 1997) demon- (Pelissier et al., 2000). Inmate-led self-help strated short-term (1-year) benefits of this approaches are not present in such facili- interventions TC treatment continuum, many of the pos- ties. The following discussion will present using standard itive improvements between the second an evaluation of long-term residential treat- and third stage clients appeared to disap- ment, as well as one specific evaluation treatment pear in 3-year followup studies (Martin project. approaches. et al., 1999). However, when less conser- vative analytical models were applied (the From 1990 to 1993, the National Institute new analysis examined Crest dropouts, on Drug Abuse funded the Drug Abuse Crest completers, and Crest completers Treatment Outcome Study (DATOS), with aftercare), significant findings which included 96 programs in 11 cities. emerged. When compared with the com- Positive outcomes were reported in multi- parison group, Crest dropouts were more ple treatment modalities, including long- than three times as likely to be drug free term residential treatment (Simpson et al., (as measured by initial self-reports and 1997). DATOS found that individuals in subsequent urinalysis); Crest completers long-term residential treatment reduced were more than five times as likely to be weekly or more frequent use of cocaine drug free; and Crest completers with after- from 66 percent in the year prior to treat- care were seven times more likely to be ment to 22 percent in the year following drug free. Rearrests on a new charge treatment (see exhibit 1). This same group showed a similar pattern, with Crest drop- reported a 26-percent drop (from 41 per- outs having the same rate of rearrests as cent down to 16 percent) in predatory illegal activity during that same time period

131 SPECIAL REPORT / JULY 03

(Fletcher, Tims, and Brown, 1997). Similar- phase, inmates were required to partici- ly dramatic reductions in self-reported pate in community transitional services in cocaine use were also found for short- which they received individual, group, term residential treatment. and/or family counseling from community- based drug treatment providers. Three- Using one of the most methodologically year followup findings indicated that men rigorous research designs to date, the and women who were motivated to Federal Bureau of Prisons (BOP) recently change were more likely to enter and conducted a 3-year, 20-site evaluation of complete treatment. Findings on both its residential drug treatment programs recidivism and post-treatment drug use (Pelissier et al., 2000). During the three- were significant for men but not for phase Treating Inmates’ Addiction to women.4 Specifically, men who entered Drugs (TRIAD) Drug Treatment Evaluation and completed in-prison residential treat- Project, more than 1,000 inmates first ment were 16 percent less likely to recidi- voluntarily participated in a 9- or 12-month vate when compared with untreated residential treatment program. Treatment inmates at 3-year postrelease followup. group results were compared with a true In addition, participants who entered and comparison group as well as a control completed treatment were 15 percent group, neither of whom received any drug less likely to use drugs than untreated treatment services. A second phase re- inmates within 3 years after release. quired inmates to continue drug abuse These findings are particularly significant booster sessions (including relapse pre- because the selection process actually vention and review of treatment tech- attracted riskier offenders into the treat- niques) for 1 year following their return to ment programs. In addition, this study the general community. During the final carefully addressed the issue of selection bias by comparing results using two different bias correction methods. Exhibit 1. Self-reported cocaine use among addicts participating in treatment Day reporting centers. As noted previously, many offenders are serving time be- Before treatment cause of nonviolent drug convictions. 1 year after treatment To deal with prison overcrowding and Percentage of DATOS sample the prohibitive costs associated with incarceration-based treatment programs, 80 some correctional facilities have devel- 66 67 oped day reporting centers (DRCs). DRCs 60 are a form of intermediate sanction in which offenders attend highly structured, 42 42 40 nonresidential programs where a variety of services and supervision are provided. 22 21 22 First introduced in the United States in 20 18 1986, DRCs can be operated by a wide range of public, government, and private 0 agencies, such as residential community Long-term Short-term Outpatient Outpatient resident resident methadone drug free corrections centers, work release pro- treatment grams, jails, TASCs, and treatment pro- Type of treatment grams (Parent, 1990; McBride and VanderWaal, 1997). Services such as drug Source: Chart reproduced from Taxman (1998). treatment and education, GED courses, English as a Second Language and life

132 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

skills are often supervised by both correc- settings, services, eligibility criteria, moni- tions and case management personnel. toring procedures, levels of supervision, A DRC has three primary goals: enhanced and termination policies (Diggs and Pieper, supervision and decreased liberty for 1994). While some studies have shown offenders, treatment of offender prob- initial evidence of cost savings (Craddock, lems, and reduced crowding of incarcera- 2000) and lower rearrest rates (Diggs and tion facilities (Parent, 1990). The concept Pieper, 1994; McBride and VanderWaal, has been adapted in a number of ways, 1997), evidence of program effectiveness including: was not as great in programs that lasted 12 months or longer5 (Marciniak, 1999). ■ Providing enhanced treatment and Marciniak (2000) found high rates of pro- supervision to probationers or sen- gram termination for drug violations and tenced offenders not on probation. rearrests. Several authors (Blomberg and Lucken, 1994; Marciniak, 1999; Tonry, ■ Monitoring inmates on early release 1990, 1997) have also expressed concerns from jail or prison. of “net widening” since many offenders who would otherwise be sentenced to ■ Monitoring arrested persons prior to probation are placed in DRCs where they trial. are watched more closely and are there- ■ As a halfway-out step for inmates who fore more likely to be rearrested. Given have shown progress in community- the relatively recent emergence of this based corrections or work release form of intermediate sanctioning, future centers. studies should focus on success indicators such as program completion, drug ■ As a halfway-in step for offenders who use, rearrests, and cost-effectiveness, par- have violated their probation or parole ticularly in longer term programs. Program (Curtin, 1990, as cited in Diggs and success indicators should be based on Pieper, 1994). comparisons with offenders who would have been incarcerated as opposed to These programs are probably most appro- those traditionally found in probation to priate for nonviolent offenders whose be- avoid a net-widening bias (Diggs and haviors have not been improved through Pieper, 1994). probation and/or who need greater structure and treatment services than could be Outpatient and intensive outpatient treat- provided in a less restrictive setting. While ment. Taxman (1998) notes that the loca- attending the center, participants are often tion of drug treatment does not always required to submit to random drug testing relate to the intensity of services provided and participate in counseling, education, to the client. Instead, the number of serv- and vocational placement assistance. ice hours is often a better indicator. As Graduated sanctions are applied when par- such, community-based outpatient and ticipants are found to have violated the intensive outpatient treatment services terms of their sentence. are often used as a transition from TCs and other more intensive corrections- Relatively few studies have been conduct- based services. Such services are pared to assess predictors of program com- ticularly important to drug courts, who pletion or termination in DRCs. Studies primarily use treatment alternatives within which have been conducted are difficult the community. The setting is generally to compare due to the wide variability of less important than the quality and quantity of services provided to clients, although

133 SPECIAL REPORT / JULY 03

the organization providing the services Based on a comprehensive review of clini- must be supportive of delivering interven- cal and health services research on drug tions to correctional populations (Pogrebin, abuse, ONDCP (1996) made the following 1978). The DATOS study introduced in the recommendations regarding critical ele- previous section (regarding long-term resi- ments for successful treatment in any set- dential treatment) also included positive ting (e.g. prison based, residential, or outcomes for outpatient drug-free treat- outpatient): ment: self-reported cocaine use dropped from 42 percent before treatment to 18 ■ Complete and ongoing assessment of percent at 1-year followup (see exhibit 1). the client.

Treatment intervention approaches. ■ A comprehensive range of services, The previous section reviewed outcome including pharmacological treatment (if studies on a variety of drug treatment set- necessary), counseling (either individual tings, based on a range of restrictiveness. or group, in either structured or unstruc- Each of these settings often includes such tured settings), and HIV-risk reduction intervention approaches as life-skills train- education. ing, group and individual counseling, re- ■ lapse prevention training, and educational A continuum of treatment interventions. and vocational skills training. In addition, a ■ Case management and monitoring to variety of theoretical models influence the engage clients in services of appropriate content and approach to such interven- intensity. tions. It is beyond the scope of this paper to review these approaches and theories. ■ Provision and integration of continuing As mentioned earlier, however, NIDA has social supports. conducted a number of large-scale research evaluations on a variety of inter- NIDA came to many similar conclusions ventions (e.g. DARP, TOPS, DATOS), and in their research-based guide, Principles readers are referred to those studies to of Drug Addiction Treatment (NIDA, review intervention effectiveness. In addi- 1999). This guide also reviews scientifical- tion, NIDA is currently conducting con- ly based approaches to drug treatment trolled, multisite tests of emerging and makes recommendations. A full listing science-based drug abuse treatments of the NIDA recommendations is found in such as the use of buprenorphine/nalox- appendix C. one treatments for detoxifying opiate- It is important dependent patients and incorporating In addition to the recommendations and motivational enhancement therapy into principles listed by ONDCP and NIDA, it is to recognize standard treatments (Mathias, 2001). important to recognize the importance of the importance Motivational enhancements offer absti- matching the drug-using offender with nent clients a chance to win small prizes the appropriate treatment. This simple of matching the such as candy bars, Walkmans, or gift concept is, at times, especially difficult to drug-using certificates to local restaurants by testing employ in jurisdictions that may lack ade- negative for various illicit drugs. As the quate resources to provide a full continu- offender with number of abstinent weeks increases, um of services. Essentially, treatment the appropriate so do the number and value of the incen- matching recognizes that no single treat- tives. It is anticipated that such evalua- ment is universally applicable. Levels of treatment. tions will provide preliminary evidence of restriction and supervision, treatment effectiveness and efficacy so that knowl- modalities, and psychopharmacological edge about treatment effectiveness can treatment options (such as methadone) be improved. must be assessed on a case-by-case

134 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

basis. The ramifications of this issue Aftercare include the need for training system personnel on treatment continuum issues, Aftercare (or continuing care) is defined realistic expectations by both treatment as “a set of supportive and therapeutic and criminal justice systems regarding the activities designed to prevent relapse and potential impacts of available services, and maintain behavioral changes achieved in the potential need to educate the commu- previous treatment stages” (Fortney et al., nity on what can be expected from avail- 1998, as cited in Inciardi et al., 2001). The able resources. aftercare phase of the treatment continuum is often neglected for drug-using Gender differences in treatment. offenders. As noted previously, most drug- Pelissier and her colleagues (2000) com- using offenders have high relapse rates pleted a comprehensive review of litera- and therefore require extended periods of ture on gender differences among treatment exposure and ongoing support substance abusers (for supporting litera- to achieve and maintain sobriety. In addi- ture documentation of this summary para- tion, most treatment graduates are ill graph, see Pelissier et al., 2000). Although equipped to integrate back into their old much of the current increase in the num- neighborhoods (Berman and Anderson, ber of incarcerated women is linked to 1999). For these reasons, providing after- substance abuse (Kassebaum, 1999), care as a followup to more restrictive few studies have examined gender differ- treatment may improve treatment effec- ences among substance-abusing inmates. tiveness. Cross-systems case manage- Studies primarily on nonoffending sub- ment and collaboration are critical at this stance abusers show that women general- phase in the treatment process to mainly have different social, psychological, and tain an integrated continuum of care for economic circumstances; different initia- clients as they transition back into the tion and drug use patterns; and different community. criminal histories than men. Most discussions of treatment approaches for women Martin et al. (1999) recommend that treat- include a strong focus on ancillary services ment interventions at this stage include such as health care, child care, and female continued monitoring by previously in- treatment staff. Therapeutic recommen- volved treatment counselors (such as TC dations include a focus on relationship counselors). Interventions at this stage issues, support, skill building, and identifi- could include regular outpatient counsel- cation of strengths as opposed to the con- ing, support groups such as Alcoholics frontation strategies that are common for Anonymous, group therapy, and family men (for a summary of treatment effec- therapy sessions. In addition, Tauber tiveness studies for men and women, see (1994) calls for educational opportunities, Landry, 1997). Despite these differences, job training and placement, and health and however, few treatment programs focus housing assistance. heavily on women’s issues, particularly in Several studies (Lash, 1998; McKay et al., correctional facilities. Not surprisingly, few 1998; Rychtarik et al., 1992) with noncor- studies have looked at outcomes of treat- rectional populations have suggested that ment programs designed specifically for improved treatment outcomes can result women (Landry, 1997), in part due to the from aftercare (most of these studies are relatively small numbers of female drug correlational in nature). In such settings, it treatment participants (Moras, 1998). is possible that selection bias is present,

135 SPECIAL REPORT / JULY 03

since motivated clients may make better control of self-selection bias, and careful use of aftercare services (Inciardi et al., analysis of other intervening variables. 2001). However, recent studies with corrections-based treatment followed by Summary aftercare have also shown preliminary indications of success (DeLeon et al., 2000; Current research suggests that successful Wexler et al., 1999). Offenders in the programmatic efforts to intervene in the California-based Amity Right Turn Project drugs-crime relationship are based on a received voluntary TC treatment followed continuum of integrated services stretch- by community-based aftercare program- ing from assessment through aftercare. ming. No-treatment control groups were Although research has evaluated the compared with TC dropouts, TC gradu- various components that might be most ates, and aftercare completers after 12, beneficial for inclusion in a successfully in- 24, and 36 months. Although recidivism tegrated system, we know of no studies rates increased for all groups as time that have attempted to measure the suc- increased, those who completed both the cess or lack of impact of such integrated treatment and aftercare phases had the approaches. lowest rearrest rates. Inciardi and colleagues (2001; see also Martin et al., 1999) conducted a similar aftercare study Suggestions for future with Key-Crest participants. Voluntary research clients were randomly assigned and pur- posively sampled across four groups: a In any field of scientific inquiry, one of the no-treatment comparison group, treat- easiest things to do is to call for more ment dropouts, treatment graduates, and research. Not surprisingly, that is exactly treatment graduates with aftercare. Re- the most appropriate thing to do with searchers conducted followup interviews regard to the drugs-crime relationship. at 18 and 42 months and collected infor- New conceptual and mathematical models mation on drug use (interview and urine have emerged recently in the social sci- screen) and rearrest rates (interview ences that will allow a fresh perspective compared with official prison records). on many of the questions that have been Eighteen-month followups indicated that addressed in the past and provide a new treatment dropouts and graduates were baseline for the 21st century. Human cul- twice as likely than the comparison group tures change, some fairly rapidly, and even to be drug free, and treatment graduates a brief review of the past 25 years in the with aftercare were three times more like- United States with regard to drugs and ly to be drug free. Preliminary data from crime would indicate that ours has changed the 42-month followup were even more dramatically. In the area of the drugs-crime impressive. Although only 25 percent of relationship, one illustration of this change the comparison group were arrest free, is the apparent reduction in the violence more than half of the graduates with after- associated with cocaine/crack distribution. care remained arrest free. Similarly, 25 Such changes require fresh examinations percent of comparison cases remained of previously collected data and more rig- drug free, compared with 36 percent of orous evaluations of current programs and the treatment-with-aftercare group. Such policies. Although there are certainly many studies could be further strengthened areas of potential further inquiry, the fol- with larger sample sizes, evaluating suit- lowing areas are suggested: ability of clients for treatment, more careful

136 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Using secondary data analyses could obtain parallel geocoding data for to provide a new empirical the ADAM dataset, the number of ques- baseline for understanding the tions that could be addressed about the drugs-crime relationship drugs-crime relationship would expand geometrically. We need to integrate The Federal Government, other agencies, advances in analytical models with ad- and universities have collected enormous vances in neurobiology, personality, family amounts of data that are directly relevant systems, and peer influence studies as to many key drugs-crime questions. These well as include broader contextual vari- data include the National Household Sur- ables (including ecosystems theory, social vey on Drug Abuse (NHSDA), the Monitor- capital, economic opportunity, drug prices ing the Future (MTF) study, the Arrestee and market variables, drug laws/policy, Drug Abuse Monitoring Program, and and geographical data). the Treatment Episode Data Set (TEDS). These data could be used to provide a new baseline of knowledge about certain Consider using computer simu- statistical elements of the drugs-crime lation modeling to examine relationship across the lifespan and in key research questions many different segments of the popula- Some of the etiological ideas that re- tion. In addition, these data could be used searchers are examining may be applica- to demythologize many policy and popular ble to computer modeling in the future. conclusions about the drugs-crime rela- For example, it might be useful, in a simu- tionship. For example, data from some of lated model, to manipulate reductions in these systems call into question some supply, increases in price, changes in beliefs about the cocaine-violence connec- policy (such as treatment on demand and/ tion as well as suggest that the criminal or marijuana decriminalization/medicaliza- justice system may primarily direct mari- tion) to examine how such issues would juana users to the treatment system to affect drug use, crime, and their interrela- the exclusion of other drug users. tionship. Although the data entered in a simulation would be based on the types Further studying the nature and of research previously noted, and the pit- complexity of the drugs-crime falls and complexities of undertaking this relationship using the latest approach have not been thought out, it interdisciplinary conceptual may be time for the drugs-crime field to and analytical models begin considering the use of computer simulation technology to address the criti- Many of the interventions that have been cal issues facing many communities. applied to breaking the drugs-crime cycle have involved a fairly narrow focus on drug treatment and have not sufficiently Evaluating State changes in recognized the complex origins of both drug policy to examine different behaviors. Further, there is increasing attempts to address the drugs- evidence of a need to include multilevel crime relationship at a macro variables in order to understand how but yet subnational level crime and drugs are connected. This was Throughout this document, it has been not possible previously due to the statisti- noted that while there has been relatively cal precision needed. In addition, the 2000 little modification of drug law and policies Census and geocoding provide an oppor- at the national level, there has been con- tunity to add another data dimension to siderable legislative action in many States drugs-crime analyses. For example, if we

137 SPECIAL REPORT / JULY 03

and communities. Model State drug laws criteria and options. It is critical that we have been proposed. Many States are evaluate such changes early so that les- moving towards allowing medical marijua- sons learned from them may be used na, and many States have decriminalized strategically to change later interventions. marijuana possession (or at least removed incarceration penalties for the first marijua- Considering the need to na possession conviction). Other States establish research field stations are changing club-drug laws to increase in high-risk communities scheduling and penalties. In addition, there are significant differences between One idea that has been discussed episo- States (and communities) regarding treat- dically in the drug field for the past two ment availability and budgets. For many decades involves the use of a research years, there have been calls for interna- field station approach. Although there tional research comparing the impact of have been some attempts to undertake different national drug policies. However, such an endeavor, these efforts generally given significant differences between have been limited in time and/or place. national cultures, these comparisons are Existing data (combined with geocoding) difficult. Variance in State law and policy could be used to identify communities provides a more readily available opportu- with high rates of drug use and crime. nity to examine variance between entities Theoretically based multivariate research (the 50 States) with differing laws and poli- projects could then be conducted in these cies. These changes suggest a number of targeted communities from a qualitative, possible research areas. For example, on-the-ground perspective. Such an comparing differences in marijuana use approach might permit researchers to (or drug use in general), perceptions of understand some of the changes in risk, and peer disapproval in States that violence associated with crack distribu- have medical marijuana and/or marijuana tion that seem to have occurred in decriminalization with States with high- recent years. deterrence prohibition policies could provide an excellent foundation for evaluating Examining the relationship changing drugs-crime policies. between particular enforcement strategies and drug markets Evaluating attempted inter- Recent modifications to the ADAM study ventions in the drugs-crime (including asking subjects about access to cycle for net widening drugs and conditions that they perceive as As noted, the increasing availability of affecting access) provide the possibility of drug courts and other mandatory treat- empirically modeling the effects of specif- ment programs may encourage law ic enforcement strategies on specific drug enforcement to intervene earlier and more markets (cocaine, crack, and heroin) and formally in the lives of individual drug drug prices. In particular, researchers may users. This change in strategy and tactics be able to evaluate a particular enforce- could begin a formal criminal justice label- ment strategy’s impact on drug market ing process that may exacerbate, rather location (moving it indoors or to more than ameliorate, the relationship between urban settings), the number of dealers typ- drug use and crime. It may also result in ically used, the amount of time searching changing definitions of law violation and for drugs, or the price of that drug (from increase the number of those arrested STRIDE [System to Retrieve Information and incarcerated due to new placement from Drug Evidence] or other sources) and

138 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

more. This could provide researchers with and feasibility? How does drug moni- very important information about how drug toring alone compare with more com- markets operate in local areas in response prehensive systems and treatment to enforcement strategies. interventions in terms of outcomes such as drug use and recidivism?

Comprehensively evaluating ■ What assessment protocols can most current programs designed accurately be used to place offenders in to intervene in the drugs- the safest, least restrictive, and most crime cycle effective treatment settings? Many programs exist that attempt to intervene in the drugs-crime cycle from the ■ What level and intensity of drug treat- juvenile to adult level. Although there have ment services are most appropriate for been significant attempts to evaluate which offender types and settings? these programs, most of these efforts ■ What forms and mixtures of the have been descriptive or have used fairly reviewed programmatic interventions simple analytical designs (often quasi- (e.g. graduated sanctions, supervision/ experimental). What is needed are large- monitoring, various drug treatment scale, carefully controlled studies that services and settings, aftercare, etc.) focus on long-term program outcomes predict program completion or termina- using multiple indicators of success and tion (or other specific outcomes) with that identify program elements related to which populations and under which outcomes. These evaluations should focus conditions? on what the literature might call best-case program models that generally involve comprehensive assessment, needed serv- Using interdisciplinary teams ice provision based on that assessment, to conduct research on the case management, graduated sanctions, drugs-crime relationship and aftercare. Most outcome studies A review of the literature shows that indi- examine such factors as rearrest rates or viduals from a variety of disciplines have drug relapses. Additional successful out- examined the drugs-crime relationship. come measures might include such non- Each discipline has approached the rela- crime-related outcomes as payment of tionship from its particular perspective, child support, family formation and stabili- and each discipline likely has an important ty, employment stability, and residential and unique perspective on understanding stability. In addition, it is important to the relationship. Some of the critical re- examine how these programs vary in their views of conceptualization, methodology, impact by gender, ethnicity, and age as and conclusions in drugs-crime research well as provision context (prison to com- are often based on particular disciplinary munity). Finally, it is crucial to examine perspectives. To broaden the perspectives program costs relative to the cost of incar- of these disciplines, the types of research ceration and the cost of no intervention. issues/questions that have been proposed Although specific recommendations for require the efforts of an interdisciplinary further research were included at the end team. If there is to be clear definition, of each program intervention section in development, and operationalization of this chapter, the following research ques- treatment program elements, treatment tions are of high priority: providers must provide input. Researchers ■ Which drug testing methods offer the trained in experimental or quasi-experimental best combination of accuracy, privacy, design are crucial in developing and carrying out the needed scientific designs.

139 SPECIAL REPORT / JULY 03

Social scientists (survey researchers, no significant differences between the coefficient geographers, and ethnographers) are for men and women. This lack of significance for needed if issues of gender, ethnicity, and women is likely a reflection of the smaller sample size for this population (Pelissier, 2001; personal other sociocultural and spatial characteris- communication). tics are to be included in the design and data interpretation. Given today’s strong 5. The issue of length of time in treatment as indica- social concern relative to cost-benefit out- tive of stronger gains in treatment was raised previ- comes, it is crucial to include economists ously in this paper. This issue is debated in the field. Marciniak (1999) argues that longer may be better on research teams. Drugs-crime research only up through 9 to 12 months; treatment deteriora- has clearly reached the stage where inter- tion may then begin. Other researchers argue that disciplinary research teams are required. this outcome needs more study.

Establishing interagency cooperation in funding research References An examination of the various governmen- Arrestee Drug Abuse Monitoring Program tal reports and our conversations with col- (2000). 1999 annual report on drug use leagues about this project suggest that among adult and juvenile arrestees (NCJ many different agencies focus on and 181426) [Online]. Available: www.adam- issue reports about the drugs-crime rela- nij.net/files/INTO.pdf. tionship. It appears that the authors of Alemagno, S.A. (2001). Women in jail: Is many of these reports are not aware of substance abuse treatment enough? the excellent research funded by other American Journal of Public Health, 91(5), agencies. Given the limited resources in 798–800. any given funding agency and the different research traditions in various agencies, Anglin, M., and Hser, Y. (1990). The integrated research will require significant treatment of drug abuse. In M. Tonry and interagency cooperation. Such cooperation J.Q. Wilson (eds.), Drugs and crime (pp. could make sufficient resources available 393–460). Chicago: University of Chicago to address the types of complex research Press. needed in drugs-crime analysis. Anglin, M., Longshore, D., and Turner, S. (1999). Treatment alternatives to street Notes crime: An evaluation of five programs. Criminal Justice & Behavior, 26(2), 1. For economists, the term “open market” has a very precise meaning. In this paper, however, we use 168–195. the term in a general sense to indicate low levels of government regulation. Anglin, M., and Maugh, T. (1992). Over- turning myths about coerced drug treat- 2. The companion papers to this work discuss the ment. California Psychologist, 19–22. psychopharmacological and economic components of the drugs-crime relationship. Anslinger, H., and Tompkins, W. (1953). 3. The focus will be primarily on adult intervention The traffic in narcotics. New York: Funk & strategies since other recent reports have completed Wagnalls. a comprehensive literature review and offered program guidelines focusing specifically on juveniles Apospori, E., Vega, W., Zimmerman, R., (McBride et al., 1999). Warheit, G., and Andres, G. (1995). A longitudinal study of the conditional effects of 4. Although Pelissier and her colleagues did not find a significant treatment effect on postrelease drug deviant behavior on drug use among three use and crime for women, further analyses indicated racial/ethnic groups of adolescents. In

140 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

H.B. Kaplan (ed.), Drugs, crime, and other Beck, A., and Mumola, C. (1999). deviant adaptations: Longitudinal studies. Prisoners in 1998 (BJS Bulletin, NCJ New York: Plenum Press. 175687) [Online]. Available: www.ojp. usdoj.gov/bjs/abstract/p98.htm. Arizona Supreme Court (1999). Drug treatment and the education fund [Online]. Beckett, J., and Johnson, H. (1995). Available: www.supreme.state.az.us/asd/ Encyclopedia of social work: Human dteftoday.pdf. development. In R.L. Edwards (ed.), Encyclopedia of social work (19th ed., vol. Austin, G., and Lettieri, D. (1976). Drugs 2, pp. 1385–1405). Washington, DC: and crime: The relationship of drug use National Association of Social Workers. and concomitant criminal behavior (DHHS Publication No. ADM 77–393). Rockville, Belenko, S.R. (1998). Research on drug MD: U.S. Department of Health and courts: A critical review. National Drug Human Services, National Institute on Court Institute Review, 1(1), 1–42. Drug Abuse. Belenko, S.R. (2000). Drugs and drug poli- Ballew, J.R., and Mink, G. (1996). Case cy in America. Westport, CT: Greenwood management in social work: Developing Press. the professional skills needed for work with multiproblem clients (2d ed.). Berman, G., and Anderson, D. (1999). Springfield, IL: Charles C. Thomas. Drugs, courts and neighborhoods: Com- munity reintegration and the Brooklyn Babor, T.F., del Boca, F.K., McLaney, M.A., treatment court [Online]. Available: Jacobi, B., Higgin-Biddle, J., and Hass, W. www.courtinnovation.org/pdf/dru_cour_ (1991). Just say Y.E.S.: Matching adoles- neigh.pdf. cents to appropriate interventions for alcohol and other drug-related problems. Billings, J.R. (2000). Community develop- Alcohol, Health & Research World, 15(1), ment: A critical review of approaches to 77–86. evaluation. Journal of Advanced Nursing, 31(2), 472–480. Bartollas, C. (2000). Juvenile delinquency (5th ed.). Boston: Allyn and Bacon. Birmingham, L., Gray, J., Mason, D., and Grubin, D. (2000). Mental illness at recep- Baum, F. (1997). Public health and civil tion into prison. Criminal Behaviour & society: Understanding and valuing the Mental Health, 10(2), 77–87. connection. Australian and New Zealand Journal of Public Health, 21(7), 673–675. Blomberg, T., and Lucken, K. (1994). Stacking the deck by piling up sanctions: Baum, F. (2000). Social capital, economic Is intermediate punishment destined to capital and power: Further issues for a fail? Howard Journal, 33(1), 62–80. public health agenda. Journal of Epidemi- ology & Community Health, 54(6), 409–410. Bokos, P., Mejta, C., Mickenberb, J., and Monks, R. (1992). Case management: An Bazemore, G., and Umbreit, M. (2001). alternative approach to working with intra- A comparison of four restorative confer- venous drug users. In R.S. Ashery (ed.), encing models (OJJDP Juvenile Justice Progress and issues in case manage- Bulletin, NCJ 184738) [Online]. Available: ment (NIDA Research Monograph 127, www.ncjrs.org/pdffiles1/ojjdp/184738.pdf. DHHS Publication No. ADM 92–1946, pp. 92–111). Rockville, MD: U.S. Department

141 SPECIAL REPORT / JULY 03

of Health and Human Services, National Bureau of Justice Assistance (1995). Institute on Drug Abuse. Treatment accountability for safer communities (Fact Sheet, FS 000044) [Online]. Boot, B.P., McGregor, I.S., and Hall, W. Available: www.ncjrs.org/pdffiles/tasc.pdf. (2000, May 20). MDMA (ecstasy) neuro- toxicity: Assessing and communicating Bureau of Justice Statistics. (1999). the risks. Lancet, 355(9217), 1818–1821. Federal criminal case processing, 1998, with trends 1982–98 (NCJ 169277) Boston Globe. (2000, November 11). [Online]. Available: www.ojp.usdoj.gov/ Editorial: Drug treatment worth investment. bjs/pub/pdf/fccp98.pdf.

Brady, K.T., and Randall, C.L. (1999). Burke, P. (1997). Policy-driven responses Gender differences in substance use to probation and parole violations. Wash- disorders. Psychiatric Clinics of North ington, DC: U.S. Department of Justice, America, 22(2), 241–252. National Institute of Corrections.

Brannigan, A. (1997). Self-control, social Bursik, R.J., Jr. (1999). The informal con- control and evolutionary psychology: trol of crime through neighborhood net- Towards an integrated perspective on works. Sociological Focus, 32(1), 85–89. crime. Canadian Journal of Criminology, 39(4), 403–431. Cartwright, W.S. (2001). The economics of drug treatment: Bibliography [Online]. Brownstein, H.H. (1991). The social con- Available: http://165.112.78.61/HSR/ struction of public policy: A case for par- other/EconBiblio.htm. ticipation by researchers. Sociological Practice Review, 2(2), 132–145. Caspi, A., Moffitt, T.E., Wright, B.R.E., and Silva, P.A. (1998). Early failure in the labor Brownstein, H.H. (1996). The rise and fall market: Childhood and adolescent predic- of a violent crime wave: Crack cocaine and tors of unemployment in the transition to the social construction of a crime problem. adulthood. American Sociological Review, Guilderland, NY: Harrow and Heston. 63(3), 424–451.

Brownstein, H.H. (2000). The social reality Crime Prevention Act of 2000, Proposition of violence and violent crime. Boston: 36, California (November 7). Allyn and Bacon. Caulkins, J.P., Chiesa, J., and Everingham, Brownstein, H.H., Spunt, B.J., Crimmins, S.M.S. (2000). Response to NRC assess- S.M., Goldstein, P.J., and Langley, S.C. ment of RAND’s controlling cocaine study. (1994). Changing patterns of lethal Santa Monica, CA: RAND. violence by women: A research note. Women and Criminal Justice, 5(2), Caulkins, J.P., Rydell, C.P., Schwabe, W., 99–118. and Chiesa, J.R. (1998). Are mandatory minimum drug sentences cost-effective? Bui, K.V.T., Ellickson, P.L., and Bell, R.M. Corrections Management Quarterly, 2(1), (2000). Cross-lagged relationships among 62–73. adolescent problem drug use, delinquent behavior, and emotional distress. Journal Caulkins, J.P., Rydell, C.P., Schwabe, W., of Drug Issues, 30(2), 283–303. and Chiesa, J.R. (1997). Mandatory minimum drug sentences: Throwing away the key or the taxpayers’ money? (MR–827–DPRC). Santa Monica, CA: RAND.

142 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Cherrington, E. (1920). The evolution Collison, M. (1996). In search of the high of prohibition in the United States of life: Drugs, crime, masculinities and con- America. Westerville, OH: American sumption. British Journal of Criminology, Press. 36(3), 428–444.

Clark, M.D. (1997). Strength-based prac- Cooper, C. (1997). 1997 drug court survey tice: The new paradigm. Corrections report: Executive summary. Washington, Today, 59(April), 110–111. DC: American University, Justice Pro- grams Office. Coker, A., Smith, P., McKeown, R., and King, M. (2000). Frequency and correlates Covington, J. (2001). Appendix E. Linking of intimate partner violence by type: treatment to punishment: An evaluation of Physical, sexual, and psychological batter- drug treatment in the criminal justice sys- ing. American Journal of Public Health, tem. In National Research Council (author) 90(4), 553–559. and Committee on Data and Research for Policy on Illegal Drugs: C.F. Manski, J.V. Coleman, J.S. (1988). Social capital in Pepper, and C.V. Petrie (eds.), Informing the creation of human capital. American America’s policy on illegal drugs: What we Journal of Sociology, 94(Supp.), S95–S120. don’t know keeps hurting us (pp. E1–E27). Washington, DC: National Academy Press. Collins, J. (1993). Drinking and violence: An individual offender focus. In S.E. Martin Cox, J.F., Landsberg, G., Paravati, M.P. (ed.), Alcohol and interpersonal violence: (1989). The essential components of a cri- Fostering multidisciplinary perspectives sis intervention program for local jails: The (Research Monograph 24). Rockville, MD: New York Local Forensic Suicide Preven- U.S. Department of Health and Human tion Crisis Service Model. Psychiatric Services, National Institute on Alcohol Quarterly, 60(2), 103–117. Abuse and Alcoholism. Craddock, A. (2000). Exploratory analysis Collins, J. (1990). Summary thoughts of client outcomes, costs, and benefits of about drugs and violence in the distribu- day reporting centers: Final report (NCJ tion of crack. In M. De La Rosa, E.Y. 182365). Washington, DC: U.S. Depart- Lambert, and B. Gropper (eds.), Drugs and ment of Justice, National Institute of violence: Causes, correlates, and conse- Justice. quences (NIDA Research Monograph 103, DHHS Publication No. ADM 90–1721, pp. Craddock, A., Collins, J., and Timrots, A. 265–275). Rockville, MD: U.S. Department (1994). Fact sheet: Drug-related crime of Health and Human Services, National (NCJ 149286). Washington, DC: U.S. Institute on Drug Abuse. Department of Justice, Bureau of Justice Statistics. Collins, J.J., Hubbard, R.L., Rachal, J.V., Cavanaugh, E.R., and Craddock, S.G. Curtin, E. (1990). Day reporting centers, a (1982). Criminal justice clients in drug promising alternative. IARCA Journal, 3, 8. treatment (NCJ 124130). Washington, DC: U.S. Department of Justice, National Curtis, R. (1999). The ethnographic Institute of Justice, and U.S. Department approach to studying drug crime. In of Health and Human Services, National National Institute of Justice, Looking at Institute on Drug Abuse. crime from the street level (NIJ Research Forum, NCJ 178260, pp.13–22) [Online]. Available: www.ncjrs.org/pdffiles1/nij/ 178260.pdf.

143 SPECIAL REPORT / JULY 03

Dawkins, M.P. (1997). Drug use and vio- Doherty, M.C., Garfein, R.S., Monterroso, lent crime among adolescents. Adoles- E., Latkin, C., and Vlahov, D. (2000). Gen- cence, 32(126), 395–405. der differences in the initiation of injection drug use among young adults. Journal of De Leon, G., Melnick, G., Thomas, G., Urban Health, 77(3), 396–414. Kressler, D., and Wexler, H. (2000). Moti- vation for treatment in a prison-based ther- Dorsey, T.L., and Zawitz, M. (1999). Drugs apeutic community. American Journal of and crime facts (NCJ 165148) [Online]. Drug & Alcohol Abuse, 26(1), 33–46. Available: www.ojp.usdoj.gov/bjs/pub/pdf/ dcf.pdf. De Li, S., Priu, H., and MacKenzie, D. (2000). Drug involvement, lifestyles, and Drug Abuse Education, Prevention, and criminal activities among probationers. Treatment Act of 2001, S. 234, 107th Journal of Drug Issues, 30(3), 593–620. Cong., 1st Sess. (2001).

Defleur, L., Ball, J., and Snarr, R. (1969). Drug Abuse Warning Network (DAWN) The long-term social correlates of opiate (2000). The DAWN report: Club drugs addiction. Social Problems, 17(2), 225–233. [Online]. Available: www.samhsa.gov/oas/ clubdrug.pdf. Dembo, R., Williams, L., Wothke, W., and Schmeidler, J. (1994). The relationship Drug Enforcement Administration Intelli- among family problems, friends’ troubled gence Division (2000). An overview of behavior, and high risk youths’ alcohol/ club drugs (DEA–20005). Alexandria, VA: other drug use and delinquent behavior: U.S. Department of Justice, Drug En- A longitudinal study. The International forcement Administration [Online]. Journal of the Addictions, 29(11), Available: www.usdoj.gov/dea/pubs/ 1419–1442. intel/20005intellbrief.pdf.

Dennis, M., Karuntzos, G., and Rachal, J. Drug Medicalization, Prevention and (1992). Accessing additional community Control Act of 1996, Proposition 2000, resources through case management to Arizona (December 6). meet the needs of methadone clients. In R.S. Ashery (ed.), Progress and issues in Drug Strategies (1997). Cutting crime: case management (NIDA Research Mono- Drug courts in action. Washington, DC: graph 127, DHHS Publication No. ADM Drug Strategies. 92–1946, pp. 54–78). Rockville, MD: U.S. Department of Health and Human Serv- Eisenburg, M., and Fabelo, T. (1996). ices, National Institute on Drug Abuse. Evaluation of the Texas correctional substance abuse treatment initiative: The Diggs, D., and Pieper, S. (1994). Using day impact of policy research. Crime & reporting centers as an alternative to jail. Delinquency, 42(2), 296–308. Federal Probation, 58(1), 9–12. Ellickson, P.L., and McGuigan, K.A. (2000). Dogwill, N. (1998). The burning question: Early predictors of adolescent violence. How will the United States deal with the American Journal of Public Health, 90(4), medical-marijuana debate? Detroit College 566–572. of Law at Michigan State University Law Review, 247 [Online]. Available: www.dcl. Elliott, D.S., and Huizinga, D. (1984). The edu/lawrev/98-1/dogwill.htm. relationship between delinquent behavior and ADM problems. Paper presented at

144 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

the ADAMHA/OJJDP State-of-the-Art Field, G. (1989). A study of the effects of Research Conference on Juvenile Of- intensive treatment on reducing the crimi- fenders With Serious Drug Alcohol and nal recidivism of addicted offenders. Fed- Mental Health Problems. Rockville, MD. eral Probation, 53(10), 51–56.

Elliott, D.S., Huizinga, D., and Menard, Field, G. (1992). Oregon prison drug treat- S.W. (1989). Multiple problem youth: ment programs. In C.G. Leukefeld and Delinquency, substance use, and mental F.M. Tims (eds.), Drug abuse treatment health problems (Research in Crimin- in prisons and jails (pp. 142–155). Wash- ology). New York: Springer-Verlag. ington, DC: U.S. Government Printing Office. Enos, R., and Southern, S. (1996). Correc- tional case management. Cincinnati: Fine, M. and Weis, L. (1998). Crime sto- Anderson Publishing Company. ries: A critical look through race, ethnicity, and gender. Qualitative Studies in Educa- Ernst, A., Nick, T., Weiss, S., Houry, D., tion, 11(3), 435–459. and Mills, T. (1997). Domestic violence in an inner-city. Annals of Emergency Fishbein, D. (1998). Differential suscepti- Medicine, 30(2), 190–197. bility to comorbid drug abuse and violence. Journal of Drug Issues, 28(4), Fagan, J. (1987). Neighborhood education, 859–890. mobilization, and organization for juvenile crime prevention. The Annals of the Fletcher, B.W., Tims, F.M., and Brown, American Academy of Political and Social B.S. (1997). Drug abuse treatment out- Science, 494(Nov.), 54–70. come study (DATOS): Treatment evaluation research in the United States. Fagan, J., and Chin, K. (1990). Violence as Psychology of Addictive Behaviors, 11(4), regulation and social control in the distri- 216–229. bution of crack. In M. De La Rosa, E.Y. Lambert, and B. Gropper (eds.), Drugs Fortney, J., Booth, B., Zhang, M., and violence: Causes, correlates, and Humphrey, J., and Wiseman, E. (1998). consequences (NIDA Research Mono- Controlling for selection bias in the evalua- graph 103, DHHS Publication No. ADM tion of Alcoholics Anonymous as aftercare 90–1721, pp. 8–43). Rockville, MD: U.S. treatment. Journal of Studies on Alcohol, Department of Health and Human 59(6), 690–697. Services, National Institute on Drug Abuse. French, M.T., and Martin, R.F. (1996). The costs of drug abuse consequences: A Farabee, D., Prendergast, M., and Anglin, summary of research findings. Journal M. (1998). The effectiveness of coerced of Substance Abuse Treatment, 13(6), treatment for drug-abusing offenders. 453–466. Federal Probation, 62(1), 3–10. Gallagher, J. (2001, January 18). Rocke- Farabee, D., Prendergast, M., Cartier, J., feller drug laws: Pataki unveils reform Wexler, H., Knight, K., and Anglin, M. plan. Binghamton Press & Sun Bulletin. (1999). Barriers to implementing effective correctional drug treatment programs. Gandossy, R.P., Williams, J.R., Cohen, J., Prison Journal, 79(2), 150–160. and Harwood, H.J. (1980). Drugs and crime: A survey and analysis of the litera- Field, G. (1985). Cornerstone program: A ture (NCJ 159074). Washington, DC: U.S. client outcome study. Federal Probation, 49(2), 50–55.

145 SPECIAL REPORT / JULY 03

Department of Justice, National Institute Greenfeld, L.A. (1998). Alcohol and crime: of Justice. An analysis of national data on the prevalence of alcohol involvement (NCJ Germain, C., and Gitterman, A. (1995). 168632) [Online]. Available: www.ojp. Ecological perspective. In Encyclopedia usdoj.gov/bjs/pub/pdf/ac.pdf. of social work (vol. 1, pp. 816–824). Washington, DC: National Association of Greenwood, P., and Turner, S. (1993). Social Workers Press. Evaluation of the Paint Creek Youth Center: A residential program for serious Gerstein, D., Johnson, R., Larison, C., delinquents. Criminology, 31, 263–279. Harwood, H., and Fountain, D. (1997). Alcohol and other drug treatment for par- Grogger, J., and Willis, M. (2000). The ents and welfare recipients: Outcomes, emergence of crack cocaine and the rise costs, and benefits (HHS–100–95–0036) in urban crime. Review of Economics & [Online]. Available: http://aspe.hhs.gov/ Statistics, 82(4), 519–530. hsp/caldrug/calfin97.htm. Hagan, J. (1995). Current perspectives Gibbens, T.C.N. (1979). Literature over- on aging and the life cycle: Vol. 4. Delin- view on recent European and North quency and disrepute in the life course. American research on the relationship Greenwich, CT: JAI Press. between mental illness and criminality. In L. Beliveau, G. Canepa, and D. Szabo Hagan, J. (1997). Crime and capitalization: (eds.), Today’s problems in clinical crimi- Toward a developmental theory of street nology—Research on diagnosis and treat- crime in America. In T.P. Thornberry (ed.), ment. Montreal, Canada: Universite de Advances in criminological theory: Vol. 7. Montreal. Developmental theories of crime and delinquency (pp. 287–308). New Bruns- Godley, S.H., Godley, M.D., Pratt, A., and wick, NJ: Transaction Publishers. Wallace, J.L. (1994). Case management services for adolescent substance abusers: Hamid, A. (1998). Drugs in America: A program description. Journal of Sub- Sociology, economics, and politics. stance Abuse Treatment, 11(4), 309–317. Gaithersburg, MD: Aspen Publishers.

Gold, L.H. (1973). Discovery of mental Hanlon, T.E., Nurco, D.N., Kinlock, T.W., illness and mental defect among offend- and Duszynski, K.R. (1990). Trends in crim- ers. Journal of Forensic Sciences, 18(2), inal activity and drug use over an addiction 125–129. career. American Journal of Drug and Alcohol Abuse, 16(3–4), 223–238. Goldkamp, J.S. (1994). Justice and treatment innovation: The drug court move- Harlow, C.W. (1998). Profile of jail inmates ment (NCJ 149260). Washington, DC: U.S. 1996 (BJS Special Report, NCJ 164620) Department of Justice, National Institute [Online]. Available: www.ojp.usdoj.gov/ of Justice. bjs/pub/pdf/pji96.pdf.

Goldstein, P. (1985). The drugs/violence Harrell, A. (producer). (1998). Drug courts nexus: A tripartite conceptual framework. and the role of graduated sanctions (NCJ Journal of Drug Issues, 15(4), 493–506. 169597) [Film]. Washington, DC: U.S. Department of Justice, National Institute Goode, E. (1997). Between politics and of Justice. reason: The drug legalization debate. New York: St. Martin’s Press.

146 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Harrell, A., Cavanagh, S., and Roman, J. Hser, Y.I., Anglin, M.D., and Powers, K. (2000). Evaluation of the D.C. Superior (1993). A 24-year follow-up of California Court drug intervention programs (NCJ narcotics addicts. Archives of General 178941) [Online]. Available: www.ncjrs. Psychiatry, 50(7), 577–584. org/pdffiles1/nij/178941.pdf. Hubbard, R.L., Marsden, M.E., Rachal, Harrell, A., and Peterson, G. (1992). Drugs, J.V., Harwood, H.J., Cavanaugh, E.R., and crime, and social isolation: Barriers to Ginzburg, H.M. (1989). Drug abuse treat- urban opportunity. Washington, DC: Urban ment: A national study of effectiveness. Institute Press. Chapel Hill, NC: University of North Carolina Press. Harrison, L., and Gfroerer, J. (1992). The intersection of drug use and criminal be- ImpacTeen Illicit Drug Team (2002). Illicit havior: Results from the National House- drug policies in the United States: Laws hold Survey on Drug Abuse. Crime & and penalties from the 50 States. Berrien Delinquency, 38(4), 422–443. Springs, MI: Andrews University.

Healey, K.M. (1999). Case management in Inciardi, J.A. (1994). Screening and the criminal justice system (NCJ 173409) assessment for alcohol and other drug [Online]. Available: www.ncjrs.org/pdffiles1/ abuse among adults in the criminal justice 173409.pdf. system. (Treatment Improvement Protocol [TIP] Series No. 7, DHHS Publication No. Hickman, T.A. (2000). Drugs and race in (SMA) 94–2076). Rockville, MD: U.S. De- American culture. American Studies, partment of Health and Human Services, 41(1), 71–90. Center for Substance Abuse Treatment.

Hirschfield, A., and Bowers, K.J. (1997). Inciardi, J.A. (2001). The war on drugs III: The effect of social cohesion on levels of The continuing saga of the mysteries and recorded crime in disadvantaged areas. miseries of intoxication, addiction, crime, Urban Studies, 34(8), 1275–1295. and public policy. Boston: Allyn and Bacon.

Hobson, R. (1928). The struggle of man- Inciardi, J.A., Horowitz, R., and Pottieger, kind against its deadliest foe. Narcotic A.E. (1993). Street kids, street drugs, Education, 1, 51–54. street crime: An examination of drug use and serious delinquency in Miami. Hora, P., Schma, W., and Rosenthal, J. Belmont, CA: Wadsworth Publication (1999). Therapeutic jurisprudence and the Company. drug treatment court movement: Revolu- tionizing the criminal justice system. Notre Inciardi, J.A., Martin, S.S., Butzin, C.A., Dame Law Review, 74(2), 439–537. Hooper, R.M., and Harrison, L.D. (1997). An effective model of prison-based treat- Hser, Y.I. (1997). Self-reported drug use: ment for drug-involved offenders. Journal Results of selected empirical investiga- of Drug Issues, 27(2), 261–278. tions of validity. In L. Harrison and A. Hughes (eds.), The validity of self-reported Inciardi, J.A., and McBride, D.C. (1991). drug use: Improving the accuracy of sur- Treatment alternatives to street crime: vey estimates (NIDA Research Mono- History, experiences, and issues (DHHS graph 167, NIH Publication No. 97–4147, Publication No. ADM 91–1749). Rockville, pp. 320–343). Rockville, MD: U.S. De- MD: U.S. Department of Health and partment of Health and Human Services, National Institute on Drug Abuse.

147 SPECIAL REPORT / JULY 03

Human Services, National Institute on Kawachi, I., Kennedy, B.P., and Glass, R. Drug Abuse. (1999). Social capital and self-rated health: A contextual analysis. American Journal of Inciardi, J.A., McBride, D.C., and Rivers, Public Health, 89(8), 1187–1193. J.E. (1996). Drug control and the courts. Thousand Oaks, CA: Sage Publications. Kawachi, I., Kennedy, B.P., Lochner, K., and Prothrow-Stith, D. (1997). Social capi- Inciardi, J.A., Surratt, H., Martin, S., and tal, income inequality, and mortality. Hooper, R. (2001). The importance of American Journal of Public Health, 87(9), aftercare in a corrections-based treatment 1491–1499. continuum. In C.G. Leukefeld, F. Tims, and D. Farabee (eds.), Treatment of drug of- Kawachi, I., Kennedy, B.P., and Wilkinson, fenders: Policies and issues (pp. 204–216). R.G. (1999). Crime: Social disorganization New York: Springer Publishing Company. and relative deprivation. Social Science & Medicine, 48(6), 719–731. Jacobs, W., DuPont, R., and Gold, M. (2000). Drug testing and the DSM-IV. Kennedy, B.P., Kawachi, I., Prothrow-Stith, Psychiatric Annals, 30(9), 583–588. D., Lochner, K., and Gupta, V. (1998). Social capital, income inequality, and Jobes, P.C. (1999). Residential stability and firearm violent crime. Social Science & crime in small rural agricultural and recre- Medicine, 47(1), 7–17. ational towns. Sociological Perspectives, 42(3), 499–524. Khantzian, E.J. (1985). The self-medication hypothesis of addictive disorders: Focus Johnson, B.D., Wish, E.D., Schmeidler, J., on heroin and cocaine dependence. and Huizinga, D. (1991). Concentration of American Journal of Psychiatry, 142(11), delinquent offending: Serious drug in- 1259–1264. volvement and high delinquency rates. Journal of Drug Issues, 21(2), 205–229. Kirst-Ashman, K. (2000). Human behavior, communities, organizations, and groups Johnston, L., O’Malley, P., and Bachman, in the macro social environment: An J. (2001). Monitoring the Future national empowerment approach. Belmont, CA: results on adolescent drug use (NIH Pub- Wadsworth Publications Company. lication No. 01–4923) [Online]. Available: http://monitoringthefuture.org/pubs/ Kolden, G.G., Howard, K.I., Bankoff, E., monographs/overview2000.pdf. Maling, M., and Martinovich, Z. (1997). Factors associated with treatment continu- Join Together (1999). Results of the fourth ation: Implications for the treatment of national survey on community efforts to drug dependence. In L.S. Onken, J.D. reduce substance abuse and gun violence Blaine, and J.J. Boren (eds.), Beyond the [Online]. Available: www.jointogether.org/ therapeutic alliance: Keeping the drug- sa/files/pdf/survey98.pdf. dependent individual in treatment (NIDA Research Monograph 196, NIH Publication Kassebaum, P. (1999). Substance abuse No. 97–4142, pp. 110–130) [Online]. Avail- treatment for women offenders (Technical able: www.nida.nih.gov/pdf/monographs/ Assistance Publication Series No. 23, monograph165/000_TOC165.pdf. DHHS Publication No. SMA 00–3454). Rockville, MD: U.S. Department of Health Kraft, M., and Dickinson, J. (1997). Part- and Human Services, Substance Abuse nerships for improved service delivery: and Mental Health Services Administration. The Newark Target Cities Project. Health & Social Work, 22(2), 143–148.

148 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Landry, M. (1997). Overview of addiction Problems and prospects (pp. 1–19). treatment effectiveness. Rockville, MD: Albany: State University of New York Substance Abuse and Mental Health Press. Services Administration. Lurigio, A.J. (2000). Drug treatment avail- Langan, P., and Cunniff, M. (1992). Re- ability and effectiveness: Studies of the cidivism of felons on probation, 1986–89 general and criminal justice populations. (NCJ 134177). Washington, DC: U.S. Criminal Justice and Behavior, 27(4), Department of Justice, Bureau of Justice 495–528. Statistics. Lurigio, A.J., and Lewis, D.A. (1987). The Larkin, M. (2000). Thwarting the agony of criminal mental patient: A descriptive ecstasy? Lancet, 356(9231), 779. analysis and suggestions for future research. Criminal Justice and Behavior, Lash, S.J. (1998). Increasing participation 14(3), 268–287. in substance abuse aftercare treatment. American Journal of Drug & Alcohol Lurigio, A.J., and Swartz, J.A. (2000). Abuse, 24(1), 31–36. Changing the contours of the criminal justice system to meet the needs of persons Lee, M.R. (2000). Community cohesion with serious mental illness. In J. Horney and violent predatory victimization: A theo- (ed.), Criminal justice 2000: Volume 3: retical extension and cross-national test of Policies, processes, and decisions of the opportunity theory. Social Forces, 79(2), criminal justice system (NCJ 185529, 683–706. pp. 45–108). Available: www.ncjrs.org/ criminal_justice2000/vol_3/03c.pdf. Lerner, R.M. (2000). Developing civil society through the promotion of positive youth MacCoun, R. and Reuter, P. (1997, development. Journal of Developmental & October 3). Interpreting Dutch cannabis Behavioral Pediatrics, 21(1), 48–49. policy: Reasoning by analogy in the legalization debate. Science, 278(5335), 47–52. Leshner, A. (2001). Treatment, education, and prevention: Adding to the arsenal in Macmillan, R. (1995). Changes in the the war on drugs: Hearing before the structure of life courses and the decline of Senate Committee on the Judiciary social capital in Canadian society: A time [Online]. Available: www.nida.nih.gov/ series analysis of property crime rates. Testimony/3-14-01Testimony.html. Canadian Journal of Sociology, 20(1), 51–79. Linnever, J., and Shoemaker, D. (1995). Drugs and crime: A macro-level analysis. Mahoney, B., Carver, J., Cooper, C., Paper presented to the Society for the Polansky, L., Weinstein, S., Dyer Wells, J., Study of Social Problems. and Westfield, T. (1998). Drug court monitoring, evaluation, and management infor- Lipton, D.S. (1998). Therapeutic communi- mation systems (NCJ 171138) [Online]. ties: History, effectiveness and prospects. Available: www.ojp.usdoj.gov/dcpo/monitor. Corrections Today, 60(6), 106. Manski, C.F., Pepper, J.V., and Thomas, Liska, A., Krohn, M., and Messner, S. Y.F. (1999). Assessment of two cost- (1989). Strategies and requisites for theo- effectiveness studies on cocaine control retical integration in the study of crime and policy. Washington, DC: National Academy deviance. In S.F. Messner, M.D. Krohn, Press. and A.E. Liska (eds.), Theoretical integration in the study of deviance and crime:

149 SPECIAL REPORT / JULY 03

Marciniak, L. (1999). Use of day reporting NIDA Notes, 15(6) [Online]. Available: as an intermediate sanction: A study of www.nida.nih.gov/NIDA_Notes/NNVol15N6/ offender targeting and program termina- Index.html. tion. Prison Journal, 79(2), 205–225. Maxwell, S., and Maxwell, C. (2000). Marciniak, L. (2000). Addition of day Examining the “criminal careers” of prosti- reporting to intensive supervision proba- tutes within the nexus of drug use, drug tion: A comparison of recidivism rates. selling, and other illicit activities. Crimin- Federal Probation, 64(1), 34–39. ology, 38(3), 787–810.

Markoff, S. (1997). State-by-State medici- Mazza, D., and Dennerstein, L. (1996). nal marijuana laws. Washington, DC: Psychotropic drug use by women: Marijuana Policy Project Foundation. Could violence account for the gender difference? Journal of Psychosomatic Markou, A., Kosten, T.R., and Koob, G.F. Obstetrics and Gynaecology, 17, 229–234. (1998). Neurobiological similarities in depression and drug dependence: A self- McBride, D.C., and McCoy, C.B. (1982). medication hypothesis. Neuropsycho- Crime and drugs: The issues and litera- pharmacology, 18(3), 135–174. ture. Journal of Drug Issues, 12(2), 137–152. Marlowe, D. (2001). Coercive treatment of substance abusing criminal offenders. McBride, D.C., and McCoy, C.B. (1993). Journal of Forensic Psychology Practice, The drugs-crime relationship: An analytical 1(1), 65–73. framework. The Prison Journal, 73(3–4), 257–278. Martin, S., Butzin, C., and Inciardi, J. (1995). Assessment of a multistage therapeutic McBride, D.C., VanBuren, H., Terry, Y.M., community for drug-involved offenders. and Goldstein, B.J. (2000). Depression, Journal of Psychoactive Drugs, 27(1), drug use, and health services need, utiliza- 109–116. tion, and cost. Advances in Medical Sociology, 7, 67–69. Martin, S., Butzin, C., Saum, C., and Inciardi, J. (1999). Three-year outcomes of McBride, D.C., and VanderWaal, C.J. therapeutic community treatment for drug- (1997). Day reporting centers as an alter- involved offenders in Delaware: From native for drug using offenders. Journal of prison to work release to aftercare. Prison Drug Issues, 27(2), 379–397. Journal, 79(3), 294–320. McBride, D.C., VanderWaal, C.J., Pacula, Massachusetts Department of Public R., Terry, Y.M., and Chriqui, J. (2001). Health (2000). Substance abuse treatment Mandatory minimum sentencing and drug outcomes and system improvements. law violations: Effects on the criminal jus- Boston: Executive Office of Health and tice system. In C.G. Leukefeld, F. Tims, Human Services, Bureau of Substance and D. Farabee (eds.), Treatment of drug Abuse Services. offenders: Policies and issues, pp. 319– 334. New York: Springer Publishing Mathias, R. (2001). NIDA clinical trials net- Company. work begins first multisite tests of new science-based drug abuse treatments.

150 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

McBride, D.C., VanderWaal, C.J., Terry, review. In C.L. Wetherington and A.B. Y.M., and VanBuren, H. (1999). Breaking Roman (eds.), Drug addiction research the cycle of drug use among juvenile and the health of women (pp. 197–222). offenders: Final technical report (NCJ Rockville, MD: U.S. Department of Health 179273) [Online]. Available: www.ojp. and Human Services, National Institute on usdoj.gov/nij/drugdocs.htm. Drug Abuse.

McKay, J., McLellan, A., Alterman, A.I., Mull, E. (1998). Developing a managed Cacciola, J., Rutherford, M., and O’Brien, care response for juvenile justice: A guide. C. (1998). Predictors of participation in Chicago: Treatment Alternatives for Safe aftercare sessions and self-help groups Communities. following completion of intensive outpatient treatment for substance abuse. Mullen, R.E., and Donnermeyer, J.F. Journal of Studies on Alcohol, 59(2), (1985). Age, trust, and perceived safety 152–162. from crime in rural areas. Gerontologist, 25(3), 237–242. McLellan, A.T., Grissom, G., Zanis, D., Randall, M., Brill, P., and O’Brien, C. Mumola, C. (1999). Substance abuse and (1997). Problem-service “matching” in treatment, State and Federal prisoners, addiction treatment: A prospective study 1997 (BJS Special Report, NCJ 172871) in 4 programs. Archives of General [Online]. Available: www.ojp.usdoj.gov/ Psychiatry, 54(8), 730–735. bjs/pub/pdf/satsfp97.pdf.

McPhail, M., and Wiest, B. (1995). Com- Musto, D.F. (1999). The American disease. bining alcohol and other drug abuse treat- New York: Oxford University Press. ment with diversion for juveniles in the justice system (Treatment Improvement National Alliance for Model State Drug Protocol [TIP] Series No. 2, DHHS Publica- Laws (2001). Highlights of accomplish- tion No. SMA 95–3051). Rockville, MD: ments. Alexandria, VA: National Alliance U.S. Department of Health and Human for Model State Drug Laws. Services, Center for Substance Abuse National Association of Drug Court Pro- Treatment. fessionals (2000). Community policing Menaghan, E., and Parcel, G. (1993). and drug court/community courts project: Family social capital and children’s behav- A two-year progress report. Alexandria, ior problems. Social Psychology Quarterly, VA: National Association of Drug Court 56(2), 120–135. Professionals.

Metsch, L.R., McCoy, C.B., Shultz, J.M., National Commission on Marihuana and Page, J.B., Philippe, E., and McKay, C. Drug Abuse (1972). Marihuana: A signal (1999). Gender comparisons of injection of misunderstanding. First report of the drug use practices in shooting galleries. National Commission on Marihuana and Population Research and Policy Review, Drug Abuse. Washington, DC: U.S. Govern- 18, 101–117. ment Printing Office.

Mill, J.S. (1979). Utilitarianism; On liberty; National Institute on Drug Abuse (1999). Essay on Bentham. Glasgow: Collins. Principles of drug addiction treatment: A research-based guide (DHHS Publication Moras, K. (1998). Behavioral therapies for No. 00–4180) [Online]. Available: www. female drug users: An efficacy-focused 165.112.78.61/PODAT/PODATindex.html.

151 SPECIAL REPORT / JULY 03

National Research Council (author) and Pacula, R.L., Chriqui, J.F., Reichmann, Committee on Data and Research for D.A., and Terry-McElrath, Y.M. (2001). Policy on Illegal Drugs: C.F. Manski, J.V. State medical marijuana laws: Under- Pepper, and C.V. Petrie, eds. (2001). In- standing the laws and their limitations forming America’s policy on illegal drugs: (Research Paper Series, No. 13) [Online]. What we don’t know keeps hurting us. Available: www.impacteen.org/generalarea_ Washington, DC: National Academy Press. PDFs/medicalmarijuanapaper100301.pdf.

New York State Commission on Drugs and Paniagua, F.A. (1998). Assessing and treat- the Courts (2000). Confronting the cycle ing culturally diverse clients: A practical of addiction and recidivism: A report to guide (2d ed.). Thousand Oaks, CA: Sage Chief Judge Judith S. Kaye by the New Publications. York State Commission on Drugs and the Courts. [Online]. Available: www.courts. Parcel, T.L., and Menaghan, E.G. (1993). state.ny.us/addictionrecidivism62000.html. Family social capital and children’s behavior problems. Social Psychology Nurco, D., Cisin, I., and Balter, M. (1981). Quarterly, 56(2), 120–135. Addict careers III: Trends across time. The International Journal of Addictions, 8, Parent, D. (1990). Day reporting centers 1357–1372. for criminal offenders—A descriptive analysis of existing programs (NIJ Issues Nurco, D., Kinlock, T., and Balter, M. and Practices, NCJ 125268). Washington, (1993). The severity of preaddiction crimi- DC: U.S. Department of Justice, National nal behavior among urban, male narcotic Institute of Justice. addicts and two nonaddicted control groups. Journal of Research in Crime and Parker, R., and Auerhahn, K. (1998). Delinquency, 30(3), 293–316. Alcohol, drugs, and violence. Annual Review of Sociology, 24, 291–311. Office of National Drug Control Policy (1996). Treatment protocol effectiveness Pelissier, B. (2001). Personal communication. study (NCJ 160933) [Online]. Available: www.ncjrs.org/txtfiles/trmtprot.txt. Pelissier, B., Gaes, G., Rhodes, W., Camp, S., O’Neil, J., Wallace, S., and Saylor, W. Office of National Drug Control Policy (2000). TRIAD drug treatment evaluation (2000). The national drug control strategy: project. [Online]. Available: www.bop.gov/ FY 2001 budget summary (NCJ 180083). orepg/oretriad.html. Washington, DC: The White House. Peters, R.H., Kearns, W.D., Murrin, M.R., Office of National Drug Control Policy and Dolente, A.S. (1992). Psychopathology (2001). The national drug control strategy: and mental health needs among drug- 2001 annual report (NCJ 185396). Wash- involved inmates. Journal of Prison & Jail ington, DC: The White House. Health, 11(1), 3–25.

Owens, S., Klebe, K., Arens, S., Durham, Peyton, E.A., and Gossweiler, R. (2001). R., Hughes, J., Moor, C., O’Keefe, M., Treatment services in adult drug courts: Phillips, J., Sarno, J., and Stommel, J. Report on the 1999 National Drug Court (1997). The effectiveness of Colorado’s Treatment Survey, executive summary TASC programs. Journal of Offender (NCJ 188086) [Online]. Available: www.ncjrs. Rehabilitation, 26(1–2), 161–176. org/pdffiles1/ojp/188086.pdf.

152 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Pihl, R., and Peterson, J. (1995). Drugs Rapp, R.C., Siegal, H.A., Li, L., and Saha, and aggression: Correlations, crime, and P. (1998). Predicting postprimary treat- human manipulative studies and some ment services and drug use outcome. proposed mechanisms. Journal of Psychi- American Journal of Drug and Alcohol atry & Neuroscience, 20(2), 141–149. Abuse, 24(4), 603–615.

Platt, S. (2001). Drug court experiment: Resignato, A.J. (2000). Violent crime: A Policy choice—political decision. Maryland function of drug use or drug enforcement? Bar Journal, 34(1), 44–47. Applied Economics, 32(6), 681–688.

Pogrebin, M. (1978). Role conflict among Rhine, E. (1993). Reclaiming offender correctional officers in treatment oriented accountability: Intermediate sanctions for correctional institutions. International probation and parole violators. Laurel, MD: Journal of Offender Therapy and Com- American Correctional Association. parative Criminology, 22(2), 149–155. Rivers, J., and Trotti, T. (1995). South Popay, J. (2000). Social capital: The role of Carolina delinquent males: An 11-year narrative and historical research. Journal follow-up into adult probation and prison. of Epidemiology & Community Health, Columbia: South Carolina Department of 54(6), 401. Corrections.

President’s Commission on Model State Rose, R. (2000). How much does social Drug Laws (1993). President’s Com- capital add to individual health? A survey mission on Model State Drug Laws: study of Russians. Social Science & Executive summary. Washington, DC: Medicine, 51(9), 1421–1435. President’s Commission on Model State Drug Laws. Roth, J. (1994). Psychoactive substances and violence (NIJ Research in Brief, NCJ Putnam, R. (1993). The prosperous com- 145534) [Online]. Available: www.ncjrs. munity: Social capital and public life. org/txtfiles/psycho.txt. The American Prospect, 4(13) [Online]. Available: www.prospect.org/print/V4/ Rychtarik, R., Prue, D., Rapp, S., and King, 13/putnam-r.html. A. (1992). Self-efficacy, aftercare and relapse in a treatment program for alco- Putnam, R. (2000). Bowling alone: The col- holics. Journal of Studies on Alcohol, lapse and revival of American community. 53(5), 435–440. New York: Simon & Schuster. Rydell, C.P., Caulkins, P., and Everingham, Putnam, R.D., and Campbell, D.E. (2000, S.M.S. (1996). Enforcement or treatment? August 10). The country’s great challenge: Modeling the relative efficacy of alterna- Enticing the young to the voting booth. tives for controlling cocaine. Operations Boston Globe, A19. Research, 44(5), 687–693.

Rapp, R.C. (1997). The strengths perspec- Rydell, C.P., and Everingham, S.M.S. tive and persons with substance abuse (1994). Controlling cocaine: Supply versus problems. In D. Saleebey (ed.), The demand programs (MR–331–ONDCP/ strengths perspective in social work prac- A/DPRC). Santa Monica, CA: RAND. tice (2d ed.). New York: Longman.

153 SPECIAL REPORT / JULY 03

Salmon, R., and Salmon, R. (1983). The Sengupta, S. (2001, March 13). A new role of coercion in rehabilitation of drug plan to roll back drug terms. New York abusers. The International Journal of Times. Addictions, 18(1), 9–21. Sentencing Project (1998). Proposed Sampson, R.J., and Laub, J.H. (1990). changes in crack/cocaine sentencing laws Crime and deviance over the life course: would increase number of minorities in The salience of adult social bonds. Ameri- prison, have little impact on drug abuse can Sociological Review, 55(5), 609–627. [Online]. Available: www.sentencing project. org/brief/pub1057.htm. Sampson, R.J., and Raudenbush, S.W. (1999). Systematic social observation of Siegal, H.A. (1998). Comprehensive case public spaces: A new look at disorder in management for substance abuse treat- urban neighborhoods. American Journal ment (Treatment Improvement Protocol of Sociology, 105(3), 603–651. [TIP] Series, DHHS Publication No. SMA 98–3222). Rockville, MD: U.S. Department Sampson, R.J., Raudenbush, S.W., and of Health and Human Services, Center for Earls, F. (1997, August 15). Neighborhoods Substance Abuse Treatment. and violent crime: A multilevel study of collective efficacy. Science, 277, 918–924. Siegal, H.A., Fisher, J.H., Rapp, R.C., Kelliher, C.W., Wagner, J.H., O’Brien, W.F., San Francisco Examiner. (2000, November and Cole, P.A. (1996). Enhancing sub- 9). San Francisco plans to expand drug stance abuse treatment with case treatment. management: Its impact on employment. Journal of Substance Abuse Treatment, Savelsberg, J.J. (1999). Controlling vio- 13(2), 93–98. lence: Criminal justice, society, and les- sons from the US. Crime, Law & Social Siegal, H.A., Rapp, R.C., Kelliher, C.W., Change, 30(2), 185–203. Fisher, J.H., Wagner, J.H., and Cole, P.A. (1995). The strengths perspective of case Schlenger, W., Kroutil, L., and Roland, E. management: A promising inpatient sub- (1992). Case management as a mecha- stance abuse treatment enhancement. nism for linking drug abuse treatment and Journal of Psychoactive Drugs, 27(1), primary care: Preliminary evidence from 67–72. the ADAMHA/HRSA linkage demonstra- tion. In R.S. Ashery (ed.), Progress and Siegal, H.A., Rapp, R.C., Li, L., Saha, P., issues in case management (NIDA Re- and Kirk, K. (1997). The role of case man- search Monograph 127, DHHS Publica- agement in retaining clients in substance tion No. ADM 92–1946, pp. 316–330). abuse treatment: An explanatory analysis. Rockville, MD: U.S. Department of Health Journal of Drug Issues, 27(4), 821–831. and Human Services, National Institute on Drug Abuse. Sigmon, J., Nugent, M., Goerdt, J., and Wallace, S. (1999). Key elements of suc- Schmidt, L. (1995). “A battle no man’s but cessful adjudication partnerships (BJA God’s”: Origins of the American temper- Bulletin, NCJ 173949) [Online]. Available: ance crusade in the struggle for religious www.ncjrs.org/pdffiles1/bja/173949.pdf. authority. Journal of Studies on Alcohol, 56(1), 110–121. Simpson, D.C., and Curry, S., eds. (1997–98). DATOS first-wave findings released. Institute of Behavioral Research at Texas Christian University: Research

154 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Roundup, 7(4) [Online]. Available: www.ibr. indicators of recidivism. Journal of Drug tcu.edu/pubs/newslet/97-98winter.pdf. Issues, 31(1), 149–176.

Simpson, D., Joe, G., Broome, K., Hiller, Stephens, R. (1991). The street addict M., Knight, K., and Rowan-Szal, G. (1997). role: A theory of heroin addiction. Albany: Program diversity and treatment retention State University of New York Press. rates in the Drug Abuse Treatment Out- come Study (DATOS). Psychology of Substance Abuse and Mental Health Addictive Behaviors, 11(4), 279–293. Services Administration (2000). Treatment Episode Data Set (TEDS): 1993–1998. Simpson, D., Joe, G., and Brown, B. [Online]. Available: www.samhsa.gov/ (1997). Treatment retention and follow-up statistics/statistics.html. outcomes in the Drug Abuse Treatment Outcome Study (DATOS). Psychology of System Sciences (1979). Evaluation of Addictive Behaviors, 11(4), 294–307. treatment alternatives to street crime: National evaluation programs, phase II Slobogin, C. (1995). Therapeutic jurispru- reports (NCJ 51931). Washington, DC: dence: Five dilemmas to ponder. Psych- U.S. Department of Justice, National ology, Public Policy, and Law, 1(1), Institute of Law Enforcement and Criminal 193–219. Justice.

Smith, M.E. (2001). What future for “pub- Tauber, J. (1994). Treating drug-using lic safety” and “restorative justice” in offenders through sanctions, incentives. community corrections? (NIJ Research in Corrections Today, 56(1), 28–33. Brief, NCJ 187773) [Online]. Available: www.ncjrs.org/pdffiles1/nij/187773.pdf. Taxman, F. (1998). Reducing recidivism through a seamless system of care: Spangenberg, R., and Beeman, M. (1998). Components of effective treatment, Improving State and local criminal justice supervision, and transition services in the systems: A report on how public defend- community (NCJ 171836). Washington, ers, prosecutors, and other criminal justice DC: Office of National Drug Control Policy. system practitioners are collaborating across the country (BJA Monograph, NCJ Taxman, F. (2000). Effective practices for 173391) [Online]. Available: www.ncjrs. protecting public safety through substance org/pdffiles/173391.pdf. abuse treatment. Unpublished report commissioned by the National Institute on Spear, F., and Skala, S.Y. (1995). Post- Drug Abuse. treatment services for chemically dependent adolescents. In E. Rahdert and D. Taxman, F., and Sherman, S. (1998). Czechowicz (eds.), Adolescent drug Seamless systems of care: Using automa- abuse: Clinical assessment and therapeu- tion to improve outcomes. In L.J. Moriarty tic interventions (NIDA Research Mono- and D.L. Carter (eds.), Criminal justice graph 156, NIH Publication No. 95–3908, technology in the 21st century. Spring- pp. 341–363) [Online]. Available: www.nida. field, IL: Charles C. Thomas. nih.gov/pdf/monographs/ 156.pdf. Taxman, F., Soule, D., and Gelb, A. (1999). Spohn, C., Piper, R.K., Martin, T., and Graduated sanctions: Stepping into Frenzel, E.D. (2001). Drug courts and accountable systems and offenders. recidivism: The results of an evaluation Prison Journal, 79(2), 182–204. using two comparison groups and multiple

155 SPECIAL REPORT / JULY 03

Taylor, D.L., Chitwood, D.D., McElrath, K., United States Sentencing Commission and Belgrave, L.L. (1994). Ethnicity, social (1997). Special report to the Congress: support, and injection drug use. Journal of Cocaine and Federal sentencing policy Black Psychology, 20(1), 36–46. [Online]. Available: www.ussc.gov/ legist.htm. Teplin, L.A. (2001). Assessing alcohol, drug, and mental disorders in juvenile Vaccaro, D., and Wills, T.A. (1998). Stress- detainees (NCJ 186367) [Online]. coping factors in adolescent substance Available: www.ncjrs.org/pdffiles1/ojjdp/ use: Test of ethnic and gender differences fs200102.pdf. in samples of urban adolescents. Journal of Drug Education, 28(3), 257–282. Toborg, M., Levin, D., Milkman, R., and Center, L. (1976). Treatment alternatives Veenstra, G. (2000). Social capital, SES to street crime (TASC) projects: National and health: An individual-level analysis. evaluation program, phase I summary Social Science and Medicine, 50(5), report (NCJ 34057). Washington, DC: U.S. 619–629. Department of Justice, National Institute of Law Enforcement and Criminal Justice. Verma, S.K. (1979). Criminality as a mental health problem—a point of view. Social Tonry, M. (1990). Stated and latent func- Defense, 15(57), 28–34. tions of ISP. Crime & Delinquency, 36(1), 174–191. Walklate, S. (1998). Crime and community: Fear or trust? British Journal of Sociology, Tonry, M. (1996). Sentencing matters. 49(4), 550–569. New York: Oxford University Press. Wexler, H. (1995). The success of thera- Tonry, M. (1997). Intermediate sanctions peutic communities for substance abusers in sentencing guidelines (NIJ Issues and in American prisons. Journal of Psycho- Practices, NCJ 165043) [Online]. Available: active Drugs, 27, 56–66. www.ncjrs.org/pdffiles/165043.pdf. Wexler, H., Falkin, G., and Lipton, D. U.S. General Accounting Office (1997). (1990). Outcome evaluation of a prison Drug courts: Overview of growth, charac- therapeutic community for substance teristics, and results (GAO/GGD–97–106) abuse treatment. Criminal Justice and [Online]. Available: www.ncjrs.org/pdffiles/ Behavior, 17(1), 71–92. dcourts.pdf. Wexler, H., Lipton, D., Falkin, G., and U.S. General Accounting Office (1998). Rosenbaum, A. (1992). Outcome evalua- Drug abuse: Research shows treatment is tion of a prison therapeutic community effective, but benefits may be overstated for substance abuse treatment. In C.G. (GAO/HEHS–98–72) [Online]. Available: Leukefeld and F.M. Tims (eds.), Drug www.druglibrary.org/schaffer/govpubs/ abuse treatment in prisons and jails (pp. gao/gao74.htm. 156–175). Washington, DC: U.S. Govern- ment Printing Office. Uniform Crime Reports (1998). Crime in the United States [Online]. Available: Wexler, H., Melnick, G., Lowe, L., and www.fbi.gov/ucr/Cius_98/98crime/ Peters, J. (1999). Three-year reincarnation 98cius01.pdf. outcomes for amity in-prison therapeutic community and aftercare in California. Prison Journal, 79(3), 321–336.

156 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

White, H. (1990). The drug use-delinquency Authors’ note connection in adolescence. In R. Weisheit (ed.), Drugs, crime, and criminal justice. The authors would like to thank Helo Cincinnati: Anderson Publishing Company. Oidjarv and Shannon Bond for both their writing contributions and their assistance White, H., and Gorman, D. (2000). Dyna- with literature review, and Rachael Del Rio mics of the drug-crime relationship. In G. for her work in preparing the manuscript. LaFree (ed.), Criminal Justice 2000: In addition, the authors would like to Volume 1: The nature of crime: Continuity acknowledge the contributions of Frank and change (NCJ 182408, pp. 151–218) Chaloupka, Dick Clayton, Paul Goldstein, [Online]. Available: www.ncjrs.org/criminal_ Jamie Chriqui, and Rosalie Pacula, who justice2000/vol_1/02d.pdf. reviewed various concepts presented in the paper. Finally, in addition to the sup- Wilson, D.J. (2000). Drug use, testing, and port provided by the National Institute of treatment in jails (BJS Special Report, NCJ Justice, the authors would like to acknowl- 179999) [Online]. Available: www.ojp.usdoj. edge the ImpacTeen Project: a Policy gov/bjs/pub/pdf/duttj.pdf. Research Partnership to Reduce Youth Substance Use supported by the Robert Winters, K.C. (1998). Kids and drugs. Wood Johnson Foundation and adminis- Corrections Today, 60(6), 118–121. tered by the University of Illinois at World Bank Group. (2002). What is social Chicago. The views expressed are those capital? [Online]. Available: www.worldbank. of the authors and do not necessarily org/poverty/scapital/ whatsc.htm. reflect the views of the Robert Wood Johnson Foundation. Zehr, H. (1990). Changing lenses: A new focus for crime and justice. Scottsdale, Correspondence concerning this paper PA: Herald Press. should be addressed to Duane C. McBride, Department of Behavioral Sciences, Nethery Hall 203, Andrews University, Berrien Springs, MI 49104– 0030; e-mail [email protected].

157 SPECIAL REPORT / JULY 03

Appendix A. Model State early and periodic screening; diagnosis and treatment services; health profession- drug laws and policies als training; criminal justice treatment; The President’s Commission on Model caregiver’s assistance. State Drug Laws’ (1993) model legislation specified five main policy areas. Following Drug-free families/schools/ is a more complete list of the laws and workplaces policies within each general policy area. For drug-free families, underage alcohol consumption reduction; preventive coun- Economic remedies seling services for children of alcoholics Forfeiture reform; money laundering; and addicts; sensible advertising and fami- financial transaction reporting; money ly education; tobacco vending machine transmitter licensing and regulation; on- restriction; revocation of professional or going criminal conduct. business licenses for alcohol and other drugs.

Community mobilization For drug-free schools, drug-free school Expedited eviction of drug traffickers; drug zones; ban on tobacco use in schools; nuisance abatement; crimes code provi- intervention for students with substance sions to protect tenants and neighbors; abuse problems; State safe schools; antidrug volunteer protection; community alcohol- and drug-free colleges and univer- mobilization funding; alcohol/other drug sities; truancy, expulsion, and children out abuse policy and planning coordination. of school. For drug-free workplaces, drug-free Crimes code enforcement private-sector workplaces; drug-free workplace workers’ compensation premium Prescription accountability; State chemical reduction; employee assistance programs control; Uniform Controlled Substances and professionals; drug-free public work Act controlled substance analogs; contin- force; drug-free workplace; employee ued access by law enforcement to wire addiction recovery. and electronic communications; wiretap- ping and electronic surveillance control; driving while under the influence of alcohol and other drugs. Reference President’s Commission on Model State Treatment Drug Laws (1993), President’s Commis- sion on Model State Drug Laws: Executive Addiction cost reduction; Medicaid addic- summary, Washington, DC: President’s tion cost reduction; managed care con- Commission on Model State Drug Laws. sumer protection; family preservation;

158 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Appendix B. Critical Membership elements for collaborative As noted previously, membership should success be broad based, representing key agencies in the justice, law enforcement, and As noted in the main body of this paper, treatment systems, and a broad range of reviews of collaborative efforts have iden- other community agencies. tified several critical elements for suc- 1 cess. These elements are specified and Goals discussed below. Collaboratives should design specific goals Leadership that are clear, useful in the minds of participants, and achievable within specified There is a need for one or more key agen- timeframes, including both short- and long- cies to start the collaborative process, term goals, and with specified priorities. preferably bringing experienced leadership Successful collaborative groups have and/or supervision to the table. This body reported the existence of a strategic plan, must be willing to take the responsibility including specific goals, an outline of pro- to identify problems and help other mem- grams related to achieving those goals, bers to envision solutions, maintain the evaluation methods, and regular public support and involvement of other mem- progress updates. A description of goal bers, and work toward helping build an and program review and change was relat- atmosphere of equality. Because in many ed to successful formation and structure.2 communities the relationship between the Performance measures can be especially treatment and criminal justice systems is useful for evaluation and thus the possibili- often strained, there is a need to recog- ty of obtaining continued funding. nize differing primary responsibilities. Within the context of the courts, the jus- Team approach tice system has the primary role in monitoring offenders along the graduated Collaborative efforts should seek a team sanctions continuum; treatment systems approach for both decision planning and have the primary role in providing appropri- making. Leader agencies and/or organiza- ate and effective treatment services. tions should seek to maintain civility at Some evidence indicates that the opti- meetings and encourage flexibility. mum structure might place in the position Decisionmaking should strive to use of managing partner a “neutral” group consensus-building methods. Efforts that does not provide direct services (such toward developing a team approach can as TASC) to ensure unbiased service be assisted by making sure that each col- organization referrals, case management, laborative member has a clearly defined and collaborative organization. No matter role and responsibilities; this can be aided who holds the leadership role, this individual/ by early cross-training for collaborative agency/group must seek consciously to members in the activities and responsibili- actively involve all stakeholders from the ties of the systems involved. beginning of design and implementation of the proposed program(s) or initiative(s).

159 SPECIAL REPORT / JULY 03

Long-term view Funding Members should recognize the complexity Long-term funding sources are crucial for of collaborative goals and strategies, that the viability of any coalition. External fund- neither substance abuse nor crime has a ing sources may assist in providing incen- single solution. Realistic timelines for all tives for development of successful efforts should be set. partnerships3 such as through block grants or private foundations; in addition, com- Research and evaluation munities may have the possibility of pool- ing funds from various agencies. However, Communities considering collaborative efforts should be made to gain line-item work should use available information on legislative support for sustainability. best practices from the literature to guide collaborative and program development. In addition, methods should be developed to Notes systematically collect objective data for monitoring and evaluating collaborative 1. Sigmon, J., Nugent, M., Goerdt, J., and Wallace, S. (1999), Key elements of successful adjudication projects. partnerships (BJA Bulletin, NCJ 173949) pp. 2–4 [Online], available: http://www.ncjrs.org/pdffiles1/bja/ Broad support 173949.pdf; see also McBride, D.C., VanderWaal, C.J., Terry, Y.M., and VanBuren, H. (1999), Breaking The need to gain the support of the com- the cycle of drug use among juvenile offenders: Final munity at large is essential for sustainabili- technical report (NCJ 179273) [Online], available: http://www.ojp.usdoj.gov/nij/drugdocs.htm. ty; active efforts to seek community input can gain support, and regular communica- 2. Join Together (1999). Results of the fourth national tion about the goals and accomplishments survey on community efforts to reduce substance of the partnership can help maintain that abuse and gun violence [Online]. Available: www. support. jointogether.org/ sa/files/pdf/survey98.pdf. 3. Kraft, M., and Dickinson, J. (1997). Partnerships for improved service delivery: The Newark Target Cities Project. Health & Social Work, 22(2), 143–148.

160 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Appendix C. Principles of 8. Addicted or drug-abusing individuals * with coexisting mental disorders drug addiction treatment should have both disorders treated in NIDA (1999) developed a list of scientifi- an integrated way. cally based recommendations for drug 9. Medical detoxification is only the first treatment applicable for use across the stage of addiction treatment and by entire system of service delivery. These itself does little to change long-term principles are listed below: drug use. 1. No single treatment is appropriate for 10. Treatment does not need to be volun- all individuals. tary to be effective. 2. Treatment needs to be readily avail- 11. Possible drug use during treatment able. must be monitored continuously. 3. Effective treatment attends to multi- 12. Treatment programs should provide ple needs of the individual, not just his assessment for HIV/AIDS, Hepatitis B or her drug use. and C, tuberculosis, and other infec- 4. An individual’s treatment and services tious diseases; and counseling to help plan must be assessed continually and patients modify or change behaviors modified as necessary to ensure that that place themselves or others at risk the plan meets the person’s changing of infection. needs. 13. Recovery from drug addiction can be 5. Remaining in treatment for an ade- along-term process and frequently re- quate period of time is critical for quires multiple episodes of treatment. treatment effectiveness. 6. Counseling (individual and/or group) Reference and other behavioral therapies are critical components of effective treat- *National Institute on Drug Abuse (1999), ment for addiction. Principles of drug addiction treatment: A research-based guide (DHHS Publication 7. Medications are an important element No 00–4180), pp. 1–3 [Online], available: of treatment for many patients, espe- http://165.112.78.61/PODAT/PODATindex. cially when combined with counseling html. and other behavioral therapies.

161 Appendix A: Summary of Proceedings

Opening remarks were made by Sally T. great deal has been learned about drugs, Hillsman, deputy director of the National drug use, drug abuse, drug markets, and Institute of Justice (NIJ), and by forum drug law enforcement, the agencies and organizers Henry H. Brownstein and Lynda the scholarly community remain ill-informed Erinoff.1 Dr. Brownstein is director of the about the complexities and nuances of Drugs and Crime Research Division of NIJ. drugs-crime interrelationships. Dr. Erinoff is health science administrator at the Epidemiology Research Branch of She emphasized the need for a focused the National Institute on Drug Abuse research agenda in which researchers tar- (NIDA). Roger Conner, director of Search get specific questions and the most effec- for Common Ground in America, served tive methods. Researchers who work for as facilitator of the forum discussions. Federal agencies need to know how scarce public funds will be spent to make Dr. Hillsman explained the origins of the the most significant contributions in this collaboration between NIJ and NIDA. In area. All Federal, State, and local policy- the Omnibus Crime Control Act, as amend- makers and practitioners need to know ed in 1976, Congress asked the two agen- which policies and programs will be effec- cies to explore the relationship between tive in producing healthier and safer com- drug abuse and crime. That year NIJ and munities. Dr. Hillsman pointed out that NIJ NIDA formed an interdisciplinary study and NIDA have played essential and com- team to review state-of-the-art knowledge plementary roles in creating a solid scien- about drugs and crime and to recommend tific foundation for informed policies and a research agenda. NIJ published the practices, and that with the help of forum agenda—essentially a literature review— participants, the two agencies will take up in 1980, and the resulting research signifi- the challenge posed by NRC. That chal- cantly advanced knowledge of the drugs- lenge is clearly echoed by policymakers crime relationship. However, a great deal and practitioners throughout the country, remains to be done. Dr. Hillsman noted who are turning to the research communi- that in a 2001 report,2 the National Re- ty to contribute sound and relevant knowl- search Council (NRC) recommended that edge to the Nation’s deliberations about NIJ and NIDA collaboratively undertake drugs and crime. research to meet the challenge of informing public policy in the area of drug use. Dr. Brownstein commented that the early and mid-1980s and through the early Dr. Hillsman noted the amount of research 1990s had seen a great deal of interest in being done on drugs and crime makes the drugs and crime research. Then, in the field increasingly relevant to policy and mid- and late 1990s, research seemed to practice. She reflected that 25 years after have focused on particular applied or prac- the establishment of the initial NIJ and tical areas. He suggested there is a need NIDA collaboration, the two agencies are for more theoretical research and assimila- welcoming researchers to this forum and tion of the knowledge accumulated during are anticipating that the forum will stimu- the 20th century. He noted that the meth- late another extraordinary era of interest ods and technology to make this possible and productivity in the field. Although a are now available.

163 SPECIAL REPORT / JULY 03

Dr. Erinoff noted that Dr. Brownstein de- matrix. The five rubrics were discussed in signed the framework for the forum and relation to the ecological concept of scale, obtained the funding for it. He made cer- which defines the level at which a subject tain that the public health perspective was is studied from the microcosmic to the included and took a hard look at previous macrocosmic; that is, from genes at one research, including the tripartite model, end to policies regulating nations or global which he and Dr. Paul Goldstein had de- relationships at the other. Scale is placed veloped. Dr. Erinoff expressed hope that across the horizontal axis. Each location on all forum participants would emulate the matrix where rubrics and scale inter- Dr. Brownstein in critically reviewing cur- sect represents an area of past or current rent research. She noted that when NIDA research. Some units are filled, indicating staff had to choose someone to write the past or current research; others are empty, forum paper that presented the public indicating that little research has been health perspective, they asked Dr. James done. Evidence that crosses several cells Anthony to do so and to look “outside the is termed broadband research; that is, box.” She appealed to the attendees to research that cuts across domains and lev- do the same: to move outside their own els of scale. frameworks. Discussion: Most helpful ideas “At the Intersection of Scale: A concept or a method? Dr. Anthony was asked to elaborate on the Public Health and Criminal concept of scale and how he interpreted Justice Research on Drugs it as applying to the biological and social and Crime” aspects of his work. James C. Anthony with Valerie Forman Dr. Anthony responded with an example from the field of ecology. For ecologists, The recent focus of Dr. Anthony’s scale is the way one thinks about geocod- research has been on influences that take ing, in which trends are followed in a cen- a user from initial use of a drug toward sus tract or a metropolitan area. Using a drug dependence and on factors that technique called two-dimensional and account for that transition. In discussing three-dimensional wavelet analysis, the some of the directions in which the field ecologist allows the data to evoke the of drugs and crime research should be scale. When studying migratory birds, headed, he noted possible genetic vul- for example, ecologists infer the scale nerabilities to drug abuse, how those vul- from the migration pattern of the birds. nerabilities might influence drugs-crime Ecologists tend to work upward and out- relationships, and cognitive science appli- ward from the smaller level or scale of cations of current interest to NIDA. the organism toward the larger level or scale of the forest or the continent. In Dr. Anthony noted his work in developing Dr. Anthony’s own approach, he worked a conceptual framework for identifying inward, in the direction of methods micro- future research in this area. The frame- scopic in scale. This is in the spirit of the work consists of a matrix that presents ecological concept of scale, but is a twist the rubrics, or the main questions asked, on that concept, which has been used in in epidemiology or public health research: studies of delinquency in relation to such quantity, location, causes, mechanism, neighborhood characteristics as social and prevention and control. These rubrics cohesion. are placed along the vertical axis of the

164 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

When asked whether scale should be Dr. Anthony responded that in the immedi- viewed as a methodology, Dr. Anthony ate future, genetic polymorphisms3 would responded that it should be viewed as a be examined one at a time for effects on concept. Although scale is not strictly drug dependence, but it will probably be 5 methodological, it can guide researchers to 10 years before researchers understand toward methodologies. He applied it in the the covariation of specific polymorphisms context of drugs and crime to provide in a way that will allow them to measure guidance on how NIDA and NIJ can work shared diathesis4 with respect to drug together to foster the next generation of dependence. He expects researchers to research. find that different genes regulate responses to different drugs. He cited as an exam- NIJ, Dr. Anthony suggested, should not ple the finding that genes regulating the establish the kind of biobehavioral labora- liver’s metabolism of alcohol do not appear tories required for research on the genet- to have much of an effect on cocaine ics of the relationships between drug use metabolism. and crime. Rather, it may be wiser to reinforce NIDA’s investment in those areas, One issue to be addressed in this context with the two agencies coming to some is whether polymorphisms sort people agreement. He recommended the same into different latent classes of vulnerability approach for directing, controlling, and rather than arraying them on an underlying planning research on the pharmacological dimension of genetic vulnerability. Dr. effects of drugs on aggression and on the Anthony expressed hope that new inter- cognitive functions. Arrangements should ventions in areas like obesity, where the be made for research where substantial array of known genes and polymorphisms investment has already been made in is much broader than those for drug de- biobehavioral laboratory domains. With pendence, will open up the possibility of respect to national and global policies, studying gene-environment interactions in however, NIDA’s research agenda has not a way that can guide NIDA’s research been strong. NIJ can fill this gap. NIDA agenda on genetics and drug dependence. has supported organizations and operations research on drug dependence for A forum participant suggested that the dis- correctional officers, police, and postre- tribution of drug use in society is wide lease juvenile justice programs. In this enough that genes could in fact play a area, the agendas of the two agencies major role in drug use. However, looking overlap, and it is an area in which they at acquisitive crimes, violent crimes, and, both should be working. in particular, the intersection of drug use and crime and the types of people most Genetics of drug dependence. Dr. likely to be involved in both, we do not Anthony was asked how much progress see those activities distributed throughout has been made in identifying genes or society in a way that suggests such a role. constellations of genes that might predict It was further suggested that NIDA should dependence. He was also asked whether not place a lot of emphasis on genetic he thinks researchers will discover an research. Rather, social science research overall genetics of dependence, as op- should be emphasized because that is posed to highly specific genetic links that where researchers are more likely to find predict dependence on particular drugs, answers. each involving different genes or sets of genes and their protein products. Costs versus benefits of genetic neuropharmacology research. The intersection of the genetic neuropharmacology

165 SPECIAL REPORT / JULY 03

and social behavioral components of drug Types of crime to be addressed. One use was seen as a key element of Dr. participant commented on Dr. Anthony’s Anthony’s paper. A key issue is whether statement that criminal behavior and drug and how funding for this research can be use both affect a person’s social standing. improved. Researchers who seek funding The questioner inferred from Dr. Anthony’s from NIJ or NIDA to conduct sophisticated statement that the effect is negative and urine specimen tests, such as those using suggested that there are probably situa- gas chromatography/mass spectroscopy tions in which it is positive. Furthermore, (GC/MS), know the cost is about $56 per an implicit assumption made at the forum drug tested per GC/MS test. They also is that in discussions about the drugs- know the cost of genetic testing is much crime relationship, researchers are dealing higher, at about $500 per specimen. Given with one kind of crime. However, that rela- the huge price differential, the cost-benefit tionship (if there is one) may exist in a cor- balance becomes an important considera- relate sense, involving economic crime, tion, making it difficult to decide whether financial crime, cybercrime, and other to conduct a social science survey in com- types of crime. The consequences of that bination with biological data collection and relationship, which researchers in their genetic testing. Dr. Anthony was asked to current analyses imply exists, cannot be comment on the implications of these proven. The question posed was whether issues for funding. the forum was focusing on a particular type of crime, such as street crime. If so, He responded by encouraging the partici- it should have been specified, because it pants to think about this problem the way would influence the kind of research agen- they would think about the evolution of da that participants would want to shape. computing speed and costs. Researchers now are able to use multilevel models Drs. Brownstein and Erinoff responded in their research due to the always-increasing that, for the purposes of the forum, they inexpensive computing speed available were not defining crime in any narrow today. Because change is occurring at a sense. They wanted forum participants to similar rate in assays for genes, advances think more broadly and include the effects made by microarray technologists will of alcohol in the discussion. Forum partici- cause the prices of these tests to fall. The pants were urged to keep in mind that the time this change will take is part of the basis in biological science for a relation- reason Dr. Anthony projected it will be 5 to ship between alcohol and aggression is 10 years before researchers are able to the strongest for any drug. understand covariation between polymorphisms in the areas of drug dependence Further comment by participants focused and complex behavior such as criminal on the millions of episodes of drug use offending. Ten years might even be an that are not associated with crime. How, it optimistic projection of the amount of time was asked, could these be reconciled with needed because, ultimately, researchers the framework suggested by Dr. Anthony? want to identify the environments that The discussion also addressed whether modify the expression of the genes. This drug-related violence is instrumental; that calls for both observational studies and is, whether it serves a purpose that pro- experiments. For example, it took approxi- motes the perpetrator’s interests. Tran- mately 10 to 15 years between identifica- scripts of interviews with drug sellers, tion of apolipoprotein-4 for Alzheimer’s which a participant had reviewed before disease and recent work on interventions. the forum began, contained descriptions of violent activities that indicated the

166 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

violence is instrumental. Most of the indicate a growth area only among senior actions described in these transcripts, citizens. This growth trajectory, which which were not taken from a random sam- begins at age 60, is not what would nor- ple of drug dealers, did not appear to be mally be found. Another observed trajecto- the result of impaired functioning or intoxi- ry involves young men who were well cation. Rather, they were quite deliberate, behaved between ages 18 and 25 who instrumental acts. One participant noted begin to misbehave when they are that in discussing crime, the forum was between 30 and 40 years old. The ques- addressing property crime or violent crime tioner asked whether those trajectories are and not including other acts that society outcomes produced by attempts to inter- also defines as crime, such as the use of vene and if so, what this says about growth the drug. trajectories.

Impact of IRBs on social science re- Dr. Anthony responded that the sense in search. The impact of institutional review which he discussed growth trajectories boards (IRBs) on some lines of research was not related to the trends the ques- recommended at the forum were dis- tioner described. Rather, it was in the con- cussed in the context of potential to stifle text of the ecodevelopmental trajectory social research, especially research on model. Conceptually, this model cuts juveniles. A participant suggested that across the levels of scale. The characteris- researchers would know less about drugs tics of an individual, over time and over and crime than they do now if current development, are modulated by the social human-subject standards had been in characteristics of the peer group, the fami- effect in the past. ly, or society. Thus, there is reciprocity over time between the predispositions of Understanding the roots of misbehav- an individual and the environment. An ior. Dr. Anthony was asked about inte- example is the growth of illegal income, grating the micro and macro levels of or the proportion of annual income earned research, specifically their effects on fund- through criminal behavior, by the young ing, policies, and research on biological men the questioner cited. One might ask, influences on drugs and crime. given one set of regulatory conditions, what those growth trajectories would He responded by characterizing misbehav- look like under another set of regulatory ior as a phenomenon rooted in the origins conditions. of the human species, family heritage, and social structure. The more researchers Dr. Anthony commented further that the understand about these factors, the more paper presented by Dr. McBride noted they can use that understanding to shape that State-by-State variations in drug regu- policies and perhaps foster a more civi- lations give researchers opportunities for lized society. study. He had also discussed with Dr. MacCoun the contrasts between growth Growth trajectories and the ecodevelop- trajectories for marijuana involvement by mental trajectory model. The section of young people in Amsterdam in contrast to Dr. Anthony’s paper that dealt with tempo- young people in a comparable city in the ral relationships and growth trajectories United States. The more ready availability raised the question of the degree to which of cannabis might have an increased interventions affect growth trajectories. impact on the growth curve if the young Changes observed lately in drug markets, people smoke more. With respect to and in the crack markets in particular, the cannabis itself, the trajectory might

167 SPECIAL REPORT / JULY 03

decline if the young people segregate into as an opportune site for public health heavier or lighter users, in contrast to interventions. About 15 to 35 percent of young people subject to the current regu- all infectious disease cases, from HIV latory scheme in the United States. These infection to tuberculosis, passed through a developments would guide researchers correctional institution in the past year. Dr. toward some crossnational research to Anthony was asked whether these issues look developmentally, over time, at the should be part of the public health focus young people’s dispositions. Those dispo- on the drugs-crime nexus. sitions might be to maintain a flat trajectory in drug use, to shift from one drug to He responded in the affirmative, stating another, or to display a declining trajectory that NIDA has an active research portfolio in drug use. in interventions in criminal justice environments and is likely to increase its invest- Turning to the question of which interven- ments in that area. He suggested that this tion would make a difference, Dr. Anthony would be another area in which NIJ and commented that he had started off hoping NIDA could coordinate. to conduct policy analyses in this area but had decided that the observational data Dr. Anthony commented on a point not were not good enough for that kind of included in his paper that relates to the work. He questioned whether the time is pharmacological model for the tripartite right for social experiments that would approach. He thought that in the next 5 to allow researchers to contrast one regula- 10 years researchers will see some inter- tory condition with another. In addition, esting findings from longitudinal studies econometricians have convinced him that of cocaine-exposed children. These stud- there may be problems even with random- ies will demonstrate, he thinks, that it is ized experiments, so researchers may not not the children’s drug use that leads to ever be able to collect definitive evidence their aggressive behavior, executive dys- in this area. Whether a researcher’s stum- functions, or subsequent criminal behav- bling on something might always be bet- ior. Rather, they will demonstrate that it ter than what can be designed in advance is cocaine use by their parents or the is a problem of constructivism. lifestyle associated with cocaine use by their parents. This will be an interesting new line of research and a new way of Discussion: Problematic ideas thinking about that part of the tripartite Mortality and morbidity due to drug- model. use-related injuries and diseases. A question was raised about the usual focus Evidence linking drug use to aggres- of public health research on interventions sion. Dr. Anthony was asked about the to reduce mortality and morbidity. Dr. nature of the experimental evidence link- Anthony’s paper focused on crime as an ing the use of certain drugs to aggression outcome, but other forms of mortality and and whether that evidence is as strong as morbidity are also associated with the rela- the evidence linking aggression to alcohol tionship between drugs and crime. One is use. injury resulting from the violence inherent in the drugs-crime nexus, and one can go He responded that when studied under beyond that to infectious diseases associ- experimental paradigms, the use of drugs ated with use of drugs. It is possible to like methamphetamine, cocaine, and the view the correctional system, where a amphetamines results in aggression under great many drug users are incarcerated, certain conditions. If one looks outside the laboratory and examines comorbidities,

168 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

one of the strongest co-occurrences is perjure themselves, or otherwise break drug dependence and alcohol depend- the law to obtain convictions. ence. Cocaine dependence can be treated, for example, but if the subjects Time-lagged effects and crime. A sec- continue to drink heavily, they will still be ond neglected issue is time-lagged involved in alcohol-associated criminal effects. In applying the tripartite frame- behavior. This is a complex problem in the work, researchers usually think in terms of societal environment, but that complexity crimes that occur relatively soon after the should not blind researchers to the clear drug activity. However, people may also experimental evidence linking certain use drugs, become addicted, drop out of drugs, especially the psychostimulants the labor market, and end up homeless. and drugs like phencyclidine (PCP), to They may stop using drugs, but 2 years aggression revealed in laboratory studies. later they are picking pockets to buy food because they cannot find a job. It need not Among the complexities of this issue is be the case that economic-compulsive that the drugs have different effects at dif- crime means only stealing to obtain drugs ferent doses. That is to be expected and within the next few minutes. does not contradict the causal inference based on effects that might be observed There also are children who suffer be- at specific doses. When violence is ob- cause of abuse or neglect at the hands of served in people who are using PCP, it is addicted parents. Researchers should con- generally seen in those who have ingest- sider not only the “cocaine babies” who ed very large doses. A similar phenome- were exposed in utero, but also children non is observed in methamphetamine who were abused as 2-year-olds and com- users who have been on runs that lasted a mit crimes 20 years later. This issue has weekend or longer. The resulting paranoia, been neglected because of the focus on suspiciousness, and other effects end in the activities of users and sellers proxi- violence. mate to the drug activity. Researchers should take a broader and more holistic Dr. Anthony responded to a comment view of the types of crime they should be that violent drug users have typically also thinking about in drug-related crime re- used alcohol by citing instances in soci- search. A suggestion was made to strike eties where alcohol is not widely used the word crime, because the parent who and instances in which methamphetamine is inattentive to a child, for example, may users who are not drinkers are arrested not cross the line into criminality. for aggressive behavior. Although the co- occurrence of methamphetamine and alcohol use is a palpable association, “Research on Drugs-Crime there are exceptions. Linkages: The Next Drug-related corruption. A comment Generation” was made about important topics that Robert MacCoun, Beau Kilmer, and appear to have been neglected in research Peter Reuter conducted in the 1980s and 1990s. One such topic is drug-related corruption. It is Dr. MacCoun acknowledged the impor- a crime and it is related to drugs, but re- tance of the need to define crime more searchers do not write much about it. broadly and to include the study of corrup- Examples cited included instances in tion in future research. He and his coau- which officials are involved in bribery, thors had focused on street crime and

169 SPECIAL REPORT / JULY 03

proposed some consensus principles on could use such technologies as wiretaps. causal directions that they thought would The other comment was that incapacita- be widely accepted in the drugs and crime tion is largely a function of supply. That is, research community. He also acknowl- eliminating dealers may reduce drug avail- edged that the research reviewed repre- ability, although new dealers may take sented enormous bodies of work. their place. The Federal antidrug effort does not make a large dent in the supply. Discussion: Most helpful ideas One operation conducted in the 1980s that targeted a jungle laboratory seized Victimization and the tripartite frame- several tons of cocaine but had no impact work. Dr. MacCoun offered a clarification on the cocaine supply because the dealers of the victimization issue as it relates to had five or six other laboratories. the tripartite framework. Victimization was initially included in the framework as a Clarifying legalization. Dr. MacCoun subcategory within the category of psy- responded to a favorable comment about chopharmacology. But the concept of vic- his raising the issue of legalization by clari- tims is difficult to establish in the real fying his use of the term. It is very diffi- world. In one study, for example, partici- cult, he said, to discuss alternatives to the pants in 40 percent of the violent events current system because the debate tends were classified as codisputants. The to focus on two models that are at polar researchers could not determine who was extremes: a free market in drugs and a victim and who was a perpetrator. some version of prohibition. A range of possibilities exists between these ends of The four-cell scheme for classifying the spectrum, and in examining European drug markets. Reintroduction of the four- models, researchers are looking at coun- cell design for describing drug markets tries that have legal prohibitions yet are was considered a strength of the paper. In signatories to international agreements on that design, markets are classified accord- drugs. The word legalization must be used ing to whether buyers and sellers live in or cautiously because it implies commercial- outside the area where drugs are sold. ization. The Swiss model, for example, is Although there has not been a great deal an incremental model and is heavily regu- of research on the operationalization of lated, thereby costly to apply. markets and the consequences of each type of market, the approach was re- Rather than studying the issue of drugs, garded as a useful policy paradigm. If crime, and their connections, more time researchers can work with local law could be spent on the connections among enforcement agencies to identify the drugs, crime, and policy and the effect of distribution of those types of markets and their interaction. There are opportunities their locations, they will be one step closer internationally to examine innovations in to helping the agencies implement poli- policy, which by no means constitute cies appropriate for the markets in particu- legalization in the sense of commercializa- lar communities. tion but are nevertheless more substantial than the policy variations typically observ- Drug supply. Two comments were of- ed in the United States. If one accepts the fered about drug supply. One addressed premise that drugs, crime, and policy all outdoor versus indoor markets. The sug- interact, researchers could learn from gestion was that if markets moved in- instances in which policy varies. They doors, the ability of law enforcement could, for example, conduct empirical data agencies to drive down the supply of drugs collection on experiments conducted in would increase because the agencies other countries in an effort to understand drugs-crime-policy links.

170 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Dr. MacCoun offered an example of an Discussion: Problematic ideas opportunity of this kind that has not been addressed. In the 1970s, Italy depenalized Flawed methodology. A general question (that is, removed the penalties for but did about methodology was prompted by Dr. not legalize) personal possession of all MacCoun’s statement that the Swiss drugs that are prohibited in the United experiment is flawed methodologically. Dr. States. Italy maintained depenalization McBride had made a similar point in his until 1990, repenalized that year, then paper, noting that research on interven- depenalized again in 1993. Researchers tions contains methodological problems, could conduct archival research to exam- but that the treatments work neverthe- ine the effects of these policy changes. less. The question for Dr. MacCoun was what policymakers should make of this A participant offered two examples of discussion. other kinds of opportunities for international research. One opportunity exists be- Dr. MacCoun addressed his skepticism cause of the externalities of U.S. drug about interventions, specifically about the policy in relation to drug use by young Swiss experiment with heroin mainte- people outside the United States (in Latin nance. He noted that the Swiss results America, for example). The United States were ambiguous because they lacked true has a global impact, because of what it random assignment and because heroin does domestically. This has been neglect- maintenance was confounded with provi- ed in research that focuses on the United sion of other forms of treatment. How- States. The second example is societies ever, the Swiss experiments demonstrate where the use of intoxicating substances that heroin maintenance is logistically fea- is common and the link to criminal behav- sible and provide at least tentative evi- ior is absent. This presents another oppor- dence for its benefits. tunity for international research. Although A forum participant asked whether the policy variations are more diverse outside Swiss heroin maintenance experiment the United States, other countries lack the serves as a lever for getting addicts into data infrastructure we have here, which treatment and, if so, how the Swiss meas- complicates research efforts. ure treatment outcomes. For example, is Researchers are on the brink of being able success measured strictly in terms of to capitalize on research on the temporal abstinence, or is rehabilitation a positive sequencing of policy interventions. The outcome? Dr. MacCoun responded that as analytic framework for drug policy pits the a researcher he criticizes the confounding various components of policy against each of heroin maintenance with other inter- other in a battle for resources. Conflicts ventions in the Swiss study. However, on about implementation can be found at humanitarian grounds, he might celebrate the Federal, State, and local levels. But that weakness because it implies that the researchers can develop first-order mod- heroin maintenance program encouraged els and simulations to anticipate epi- Swiss addicts to seek other needed demics and collateral problems that may treatments. be associated with epidemics, prevent Failed interventions. The results of a epidemics, address epidemics early on in National Academy of Sciences (NAS) a cycle, and address them later on in a study suggest that researchers have not cycle. This would allow consideration of done a good job documenting which inter- more dynamic policy/resource allocation. ventions work, but they can scientifically document those that do not work. In the

171 SPECIAL REPORT / JULY 03

last quarter of the 20th century, one of the course over which researchers have stud- most important interventions has been to ied this phenomenon is 1.5 cycles, which lock up drug offenders and “throw away is not long enough to attempt a fit with the key.” Research projects that use this any cyclical model. If drug epidemics are kind of intervention as a control almost cyclical, researchers should examine what always indicate better results for treat- has occurred before the next cycle. ment. Foreign interdiction is another failed intervention, as price and purity data indi- Event dynamics of the tripartite frame- cate. Data are also available on 25 other work. The tripartite framework could be interventions that do not prevent youths applied to research on why a drug user from starting drug use or other deviant becomes violent in one circumstance but behaviors. A key element in a research not in another and under what conditions agenda may be to develop a list of inter- violence does or does not ensue during ventions that are popular politically but do episodes of drug use. One study bor- not work. rowed methods from symbolic interaction- ism and game theory to examine drug Successful interventions? At times in transactions. Researchers have tried to recent history, source-country interdiction understand the sequence of potential to has been reported as effective. Examples motivation and motivation to action. In are the Turkish opium ban and the so- other words, they tried to observe a cas- called French Connection. There have also cading effect in which drugs may take the been spikes in prices to which the market user from a stable state to an aroused adapted immediately. These observations state and then to an aggressive state. gave rise to the question of whether data indicate reductions or increases in crime This particular study of drug transactions correlating with the price fluctuations. revealed that drugs have strong psychoac- Impulse-response analyses of these tive effects. The researchers examined events, exemplified by an NAS review, the mediating mechanisms through sub- indicate that little is known about the sug- jects’ own narratives, in which they dis- gested correlations. closed how they would change their behavior in a confrontation. When they Lack of direct measures of deterrence. were under the influence of drugs, they Measurement of the effects of deterrence said, the stakes would rise: They would on retail operations is also lacking. Re- become more boastful, their language searchers primarily use price fluctuations would change, and they would misread as their core index, and that measure is perceptions of danger or the cues from weak. There is no probabilistic sample and another person. This research demon- researchers rely on Drug Enforcement strates that thinking about event dynamics Administration (DEA) data, which are as a framework in which causal factors designed for a different purpose. This calls unfold over time is a promising method for into question the usefulness of studying examining this issue. price fluctuations as they relate to crime in drug markets. Limitations of the tripartite framework. Researchers need to be aware of prob- Are drug epidemics cyclical? The lems in applying the tripartite framework dynamics of the drug scene and the con- as a measurement tool. Although this stantly changing nature of drug markets framework is still important, it was de- raise the fundamental question of whether signed to explain connections between drug epidemics are cyclical. The time drugs and violent crime. In examining

172 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

relationships between drugs and nonvio- Researchers’ impact on State-level lent crime, we need to transcend the policies. Researchers’ results can influ- framework and also use it to understand ence policymakers and State budgets. particular events. With respect to the lat- Delaware, for example, actively tried to ter, multidimensional has been the most introduce drug treatment into its prison important category of analysis because system in the late 1990s. Delaware’s things rarely fit neatly into any of the attempts were closely associated with others. drug courts, which are based on the premise that offenders who have extensive Researchers who apply the tripartite criminal histories and signs of addiction framework also need to go beyond New will commit less crime if their addiction York City to locations throughout the coun- can be halted. By making treatment part try to identify a reliable source of data on of corrections, the State has changed the the drugs-homicide link. The Arrestee definition of crime. Drug Abuse Monitoring (ADAM) instrument contains a drug market section that An example of the way the definition of addresses the characteristics of markets crime can change involves urinalysis con- over time, by location, and by comparing ducted among offenders on probation to indoor with outdoor. Another potential test for drug use. A positive test counts data source is a NIDA-funded international as a technical violation for which proba- study of the psychoactive and sociobe- tioners may be returned to jail. With the havioral effects of marijuana use in Amster- introduction of the therapeutic approach dam, San Francisco, and Melbourne. into prisons, Delaware’s Department of Corrections and judges have become sen- Decriminalization and the link between sitive to the implications of positive uri- drug use and crime. What is the link nalysis results. Although classified as between drug use and nondrug crime? relapses, positive urinalysis results have Because drug possession is itself a crime, become the equivalent of crimes. In this various types of decriminalization (which way, research has affected policy as the are not necessarily the same as free corrections system introduced treatment access) will be needed to sever the link and changed the way crime is defined. between drug use and crime. This point is important from a research perspective Corruption in models of drug distribu- because the illegality of drugs is a con- tion. A comment was made about cor- stant in all research conducted in the ruption in connection with models of United States. Various models address distribution. It is essential, in the view of the implications of illegality, such as the participant who made the comment, corruption and market-oriented violence. that these models address street-level Opportunities for comparative research ethnographic research that found some should be sought in countries that have police on the street to be involved in the decriminalization policies, countries where drug trade to the extent of having a dra- some drugs are part of the culture and matic impact on how the trade operates. used freely, and in historical work on peri- Police corruption plays a dramatic role in ods when drugs were legal in the United the drug trade on both the micro and sys- States. Other research opportunities temic levels. One reason is that public pol- include the study of legal drugs, such as icy sometimes permits seized assets to alcohol, to see how society manages the devolve to police departments on the effects of these widely used substances. basis of their own enforcement activities. A related issue is the way in which police

173 SPECIAL REPORT / JULY 03

affect or do not affect particular drug mar- How DEA data on such elements as price, kets depending on how they choose to purity, quality, and signature information enforce the law. can be folded into programs like ADAM is a key issue. Apparently signature analysis can be conducted as easily on urine speci- Reports and feedback from mens as on actual samples of drugs. If so, it would enable additional information to the roundtable discussions be integrated with the ADAM data. Methods, measurements, and datasets Ethnicity and race There is a need for integrated data collec- Family and genetic issues as well as race tion and better measures. Suggested and ethnicity are important for future areas for further research include identify- study. The Human Genome Project may ing particular local areas for saturation provide data relevant to studying drugs testing of multiple measurement methods and crime. Researchers must do a better and determining how current measures job of articulating the importance of these overlap. In the National Household Survey factors. The public should not be led to on Drug Abuse (NHSDA), data are gath- believe, for example, that there are single ered from the general population, but genes related to poverty or violence. The other data collections, such as Monitoring need for more attention to diversity within the Future and ADAM, are more narrowly and among ethnic and racial groups was targeted. We do not fully understand how recognized as a priority. Acknowledging the methods used in these surveys over- that caution must be used in analyzing lap. The ADAM program has attempted to the concept of race in scientific research, include questions in the survey instrument participants recommended the study of that enable analysts to link ADAM data to ethnic variation, learned behavior, and data in other surveys. culture. The addition of discussions of culture to the three forum papers was also The validity of self-reports of both crime recommended. and drug measures is a significant issue for research. Such issues as subjects’ We can expect diversity in drug use recall and telescoping are addressed in among different groups, such as Hispanic parts of larger studies, but funding en- Americans and blacks. An example of ables the testing of the validity of self- diversity within groups is the drug use reports in only a few studies. The wording patterns of Mexican Americans in the of questions, the order in which they are Hispanic American population. Diversity asked, and the effects of the urinalysis raises complex issues that involve itself (in the ADAM program) are possible conceptualization in measuring environ- areas of investigation. Experiments involv- mental circumstances, conditions, and ing random assignment of survey ques- processes. These issues affect research tions offer an opportunity for study as on ethnicity and on gene-environment does altering the sequence of urinalysis interactions. The research community and survey administration. Interviews with must address the sensitivity of the combi- study subjects might produce different nation of the issues of genes and race, results depending on whether the urine possibly through formation of a NIDA-NIJ specimen is taken before or after the working group. interview.

174 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Several specific issues or phenomena drugs, such as economic, social, and involving ethnicity and drugs offer oppor- health policies. NIJ and NIDA could con- tunities for study. An example is the sider commissioning a series of multidisci- National Development and Research plinary review papers that focus on the Institutes study of the “blunt generation,” potential impact of policies on outcomes. which revealed that black youths in New The policy community would be the target York City are shifting from crack to mari- audience. Currently, no mechanisms are juana and tobacco. This research can be available to examine policies at the Fed- used as a model to study whether the eral, State, and local levels. We need same transition is occurring outside New a database that would enable researchers York City. to examine variations in policy among jurisdictions. Researchers should be mindful of the factor of religion because, like ethnicity, it has Another important and related area for different levels of importance for different research is the development of models ethnic groups and a bearing on criminal that transcend econometric models in behavior as well as drug use. Another area examining the impacts of policies on out- for study is the differential effects of comes. Researchers have not developed methamphetamine on different ethnic models that examine the impact of public communities. There are few examples in policies on behavioral outcomes and the which blacks are represented among relationship between drugs and crime. either methamphetamine users or casual- Input from practitioners about the impacts ties of its use. However, research has of policies on their constituents may revealed users and casualties among advance this line of investigation. members of other ethnic groups, particularly Pacific Islanders. Drug markets Two research design issues were raised in How do researchers define and measure the roundtable discussion. One had to do drug markets? Among the issues consid- with the false belief that there is variation ered in the roundtable discussion were by ethnic group in the extent to which the usefulness of such measurements information from self-reports differs from and the benefits to law enforcement agen- bioassay results. Recent research might cies from this kind of research. The dis- be developing evidence that will contradict cussion covered how researchers might some of these false beliefs. The other measure the harmful effects of drug mar- involved IRBs, protection of human sub- kets and how to detect changes in those jects, and the differences in the confiden- harmful outcomes over time. For example, tiality certificates issued by NIJ and the how would researchers compare the U.S. Department of Health and Human effects of crack markets that proliferated Services (HHS). Discussions between NIJ 10 years ago with the effects of the blunt and HHS about checks and balances in the generation today? Nonharmful outcomes IRB processes would allow researchers to and the need to examine how they change learn more about pressing social problems over time were also considered. Social that do not necessarily fit the HHS bio- control mechanisms operating in markets, medical research model. the question of whether market stability is desirable, and health issues associated Policy issues with market stability were also suggested as topics for research. Researchers need to examine policies other than those specifically directed at

175 SPECIAL REPORT / JULY 03

Treatment in the criminal behavior starts reverting to baseline when justice system rewards stop.

The treatment roundtable focused on two The impact of welfare restructuring on the measurement issues and two potential drugs-violence nexus is also a topic for interventions. The first measurement future research. Some inmates were issue was dropouts. When drug users receiving Medicaid benefits, which they undergoing treatment drop out of the pro- were using to pay for drug treatment. gram, this affects any evaluation under Since they no longer receive Medicaid, way because the numbers change. Other corrections-based treatment plays a larger disciplines have dealt with the problem by role. Outcomes other than refraining from using econometric and other statistical drug use, such as payment of child sup- techniques. The participants thought port, family formation, employment stabili- some of these tools should be brought ty, and residential stability, may also be into the drug treatment literature. useful as indicators that a former prisoner has addressed problems associated with They also recommended comparing the drug use. effectiveness of different types of treatments used in the criminal justice system, In assessing the effectiveness of treat- an undertaking for which there is currently ment programs, the fundamental problem no common measure. The Addiction for researchers is obtaining the kind of Severity Index (ASI) was discussed, but post-treatment and postrelease data they using the ASI poses problems because of need. A study under way in Florida is the “past 30-day” questions it includes. examining this issue in a nonprison treat- Opportunities provided by new technolo- ment setting where researchers have gies for detecting drug use, such as hair access to measures of criminality and testing and sweat patches, were also other data. However, these measures may considered. not enable the researchers to effectively differentiate between the treatment pro- Contingency management is an interven- grams in which the subjects were en- tion examined in the Greenwood study, rolled, which include most programs in which revealed that paying students to fin- Florida. ish high school is cost-effective. If this approach is applied to encourage treated The reporter for the roundtable responded prisoners to receive more treatment or to to a followup question about the possibili- refrain from using drugs once they leave ty of requiring treatment providers to track prison, it may offer opportunities for data. He noted that the mandate for the research. Florida study came from the State legisla- ture, which requires evidence that money The problem of treating drug-using offend- spent on treatment produces a result that ers after release from prison is another is more economically valuable to the State intervention issue that could benefit from than the current expenditures. However, research. In addition to using contingency treatment providers cannot respond very management, some States hire case man- effectively to the legislators’ mandate agers to encourage prisoners to continue because they do not have the resources receiving treatment. Texas, among other to track all the data. States, makes such additional treatment a condition of parole. Research indicates The facilitator told forum participants to that contingency management has not imagine a situation in which they are ap- worked well in the long run because proached by a philanthropist who claimed

176 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

that he could raise large sums of money if with a defensible number, and tell him researchers could develop agreed-upon what it would cost to net 50 successes measures of outcomes that were valuable from the 100 released offenders, he and whose cost was lower than the cost would raise the money for those 50 peo- of producing the results. The participants ple. He would have to know and be able were challenged to find out if they could to tell his donors, however, the numerical do so with the interventions they had value of those 50 successes. tested. One participant who responded comment- In response to the challenge, one partici- ed that for almost any intervention, re- pant noted that in some studies of the searchers could produce a calculation valuing of drug abuse treatment out- indicating that the resultant number is bet- comes, people who are not drug users are ter than doing nothing, although there asked how much it is worth to them to probably is no drug control intervention for live in a drug-free and crime-free commu- whose effectiveness researchers could nity. It is not possible to put a monetary provide definitive proof 20 years hence. value on such issues. Other studies exam- This approach, however, is not a construc- ine outcomes like abstinence or a range tive way to make practical managerial of outcomes involving improvements in decisions and is not the way that busi- health, social functioning, and criminal nesses, for example, think about such behavior. To provide guidance on improv- matters. Another responder pointed out ing drug abuse treatment at the program that a similar question is not asked about level, NIDA has tried to shift the focus of dialysis treatment for end-stage renal dis- the research to what the program does, ease or liver transplants for people with what is unique about it, how it is organized cirrhosis, although they might persist in and managed, and what is unique about behaviors that promote their diseases. the treatment delivery system. Another Thus, researchers are imposing a standard responder stated that if the philanthropist on drug treatment that they do not impose could find a way to support graduates on other medical treatments. from drug treatment programs who live either in prison or in the community and are otherwise unemployable, guarantee “The Drugs-Crime Wars: them jobs, and assure them of an income of about $20,000 per year, there would Past, Present, and Future be much better results than those that Directions in Theory, Policy, researchers are seeing now. and Program Interventions” In a followup scenario, the philanthropist is Duane C. McBride, Curtis J. VanderWaal, prepared to supply the money for whatev- and Yvonne M. Terry-McElrath er it takes to produce a graduate of a treatment program who was formerly a drug Current drug policies have not always user and in prison and to assure him or been in effect and may not always be in her an income of $20,000 per year. The place, and some policy changes have question for researchers is what is the been dramatic. Dr. McBride discussed the value to society of 50 of 100 people leav- historical context in which drugs-crime ing prison, acquiring job skills, and earning relationships should be examined. In the $20,000 per year 2 years after release? If 19th century, drug policies in the United researchers could specify for the philan- States varied enormously. Distribution was thropist the value to society, backing it up relatively open: Imports were regulated but domestically there was some access

177 SPECIAL REPORT / JULY 03

and even commercialization. Drugs, nee- Collaboration among agencies. As dles, and syringes were available through States develop comprehensive systems the Sears-Roebuck catalogue. In reaction to address drug problems in their criminal to this openness, many States began to justice systems, collaboration among heavily regulate drugs. Officials from agencies becomes more crucial to policy some States complained that other States formulation. For example, in the California openly sold drugs that they themselves Department of Corrections’ treatment sys- were trying to regulate. The labeling of tem, which currently has 7,000 beds, the drug content was instituted and States treatment providers and the corrections made many changes in their laws and unit that operates the system have been policies. meeting regularly for years. Parole officials began attending the meetings only in the Discussion: What ideas from the past 6 months, however. Parole is an paper are most helpful? essential policy element because the program includes an aftercare component. Cause and effect in the drugs-crime relationship. The section of the paper Using graduated rewards and clients’ dealing with the cause- effect relationship strengths in drug treatment. Sug- of drugs and crime suggested an interest- gestions were made that graduated ing line of inquiry. The forum discussions rewards, as well as the graduated sanc- had looked primarily at the relationship in tions mentioned in the paper, should be terms of drug use preceding crime. From studied for their use in drug treatment that perspective, interventions were as- programs. Further, more attention should sessed on the basis of the effects they be paid to the clients’ strengths in addition might have on drug-related crime that to their needs, problems, and resources. immediately follows drug use. There was Dr. McBride agreed that inclusion of a no consideration of the early antecedents strengths-based case management sys- of drug use and criminal behavior. tem is crucial to treatment.

Study findings on the antecedents of drug Comorbidity issues. The comorbidity use and criminal behavior reviewed in Dr. issues covered in Dr. McBride’s paper McBride’s paper, as well as research con- were considered relevant to the forum dis- ducted by Dr. Anthony on aggression in cussions of policy and of treatment in the first-grade students, were cited as exam- correctional system. As a result of high ples of areas where further research is rates of comorbidity and of deinstitutional- needed. Evaluation of classroom interven- ization in the mental health field, some tions revealed that addressing conduct prisons are the major mental health serv- and aggression problems reduced the risk ice providers in large urban counties. This of future drug use, which suggests that situation influences the effectiveness of deviant behavior may precede drug use. treatment in correctional settings, and is a situation in which some medical care History of drug policy. Participants identi- providers feel more like law enforcement fied the attention given to the history of personnel. drug policy as a strength of Dr. McBride’s paper. The conflict between the puritan Ballot initiatives and research. Ballot and libertarian traditions, which he cited, measures such as California’s Proposition is played out in current drug policy 36 address drug policy, and are frequently discussions. supported by advocacy organizations that are also interested in research. Social scientists should try to gain currency with

178 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

these organizations and open an avenue that should be addressing such policy- through which the research community driven issues as alternative enforcement can examine these policy experiments and strategies. their outcomes. Researchers previously have not made strong connections to Publishing policy research. Studies of those who propose public policy reforms Proposition 36, changes in the Rockefeller from the perspective of the political right, drug laws, or the Swiss heroin experiment and the same may now be true for the do not have perfect control groups and political left. random assignment of subjects because they examine real-world situations. This Health versus criminal justice research may limit researchers’ ability to publish in funding. The Robert Wood Johnson the better journals. In addition to providing Foundation is funding a study of Proposi- funding, NIJ and NIDA could increase the tion 36 that may become a model for demand for policy research by fostering reporting that could affect public policy. publication outlets. The Robert Wood Johnson Foundation focuses on health issues, and its sponsor- To obtain funding, grant applicants are re- ship of this project indicates that the quired to address scientific design issues. criminal justice research community is Poor designs submitted to the National underfunded because a health funding Institutes of Health (NIH) may not be organization is implementing portions of acceptable to epidemiologists, but the the criminal justice research agenda. New England Journal of Medicine has published comparative studies of the Building collaborations between NIDA and impact of handgun regulations on homi- NIJ to study drug enforcement would be cide and suicide rates in Vancouver and an important part of a future research Seattle. Dr. McBride cited these studies, agenda. Such collaboration could address which compared different populations, as major policy issues, such as variations examples of flawed designs that would among States in the intensity of drug en- not have received NIH funding but were forcement and how strongly they enforce nonetheless published in a quality journal. drug prohibitions. Funding for secondary data analysis. Dr. Researchers could also evaluate the McBride’s paper was praised for its list of effects of different kinds of enforcement, suggestions for future research, particular- but would need to identify appropriate ly because of the proposal that secondary outcome measures to do so. State-level data analysis could provide a new empiri- measures are being developed for NHSDA, cal baseline for study of the drugs-crime but would be inadequate for these purpos- relationship. Securing funding for the es because they focus on the prevalence analysis of NHS data, to assess the extent of addiction. The ADAM sample frame is of drug use or the gateway model has not suitable for this type of project, which been difficult. The only sources of funding would address how enforcement affects for analysis of ADAM data have been drug use. The relationship between drugs small grants from NIJ or organizations like and crime, in and of itself, is not as useful the Robert Wood Johnson Foundation. as is research that will inform drug policy. Funding agencies spend large sums of NIJ is primarily a policy research agency money supporting new data collection and relatively small amounts supporting secondary analysis.

179 SPECIAL REPORT / JULY 03

Discussion: Problematic ideas morality on the other resulted in a lack of historical background needed for under- Policy implementation and evaluation. standing the current situation. Missing ele- Without effective enforcement and imple- ments include the harm that drugs cause, mentation, it does not matter which poli- status battles among people who want cies have been adopted. With respect to their moral beliefs adopted as official poli- tobacco, there was a great deal of policy- cy, and the agendas of interest groups. making at the State level, but until there was enforcement, the policies did not In citing an example of the effects of make a difference. Dr. McBride’s paper interest-group issues, one participant suggests that policies directed at club suggested that if asset forfeiture laws drugs (for example, changes in metham- changed so that seized assets were spent phetamine penalty structures) offer impor- on drug treatment rather than enforce- tant research opportunities. Researchers ment, the statistical portrait of drug use have an opportunity to evaluate the might change. A better understanding of effects of these laws and policies from the historical roots of current policies both the criminal justice and the public should be included in the research com- health perspective. munity’s policy research agenda. Also suggested for inclusion in Dr. McBride’s Dr. McBride noted in his paper that model paper were more material about the racial laws developed by the National Alliance dimensions of Prohibition, its 19th-century for Model State Drug Laws have not been roots, and its current manifestations; and a examined for their effectiveness. This indi- reference to Tonry’s Sentencing Matters in cates a need for studies of implementa- the paper’s discussion of mandatory mini- tion and enforcement. One participant mum sentencing. suggested that the forum should be open to the possibility that not implementing Continuum-of-care treatment models. current laws might be advantageous in Studies of drug addiction as a chronic dis- some situations. order have implications for treatment models, such as continuum-of-care programs. References in the paper to the moral ten- Current research has established the need sions surrounding drug policies illustrate for continuum of care, and future research how values affect assessment of those could systematically address the elements policies. An example of those tensions is of a continuum-of-care model rather than the different standards of evidence used considering adaptation of current models. in assessing new pharmaceutical products and in evaluating controversial new drug In order to provide good continuum of policies, such as those based on relaxed care, medical and social services need to enforcement. There is also a reciprocal be linked. This would involve coordination relationship between drug policy and drug among social service agencies, public use, because the public, to whom laws health agencies, and corrections or other and policies are directed, includes the vot- criminal justice agencies. Issues concern- ers who elect the legislators who in turn ing the reintegration of treated drug users make the policy. Policy research must take into the community should also be ad- into account that the consumers of policy, dressed in future continuum-of-care or the public, also influence policy. research.

Historical roots of current policies. In Computer simulations. Dr. McBride sug- Dr. McBride’s paper, the juxtaposition of gested that researchers start thinking libertarianism on one side and puritan about computer simulations. A participant

180 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

interpreted this to mean creating broad marijuana, for example) indicates that this models with many parameters, which behavior is normative for adolescents. would produce many research questions These drug-using adolescents do not com- that could be used to generate useful poli- mit many crimes other than using the ille- cy analyses. The drugs-crime research gal substance. If the drug use continues field would benefit from a macro effort in as they grow older and they also move on multiple places, with multiple perspec- to using harder substances, problems with tives, which would examine policy con- other kinds of crime are then observed. cerns with research backing. Modeling would stimulate further work in all areas The role of the family in shaping be- of drugs-crime research. havior. Another topic not adequately addressed was the critical role of the family A comprehensive surveillance system. in shaping behavior. In developmental psy- In his paper, Dr. McBride did not address chology, interventions are family-based, the need for a comprehensive surveillance rather than broad-based population inter- system that would enable researchers to ventions. (An example of the latter is detect when peaks in drug-related vio- keeping offenders in treatment.) The im- lence begin. Such a system would allow portance of interactive relationships within researchers to study the peaks as they the family was illustrated by research on form and also understand why they form. children who have attention deficit hyper- A big peak of violence occurred in the activity disorder (ADHD). When children United States in the late 1970s and early with ADHD are medicated, the behavior of 1980s, and another occurred in the late the parents changes, even if that of the 1980s and early 1990s. Between those children does not. peaks was a valley, and a very deep chasm began in the 1990s. Changes in heritability. Quantitative- genetics research projects, such as Many explanations have been offered for twin-sibling studies, have revealed that the peaks, such as changes in drug mar- changes in heritability occur over time kets, incarceration rates, community polic- and with respect to gender. Thus, if re- ing, and enforcement of quality-of-life searchers identify a gene associated with offenses. Ethnographic field stations in substance use, the association may not high-risk communities, enhanced data col- hold for all age groups or populations. In lection by police, and a study of medical addition, the concept of a single gene examiner data were proposed as research determining complex behaviors like sub- topics in this area. These enhancements stance use is unsound. should be incorporated into a routine surveillance system that would facilitate Missing data estimation. Many re- study of the next peak in violence. Such a searchers are working with techniques system would allow analysis of qualitative for estimating missing data. These tech- observations in conjunction with overdose niques may produce results more repre- and arrest data. sentative of reality than those produced by other methods. Biostatisticians use miss- Use of mild drugs is normative for ado- ing data techniques to model longitudinal lescents. Developmental factors that data, such as tracking youth drug use and affect drug use and its relationship with transitions through different stages of crime had been alluded to in previous drug use. For example, researchers do not forum discussions, but had not been a assume that dropping out of a study is a topic of discussion. Developmental litera- random phenomenon. They try to account ture on use of mild drugs (alcohol and for it by modeling the dropout process.

181 SPECIAL REPORT / JULY 03

Many standard statistical methods are spousal abuse on drug use and drugs- built on the assumption that missing data crime interrelationships. The original state- are random, thus young people who drop ment was amended to eliminate a clause out would be no different from those who on intergenerational effects of abuse on stay in the study. However, in studies of crime and drug use because these effects drug use, dropouts often may be incarcer- have not been sufficiently characterized. ated, in low-income families, or moving. We know that trends in drug use do not parallel rates of incarceration. The number What do and don’t we know of people in the United States who are incarcerated has tripled since 1983. There about the relationship was no consensus on whether most of between drugs and crime? the increase was due to enforcement of Reaching for consensus drug laws. The statement was amended to reflect the consensus that we do not The facilitator led a consensus-building know whether the increases in incarcera- exercise in which statements by forum tion have resulted in decreases in drug participants about the drugs-crime relation- use in particular or crime in general. ship were accepted, amended, or rejected. Decisions were made according to an We know that drug use is neither a neces- iterative process; only ideas on which sary nor a sufficient cause of nondrug there was unanimous agreement were criminal behavior. The statement was approved. The statements concerned accepted without amendment. either what is known or what is not known about the relationships between drugs We know that we need longitudinal data and crime. to sort out the relationships between drugs and a host of other causal factors. The participants could not arrive at a con- Consensus statements on what sensus on a statement that cause-and- we know about drugs-crime effect questions can be addressed only by relationships using longitudinal data. It was modified to The complexity of drugs-crime relation- read that longitudinal studies are important ships is widely accepted and means that in making cause-and-effect statements. the research tools we have been using to This modification was not accepted, how- study that relationship cannot get us very ever, and consensus was achieved only on much further in the next decade. The the need for using longitudinal data to elu- complexity of the drugs-crime relationship cidate relationships between drugs and was accepted as fact. Forum participants many causal variables. arrived at a consensus on the need for new tools, which would include both para- We know that illegal drug choices tend to digms and methods, if future research is vary with social position. This proposition to elucidate the relationship. The belief began as a statement that illegal drug pref- that 10 more years of the same type of erences tend to vary with social position. research currently being conducted would A participant objected, saying that what is not advance the understanding of the available in different markets may deter- drugs-crime relationship was disputed mine what options are available to be pre- because some progress has been made. ferred. The statement was accepted after “preferences” was changed to “choices” We know that we need to know more and the meaning of the phrase “social about the effects of child abuse and position” was clarified. Social position may

182 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

be affected by, but is not synonymous among policymakers and the forum publi- with, either social class or ethnicity. cation will be a means for doing so.

We know that we urgently need local as We know that a person’s drug taking well as national data (on drug use pat- makes him or her more likely to be either terns) to augment the utility of those data a victim or perpetrator of a crime that for practitioners. A participant proposed otherwise would not be committed. The that data on drug-use patterns must be statement was accepted without local to be useful because ethnographic amendment. and ADAM data show differences by site. For example, methamphetamine use is We know that incarceration of drug not a problem in most U.S. communities, sellers is in large measure offset through but it is the major drug problem in Hawaii replacement by other sellers. As originally and southern California. The audience for proposed, the statement read that incar- data may affect their usefulness; thus, for ceration of drug sellers leads to recruit- example, national data may be useful to ment of replacements for those sellers. Federal policymakers. Crossnational data There was no consensus on whether the may also be useful in addressing some replacement phenomenon is known to issues. occur or not. One suggestion was to modify the statement to read that a conse- Modifications of the consensus statement quence of incarceration of drug sellers is to the effect that drug-use data need to be recruitment of replacements. This state- local to describe local markets, to be use- ment was considered too weak because ful to practitioners generally, or to be use- there could be many consequences. ful to local practitioners were all rejected. Another proposal was to modify the state- Local data were deemed useful in detect- ment to read that the incarceration of drug ing emerging trends in drug-use patterns. sellers is in large measure offset by the The idea that trends in national data, such recruitment of additional sellers. The word as the decline in drug use identified by the “recruitment,” which implied that superior Monitoring the Future Study during the sellers were seeking replacements, was 1980s, were not useful was seen as im- eliminated in favor of “replacement,” and plausible. The statement was amended to the amended statement was accepted. reflect a need to augment national data with local data to maximize policy and We know that the interdiction efforts that practice utility. The statement was amend- have been conducted over the past 20 ed as shown above and accepted by the years have not achieved their goal of subgroup. stantially reducing street-level access to drugs. As originally proposed, the state- We know that given arrest for a drug ment read that interdiction has had mini- offense, an African-American is more likely mal effects on the availability of drugs at than a white American to be prosecuted; the street level. Researchers do not have and given conviction, an African-American a sense of the flow of drugs from source is more likely to be incarcerated and for a countries through transit and arrival zones longer time than a white American. These to markets. Since there is little empirical official results do not accurately reflect the evidence of how much drug traffic is actu- racial differences in involvement with illicit ally interdicted and how interdiction has drugs. There was general agreement that affected market availability, the statement this is known to the research community was amended as shown above and but not to society generally. There is a accepted. need to disseminate the information

183 SPECIAL REPORT / JULY 03

Consensus statements on what We do not know about community in- we do not know about drugs- volvement with and orientation toward crime relationships drugs, or how and why those change. “Community orientation” means what the Long-term, intergenerational studies are community thinks of the issue: For exam- needed to generate and test causal ple, whether the community cooperates hypotheses about drugs-crime relation- with or opposes the police. The communi- ships. A generation of grandparents as ty is divided into subgroups and subsys- well as parents has been heavily involved tems that do not interact well. However, with illegal substances and these people New York City today enjoys a collective now have children or grandchildren. Thus, consensus reflecting dislike and intoler- there are intergenerational subjects who ance of heroin injection, crack sales, and could be studied, but such studies would crack use. This consensus was absent a not necessarily establish cause and effect. decade ago. Inner-city youths in particular A suggestion was made that intergenera- routinely register disgust at intravenous tional data are needed to evaluate cause- drug use and drug selling. and-effect statements and that they could be used to generate causal hypotheses. We do not know the effect of street en- The proposed statement was amended forcement on drug market violence. The accordingly and moved to the do-not-know proposed statement, “We do not know category. whether some forms of street enforcement actually increase drug market vio- We do not know whether genes interact lence,” was accepted as amended to read with the environment to make people as shown here. The phrase “some forms more or less prone to illegal drug use or of” was removed. addiction. In the proposed statement, “We do not know how genes interact with We do not know how best to match treat- the environment to make people more or ment approaches to the individual needs less prone to illegal drug use or addic- of offenders. As originally proposed, the tion,” use of the word “how” implied that statement read that researchers do not there is in fact a relationship between know how to determine which type of genes and drug use. The statement was drug treatment is appropriate for which therefore amended and accepted as type of drug-abusing offenders. It implied shown above. that there is always an appropriate treatment. The person who proposed the We do not know the extent to which the statement responded that most studies decline in rates of violence in the 1990s do show a length-of-time-in-treatment was related to changes in the crack effect, regardless of type of treatment. cocaine market. In the proposed state- Another objection cited the many drug- ment, “We do not know the extent to dependent and alcohol-dependent persons which the decline in rates of violence in who mature out of their addiction without the 1990s was related to maturation of the formal treatment: Researchers do not crack cocaine market,” there was a risk of know why this happens. Many people tautology in conveying the notion that a arrested as dealers falsely claim that they market is mature if it is no longer violent. are users and are offered treatment. In The statement, amended to replace these cases there clearly is no appropriate “maturation” with “changes,” was treatment. The statement was amended accepted as quoted above. to propose matches of treatment services or approaches to the individual patient and was accepted.

184 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

In the aggregate, we do not know if in- We do not know how the different policies creases in incarceration have resulted in implemented in various jurisdictions have decreases in illegal drug use by the per- produced different outcomes. Policies sons incarcerated. The initial statement, vary nationally and by State, and research- “We do not know if increases in incarcera- ers need to know how those variations tion have resulted in decreases in illegal produce different impacts. drug use or crime,” was considered too broad. It appeared to mean that there is We do not know what etiologically differ- no class of persons researchers can de- entiates drug-using offenders from other scribe for whom incarceration results in a offenders. The statement was accepted decrease in subsequent criminal behavior as representing the group’s consensus. or illegal drug use. Specific deterrent effects reported in the literature would Statements for which consensus contradict such a statement. If the state- was not achieved ment were more specific, referring only to particular deterrence effects for drug A statement to the effect that enforce- use, it could be true. The statement also ment alone will not reduce drug use or appeared to be an assault on U.S. policy, related crime was rejected as uninterest- which is to “lock them up and throw away ing because few people claim that only the key.” enforcement is effective and treatment does not matter. The statement also failed A proposed modification narrowed the to consider how much effort and re- statement to make it read that researchers sources would be applied to the problem. do not know if increases in incarceration Given enough resources, law enforcement have resulted in decreases in illegal drug agencies could reduce drug-related crime. use by those incarcerated. It was intended to include postrelease drug use by people The group rejected an assertion that crack incarcerated and then released without sales/illegal transactions are among the treatment. There was general agreement most common offenses in the United that outcomes for individual drug users States, although the assertion was sup- cannot be predicted, but in the aggregate, ported by calculations indicating that they postrelease recidivism and relapse rates swamp other kinds of felony offenses. A return drug use to roughly the levels it had participant pointed out that each sale pro- been before incarceration. The statement, duces at least one incident of drug use, so as clarified and amended, was accepted by definition there are at least as many as reflecting the group’s consensus. cases of crack use as there are of sales. Another participant cited research in which We do not know enough about the co- crack metabolites were detected in ADAM occurrence of alcohol and other drugs in samples. The data show that roughly 90 the drugs-crime relationship. There is a percent of the cocaine-positive urine great deal of statistical evidence for this specimens were positive for crack. If kind of co-occurrence, but a lack of under- researchers could estimate the aggregate standing of its effects on criminal behavior. number of crimes, especially felonies, then the crack sales would probably We do not have accurate price or sellers’ swamp even thefts, and robberies would income data for illegal drug sales. The be negligible relative to the number of statement was accepted without crack sales. Marijuana sales would not amendment. swamp crack sales because many crack users engage in 5 to 10 transactions each day. Other participants disputed these

185 SPECIAL REPORT / JULY 03

contentions, citing work indicating that the Effective, ineffective, and promising entire universe of cocaine-related transac- policies. Researchers should conduct tions would not account for the supposed studies that document policies that work, large number of crimes. that do not work, and that show promise. Long-term incarceration was identified as Statements asserting that cause and the most significant policy among those effect can be determined only by using that researchers believe do not work. longitudinal studies, that current beliefs about the effects of drug policies are A multicity, multimethod surveillance mainly expressions of ideological prefer- system. Prospective approaches that ences rather than scientific evidence, and combine ethnographic observations with that development of low-toxicity substi- arrest, drug pricing, and health data could tutes for marijuana and alcohol are pre- be employed by using field stations in cluded by the Schedule I requirements in high-risk communities. An ongoing multi- Federal law were also stricken. city, multimethod surveillance system should be set up. It would focus on drugs and crime by using police data, medical Discussion of areas for examiner data, and public health data (such as those documenting overdoses, future research deaths, homicides, and HIV and other sex- ually transmitted diseases). A suggestion Discussion: What research in that the crime data be limited to homi- this area do you think is most cides (because the researcher can identify important? the endings of cases) was rejected be- The following topics are areas of research cause researchers would want to track that the forum participants considered the changes in drug-use patterns in various most important objectives for future study. places, including those where few homicides occur. Long-range cost-benefit analyses. Long- range cost-benefit analyses of policies on Ethical implications of genetics re- drugs and crime are needed. Such work search. The ethical implications of policies would address various interventions, in- based on genetics research in the areas cluding those that have already been of alcohol, drugs, and crime should be attempted (such as incarceration). In- investigated. Scientists have not been clusion of policy simulation studies was responsible about addressing the ethical suggested because they could be used to implications of their research; they should simultaneously produce cost-benefit analy- be proactive about the issues raised by ses as well as many other insights. genetics research. An agenda or process for bringing experts together to produce Secondary analyses of ethnographic a consensus on ethical issues was data. Secondary analyses of ethnographic recommended. databases from different cities should be conducted to examine data collected dur- An example of such proposed research is ing the peak years of the crack markets. identifying links between genetic suscepti- These studies should include comparisons bility to drugs or alcohol abuse and various among drug markets in different neighbor- outcomes in the criminal justice system. hoods or cities as well as prospective Researchers need to think in advance studies describing the criminal justice and about what the policy implications might public health impacts of illicit drugs on sell- be, and the ethical implications of those ing and using communities. developments should be discussed.

186 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Drugs, mental illness, and crime. More no-treatment arms. However, there may work should be done on the relationships creative, equitable approaches to these among drugs, mental illness, and crime kinds of studies, such as conducting as well as the appropriate interventions. research in locations where scarcity of Cognitive dysfunctions were also suggest- treatment slots and randomization may ed for study, making this a proposal to provide a fair way to allocate treatment. conduct research on mental illnesses, cog- Random assignment to further treatment nitive dysfunctions, and drugs (in combina- for previously treated subjects was sug- tion) and their relationships with crime. gested as an ethical approach to controlled studies of the effects of length of The structure of drug marketing in eth- time in treatment. nic communities. Comparative research on the structure of drug marketing and its Alcohol and marijuana: Complements implications in different ethnic communi- or substitutes? Meta-analyses that ties should be conducted at multiple sites. assess whether alcohol and marijuana are One of the implications to be studied is complements or substitutes would be the extent to which drug marketing results useful for modeling policy alternatives. in penetration of sales into the ethnic communities. Drug “consumer price index.” Metho- dological research on a so-called con- Operational research to improve sumer price index for illegal drugs was treatment-outcome studies. Opera- suggested as an important research proj- tional research should be conducted to ect. The proposed index would cover retail bring more rigor to therapeutic-justice, and wholesale prices and would comple- treatment-outcome studies. Estimates ment the DEA’s System to Retrieve indicate that up to half of hardcore drug Information from Drug Evidence (STRIDE) users are nominally in the criminal justice database. Reorganization of current data system, either as parolees, as probation- collection or expansion of STRIDE through ers, or in pretrial release status. This sit- random drug purchases may contribute to uation has significant implications for production of an index. national policy and budget. However, the relevant research literature is inadequate Survey policymakers. A survey of because the programs vary significantly in Federal, State, and local policymakers to characteristics and eligibility requirements; assess their research needs was pro- many studies measured recidivism rather posed. The survey would focus on ques- than relapses, limiting their utility; and tions that need to be answered in order to some studies are based on nonrandom- make better policy decisions. The survey ized comparison groups, which results in could also help build bridges between the data interpretation problems. policy and research communities.

Randomized studies employing no- Relationships between distributors and treatment arms are freighted with ethi- consumers. Research on changes in the cal, legal, and analytic problems that must relationships between drug distributors be addressed within the context of the and consumers was proposed. The atti- Code of Federal Regulations (CFR). Under tudes and orientations of distributors CFR, with its minimal risk requirement, it would be investigated with an emphasis might not be practical or feasible to con- on how distribution affects consumption. duct these experiments because international review boards (IRBs) will interpret Event dynamics in drug markets. The mini-mal risk stringently with respect to event dynamics in drug markets should be

187 SPECIAL REPORT / JULY 03

studied with an emphasis on the role of how people with different susceptibilities peer groups and associations. Peer asso- function in different environments. ciations may vary in different ethnic communities and thus affect drug-selling Drugs have certain physiological effects behaviors. The influence of ethnic commu- and there probably are certain genetic pro- nities may or may not affect sales in those clivities affecting susceptibility to those communities and could be instrumental effects. However, the consequences of for marketing in other communities or those proclivities differ according to inter- cities. actions with the environments in which the drugs are used. The effects on crime The market research could focus on ana- and other behaviors may vary in different lyzing when drugs cause delinquents to communities and subcultures within the engage in crime, rather than analyzing the larger society. There may be stigmatization actual buying and selling events. This and other consequences that result from approach would emphasize the drugs- interactions between people’s genetic crime relationship rather than the buyer- makeup and the environments in which seller relationship. Drug-selling peer they live, but people with similar drug groups observed at different times have genes may respond differently in different been observed first encouraging, then environments. subsequently discouraging their members’ violent activities. This indicates that peer Not all drug users need treatment. relationships affect the types and frequen- Allocation of limited resources should be cy of crime. based on research that examines which drug users truly need treatment rather Extending the focus of the proposed than those whose behavior should be research beyond cocaine to, for example, addressed through law enforcement. the marijuana market, was suggested. Researchers should study methods to Marijuana use has been an epidemic for identify users who require treatment as a 30 years in the United States but little is way to avoid the negative social conse- known about how it is sold. quences of drug use.

Effects of felony disenfranchisement on Drug hackers. Researchers need to inves- minority communities. The political and tigate a growing group of sophisticated social ramifications of felony disenfran- drug users and the more specialized sub- chisement laws, which are driven by large stances available for their consumption. numbers of drug-related felony convic- The cocaine problem may diminish sub- tions, should be examined for their effects stantially as more pharmacologically savvy on minority communities. drug users become more numerous. Amphetamine and barbiturate use has Effects of genes and the environment become commonplace; researchers may on drugs-crime relationships. Studies have to consult pharmaceutical manufac- of interactions between drug-using and turers and experts in pharmacology to -selling environments and the psychophar- address this phenomenon. macological and genetic aspects of drug use should be placed on the research A large segment of mainstream America is agenda. As genetics research becomes involved in using illicit drugs in new ways. more important in the drugs-crime field, Researchers need to change their para- researchers may start touting “drug digm of who drug users are and how they genes” without conducting research on behave. A participant volunteered the term “drug hackers” to describe new,

188 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

pharmacologically savvy drug users. They Middle-class addicts. The criminal activi- are similar to computer hackers in that ties of middle-class addicts, and the social they use drugs in unintended ways, com- and legal consequences of those activi- bining many different substances to mix ties, should be studied and compared to effects. the criminal activities of low-income addicts. This research would elucidate and Polypharmacy. Research on polypharma- deracialize the issues related to the con- cy, with emphasis on the interactions of sequences of drug use and crime. The licit and illicit drugs, should be included in differences between middle-income and the research agenda. Studies of interac- low-income addicts in use-to-addiction lev- tions between illicit drugs and alcohol are els and crime-commitment levels (for both particularly important. undetected and detected crime) are not known. The dark side of drug enforcement. Research should be conducted on the Ethical issues and genetics research. In negative aspects of drug enforcement. the forum on genetics research, partici- This includes studies of the flow of asset pants expressed apprehension about ethi- forfeiture funds across enforcement agen- cal issues raised by genetics research and cies and the degree to which that flow the need for further study of those issues. affects enforcement policies. Stigmatization and labeling of drug users are major concerns. However, the ability Enforcement-induced demand shifts. to give patients a small dose of a prescrip- Researchers should examine the shift in tion drug, measure a protein encoded by a demand from one drug to another as law gene whose expression is a secondary enforcement focuses on particular drugs, response to the drug, and use that meas- and the degree to which that shift is help- urement to predict whether the individual ful or detrimental. In other words, they is likely to become dependent on the should examine whether demand is shift- drug, would aid a physician working in a ing to drugs that are less or more serious therapeutic context. Researchers should in their marketing or crime potential. not oppose taking advantage of these kinds of benefits of the Human Genome Consequences other than crime. The Project. Ethical considerations are an research community should study conse- important part of good research and quences of drug use other than crime, should not be considered an impediment, such as mental health effects. The health but they also should not be the only effects of cocaine and methamphetamine consideration. have already been examined. Effect of economic development on the Early deviant behavior and drug use. drugs-crime nexus. A study of the effect Researchers need to study how parent of changes in the economy on drug use, monitoring, family cohesion, and family drug trafficking, and the drugs-crime nex- structure affect early deviant behavior and us would be useful. An example is the how that might in turn affect affiliation economic boom of the 1990s as an expla- with drug-using peers and drug use. Ex- nation for the decline of crime and vio- amination of the onset of criminal behavior lence during that period. following the onset of drug use in mono- 5 zygotic twins would illuminate the links The developmental role of the family in between drug use and crime. If there is a shaping behavior. There is a need for causal relationship, researchers should more research with a developmental see crime starting earlier in the twin who focus that assesses the influence of family starts drug use earlier.

189 SPECIAL REPORT / JULY 03

life on drug use and the drugs-crime rela- size of the disparity in racial impact, where tionship. One approach to understanding it occurs, and what factors contribute to it. the role of the family in shaping behavior Income level, for example, is a strong involves genetic influences on parenting. covariate with race, and the way offenders Some genetically influenced characteris- are treated in the criminal justice system tics of children, such as their tempera- varies by income level. Researchers must ment, affect how their parents treat them. separate a variety of race correlates from Thus, examination of genetic influences race itself as factors in racial discrimination and family life are critical because these in order to determine how much racial dis- interactions are frequently dyadic. parity is not due to racism and how much is a residual that is directly attributable to Continuum-of-care treatment models. racism. Treatment researchers think that the length of time drug users remain in treatment is There are also difficulties in classifying the best predictor of treatment success as people by race. For some research ques- defined by recidivism or drug use. Many tions, what may be more important is how people drop out of treatment at some point people are viewed by the police. For ex- in the process. Researchers should meas- ample, is their skin color dark enough to ure the impacts of continuum-of-care mod- be viewed as black, regardless of how els on treatment effectiveness. they self-identify culturally.

Racially disproportionate impacts of Underreporting by racial and ethnic drug policies. The participants discussed groups. The disproportionate underreport- whether the research community should ing of drug use by members of some eth- address variations in the effects of drug nic groups and how this affects research policies on different racial and ethnic findings are important topics for future groups in American society. The incarcera- research. tion rate is racially disproportionate, but whether the process leading to that im- Treatments whose effects differ by race pact involved race-influenced decision- or sex. Researchers need to know more making remains controversial and difficult about how the effect of treatment differs to investigate. Race is often a covariate in by race. They do not know if there are analyses of ADAM data but it is not a pow- specific ways to administer treatments erful covariate in explaining dependent that are more effective depending on race variables. Like gender, it frequently wash- or gender. Recent Federal Bureau of es out when multiple-level controls are Prisons studies indicate that treatments used. that are effective for men are not working for women. Comparisons of data from incarceration or other criminal justice processes with drug- Intergenerational discontinuities in use data reveal gross racial disproportion- drug taking. Research on intergenera- ality. However, it is by involvement in tional discontinuities in drug-taking would marketing, rather than in drug use, that be useful and would relate to such issues people become involved with the criminal as the blunt generation phenomenon. justice system. Research that includes Researchers are observing similarities in controls for participation in drug-market the degree to which new generations activity would be useful in identifying the buck trends or defy expectations.

190 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Discussion: What research in this should conceptualize behavior problems area is urgently needed? as latent dimensions. This would involve arraying people along a continuum of prob- Forum participants were asked to describe lem behavior or as manifestations of dif- areas of research that they think are most ferent classes, such as drug users who urgently in need of study. do or do not commit violent or property crimes. The statistical tools needed to Ethical issues in drugs and crime re- clarify uses of dimensions, classes, cate- search. Theoretical studies of ethical gories, and continua are evolving rapidly, issues are needed to address the impact and are interrelated with the missing of IRBs on drugs-crime research. Re- data issue and selection bias problems. searchers must do some rigorous thinking Approaches used by quantitative sociolo- about sound ethical models rather than gists and psychologists are already be- slavishly borrowing from the clinical trial coming mainstream biostatistical methods. model used in medical research. That clinical model emphasizes autonomy and in- Scientific justification for mandatory formed consent in ways that may not minimum sentences. Research on scien- be realistic in drugs-crime research. The tific justification for mandatory minimum research community could convene a sentences was suggested, with particular consensus-seeking meeting of social and emphasis on studying different mandatory medical scientists to discuss how existing minimums by type of substance. Manda- IRB criteria should be modified for social tory minimums have a direct bearing on science research. the racial disproportionality of drug policy impacts. There is an urgent need for re- More empirical research is needed to sup- search on the marginal cost-effectiveness plement the work of professional ethicists, of mandatory minimums and whether whose background in philosophy may not there are sociological justifications for reflect the values of ordinary people, in- them. This research would involve factors cluding drug users, and the way in which such as the way markets are structured. the latter regard the ethical and moral implications of research conducted with or for them. Survey research could be con- Discussion: What research in this ducted with the subjects and beneficiaries area would be recommended to of drugs-crime research to increase re- the best and brightest graduate searchers’ understanding of the ethical students? perspectives of various stakeholders. Participants were asked to think about Social scientists should be included on areas of research that would offer direc- Federal panels that produce regulations tion to researchers just starting their governing research. Although these pan- careers. els consist of physicians and laboratory Interventions for high-risk youths. researchers, the regulations they formu- Almost no research has been conducted late are applied inappropriately to social on interventions for high-risk youths who science. have been arrested. Most treatment out- New statistical methodologies. come studies focus on adult offenders Advances in statistical methodology who are already deeply involved in drugs should be used more widely in drugs- and crime. A large body of evidence crime research. They could be applied to assembled over the past 20 years indi- such issues as whether researchers cates a progression in drug use among arrested youths. At age 12, only a few

191 SPECIAL REPORT / JULY 03

arrestees test drug-positive. At a slightly expertise in a specialized field by publish- older age, drug tests might detect marijua- ing some lead author or sole author arti- na, and cocaine or heroin often are detect- cles in the journals of a particular discipline. ed in older youths. Thus, there is a need for research on interventions for younger The criminal addict paradigm. Empirical arrestees during their initial contacts with studies of nonuse crimes committed by the criminal justice system. Develop- drug users would follow up on research mental psychology literature on antisocial that suggests the major crime committed behavior, although not specific to drug by addicts is selling drugs. Researchers use, may be a source of information about found that people who did not have a prevention strategies for intervention with criminal history before becoming addicted early starters. did not adopt criminal behaviors other than drug selling after becoming addicted. The Linking policy interests with estab- proposed studies may reveal that the lished disciplines. The difficulty in recruit- amount of crime committed by drug ad- ing graduate students who are interested dicts, other than drug use and drug selling, in policy research may be ameliorated by is dramatically lower than conventional linking policy interests to the traditional wisdom would indicate. Hypothetically, concerns of existing academic disciplines, the results would fit a bell curve, with a such as economics, psychology, and soci- few people at one end who commit many ology. For example, research that affects crimes, a few at the other end who com- policy might involve the study of labor mit a small number, and most subjects in markets and address such topics as the the middle only selling drugs. Researchers relationship between licit and illicit wages. need to define this paradigm because of its policy implications. Policies that affect youth behavior. Students could study policies that affect The role of cognitive dysfunctions in young people’s behavior, including their drugs-crime relationships. The effects of involvement in the macro-educational job cognitive dysfunctions (whether they pre- market and labor opportunities available to cede or are induced by drug use) on drug them. One area of research to pursue is users’ decisions related to crime and their the possibility that economically, the job responses to interventions could be stud- market may be better in the illegal than in ied. Responses to interventions such as the legal domain. Students could study incarceration or treatment, for example, policies ranging from economics to educa- may be a function of cognitive deficits that tion as well as drug-education prevention. either precede or follow drug use. Various aspects of the research could be assigned to members of interdisciplinary Rational choice models of drug use. teams. Students should consult economists and others who study consumer choice behav- Interdisciplinary or comparative re- ior for assistance in developing research search. Graduate study is typically individ- that examines drug use as a choice among ualistic, which is not consistent with the various behaviors. Studies could address way research is conducted after gradua- the degree to which a young person’s de- tion. Students should seek interdisci- cision to use, or not use, drugs is based plinary work or the opportunity for on benefits to be obtained immediately or comparative studies and not be overly in the future. concerned about the topic. They could work on these projects and still establish Analyze ADAM data. In the new ADAM survey, large amounts of data have been

192 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

collected on drug treatment and crime; prevalence of lifetime drug use is lower this information also is connected to cen- among black Americans than among white sus tract information. This valuable data Americans. The models and methods collection presents an opportunity for stu- researchers use to study family structure dents to conduct data analysis without are not useful for explaining drug use collecting data. and crime in black or Hispanic-American communities. Secondary data analysis. The best use of graduate students’ time might be sec- Club drugs and crime. Almost everything ondary data analysis using ethnographic researchers know about drugs and crime or quantitative data. With mentoring and is based on past epidemics of cocaine, analytical experience, students could heroin, and marijuana use. Little is known become accustomed to working with data about the relationship of club drug use to and could gain experience with data col- other kinds of drug use and crime. This lection after graduation. potential epidemic involves mainly white, educated, 18- to 25-year-old users. The effects of moderating factors on accepted theories. When theories be- Club drugs and the Internet. People nor- come established or findings are mixed, mally start using drugs in the context of students should focus on the moderating their peer group. The Internet may influ- effects or interaction effects. They should ence use of club drugs. study conditions under which theories offer better or worse explanations for Inhalants. Inhalants/huffing is another cat- research findings. Moderating factors may egory of drug use that should be studied cause existing theories to work in some by young researchers. settings but not in others. Marijuana markets. Research on marijua- Theory integration within or across dis- na markets was suggested as a separate ciplinary boundaries. Students should research topic. consider integrating theories within or across disciplinary boundaries by examin- Policy implications of research findings. ing how their own theories fit with those Having investigators discuss the policy of other drugs-crime researchers or theo- implications of their findings and address- ries formulated in other disciplines. ing the implications of scaling up successful interventions were suggested as Methodological integration in drugs- research topics. crime research. Methodological integration, which employs techniques from History of drug policy. Study of the histo- other disciplines such as epidemiology or ry of drug use and its relationship to crime geographical information systems, may be was suggested as a way to provide per- useful in drugs-crime research. spective on the origins of current drug policies and acceptance of the fact that Family and social networks in minority policies can change over time. neighborhoods. The research community needs new models of what constitutes a Comparative international research. healthy family. Development of such mod- Graduate students should develop fluency els could focus on family and social net- in one or more foreign languages, quanti- work protective factors for reducing crime tative and methodological skills, and and drug use in high-risk neighborhoods. expertise in comparative research in order For example, 35 percent of black house- to conduct dissertation research in a for- holds are headed by women, but the eign country. Examination of entire drug

193 SPECIAL REPORT / JULY 03

enforcement regimes will require compar- Prostitution and drugs. The study of ative international research and developing prostitution as a criminal activity related to the capacity to conduct this kind of re- drugs was suggested as a research topic. search will be advantageous. Future trends. Students should look Comparative research across drug beyond the issues that researchers have types. Comparative research across drug been studying for the past 20 years and types was encouraged. Focusing on the try to assess future trends in the drugs- nexus of drugs, crime, and violence, the crime nexus. work would examine which aspects of drugs and their markets give rise to differ- The flexibility and mobility of drug mar- ent levels of pharmacological and market- kets. A study of drug markets could focus related crime. The ethnographic literature on their flexibility and mobility. will be important in this kind of research. High-functioning drug users. Studies of Onset, popularity, and termination of drug users who live routine lives and are markets for illicit drugs. Students could not involved in crimes other than taking study various criminal career paradigms illicit drugs would be an interesting re- and use them to analyze the creation, search topic. duration, and termination of drug markets. This would include a study of the preva- Models and simulations. Researchers lence of drugs that suddenly appear on should have graduate students create the market and would produce a history models and conduct simulations of the of a particular drug market. It would also effects of alternative drug policies on extend the study to a number of markets crime. The work would be methodological- to determine the factors that contribute to ly challenging because students would onset, popularity, and termination of mar- have to understand statistics, economet- kets for new illicit drugs. rics, and simulation software and conduct a literature review for each base estimate The noncriminal drug user. A suggestion to determine whether it is high or low. was made that students conduct research on noncriminal drug users. Closing remarks Interdisciplinary work involving genetics. Interdisciplinary research encom- Dr. Brownstein commented on the need passing fields such as genetics was to involve more researchers who are suggested. The methods of other disci- members of minority groups in future dis- plines could be applied in novel ways in cussions of drugs-crime interrelationships. the drugs-crime field. Practitioners and policymakers also could be more involved in the process; the The effects of interdiction programs. forum would be the first of many discus- Exhaustive analysis of interdiction pro- sions about these issues. grams and their effects may result in savings on interdiction expenditures. Dr. Erinoff thanked Roger Conner for acting as forum facilitator and reminded par- Extradisciplinary knowledge. Graduate ticipants that they may submit additional students should acquire some knowledge comments on drugs and crime through an outside their discipline in fields such as e-mail listserv. She and Dr. Brownstein will pharmacology. moderate the submissions.

194 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Bennett Fletcher, research psychologist at Notes NIDA, reminded the group that drugs and crime research has been important to 1. The summary was prepared by CSR, Inc. NIDA since the agency’s founding in 1974. 2. Manski, F., John V. Pepper, and Carol V. Petrie, He encouraged broadening the agency’s Informing America’s Policy on Illegal Drugs: What We criminal justice initiative and vigorous fol- Don’t Know Keeps Hurting Us, Washington, DC: lowup of the forum. National Academy Press, 2001.

Dr. Hillsman was pleased that the relation- 3. Genetic polymorphisms are differences in DNA ship between NIJ and NIDA had been sequences among individuals, groups, or populations. Genes for blue or brown eyes are an example. strengthened. She thanked the facilitator, organizers, authors, and participants for their 4. Diathesis is a condition of the body that makes tis- efforts and then adjourned the meeting. sues react in certain ways to certain external stimuli and thus makes them more than usually susceptible to other conditions.

5. Identical twins.

195 Appendix B: Forum Agenda

Thursday, April 19, 2001

8:30–9:00 a.m. Registration and Coffee Service

9:00–9:30 a.m. Opening Remarks Sally T. Hillsman, Deputy Director, National Institute of Justice

Henry H. Brownstein, Director, Drugs and Crime Research Division, National Institute of Justice

Lynda Erinoff, Health Scientist Administrator, Epidemiology Research Branch, Division of Epidemiology, Services and Prevention Research, National Institute on Drug Abuse

9:30–9:45 a.m. Facilitator Comments Roger Conner, Director, Search for Common Ground in America

9:45–10:00 a.m. Group Exercise

10:00–11:00 a.m. At the Intersection of Public Health and Criminal Justice Research on Drugs and Crime James C. Anthony with Valerie Forman

10:00–10:15 a.m. Presentation

10:15–11:00 a.m. Discussion

11:00–11:15 a.m. Break

11:15–12:15 p.m. Research on Drugs-Crime Linkages: The Next Generation Robert MacCoun, Beau Kilmer, and Peter Reuter

11:15–11:30 a.m. Presentation

11:30 a.m.-12:15 p.m. Discussion

12:15–12:30 p.m. Discussion and Planning Topics for lunch: “One-sentence statements”

197 SPECIAL REPORT / JULY 03

12:30–1:30 p.m. Working Lunch—Roundtable Discussions 1. Topic 1 2. Topic 2 3. Topic 3 4. Topic 4

1:30–1:45 p.m. Break

1:45–1:55 p.m. Remarks Richard A. Millstein, Acting Director, Division of Epidemiology, Services and Prevention Research, National Institute on Drug Abuse

1:55–2:15 p.m. Reports and Feedback From the Roundtable Discussions

2:15–3:15 p.m. 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

2:15–2:30 p.m. Presentation

2:30–3:15 p.m. Discussion

3:15–3:30 p.m. Break

3:30–5:00 p.m. Discussion What don’t we know about the relationships between drugs and crime? Reaching for consensus

Friday, April 20, 2001

8:30–9:00 a.m. Registration and Coffee Service

9:00–10:30 a.m. Discussion What research in this area do you think is most important?

What research in this area do you think is urgently needed?

10:30–10:45 a.m. Break

10:45–11:45 a.m. Discussion What research in this area would you recommend to the best and brightest graduate students?

11:45 a.m.–12:00 noon Closing Remarks

198 Appendix C: List of Participants

James C. Anthony Terence Dunworth Department of Mental Hygiene Abt Associates Inc. School of Public Health Johns Hopkins University Robert B. Eiss, Director Office of Programs, Budget, Research, Alfred Blumstein and Evaluation The Heinz School of Public Policy Office of National Drug Control Policy Carnegie Mellon University Lynda Erinoff, Health Scientist Henry H. Brownstein, Director Administrator Drugs and Crime Research Division Epidemiology Research Branch National Institute of Justice Division of Epidemiology, Services and Prevention Research Jonathan Caulkins National Institute on Drug Abuse RAND Jeffrey Fagan Jamie F. Chriqui, Technical Vice President Columbia Law School Center for Alcohol and Drug Policy The MayaTech Corporation Nora Fitzgerald, Social Science Analyst National Institute of Justice Roger Conner, Director Search for Common Ground in America Jerry Flanzer, Acting Chief Services Research Branch Christine R. Crossland, Program and Division of Epidemiology, Services and Policy Analyst Prevention Research Drugs and Crime Research Division National Institute on Drug Abuse National Institute of Justice Bennett Fletcher, Research Psychologist Richard Curtis Analytic Staff Department of Anthropology Office of the Director John Jay College of Criminal Justice Division of Epidemiology, Services and Prevention Research Karen Cyrus, Program Assistant National Institute on Drug Abuse National Institute of Justice Patrick Gartin, Chief Kelly Damphousse Statistical Services Division Department of Sociology Drug Enforcement Administration University of Oklahoma Sidra Gifford, Statistician Ross Deck, Deputy Director Bureau of Justice Statistics Office of Programs, Budget, Research, and Evaluation Paul J. Goldstein Office of National Drug Control Policy Great Cities Institute & School of Public Health University of Illinois at Chicago

199 SPECIAL REPORT / JULY 03

Lana D. Harrison Tom Mieczkowski Center for Drug and Alcohol Studies Department of Criminology University of Delaware University of South Florida

Sally T. Hillsman, Deputy Director Arnold Mills, Public Health Advisor National Institute of Justice Epidemiology Research Branch Division of Epidemiology, Services and Dana Hunt Prevention Research Abt Associates Inc. National Institute on Drug Abuse

Bruce Johnson Richard A. Millstein, Acting Director National Development and Research Division of Epidemiology, Services and Institutes, Inc. Prevention Research National Institute on Drug Abuse Coryl Jones, Research Psychologist Epidemiology Research Branch Janae M. Neiderhiser Division of Epidemiology, Services and Center for Family Research Prevention Research George Washington University National Institute on Drug Abuse Diana Noone, Social Science Analyst Beau Kilmer Drugs and Crime Research Division Drug Policy Research Center National Institute of Justice RAND John P. O’Connell Richard Lempert, Professor Office of the Budget University of Michigan Law School Statistical Analysis Center State of Delaware Natalie Lu, Drug Testing Technology Specialist Robert Nash Parker National Institute of Justice Robert Presley Center University of California, Riverside Robert MacCoun Goldman School of Public Policy Michael Prendergast & Boalt Hall School of Law Drug Abuse Research Center University of California, Berkeley UCLA Integrated Substance Abuse Programs Susan E. Martin Prevention Research Branch David Pyle National Institute on Alcohol Abuse and Drugs and Alcohol Research Unit Alcoholism Home Office London, England Harry Matz, Senior Trial Attorney Narcotic and Dangerous Drug Section Beth A. Reboussin Criminal Division Department of Public Health Sciences U.S. Department of Justice Wake Forest University School of Medicine Duane C. McBride, Director Institute for Prevention of Addictions Andrews University

200 TOWARD A DRUGS AND CRIME RESEARCH AGENDA FOR THE 21ST CENTURY

Peter H. Reuter Yonette Thomas, Health Scientist School of Public Affairs and Department Administrator of Criminology Epidemiology Research Branch University of Maryland, College Park Division of Epidemiology, Services and Prevention Research K. Jack Riley, Director National Institute on Drug Abuse Criminal Justice Program RAND Curtis J. VanderWaal, Associate Professor of Social Work Frank R. Shults, Senior Advisor Andrews University Office of the Deputy Attorney General U.S. Department of Justice William Vega Robert Wood Johnson Medical School Karen P. Tandy, Associate Deputy University of Medicine and Dentistry of Attorney General New Jersey–University Behavioral Office of the Deputy Attorney General HealthCare U.S. Department of Justice Helene Raskin White Patrick H. Tarr, Senior Policy Advisor Center of Alcohol Studies Office of Policy Development Rutgers University U.S. Department of Justice Eric Wish, Director Bruce Taylor, Senior Social Science Center for Substance Abuse Research Analyst National Institute of Justice

Yvonne M. Terry-McElrath Institute for Social Research University of Michigan

201 About the National Institute of Justice NIJ is the research, development, and evaluation agency of the U.S. Department of Justice. The Institute provides objective, independent, evidence-based knowledge and tools to enhance the administration of justice and public safety. NIJ’s principal authorities are derived from the Omnibus Crime Control and Safe Streets Act of 1968, as amended (see 42 U.S.C. §§ 3721–3723).

The NIJ Director is appointed by the President and confirmed by the Senate. The Director establishes the Institute’s objectives, guided by the priorities of the Office of Justice Programs, the U.S. Department of Justice, and the needs of the field. The Institute actively solicits the views of To find out more about the National criminal justice and other professionals and researchers to inform its search for the knowledge Institute of Justice, please visit: and tools to guide policy and practice. http://www.ojp.usdoj.gov/nij Strategic Goals NIJ has seven strategic goals grouped into three categories: or contact:

Creating relevant knowledge and tools National Criminal Justice Reference Service 1. Partner with State and local practitioners and policymakers to identify social science research P.O. Box 6000 and technology needs. Rockville, MD 20849–6000 2. Create scientific, relevant, and reliable knowledge—with a particular emphasis on terrorism, 800–851–3420 violent crime, drugs and crime, cost-effectiveness, and community-based efforts—to enhance e-mail: [email protected] the administration of justice and public safety. 3. Develop affordable and effective tools and technologies to enhance the administration of justice and public safety.

Dissemination 4. Disseminate relevant knowledge and information to practitioners and policymakers in an understandable, timely, and concise manner. 5. Act as an honest broker to identify the information, tools, and technologies that respond to the needs of stakeholders.

Agency management 6. Practice fairness and openness in the research and development process. 7. Ensure professionalism, excellence, accountability, cost-effectiveness, and integrity in the management and conduct of NIJ activities and programs.

Program Areas In addressing these strategic challenges, the Institute is involved in the following program areas: crime control and prevention, including policing; drugs and crime; justice systems and offender behavior, including corrections; violence and victimization; communications and information technologies; critical incident response; investigative and forensic sciences, including DNA; less-than-lethal technologies; officer protection; education and training technologies; testing and standards; technology assistance to law enforcement and corrections agencies; field testing of promising programs; and international crime control.

In addition to sponsoring research and development and technology assistance, NIJ evaluates programs, policies, and technologies. NIJ communicates its research and evaluation findings through conferences and print and electronic media.