Drugs and Violence: Causes, Correlates, and Consequences Editors: Mario De La Rosa, Ph.D. Epidemiology Research Branch National Institute on Drug Abuse Elizabeth Y. Lambert, M.S. Epidemiology Studies and Surveillance Branch National Institute on Drug Abuse Bernard Gropper, Ph.D. Drugs, Alcohol, and Crime Research Program National Institute of Justice NIDA Research Monograph 103 1990 U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Public Health Service Alcohol, Drug Abuse, and Mental Health Administration National Institute on Drug Abuse 5600 Fishers Lane Rockville, MD 20857 Drugs and Violence: Causes, Correlates, and Consequences ACKNOWLEDGMENT This monograph is based on the papers and discussion from a technical review on "Drugs and Violence" held on September 25 and 26, 1989, in Rockville, MD. The review meeting was sponsored by the Office of Science and the Division of Epidemiology and Prevention Research, National Institute on Drug Abuse. COPYRIGHT STATUS The National Institute on Drug Abuse and National Institute of Justice have obtained permission from the copyright holders to reproduce certain previ- ously published material as noted in the text. Further reproduction of this copyrighted material is permitted only as part of a reprinting of the entire publication or chapter. For any other use, the copyright holder's permission is required. All other material in this volume except quoted passages from copyrighted sources is in the public domain and may be used or reproduced without permission from the Institute or the authors. Citation of the source is appreciated. Opinions expressed in this volume are those of the authors and do not necessarily reflect the opinions or official policy of the National Institute on Drug Abuse, National Institute of Justice, or any other part of the U.S. Department of Health and Human Services or U.S. Department of Justice. The U.S. Government does not endorse or favor any specific commercial product or company. Trade, proprietary, or company names appearing in this publication are used only because they are considered essential in the context of the studies reported herein. DHHS publication number (ADM)9O-1721 Printed 1990 NIDA Research Monographs are indexed in the Index Medicus. They are selectively included in the coverage of American Statistics Index, BioSciences Information Service, Chemical Abstracts, Current Contents Psychological Abstracts, and Psychopharmacology Abstracts. iv Contents Page Foreword vii Charles R. Schuster Introduction: Exploring the Substance Abuse-Violence Connection 1 Mario De La Rosa, Elizabeth Y. Lambert and Bernard Gropper Violence as Regulation and Social Control in the Distribution of Crack 8 Jeffrey Fagan and Ko-lin Chin Violence Associated With Acute Cocaine Use in Patients Admitted to a Medical Emergency Department 44 Steven L. Brody The Operational Styles of Crack Houses in Detroit 60 Tom Mieczkowski The Crack-Violence Connection Within a Population of Hard-Core Adolescent Offenders 92 James A. Inciardi v Page The Relationship Between Cocaine Use, Drug Sales, and Other Delinquency Among a Cohort of High-Risk Youths Over Time 112 Richard Dembo, Linda Williams, Werner Wothke, James Schmeialer, Alan Getreu, Estrellita Berry, Eric D. Wish, and Can dice Christensen The Drug Use-Violent Delinquency Link Among Adolescent Mexican-Americans 136 W David Watts and Loyd S. Wright Gangs, Drugs, and Violence 160 Joan Moore The Interrelationships Between Alcohol and Drugs and Family Violence 177 Brenda A. Miller Drug-Related Violence and Street Prostitution 208 Claire E. Sterk and Kirk W Elifson Drug Disorder, Mental Illness, and Violence 222 Karen M Abram and Linda A. Teplin Who's Right: Different Outcomes When Police and Scientists View the Same Set of Homicide Events, New York City, 1988 239 Patrick J. Ryan, Paul I Goldstein, Henry H Brownstein, and Patrica A. Bellucci Summary Thoughts About Drugs and Violence 265 James I Collins List of NIDA Research Monographs 276 vi Foreword Drug abuse and drug-related violence are among the greatest concerns of our citizens. There is a growing interest on the part of researchers, the public, and all levels of our government in the causes, correlates, and consequences of drugs and violence-- for better understanding of these phenomena and for improving our efforts at converting understanding into more effective prevention and control programs. Many factors, such as the emergence of relatively cheap and widely avail- able crack cocaine and widespread violence in drug trafficking, influence the increase in drug-related violence within and outside the United States. The challenge to public health and law enforcement communities is to develop strategies for intervention and control that work. These are priority issues within the missions and research agendas of both the National Institute on Drug Abuse (NIDA) and the National Institute of Justice (MJ). On September ~27, 1989, MDA, with the collaboration of MJ, held a Technical Review meeting on "Drugs and Violence." The focus of this meeting was to review recent research advances made in the study of the relationships between drugs and violence. Data from a number of NIDA- and MJ-funded research projects addressing different aspects of these relationships were presented and are included in this monograph. This meeting and monograph underscore the continuing collaborative research efforts by MDA and NIJ to explore the linkages between drug use and violence and other criminal behaviors. The studies presented here represent only a sampling of the types of basic and applied research efforts that contribute to development of a sound infor- mational base from which health providers and law enforcement officials can develop more effective strategies and programs to combat these prob- lems. It is hoped that this monograph will serve as a framework for further efforts in these areas and help us reach our common goals of making our society-and our world a safer and healthier place to live. Charles R. Schuster, Ph.D. Director National Institute on Drug Abuse vii The National Drug Control Strategy indicates that our Nation's success in overcoming the problems of drug abuse and drug-related crime depends upon the efforts of all our citizens and all agencies of Government The strategy provides a comprehensive framework and a balanced approach for our priorities in prevention and control, supply and demand reduction, law enforcement and treatment, research, and evaluation. Research plays a vital role in supporting these efforts by assessing the nature and extent of the problems, developing the scientific and technical bases for effective public policies, identifying and assessing programrnatic options, and evaluating the impacts of our drug control interventions. The studies on drugs and violence within this volume exemplify the collaborative relationships between the research programs of the National Institute of Justice in the Department of Justice and the National Institute on Drug Abuse in the Department of Health and Human Services on these priority issues. The chapters encompass both qualitative and quantitative approaches to research, including: the development of conceptual frameworks; the observation, description, treatment, prevention, and prediction of drug abuse and related violence; and the translation of definitions into practice for statistical databases and other applications. Though they focus on U.S. populations and conditions, foreign researchers and governments express continuing interest in our research methods, data, and approaches to addressing drug abuse, drug-related crime, and drug-related violence. This evidence reflects significant potential for making greater contributions to solving these drug problems that affect our Nation and the world Charles B. DeWitt Director Designate National Institute of Justice viii Introduction: Exploring the Substance Abuse-Violence Connection Mario De La Rosa, Elizabeth Y. Lambert, and Bernard Gropper The complex relationships between substance abuse and violence have posed challenges to the research community and public health professions for decades. Research literature on drugs and violence abounds and continues to grow, with broad representation from the disciplines of education, medi- cine, sociology, criminology, epidemiology, and psychology. Understanding the causes, correlates, and consequences of drugs and violence is necessary to develop effective public health and law enforcement strategies for preven- tion and control. Some may despair, believing the links between substance abuse and violence to be inseparable and complex, and, therefore, believing that effective solutions cannot be found. Efforts to understand these rela- tionships can contribute to a process for identifying ways to prevent their occurrence or to reduce their magnitude, severity, and their recent apparent intensification. Links between alcohol abuse and violence have been recognized for years. Recently, new varieties of violence have emerged, largely in relation to the abuse and distribution of crack cocaine. The 1980s have seen a growing number of apparently "random" or "impersonal" homicides--that is, homi- cides of persons unknown or hardly known to their assailants. These so- called "hit men" style slayings have been linked to the crack trade, with drug dealers competing against other dealers to corner the market or pre- serve their territories. Victims are typically young boys or men and are often minorities living in inner cities. Occasionally, distinct patterns of injury can be recognized: drug runners, young teenagers who carry drugs and money between sellers and buyers, are being seen in emergency rooms more frequently with gunshot wounds to the legs and knees; a more vicious style of drug-related injury has emerged in the western part of the United States. In this injury, known as "pithing," the victim's spinal cord is cut, and he or she is left alive, but paraplegic. In the 1950s and 1960s, research on the relation between drug abuse and violence focused on criminal behaviors of narcotic addicts. It was generally accepted that opiate or heroin users were more likely to engage in nonvio- lent property crimes than in other types of crime. This was supported by data from major metropolitan areas showing a positive correlation between rates of heroin addiction and property crimes and a negative correlation between rates of heroin addiction and crimes against persons. Later re- search by Kozel and Dupont, Inciardi, Chambers, and Nurco lent additional support to these findings. From research spanning the late 1960s to today, the primary substance implicated in violent crimes has been alcohol, far more often than illicit drugs. Alcohol abuse and violence are endemic to America's culture, dating back to the days of Prohibition, with violent rivalries between bootleggers, and existing as intensely today, with high rates of alcohol-mediated domes- tic violence, homicides, vehicular accidents, and traumatic injuries. With increased use of amphetamines in the 1970s and the development of cocaine distribution networks in the 1980s, research began to focus on the relation between psychopharmacologic effects of drug use and violent behavior and on systemic violence associated with drug-dealing lifestyles. The findings from these studies indicated that, although certain types of illicit drugs, e.g., stimulants, hallucinogens, may be associated with violent behavior, most psychopharmacologically induced violent crimes continued to involve alco- hol. Violent crimes involving illicit drugs were more likely from trade transactions between drug dealers and drug users. In the mid- to late 19&~, reports of increased violence from crack use and distribution networks among inner-city minority communities made it urgent for public health officials and epidemiologists to look into the intensifying problem and develop strategies for intervention and control. A product of this renewed research activity was Paul Goldstein's development of a con- ceptual framework to explain complexities between violence and drugs. Goldstein's tripartite theoretical model distinguishes three dimensions for understanding drugs and violence: psychopharmacologic, economic compul- sive, and systemic. The psychopharmacologic dimension refers to effects of substances on behavior, as when consumers become irrational, excited, agi- tated, or unable to control their anger and violent impulses. The economic compulsive dimension refers to violent crime committed to obtain money or other forms of currency to purchase drugs for personal use. The systemic dimension addresses violence intrinsic to the lifestyles and business methods of drug distributors and traffickers. This monograph reports findings from a variety of studies on aspects of the drug and violence nexus. Its chapters address a broad spectrum of issues, 2 including studies on violence and aggression in crack distribution networks in New York City and Detroit, MI; correlations between illicit drug abuse and domestic violence; links among gangs, drugs, and violence; crack and violence among juvenile delinquents, including delinquents who are Hispan- ic; the relation between prostitution, drugs, and violence; problems with using police data for research on drug-related violent crimes; neuropsycho- logical effects of acute cocaine abuse on violent behavior; and links be- tween mental illness, drugs, and violence. The monograph concludes with an examination of Goldstein's tripartite model as a conceptual framework for exploring, understanding, and predicting the causes, correlates, and con- sequences of drugs and violence. The themes embodied within these manuscripts reflect both qualitative and quantitative approaches. The analytic focus is on the individual and small to mid-sized groups such as the gang and the family. These are the most relevant units of inquiry for most problems of the etiology of drug use and abuse and the types of drug-related violence experienced at the neighbor- hood, local, and State level. Although the focus of this volume is on cur- rent U.S. conditions, population, and subgroups, the underlying nature of the phenomena and the principles embodied in the research methods and find- ings are to some extent relevant to situations outside of the United States. Drs. Fagan and Chin examine violence and aggression among a cohort of crack dealers and other illicit drug dealers in New York City. Unlike pre- vious research on drug-related violence, Fagan and Chin's work explores possible origins of violence in drug selling. Specifically, it addresses whether violence in crack distribution networks tends to be contingent on drug-selling activities or is more reflective of a generalized pattern of crime and violence among those individuals involved in the crack or drug trade. Their results suggest that both conditions influence violence in crack selling. In general, crack sellers were more likely than other drug sellers to use violence for regulation and control, to be involved in other types of violent crimes, and to be immersed in a violent social world. Drs. Fagan and Chin hypothesize from these data that violence among crack sellers may be less a function of risks associated with the settings in which crack is sold, and more a function of individual predispositions toward violent lifestyles, even before such persons become involved in drug dealing. Dr. Brody's chapter examines the relation between acute cocaine intoxica- tion and aggression and violent behavior among a group of patients admit- ted to a hospital emergency department. The data presented suggest that more than half the cocaine-involved patients were combative and agitated, with symptoms of paranoia and delirium at the time of admission. As the effects of cocaine tend to subside rapidly, acute pharmacologic therapy for these individuals was rarely indicated. When it was necessary to treat cocaine-induced cases of acute psychosis, however, haloperidol was found to be relatively effective. 3 Dr. Mieczkowski further explores crack and violence among a group of drug dealers and users in Detroit. The objective was to identify methods used by local crack dealers to distribute cocaine at the retail level and to describe principles of management and organization that typify these meth- ods. Results indicate that crack dealers and users practice three distinct but overlapping distribution and selling methods. These are (1) the street-corner or walkup sale system, (2) the beeper and runner system, and (3) the crack- house system, the preferred method of most crack dealers. The operational styles associated with each method reflect economic principles and practices seen in legitimate businesses. For example, Mieczkowski reports that crack- house operations ranged from very austere, in which social interaction be- tween those conducting the crack transaction was severely restricted, to "tavern style" crack houses, in which socialization extended beyond the ex- change of money for crack and often included exchanges of sex for drugs. Data further suggest that violence is more endemic to the street-corner or walkup sale system than to other methods of crack sale. One possible reason is that the social setting of the street drug-sale scene is less protected than either crack houses or the runner and beeper systems, leading dealers to gravitate toward violence to regulate and control their drug territories. Dr. Inciardi's chapter explores patterns of violent criminal behavior and crack use among a cohort of seriously delinquent youth in Miami, FL. Dr. Inciardi found that youth who were more deeply involved in crack- dealing activities were more likely to commit violent crimes than those who were less involved with drugs and the crack distribution network. The majority of violent crimes either committed by the youths in the study or perpetrated against them (robberies) were to purchase drugs, followed by violent crimes related to drug trafficking and by drug-induced or psycho- pharmacologic violence. The relation between drugs and violence is further examined by Drs. Dembo, Watts, and Wright, who, like Inciardi, address cocaine use, drug sales, and delinquency. Data from Dr. Dembo's research with a cohort of high-risk youths held in a regional detention center in Tampa indicate that involvement in drug sales was significantly and positively related to both nonviolent and violent crimes. The study found violent crimes to be signif- icantly and directly related to involvement in drug sales. Results suggest that most violent and nonviolent crimes involving youth in this sample were related to the business of drug selling, as would be predicted by the sys- temic dimension of Goldstein's tripartite model. The study by Watts and Wright explores correlations between drugs and violence among a cohort of Mexican-American youth. Correlation analysis of factors on violent delinquency among these youth revealed that illegal drug use contributed the greatest amount of variance, followed by friends' drug and tobacco use, lack of parental supervision, and family drug use. The interview results suggest that acculturation-related stress and familial 4 fragmentation between parents and children may contribute to both drug use and violent behavior among some youth. Dr. Moore's chapter addresses a topic of increasing attention in our society today, the link between gangs, drugs, and violence. She argues that what- ever we know about drug-related gang violence "comes almost entirely from the media and police, and it is almost always sensationalized." Her study of traditional Mexican-American gangs like the "White Fences" in East Los Angeles indicates that gang-related violence is in fact inherent to normal gang activities. But this violence is more often a function of intergang conflict than it is related to the drug trade. Although some youth gangs were found to be involved in drug-related violent criminal activities, this was not the norm for most. Drug-related violence stemmed from drug- dealing activities of individual gang members or former gang members more than from activities of the youth gang as an organized entity. Further, Dr. Moore and her associates argue that most of the gangs identified by law enforcement officials as heavily involved in drug-related criminal activities did not emerge from traditional youth gangs established in black and Hispanic communities before the onset of the crack epidemic. Rather, these groups grew out of criminal organizations formed solely for crack distribu- tion and trafficking. As such, they have few if any of the behavioral char- acteristics found among more traditional youth gangs. The interrelation between alcohol, illicit drugs, and family violence comprise Dr. Miller's research focus. Analysis of data from a series of studies on family violence among a group of male parolees and their spouses suggest that alcohol abuse is a greater risk factor than is illicit drug use. Dr. Miller's studies suggest that alcohol and illicit drug problems experienced by parolees related directly to the level of violence experienced by their spouses. "For those parolees who reported no drug problems, alcohol prob- lems increased the level of violence. However, when the parolee had drug problems, alcohol problems did not increase the level of violence." Addi- tional findings indicate that alcohol use rather than drug use was more likely related to child abuse. Drs. Sterk and Elifson focus on the relation between male and female pros- titution, drug use, and violence. Violence and drug use are shown to be intrinsic to the world of prostitution. Key findings from their work in Atlanta and New York are that males tend to work as prostitutes prior to initiating drug use, while females are more often drug users first and later resort to prostitution, often in direct exchange for drugs or for money to buy drugs. The research shows that the dynamics of the street scene have been dramatically affected by the emergence of crack cocaine. The world of street prostitution, always dangerous and unpredictable, has become even more so. 5 The chapter by Drs. Teplin and Abram presents findings on interrelations between mental illness, drug use, and violent crime among a cohort of male jail detainees in Chicago, IL. Individuals assessed as having antisocial per- sonality disorders, with or without drug-use problems, were found to be more likely to be involved in violent crimes than those who had a drug or alcohol problem but did not have antisocial personality disorders. The authors point out that persons having drug-use problems are not necessarily more likely to commit violent acts than other offenders. Rather, it is the individual with antisocial personality problems, regardless of his drug prob- lem, who is more prone to violence. The accuracy of information collected by police on drug-related violent crime is addressed in detail by Dr. Ryan. He presents findings from a project that sought to: (1) develop procedures for collecting valid and reliable data about apparent motives in drug-related homicides (psycho- pharmacologic, economic compulsive, and systemic) in New York city and (2) integrate these reporting and analytic procedures into protocols for homi- cide investigations by New York city police. The results suggest that experimental implementation of the procedures in a joint police-researcher effort improved the quality of data collected on drug-related homicides and permitted analyses of drugcrime links that would otherwise not have been feasible. Dr. Collins expands upon Goldstein's tripartite conceptual framework on drugs and violence by addressing other risk factors that may be indirectly or directly implicated in their epidemiology. Such factors as early childhood injuries, abuse, or neglect; socialization experiences; lack of economic opportunity; community disorganization; and physical reactions to specific types of drugs are cited as important adjuncts to Goldstein's tripartite con- ceptual framework. Dr. Collins suggests that the integration of these factors into Goldstein's model would enhance its power to explain and to predict phenomena associated with drugs and violence. The chapters in this monograph represent a diversity of disciplines and research areas concerned with the causes, correlates, and consequences of drugs and violence. Yet, despite the breadth of information presented here, many unknowns remain. It is hoped that, by bringing together some of the issues associated with drugs and violence and the consequences they have on our society, this monograph will inform and inspire others to contribute to the epidemiologic knowledge base. More important perhaps, it will lead to the formation of educational, social, judicial, and medical strategies to reduce and prevent drug abuse and violence. One point remains very clear: the problems of drugs and violence are complex and seem to be intensify- ing, underscoring the urgency for effective public health, legal, and social interventions. 6 AUTHORS Mario De La Rosa, Ph.D. Social Science Analyst Epidemiology Research Branch Elizabeth Lambert, M.S. Statistician Epidemiology Studies and Surveillance Branch Division of Epidemiology and Prevention Research National Institute on Drug Abuse Parklawn Building, Room 1 1A-55 5600 Fishers Lane Rockville, MD 20857 Bernard Gropper, Ph.D. Program Manager Drugs, Alcohol, and Crime Research Program National Institute of Justice 633 Indiana Avenue, N.W. Washington, DC 20531 7 Violence as Regulation and Social Control in the Distribution of Crack Jeffrey Fagan and Ko-lin Chin INTRODUCTION After nearly a century of research on the relation between substance use and violence, drug use and trafficking have only recently been examined separately as etiological factors in violence and aggression. Violence asso- ciated with selling illicit substances has been evident since the Opium Wars in China and illegal whiskey importation into the American colonies (Musto 1989). In this century, illicit distribution of alcohol during the Prohibition Era led to widespread violence as criminal organizations competed for mar- ket share and territory (Zahn 1989). Violence intrinsic to drug distribution has been associated with marijuana production and selling (Adler 1985), heroin selling (lanni 1974; Goldstein et at. 1984; Johnson et al 1985; McBride 1981), and, more recently, cocaine and crack selling (Goldstein et at. 1987; Goldstein et at., in press; Johnson et at., in press; Williams 1989; Adler 1985; Murphy et at., unpublished manuscript).1 Recognition of the etiological relevance of drug trafficking to violence has resulted in more careful formulation of theories of the drug-violence rela- tionship. Specifically, examination of homicides and other violent behaviors that involve drugs suggests separate explanatory frameworks for violence that occurs following intoxication, violence that occurs in the "service" of substance use, and violence that occurs during the course of drug trafficking (Goldstein 1985; Goldstein 1989). In the context of drug selling, further distinctions are evident between violence associated with wholesale distribu- tion and violence in street-level transactions (Goldstein 1989). There are several influences on violence that occurs in the context of street- level (seller-user) drug distribution. Violence may be used to enforce organizational discipline or resolve business disputes. Disputes over drugs and drug paraphernalia are commonplace among users and sellers. Territor- ial disputes are commonplace among drug sellers. Street-level sellers may skim profits from mid-level suppliers or crew bosses. In the absence of 8 legal recourse for illegal activities, such disputes are likely to be settled either by economic reprisal or by violence. Violence in drug dealing can be viewed as an extension of behaviors that are associated with efficiency and success in legitimate businesses (Black 1983). The social milieu of drug-selling and drug-buying areas also is conducive to robbery of sellers and users for either cash or drugs. The spurious connec- tion of drug use and crime suggests that drug selling will be concentrated in social areas with concentrations of the social structural features of violent crime and victimization. The reciprocal nature of the drug business and violence may influence the decision to participate in drug selling-individu- als averse to violence may avoid street-level drug transactions, leaving only those willing to engage in violent behaviors as participants. Self-selection of violent individuals for participation in the drug business also may in- crease the likelihood of violence during drug transactions. For example, Fagan (1989) found that the drug selling-violence relationship among youth gangs was strongest for gangs most frequently involved in all types of violence. This chapter examines violence and aggression among crack and other illicit drug sellers in New York City. Few studies have addressed the origins of violence in drug selling, specifically whether such violence reflects general- ized violence or violent behaviors contingent on drug selling. Aggression in crack selling appears to be commonplace and severe (Goldstein et al., un- published manuscript; Goldstein 1989; Johnson et al. 1990; New York Times 1989b) and is the focus of this study. Aggression evident in nondrug crim- inality is compared for crack sellers and other seller types. If violence in drug selling is a distinct behavior that reflects the contingencies of the un- regulated marketplace, participation of sellers in nondrug violence will be less evident. However, if violence in drug selling involves processes of self-selection of generally violent individuals, their participation in nondrug violent crimes will be extensive. This interpretation would further suggest that systemic violence in drug selling is spuriously related to other etiologi- cal factors in violence and crime commission, rather than a function of unique social processes of drug selling. The Emergence of Crack and Crack Markets The appearance of crack in New York City in 1985 has been widely associ- ated with increased violence in illicit drug markets (New York Times 1989a; Fagan and Chin, in press). Crack was introduced in New York shortly after the use and sale of powdered cocaine had reached its highest level nation- wide in 1982 (Zimmer 1987). Most cocaine users had been aware of the intensified high from smoking freebase cocaine (Siegel 1982; Siegel 1987). However, sufficient quantities of cocaine for "basing" had been beyond the economic means of most drug users. An apparent reduction in the import price of cocaine in the mid-1980s made the raw material for smokable 9 cocaine economically accessible to all users. Moreover, compared to the manufacture of freebase cocaine, the crack production process was cheaper and more efficient. Crack was produced by heating cocaine with household substances, e.g., baking soda, rather than with the volatile and expensive chemicals, e.g., ether, used to transform cocaine hydrochloride (HCl) (the powder) into its base form. Crack was marketed at a low unit cost in a rock or pebble form that was easily concealed and ingested. Its crystalline appearance conveyed an image of purity. The ingenious production and marketing strategy for crack gave it the appearance of a cheaper (albeit shorter) "high" from a purer form of cocaine. Following closely the growth in popularity of cocaine HCl and encouraged by the well-known advantages of smoking cocaine, cocaine users were quick to accept and popularize its new smokable form. As with the more expensive freebase form, compulsive use often developed follow- ing initiation into cocaine smoking (Siegel 1982; Siegel 1987; Spitz and Rosecan 1987; Washton and Gold 1987; Waldorf et al. 1990). More than 60 percent of cocaine users admitted for treatment in New York State in 1986 reported smoking crack as the primary method of cocaine use (Frank et al. 1987). At first, crack was mass marketed in inner-city neighborhoods in or near cocaine importation points such as Miami, Los Angeles, and New York (Inciardi 1987), and it spread later to other cities (Newsweek 1986). Ethno- graphic (Hopkins 1989; Bourgois 1989), government (Frank et al 1987; Mieczkowski, in press), and media reports (New York Times 1989a; News- week 1986) revealed that crack often was sold in centralized locations (crack houses) where buyers had access to crack limited only by their funds. Reports from users in treatment (Frank et al. 198~, the popular press (New York Times 1989b), and criminal justice agencies (Belenko et al., in press) also confirmed that crack was widely available throughout New York City. Within 2 years, crack use and trafficking were widespread and highly visible throughout New York City, especially in its most socially and econ- omically deprived neighborhoods (Hopkins 1989; Johnson et al, in press; New York Times 1989b; New York Times 1989c). For drug sellers, crack production was efficient, and its popularity made it extremely profitable. In short, crack was an excellent investment. Crack Selling and Aggression: Victimization or Social Control? Crack appeared in inner-city neighborhoods that had experienced profound social and economic deterioration in the decade preceding its appearance (Wacquant and Wilson 1989). The 1970s was a decade marked by labor surpluses in inner cities, created by the relocation of jobs to "satellite cities" in surrounding suburbs. Citing data from the U.S. Department of 10 Commerce, Kasarda (1989) shows that between 1970 and 1980, the number of blue-collar and clerical jobs in New York declined by over 350,000 but increased by over 75,000 in the surrounding suburbs. Technical and mana- gerial jobs in the city increased by over 250,000 during this time and by over 400,000 in the suburbs. Traditionally, African-Americans have relied heavily on blue-collar jobs in manufacturing for economic sustenance and social mobility (Farley and Allen 198~. Thus, the economic restructuring of American inner cities resulted in large-scale exclusion of their minority residents from constricting labor markets that also were transforming from manufacturing to services and shifting geographically from the inner city to the surrounding suburbs (IIochschild 1989). Similar processes, compounded by language and other cultural barriers, created severe economic dislocations for Puerto Ricans, in turn creating conditions of severe impoverishment (Farley 1987; Tienda 1989b; Kasarda 1988).2 Accordingly, the potential for high profits from selling crack attracted young initiates into drug dealing in social areas in which legitimate economic ac- tivity had decreased. For many young inner-city residents in this decade, the informal economy offered the most lucrative income opportunities (Sassen-Koob 1989). Involvement in the high-profit informal crack market offered economic opportunities to replace formal opportunities lost as capi- tal flowed out of inner-city neighborhoods in the decades preceding its emergence. Prior to crack, drug-related crimes generally were attributed to heroin use, and there was little overlap between users and sellers. Stable organized crime groups controlled heroin distribution, while drug-motivated crimes were usually attributed to heroin users, whose crimes served their drug use (Ball et al 1983; Johnson et al. 1985; Johnson et al 1990). As cocaine use increased both nationally and in inner cities from 1975 to 1982 (Kozel and Adams 1985; Siegel 1985), cocaine selling in New York City became more prevalent among drug sellers than was heroin selling (Zimmer 1987). The co-incidence of cocaine and other drug use and selling also rose during this period, as drug distribution essentially became a decentralized activity with cocaine HCl's increased availability of and decreased price (Zimmer 1987; Williams 1989). The proliferation of cocaine distribution activities in this era seemed to have two effects. First, opportunities for drug distribution by new organizations apparently increased, creating economic incentives for individuals in inner cities to participate in the informal drug economy. New York Police Department (NYPD) officials characterized the crack "industry" as "capital- ism gone mad" (New York Times 1989b), with no legal, economic, or infor- mal social controls (Adler 1985; Murphy et al., unpublished manuscript). Second, the social processes of drug distribution seemed to change, as 11 inner-city neighborhoods in New York declined economically, and new opportunities were created for users to participate in low-level selling activities. The institutionalization of drug selling among inner-city residents in this era naturally extended to crack (Johnson et al., in press; New York Times 1989c; New York Times 1989e). Drug selling increased during a time when social and economic changes weakened formal and informal social controls against violence at the neighborhood level. Thus, crack distribution systems developed in a social context in which poverty and social disorganization were intensifying (Massey and Eggers 1990; Tienda 1989a), social institutions and economic activities that tradi- tionally provided social controls against violence were weakening (Sampson 1986; Sampson 1987, there was an absence of established drug-dealing or- ganizations and territories for this product (Williams 1989; Johnson et al., in press), and a high demand existed for a product that for many initiates quickly led to compulsive use. Violence associated with drug dealing increased at the same time that crack was introduced in the drug market. Hamid (1990) showed how the political economy of drug selling changed over a 25-year period in Caribbean neigh- borhoods in Brooklyn, as drug trafficking evolved from the peaceful trade of marijuana to normative violence in crack markets. Goldstein et al. (1987; Goldstein et al. 1989) also illustrated the increase in drug-related violence associated with decentralized cocaine distribution systems. Crack's appeal as a powerful and addictive drug, together with extraordinary profits from street sales, may have intensified drug-violence links that were more tenuous and contingent before the appearance of crack. Accordingly, the appearance of crack coincided with the transformation of drug-related violence from the older patterns of economic compulsive crimes (to obtain money for drugs) to protection of economic interests (from terri- torial incursions by other sellers or robberies for cash or drugs) and regula- tion of emerging businesses (enforcement of discipline among employees). Disputes between nascent drug-dealing organizations led to reported in- creases in systemic violence during the competition for control of neighbor- hood markets (New York Times 1989b; Williams 1989; Ham id 1990; Bourgois 1989). Increases since 1987 in hospital emergency room cases involving gunshot wounds, fractures, and other wounds indicative of inten- tional injury have been attributed to violence surrounding crack, rather than increases in the base rate of violence (New York Times 1989a). The Present Study The symbolic meaning of criminal conduct may be interpreted simply as a violation of a legal or moral prohibition or as a form of self-help and social control (Black 1983). Viewed in relation to the illicit nature of drug distri- bution, violence in crack dealing would be expected to occur as a form of 12 economic regulation and organizational maintenance. Hobbesian theory would suggest that, in conditions in which law and governmental social con- trol are least developed, violence would be more evident as a form of social control. In the volatile and illicit crack market, this view implies that vio- lence should be limited to those organizational or economic situations that require regulation. Since the activity is illicit, violence also is necessary as a form of self-help; drug sellers cannot legitimately bring legal grievances for crimes within the selling context. If crime is social control and econ- omic regulation, then predatory or expressive crimes should be less evident. However, crack distribution systems developed under conditions that were conducive to criminal conduct, as well as to the specific forms of violence more commonly associated with drug distribution. The rapid growth of crack use and emergence of crack-selling organizations occurred in socially disorganized areas with few legitimate economic opportunities and strained informal social controls, conditions associated with increased rates of preda- tory and expressive violent crimes (Sampson 1986; Sampson 1987). Vio- lence thus regarded sociologically is less likely to be confined to contingen- cies that either are moralistic or instrumental and would be evident both within and outside the context of drug selling. To adequately explain violence within drug distribution, comparisons are necessary of violence both within the social and economic context of drug selling and violence that occurs in other situations. If violence within drug selling is a form of social control and economic regulation, violence not associated with drug selling should be less frequent. However, if violence within drug-selling contexts simply is a manifestation of generalized crimi- nal proclivities, there should be few distinctions between violence in the service of drug dealing and violence outside the dealing context. To test these competing explanations of violence in drug distribution, vio- lence within and apart from the context of drug dealing is compared for individuals involved in various types of drug distribution activities in New York City neighborhoods where crack use and sales have grown rapidly in the past few years. A theory of violence as social control predicts limited involvement of drug sellers in violence outside the context of selling. A generalized theory of crime predicts no distinctions between violence in the context of dealing and other varieties of crime. METHODS Samples Samples were constructed from two northern Manhattan neighborhoods with high concentrations of crack use and selling: Washington Heights and West Harlem.4 Samples included individuals from the study neighborhoods who had been arrested for drug possession or sales, residents of the study 13 neighborhoods who matched the arrested populations but who had avoided legal or social intervention for drug use or selling, and participants in resi- dential drug treatment programs. Within each group, subjects included crack users or sellers, cocaine HCl users or sellers who were not involved with crack, heroin users or sellers, and polydrug (primarily marijuana) users. Samples were recruited through chain referral or "snowball" sampling pro- cedures (Biernacki and Waldorf 1981). Since the research was part of a larger study of crack, crack users and sellers were oversampled. Crack arrestees were recruited from drug arrestees who were awaiting initial court appearances in the Manhattan central booking facility. They were identified from special charge flags recorded by arresting officers on booking slips. The arrest flags have been used by the NYPD since 1986 to identify crack offenses, since charge categories do not distinguish various types of con- trolled substances. Residential neighborhood was determined from the addresses and corresponding zip codes provided by arrestees to the interviewers. Referrals for interview were made by pretrial services interviewers during routine interviews to determine eligibility for release on their own recogni- zance. Arrestees released at arraignment were interviewed shortly after release. (Those arrestees detained were interviewed in the detention facil- ity.) Arrestees who indicated their willingness to participate in a research study were given cards that told them where and how to arrange for an interview. Their names also were given to the interview team who, in some cases, sought them out. Other subjects also were recruited through chain referral procedures: non- crack drug arrestees; nonarrested neighborhood samples who were matched to the arrested samples on age, gender, and ethnicity; and participants in two residential treatment programs in Manhattan. Several types of chain referral methods were used. Arrestees were asked to nominate potential respondents who were "like them in many ways but who have avoided arrest." Interviewers then sought out the nominees, or the nominees were referred to the field office by friends. Chains also were developed among drug users and sellers who were known to the interviewers. Interviewers were members of a street research unit that maintained ethnographic contact and did reconnaissance on drug scenes throughout the New York metropoli- tan area. Residential treatment clients were recruited from their programs based on nominations of crack and other drug users by administrators and clinical staff. Treatment residents who had been in the program for at least 1 month and had met screening criteria for each drug-user type were asked to participate in treatment. 14 A brief (10 item) screening interview was used to classify respondents and validate their reports. Respondents were classified by their primary drug involvement if they had used (or sold) that drug on more than 50 occasions in their lifetime, and if they had not used (or sold) another substance more than that amount. Multiple drug users were classified according to the most frequent drug used or sold in the past year. Interviews were conducted with 559 respondents over a 1-year period from June 1988 to May 1989. Sample characteristics are shown in table 1. Crack users or sellers (n=350) comprised 62 percent of the sample. Co- caine and heroin users comprised 15 and 14 percent, respectively; the re- mainder were polydrug users. One in four (23.6 percent) had been arrested and released, two in three respondents (67.1 percent) were neighborhood participants who had avoided arrest, and 1 in 11(9.3 percent) were in treat- ment. Crack respondents were younger than the others, and heroin users the oldest. Two in three were males. Crack users more often were African-Americans, and cocaine IlCI users most often were either of Puerto Rican or other Hispanic ethnicity. Slightly more than half were high school graduates, and about one in four had attended college. Procedures Interviews were conducted in a variety of settings that reflected criteria on appropriate interviewing conditions. The criteria required that interviews be confidential and anonymous--they could not be overheard by anyone else, and the identity of the respondent must be unknown to anyone in the imme- diate setting. The criteria also required that the conditions be sufficiently comfortable to sustain a conversation lasting as long as 2 hours. Finding locations where smoking was permitted, for example, posed some difficulty. Since urine specimens were requested as a validation measure, a locale with a bathroom was needed where the procedure could be verified. A final consideration was the safety of the interviewers, as they carried cash for interviewee stipends. Interviews lasted from 1 to 2 hours, with a short break after the first hour. Interview stipends of $25 were provided, plus $5 for the urine specimen and smaller fees for referrals of potential interviewees and location informa- tion for possible follow-up. Respondents also were given two subway tokens and a pack of cigarettes. Treatment respondents were not given the stipend; it was donated to the residential treatment program. They also were not asked for urine specimens, since they had been residing in treat- ment programs for 1 month or longer. Interview items were read aloud. Cards with the response sets were shown to respondents and the choices read aloud so that literacy problems were minimized. The interviews were conducted in both English and Spanish. 15 TABLE 1. Sample characteristics Primary Drug Used or Sold Crack Cocaine Heroin Polydrug n=350 n=85 n=76 n=48 Significance Background Factors (62%) (15%) (14%) (9%) p (chi square) Age at Interview .003 18 or Less 9.0 6.0 6.7 16.7 19-24 25.9 22.6 13.3 33.3 25-30 28.5 23.8 20.0 27.1 31 or Older 36.6 47.6 60.0 22.9 Age at Onset .000 18 or Less 19.6 56.6 50.7 72.9 19-24 24.0 28.9 35.6 20.8 25-30 28.9 10.8 12.3 4.2 31 or Older 27.5 3.6 1.4 2.1 Sex .158 Male 65.7 68.2 73.7 54.2 Female M.3 31.8 26.3 45.8 Race .000 Afro-American 69.6 38.8 48.7 70.8 Anglo 5.2 8.2 11.8 8.3 Puerto Rican 8.0 27.1 7.9 8.3 Other Hispanic 17.2 25.9 31.6 12.5 Education .004 Less than HS Graduate 49.7 38.8 40.5 21.7 HS Graduate M.9 37.6 37.8 60.9 Some College 15.4 25.5 21.6 17.4 Current Employment .000 Working/Student 15.7 32.9 25.7 66.7 Unemployed/Dropout 84.3 67.1 76.3 33.3 Legal/Social Status .000 Neighborhood 58.9 77.6 76.3 93.8 Arrested and Released 28.3 17.6 19.7 6.3 In Treatment 12.9 4.7 3.9 0.0 Marital Status .150 Married/Common Law 16.9 24.7 26.3 18.8 Single 65.3 56.5 53.9 72.9 Widow/Separated 17.8 18.8 19.7 8.3 Live With Children .008 No Children 40.5 28.6 33.3 54.2 Live With Child 16.1 21.4 13.3 25.0 Live Apart From Child 43.4 50.0 53.3 20.8 16 Variables Interview protocols included four domains of information: initiation into substance use or selling; lifetime and annual involvement with both sub- stances and nondrug crimes; the social processes of substance use or selling; and income sources and expenditures from both legitimate and illegal activi- ties. A calendar was used to record time, spent in treatment or detoxifica- tion programs, jails or prisons, or other Institutions. For initiation, respond- ents were asked to describe processes of initiation into their primary drug: how, where, and with whom did they initially use (or sell) the substance, how much money did they spend, and the time until the next use and regu- lar use (if any). Their expectations and reactions to the substance were recorded through multiple response items. Criminal career parameters were recorded through self-reports of lifetime estimates and annual frequencies of drug use, selling, and nondrug crimes from 1984 to the present. Specific estimates were recorded for several types of drugs used or sold, as well as a list of 20 nondrug crimes. items were worded in common language, e.g., "beat someone so badly they need- ed to see a doctor." A categorical scale was used to record frequencies of specific behaviors. This was chosen in lieu of self-reports of actual num- hers of crimes, to minimize distortion from the skewed distribution of responses for the small percentage of high-rate users or offenders. The response set represented an exponential scale frequency, with 9 categories ranging from "1 or 2 times" to "more than 10,000." The social processes of substance use and selling included several types of information. Respondents were asked whether they had sold drugs as part of an organization and to describe their organization using dimensions developed by Fagan (in press) in studies of drug selling among youth gangs. Items asked for reports of their participation in specific roles in drug selling, roles that were evident in their selling organization, and social processes that existed within their group. For example, respondents were asked if their group had specific prohibitions against drug use or sanctions for rule violations. "Systemic violence" (Goldstein 1985; Goldstein 1989) associated with drug dealing was operationally defined through eight items with specific types of violence. Respondents were asked whether they had experienced each of these violent events "regularly" in the course of their selling activity. The economic lives of respondents were described through questions on income and expenditures. Monthly dollar amounts were reported using a categorical scale of dollar ranges. This option was chosen over actual dol- lar reports to minimize distortion of dollar estimates and possible recall problems of long-term substance users. Dollar estimates were recorded for both legitimate and illegitimate sources of income and for expenditures both for living costs and for drugs. 17 RESULTS Patterns of Drug Sell********************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************************k to an already diversi- fled product line (Johnson et al. 1990; New York Times 1989d). According- ly, diverse patterns of drug selling were anticipated. Table 2 shows involvement in drug selling of four drugs over the course of respondents' criminal and drug-use careers. TABLE 2. Lifetime involvement in drug selling by primary drug involvement* Primary Drug Used or Sold Cocaine Type of Drug Sold Crack HCl Heroin Polydrug Significance. ("Regular" Sellers) n=350 n=85 n=76 n=48 p (chi square) Crack 26.1 9.4 3.9 8.3 .000 Cocaine HCl 29.8 35.7 27.6 2.1 .000 Heroin 22.0 17.6 38.2 6.3 .000 Marijuana 29.5 17.9 22.4 14.6 .022 Any Drug 46.3 43.5 46.1 22.9 .002 *Percentages exceed 100 owing to selling multiple drugs. For each user-and-seller sample, table 2 shows the percentage that sold each of four different substances more than 50 times in their lifetime. The per- centages of crack, cocaine HCl, and heroin users involved in drug selling were similar, but the types of drugs they sold differed according to the type of drug used. More than half (54.6 percent) sold at least one drug. Among crack users, about one in four (26.1 percent) sold crack, but similar percent- ages were involved in the sale of other drugs. For other subsamples, the highest percentages of sellers tended to sell the primary drug used. Cocaine HCl users rarely were involved in crack sales (fewer than 10 percent sold crack), while over one-third (35.7 percent) sold cocaine HU. Heroin users most often sold heroin (38.2 percent). They rarely were involved in crack sales, although more than one in four (27.6 percent) sold cocaine HCL Polydrug users were less often involved in selling drugs than the other drug-user samples. 18 The diverse patterns of drug selling from table 2 illustrate that many sellers were involved in multiple drug selling. Accordingly, cluster analytic meth- ods (Aldenfelder and Blashfield 1984) were used to develop a typology of drug selling to determine if distinct patterns of selling activity could be identified that would more accurately and sensitively describe drug-selling behaviors. Only those respondents reporting at least 50 selling events in their lifetimes (n=300) were included in the typology. Typology development used the lifetime frequencies of drug selling as the classification dimension. The categorical frequency scale was used, with values representing an exponential frequency scale, as follows: 0 (no parti- cipation), 1(1 or 2 times), 2 (3 to 9 times), 3 (10 to 49 times), 4 (50 to 99 times), 5 (100 to 499 times), 6 (500 to 999 times), 7 (1,000 to 10,000 times), and 8 (more than 10,000 times). An iterative partitioning method was used to identify patterns of drug selling. Squared Euclidean distance (Ward's centroid method) was used as the similarity measure. A k-means pass was used as the method to assign cases to clusters. The result was a nonhierarchical cluster analytic solution that optimized the minimum vari- ance within clusters.5 The six-cluster solution was chosen based on the shifts in cluster member- ship in successive iterations, and on its conceptual integrity (face validity). The selling types reflect differences between sellers in the joint distributions of selling of each of four types of drugs: heroin, crack, cocaine HCl, and marijuana. Validation procedures relied on interpretation plus the face validity and internal consistency of the aggregate behavioral characteristics of each group and the total sample classification. For example, one type specialized in heroin sales; the mean lifetime frequency of heroin sales was highest for this group and significantly lower for the other types. The results are shown in table 3 and figure 1. TABLE 3. Lifetime frequency of drug selling by type of drug and seller type Type of Drug Sold Cocaine Seller Type n Crack HCl Heroin Marijuana 1. Marijuana (49) .27 1.84 .49 4.82 2. Heroin (33) .15 1.61 5.18 .12 3. Cocaine, Heroin, and Marijuana (45) .53 5.56 5.93 4.96 4. Low-Level Crack and Cocaine (93) 2.48 1.75 .34 .58 5. Crack, Cocaine, and Marijuana (54) 4.85 4.54 2.52 5.22 6. Crack, Cocaine, and Heroin (26) 4.85 5.19 5.27 1.19 ANOVA: F 99.8 44.3 151.0 156.5 P(f) .000 .000 .000 .000 19 Ln of Lifetime Frequency 6 5 4 3 Type of Drug Marijuana 2 1 Heroin )~ Cocaine~Heroin/MJ 0 Low-level Crack 1 ______ Crack/Coke/MJ Crack Coke Heroin 0 Crack Cocaine HCL Heroin Marijuana Type of Drug Sold FIGURE 1. Typology of drug selling Table 3 and figure 1 show the distribution of cases by seller types, and the mean frequency of lifetime selling by type of drug for each seller type. Analysis of variance (ANOVA) tests for all index scores were significant (p=.00O), a confirmation of the internal validity of the classification results. Marijuana Sellers (Type 1) appropriately have the highest mean selling fre- quency for marijuana. They have relatively low mean scores for crack and heroin and moderately high scores for cocaine RU. Heroin Sellers (Type 2) have the highest mean selling frequency for heroin, moderate mean fre- quency scores for cocaine HCl, and low scores for other drugs. The other types reflect patterns of multiple drug selling. Type 3 (Cocaine RU, Heroin, and Marijuana) had low mean frequency scores for crack, but high scores for the other drugs. Type 4 (low-level Crack Sellers) had moderate frequency scores for crack and cocaine HU, but low scores for other drugs. Type 5 (Crack, Cocaine, and Marijuana) had high lifetime fre- quency scores for selling crack, cocaine HCl, and marijuana. They also had moderate scores for heroin and could alternately be classified as sellers of all drugs. Type 6 (Crack, Cocaine, and Heroin) had high scores for all drugs other than marijuana. 20 The typology of six seller types recognizes distinct selling patterns that capture the complexity of drug-selling behaviors. In turn, it provides a unique basis for comparison of the social organization of drug-selling types and a basis for interpretation of their involvement in specific varieties of violent behaviors. The Social Organization of Drug Selling Prior studies of the social organization of drug selling (Adler 1985; Williams 1989; Fagan, in press; Johnson et al. 1990; Fields 1985; Cooper 1987; Mieczkowski 1986; Mieczkowski, this volume; New York Times 1989d) suggest that selling activities vary extensively according to participa- tion in a group, as well as to the social processes, organizational structure, and internal cohesion of the group. Few of these studies have compared social organization among various seller types or drugs sold, nor have they related social organization to specific behaviors of sellers. Tables 4 and 5 compare two aspects of the business structure and social processes among sellers and within selling groups, as described by members of different sell- er types. Participation in a variety of selling roles is shown in table 4. Johnson et al. (1985) found distinct patterns of drug use and nondrug criminality among participants in the heroin trade depending on their specific role in drug selling. In this study, respondents were asked whether they had per- formed each of several types of roles, from street-level transactions to "wholesalers" and suppliers of equipment ("sell and rent works").6 Table 4 suggests that differences are evident in participation in each of the seven selling roles by seller type. These differences suggest that differences may exist in the organizational structures of drug selling according to the type of drug sold. Type 4 (low-level Crack and Cocaine) sellers least often reported involvement in formal roles and also least often reported par- ticipation in selling transactions. Similar results were apparent for Types 1 and 2, suggesting that sellers of marijuana and heroin were less often in- volved in drug-selling groups. Respondents in Types 3, 5, and 6 (high-rate multiple drug selling) most often reported participation in formal drug- selling roles. These findings suggest that cocaine HCl and crack selling are more highly organized activities, with sellers more often participating in a broader range of roles with increasing responsibility. However, the sam- pllng strategy may have influenced these results. Table 5 examines respondents' reports of their participation in drug-selling organizations or groups and examines whether their group contains each of several specific types of social organization or processes. Participants were asked if they had participated in a "group or gang" that sold drugs.7 Those who indicated that they were part of a group then were asked if their group contained any of six specific features. An index of group organization was 21 TABLE 4. Selling roles by type of seller (percentage of dealers in each role) * Type of Seller Cocaine, Crack, Crack, Heroin, and Low Crack Cocaine, and Cocaine, and Marijuana Heroin Marijuana and Cocaine Marijuana Heroin Significance Role in Drug Selling n=49 n=33 n=45 n=93 n=54 n=26 p (chi square) Selling to Customer 89.8 87.9 97.8 78.5 100 96.2 .000 Middleman 53.1 51.5 77.8 54.8 70.4 61.5 .045 Lookout 55.1 60.6 73.3 49.5 72.2 76.9 .013 Cut, Package, or Cook 77.6 57.6 91.1 53.8 81.5 92.3 .000 Lieutenant 22.4 30.3 53.3 25.8 46.3 61.5 .000 Wholesaler 55.1 33.3 68.9 30.1 59.3 61.5 .000 Sell and Rent Works 24.5 24.2 48.9 37.6 53.7 38.5 .014 *Percentages exceed 100 due to multiple selling roles. TABLE 5. Social organization of selling groups by type of seller (reports by sellers about their group)* Type of Seller Cocaine, Crack, Crack, Heroin, and Low Crack Cocaine, and Cocaine, and Marijuana Heroin Marijuana and Cocaine Marijuana Heroin Significance Organizational Feature n=49 n=33 n=45 n=93 n=54 n=26 p (chi square) A Specific Name 89.8 87.9 97.8 78.5 100 96.2 .000 leaders and Supervisors 53.1 51.5 77.8 54.8 70.4 61.5 .045 Rules and Sanctions 55.1 60.6 73.3 49.5 72.2 76.9 .013 Rules Against Use While 77.6 57.6 91.1 53.8 81.5 92.3 .000 Dealing A Specific Territory 22.4 30.3 53.3 25.8 46.3 61.5 .000 Kids Under 16 Selling 55.1 33.3 68.9 30.1 59.3 61.5 .000 Percentage in Group 10.2 30.3 24.4 33.3 59.3 50.0 Group Organization Index** .35 .91 1.04 1.29 2.54 2.12 *Percentage of group members reporting each feature. **Mean for all sellers, including sellers not in groups. constructed by summing the positive responses to each of the six features. Table 5 shows that marijuana sellers (Type 1) least often reported being part of a selling group (10.2 percent). Between 24 and 33 percent of mem- bers of Types 2, 3, and 4 reported being in a group, while over half of Types 5 and 6 reported being in a drug-selling group. The types differed significantly on cross-tabulations for each dimension. Among those reporting group participation, most said that their group had a specific name, although Type 4 respondents had a lower rate. The findings for the other dimensions reflected patterns similar to those in table 4. Parti- cipants in Types 1, 2, and 4 least often reported the presence of the several features of group. Inclusively, their reports suggested that they saw their groups as being less formally organized and having fewer unifying social processes. Respondents in Types 3, 5, and 6 who reported being in selling groups most often reported the presence of formal structures or processes. Similar patterns for these types were found for role differentiation in table 4. More than 80 percent reported prohibitions against using drugs while selling, com- pared to about half in Types 2 and 4. They more often reported having specific territory, leaders and supervisors, and formal rules and sanctions. They also more often reported using juveniles (less than 16 years of age) in drug selling. The index of group organization further showed these distinctions: respond- ents in Types 5 and 6 reported the highest scores for group organization, and respondents in Types 1, 2, and 3 had the lowest (p=.()o()). Significant- ly, these tw0 types are most often involved in selling crack. Type 3 sellers, despite their involvement in selling cocaine HCl and heroin, had lower indices of group organization. The results suggest that crack selling is a more formally organized activity: it more often occurs within selling groups, and crack-selling groups more often have a formal, hierarchical social organization. Violence in Drug Selling Although there is overwhelming evidence of an association between drugs and violence, the violence that characterizes drug use or selling actually is a heterogeneous set of behaviors. The empirical evidence of causal directions between drug involvement and violence consistently has yielded contradic- tory results (Watters et al. 1985). Thus, the drug-violence connection for now may be best understood as a probabilistic function, with uncertain causal mechanisms or temporal order (Anglin 1984). Goldstein (1985; Goldstein 1989) suggests that different theories may be needed to account for different drug-crime relationships. In his tfl- partite framework, he distinguishes "pharmacological" violence linked to 24 psychoactive effects of drug ingestion from "economic compulsive" violence in which drug users engage in crimes to support the costs of drug use. "Systemic" violence is the third type of drug-crime relationship. It is violence that is intrinsic to buying and selling any illicit substance: traditionally aggressive patterns of interaction within the system of drug distribution and use . . . disputes over territory between rival drug dealers, assaults and homicides committed within dealing hierarchies as a means of enforc- ing normative codes, robberies of drug dealers and the usually violent retaliation by drug dealers or their bosses, elimination of informers, disputes over drugs and/or drug paraphernalia, punishment for selling phony or adulterated drugs, punishment for failing to pay for one's debts, and robbery violence related to the social ecology of [buying] areas. (Goldstein 1989, p. 30) Systemic violence was expected to be greater in crack distribution than in other drug markets for two reasons. First, crack selling was concentrated in neighborhoods where social controls had been weakened by intensified social and economic dislocations in the decade preceding the emergence of crack. Second, the rapid development of new drug-selling groups following the introduction of crack brought with it competition. Accordingly, violence within new selling groups internally to maintain control and violence exter- nally to maintain selling territory and integrity (product quality) was more likely to characterize the unstable crack markets than more established drug markets and distribution systems. Table 6 examines the percent of respond- ents within types reporting "regular" systemic violence. Items were con- structed to reflect the dimensions of systemic violence defined above. For each type of systemic violence, there were significant differences in the prevalence of regular violence. Most important, each type of systemic vio- lence was reported most often by sellers in two of the three crack-seller types and least often by marijuana and heroin sellers. Type 4 (Low-Level Crack and Cocaine) sellers reported systemic violence less often than did other crack or cocaine sellers; their reports of systemic violence closely resemble the reports of heroin or marijuana sellers for nearly all items. Evidently, a wide range of violent acts is intrinsic to frequent crack or cocaine selling. For nearly all varieties of systemic violence, between 40 and 50 percent of the Type 5 and 6 respondents reported their regular occurrence. Sellers who worked in groups were compared with those who sold outside any formal or informal structure for the level of systemic violence. Re- spondents were classified according to whether they reported that their sell- ing activity was alone or in a group (see table 5). A scale of systemic violence was constructed by summing responses to the eight individual 25 TABLE 6. Systemic violence by seller type (percentage reporting "regular" occurrence) Type of Seller Cocaine, Crack, Crack, Heroin, and low Crack Cocaine, and Cocaine, and Marijuana Heroin Marijuana and Cocaine Marijuana Heroin Significance Violence in Drug Selling n=49 n=33 n=45 n=93 n=54 n=26 p (chi square) Fights With Rival Dealers 10.2 9.1 20.9 20.4 40.7 38.5 .001 Assaults to Collect Debts 12.2 12.1 20.9 18.3 44.4 38.5 .000 Fights With Other Dealers 16.3 12.1 32.6 16.1 40.7 42.3 .001 Over Quality of Drugs Robbery of Other Drug 6.1 12.1 18.6 11.8 44.4 15.4 .000 Dealers Robbery of Drug Buyers 12.2 6.1 23.8 8.6 33.3 23.1 .001 Disputes Over Paraphernalia 22.4 30.3 23.3 28.0 50.0 34.6 .028 Victimization While Selling 12.2 18.2 20.9 26.9 50.0 46.2 .000 Fights With Buyers Over 4.1 21.2 23.3 12.9 42.6 30.8 .000 Quality of Drugs items. Table 7 reports the results of analysis of covariance (ANCOVA) routines testing differences in systemic violence scale scores for selling alone or in a group.8 Covariates were introduced for the age of the respondent and self-reports of monthly income from drug selling in the past year. TABLE 7. Analysis of variance of systemic violence by seller type and group involvement Sell Alone Sell in Group Seller Type n=180 n=120 Marijuana .74 2.29 Heroin .68 2.27 Cocaine, Heroin, Marijuana .88 3.92 10w-level Crack .91 2.19 Crack, Cocaine, Marijuana 2.65 3.84 Crack, Cocaine, Heroin 1.45 3.60 All Sellers 1.03 3.08 NOTE: ANCOVA (Significance of F) Main Effects: Type, p=.000; Group, p=.000; Type x Group, p=.355. Covariates: Selling Income, p=.000; Age, p=.099. Main effects were significant (p=.()()o) for both seller type and selling group status, and there were no significant interactions. Selling income was not a significant covariate, but age as a covariate approached significance (p=.099). For each seller type, systemic violence was far greater among sellers in groups. Among those who sold alone, crack and cocaine HU sel- lers (,Types S and 6) reported the highest violence scores, although Type 5 sellers had much higher violence scores. These differences were less evi- dent for group selling. Cocaine HU sellers (as part of multiple drug sel- ling) had the highest systemic violence scores among sellers either alone or in groups, regardless of whether they sold crack concurrently. It is the fre- quency of selling cocaine products, not just selling its smokable form, that seems to best explain violence in drug selling. low-level Crack and Cocaine Sellers (Type 4) have lower violence scores compared to other crack sellers, in groups or alone. This suggests that fre- quent crack selling also may be associated with systemic violence only if it occurs concurrently with cocaine HCl selling. Although the violence poten- tial for selling crack alone is quite variable, frequent selling of any cocaine product in a group appears to be a particularly violent enterprise. This may reflect exposure during group dealing to individuals and situations for which 27 violence is commonplace or a self-selection process that determines who becomes involved in drug-dealing groups. It also may reflect the impor- tance of violence as a regulatory and management strategy within selling groups in which both internal discipline and maintenance of market share are required. The relationship between participation in a selling group and systemic vio- lence, shown in table 7, suggests that crack or cocaine HCl selling in groups involves greater involvement in systemic violence. Table 5 shows that the social organization of crack-selling groups is better developed than other groups. Accordingly, systemic violence is more evident both in group selling of cocaine products and in groups with stronger social organization. Compared to group or individual sellers of heroin or marijuana, the selling groups that have developed in the crack market appear to have a stronger social organization and are more likely to engage in a wider range of vio lent acts within the social and economic boundaries of drug transactions. Drug Selling, Drug Use, and Nondrug Crimes If systemic violence is part of a general pattern of intentional law viola- tions, then violence that occurs outside the context of drug selling should be distributed similarly to violence within those contexts. However, if systemic violence is a form of social control and regulatory behavior, then the distri- bution of systemic violence should differ from the distribution of nondrug violent acts. Moreover, since crack-selling groups developed rapidly and often in the absence of an existing market structure, systemic violence was expected to be greater among crack sellers than others. The previous sec- tion confirmed this belief. If these differences for crack sellers were not evident in other forms of violence, then systemic violence among crack sel- lers might be interpreted as an economic behavior and a form of social con- trot If crack sellers also are more often involved in violence outside the selling context, however, then systemic violence and other violence might be interpreted as indicative of part of a generalized pattern of intensified criminal behaviors among people involved in crack. Respondents were asked to indicate their lifetime involvement in each of 11 nondrug crimes, using a categorical response set for frequencies, using the previously described exponential scale (p. 19). ANOVA routines compared lifetime frequencies by seller type, controlling for group involvement in drug selling. Means for nonsellers are presented in the table, but were not included in the analyses. Age, group cohesion, and selling income were introduced as covariates. Table 8 shows that significant effects (p=.05 or less) by seller type were obtained for 5 of the 11 crime categories: robbery of persons, breaking and entering, auto theft, weapons offenses, and selling stolen goods. Results approached significance (p=.07) for three other categories: robbery of 28 TABLE 8. Analysis of variance of nondrug crimes by seller type and group selling* Type of Seller Significance of F Main Effects Covariates Cocaine, Crack, Crack, _________________ _________________ Selling Heroin, and Low Crack Cocaine, and Cocaine, Type I Nondrug Crime Status Nonsellers Marijuana Heroin Marijuana and Cocaine Marijuana and Heroin Type Group Group Age Cohesion Income Robbed Businesses Alone .30 .29 .64 1.25 .21 .59 .64 .066 .843 .979 .002 .002 .719 Group .71 .55 1.46 .64 1.00 1.00 Robbed Persona Alone .50 .74 1.45 1.88 - .52 1.88 .45 .002 .916 .122 .163 .038 .255 Group .57 .45 1.69 1.11 1.86 1.80 Broken into Homes Alone .29 .60 1.14 1.38 .32 .35 .09 .050 .447 .250 .005 .814 .164 To S,~ Group .43 .27 1.15 .39 .59 .87 Beat Someone Up Alone .12 .43 .41 .91 .39 .47 .55 .069 .318 .720 .219 .002 .072 Badly, Hurt Them Group .29 .27 1.23 .61 1.30 .67 Fighting Alone .67 1.83 1.27 2.09 1.48 2.18 1.91 .337 .314 .880 .098 .118 .198 Group 1.29 1.18 2.23 1.97 2.11 2.00 Carried Weapons Alone .86 1.57 2.23 3.63 1.57 2.59 3.27 .003 .516 .984 .002 .005 .000 Group 1.71 2.55 4.46 2.31 3.13 3.13 Stolen a Car Alone .24 .50 AS .72 .20 .76 .18 .028 .071 .710 .294 .260 .059 Group .14 .18 1.23 .14 .62 .27 Shoplifted Alone 1.27 1.79 2.41 2.75 1.32 1.71 1.09 .337 .271 .228 .015 .467 .647 Group 2.14 1.18 1.69 1.69 2.11 1.53 Stolen Money or Alone 1.02 1.24 2.41 1.81 1.04 1.65 .64 .071 .800 .029 .188 .286 .692 Valuables Group 1.14 .55 138 1.14 2.24 2.13 Stolen Things Alone .81 1.14 1.55 1.53 .88 1.59 .82 .404 .857 .415 .006 .748 .477 Worth <$50 Group 1.29 .27 1.71 .81 1.51 1.60 Sold Stolen Goods Alone .65 1.40 1.23 2.16 .86 1.35 1.09 - .035 .524 .523 .012 .000 .015 Group 1.14 1.64 2.38 1.17 2.65 1.93 5Mean score for categorical index of lifetime frequency; nonsellers excluded from ANOVA significance tea's. businesses, aggravated assault, and grand theft. In nearly all of these offense categories, lifetime frequencies were lowest for nonsellers and Type 1 and 2 sellers (sellers of other than cocaine products). Lifetime criminality was significantly higher for Type 3, 5, and 6 sellers in nearly all the of- fense categories in which the F-value was significant. These seller types were sellers of multiple drugs, including cocaine products. The trends also suggest that differences between seller types are less evident for less serious offenses: fighting, shoplifting, and petty theft. For nearly all offense categories and seller types, group sellers had greater lifetime involvement than individual sellers. Significant interactions were obtained only for grand theft: individual sellers in Types 1, 2, and 3 had higher lifetime involvement than group sellers, but the opposite trend was found in Types 4, 5, and 6. Inspection of the means for nonsellers shows that their involvement in nondrug crimes was substantially less than either individual or group sellers. Covariate effects for age were significant for several crime categories: business robbery, breaking and entering, shoplifting, weapons offenses, and selling stolen goods. Age was not significant in crimes of physical aggres- sion, nor in person robbery. Group organization was a significant covariate in business robbery, person robbery, assault, weapons offenses, and selling stolen goods. Selling income was a significant covariate only for weapons offenses and selling stolen goods. The results clearly show that involvement in nondrug violent crimes is greater for sellers of cocaine products, especially for those groups with more well-articulated organizations. Unlike the evidence on systemic vio- lence, however, there appear to be minimal differences between Types 5 and 6 (crack sellers) and the Type 3 noncrack cocaine sellers. This sug- gests that participation in multiple drug-selling groups, rather than simply crack-selling groups, is associated with involvement in a wide variety of crimes and, specifically, violence. The influence of group social organiza- tion on nondrug violence is consistent with its influence on systemic vio lence. Evidently, participation in a well-organized drug-selling group is strongly associated with involvement in violence in a variety of circum- stances and contexts. The effects of initiation into drug selling on specific forms of aggression also were compared by seller type, including nonsellers. Respondents were presented with a series of six items describing specific forms of aggression and one item about victimization from violence and asked whether their involvement had increased, decreased, or remained the same following initi- ation into crack use or selling. The percent of respondents reporting either increases or decreases is shown in table 9. 30 TABLE 9. Self-reported changes in specific forms of violence by seller type after initiation into primary drug Type of Seller (Percentage Reporting Change) Cocaine, Crack, Crack:, Specific Forms Heroin, and ILow Crack Cocaine, and Cocaine, and Significance of Violence Nonsellers Marijuana Heroin Marijuana and Cocaine Marijuana Heroin p (chi square) Stabbings .119 Involved Less 1.6 4.2 3.0 9.3 2.2 5.6 3.8 Involved More 2.8 6.3 3.0 2.3 3.2 9.3 7.7 Shootings .000 lnvolved less 0.4 2.0 6.1 11.6 4.3 11.1 11.5 Involved More 0.8 2.0 0 4.7 4.3 7.4 3.8 Assaults or Beatings .000 Involved Less 2.4 10.4 12.1 11.6 5.4 22.2 7.7 Involved More 4.7 8.3 6.1 7.0 10.8 9.3 15.4 Fighting .001 Involved less 5.9 12.2 24.2 28.6 9.7 24.1 7.7 Involved More 13.8 18.4 12.1 14.3 29.0 29.6 13.8 Robberies .000 Involved less 3.6 12.2 6.1 30.2 8.6 16.7 15.4 Involved More 12.3 16.3 21.2 14.0 20.4 35.2 23.1 Injuring Someone .199 Involved less 1.2 6.1 0 9.3 4.3 3.7 3.8 Involved More 6.7 4.1 6.1 2.3 6.5 7.4 11.5 Injured by Someone .012 Involved less 4.3 10.2 9.1 16.3 10.8 18.9 7.7 Involved More 11.0 16.3 18.2 7.0 14.0 20.8 30.8 Significant differences were found for four of the six violent acts: shooting, assault, fighting, and personal robbery. Among nonsellers, increases were reported more often than decreases for all acts. Type 3 (Cocaine, Heroin, and Marijuana) sellers report only decreases. Shootings decreased more often among all seller types, - a surprise given the higher levels of systemic violence associated with three of the seller types. Among Type 5 and 6 sellers, more respondents reported increases than decreases in robberies, stabbings, and injuring someone. Thus, violence more often increased than decreased among most crack and cocaine sellers after initiation into drug use. Since the onset of drug use preceded selling for most sellers (Fagan and Chin, in press), it seems that violence potentials may have preceded involvement in selling. Finally, drug-use patterns among sellers and nonsellers were analyzed. Recent evidence on drug selling in inner cities found that selling groups prohibited drug use among their members, especially during business hours (Chin 1986; Cooper 1987; Mieczkowski 1986; Williams 1989). Vigil (1988) reported that Chicano gang members in East Los Angeles rejected heroin users from the gang, believing that a gang member could not main- tall' loyalty to the gang and to his or her addiction at the same time. Others (Fagan 1989) found that drug use and dealing were intrinsic to gang life. Studies of drug dealers found that they "drift" into dealing from their participation in drug-using circles, rather than suddenly entering into dealing from outside drug cultures or scenes (Adler 1985; Murphy et aL 1989). Among this sample, table 5 showed that at least half of the respondents in each seller type reported prohibitions against drug use while dealing. Ac- cordingly, variation in drug-use patterns was anticipated. Crack sellers, whose organizations seemed to be well articulated, were expected to have relatively low drug use. Other seller types, whose organizations were less formal, were expected to report greater involvement in drug use. Respond- ents were asked to report their lifetime frequency of substance use for 15 substances. Since multiple drug-use patterns are commonplace among high- rate drug users, factor analyses were used to identify distinct underlying trends in drug use. Four factors were identified, accounting for 60.2 per- cent of the variance: intravenous (IV) heroin and cocaine use, cocaine (and crack) smoking or snorting, oral stimulant and depressant use (pills, psyche- delic drugs), and marijuana and alcohol use. The factor coefficients and statistics are shown in table 10. The factor scores were retained and used for comparisons of drug use among seller types. ANCOVA routines compared factor score means for each of these four di- mensions of substance use by seller type. Means factor scores for non- sellers are shown, although they were excluded from the analyses. To test for the influence of group participation, a second independent variable for group selling was included. Covariates for age also were included. Table 11 shows that significant differences by seller type were evident for all 32 TABLE 10. Rotated factor coefrienits for lifedme frequency of drug use Heroin and Cocaine and Uppers and Alcohol and Type of Drug Cocaine IV Crack Smokers Downers Marijuana Crack -.121 .781 .004 -.056 Oocaine-Snorting .252 '555 -.014 388 Speedball (IV) .896 -.055 .108 .030 Cocaine IV .846 -.035 .186 -.051 Cocaine-Freebase .072 .829 .113 .101 Heroin IV .903 -.041 .120 .068 Heroin-Snorting .716 .093 .025 .163 Methadone .788 -.020 .161 -.094 Marijuana -.060 .032 .061 .795 PCP -.144 .486 329 .057 lsD .056 .239 .645 .256 Speed, Uppers .089 .108 .823 .117 Barbiturates, Downers A34 .039 .666 - .027 Other Drugs .089 -.017 392 .021 Alcohol -.003 .099 .188 .662 Eigenvalue 4.19 2.53 1.26 1.05 % Variance Explained 27.9 16.9 8.4 7.0 dimensions of drug use except pill use. Group status was significant only for cocaine smoking and snorting. There were no significant interactions, and age covariates were significant only for the IV-drug-use dimension. Drug-use patterns tended to reflect seller type, especially for individual sellrs. The highest factor score means for heroin sellers were for IV drug use, for marijuana sellers were for marijuana use, and for cocaine or crack sellers were for cocaine smoking or snorting. There were small differences in cocaine use between individual and group sellers for Type 5 and 6 crack sellers. But cocaine use among Type 4 cocaine sellers in groups appeared to be substantially lower than among individual sellers. This may reflect organizational rules or norms opposing substance use. W drug use was evident only in those groups in which heroin was sold and was most evident in Type 3 and 4 individual sellers. Type 6 sellers (crack and other drugs) in groups had the highest factor scores for illicit pill use. Type 4 and 5 crack sellers avoided IV drug use, suggesting that they did not inject cocaine despite their high involvement in smoking or snorting it. For all four types of substance use, nonsellers were less often involved than were sellers, regardless of whether they sold individually or in groups. ll- licit pill use among marijuana sellers was the only drug use greater for indi- vidual than for group sellers. 33 TABLE 11. Analysis of varifmce of drug use factor scores by seller type and group selling controlling for age* Type of Sciler coi:aine, low Crack Crack, Cra~i~ Sigtiifieance of F Soiling Heroin, and and Cocaine, and Cocaine, jype by Drug-Use Factor 5tat:s Nonsellers Marijuana Heroin Marijuana Cocaine Marijuana and Heroin Type Onip Gmt~, Aget Heroin and Cocaine Alone -.10 .13 1.23 1.41 -.31 -.29 .01 .000 .986 .211 .000 IvUite Gmup .08 .95 .96 -.35 -.20 .40 Cocaine 5moi':ing Alone .05 .12 -.40 .39 .40 .87 .49 .000 .008 .449 .101 orSnooing'* GrOL~ -.39 -.98 -.32 .22 .75 .52 Pill Use and Alone .44 .55 .05 -.25 .47 -.22 .09 .120 .622 .219 .778 Psyclietielici' Gi':up -.20 .19 .14 .13 .07 .87 Marijuana Use Alone .52 .48 -.14 .27 .08 .51 -.22 .001 .564 .387 .341 Gtotip .78 .10 .79 -.10 .37 -.04 *Mean score for categorical index of lifetime frequency; nonsellers excluded from ANOVA significance tests. **including crack smoking. toOvariate, main effects adjusted for covariate effects. Table 11 suggests that drug use and selling jointly occur within the social worlds of specific drugs. There was little evidence that sellers avoided using the drugs that they sold; in fact, they tended to have the greatest use of those drugs they sold. Only amoilg group cocaine and heroin sellers (Type 4) was there evidence of avoidance of use of the drug they marketed. W drug use was confined to specific groups that also sold heroin, and these groups tended to avoid cocaine smoking or snorting. Cocaine smoking and snorting was evident among the groups that sold crack plus among individu- al Type 4 sellers. Despite the high proportion of sellers that reported prohibitions against use while selling, many sellers also used drugs. Evidently, these prolltibitions did not extend to personal recreational use, or they were ineffective. Use and dealing appear to be reciprocally related, with access to the substance and immersion in a drug-specific social network likely contributors to the drug-specific patterns. Substance use appears equally likely regardless of whether selling occurs alone or in groups. The provocative image of the well-disciplined dealer, whose motivations are exclusively financial and who abstains from drug use to maximize his or her dealing skills, has no grounding in these data. Drug sellers also are drug users, and their efforts as dealers and behaviors as users apparently overlap extensively. CONCLUSION Young crack sellers have been portrayed in the popular literature as young entrepreneurs, highly disciplined and coldly efficient in their business activi- ties, often using violence selectively and instrumentally in the service of profits. An ethos that rejects drug use also has been attributed to new, young crack sellers, especially those in groups, whose interest is not drug use but the material wealth that rewards the most efficient seller. Crack- selling groups have been described in the popular media as emerging organ- ized crime groups, with nationwide networks of affiliates and franchises to distribute drugs (Newsweek 1986). This image tends to attribute the spread of crack use in urban areas to a conspiracy involving cocaine importers, lcIsecent organized crime groups, and youth gangs from the inner cities of the major cocaine importation areas. This study suggests that none of these stereotypes appear to be true. Crack sellers are violent more often than other drug sellers. Further, their violence is not confined to the drug-selling context. Compared to other drug sellers and nonellers, they more often are involved in a wide rsnge of serious nondrug crnes, including both property and violent offenses. They also are involved in patterns of multiple drug use. like other drug sellers, they most often use the drugs that they sell and avoid others that may be unfamiliar. Drug-use patterns of both crack and other drug sellers suggest that drug use and dealing occur within distinct but parallel social worlds that are characterized by generic social and economic processes. 35 Violence in drug selling may be interpreted as part of a generalized pattern of crime and violence or as an economic behavior that reflects aspects of social control, good business strategy, and self-help. Cettainly, the evidence on systemic violence within crack-selling groups suggests that they are more likely to employ violence both for organizational maintenance and as a stra- tegic weapon in economic competition. However, crack sellers also are more likely to be involved in a wide range of law violations as well as regular drug use. The diversity and frequency of nondrug crimes suggests that crack sellers' behaviors are neither moralistic acts nor crimes in the pursuit of justice. That is, these appear to be neither crimes of social con- trol nor self-help. In fact, their patterns of drug use and crime suggest a pattern of spuriously related behaviors indicative of a generalized pattern of deviance. Drug use, drug selling, and violence were evident among all sel ler types. Compared to other drug sellers, crack sellers simply seem to be more deeply immersed in generic social processes of drug use, violence, and other crimes. Accordingly, both views seem appropriate. Like other offenders, drug sell- ers exhibit versatility in their patterns of violence and other crimes (Klein 1984). Crack and cocaine HU sellers are more likely than other sellers to use violence for economic regulation and control, but are also more likely to use violence in other contexts. Violence among crack sellers may reflect either processes of social selection or the contingencies of the social settings in which crack selling is concentrated. These distinctions cannot be sorted out in these data, and perhaps they are reciprocal processes that cannot be disentangled. Nevertheless, the results suggest common pathways to drug use, drug selling, and nondrug crimes. For many sellers of cocaine prod- ucts, crack has been integrated into behaviors that were evident before their involvement with crack or its appearance on New York City streets. If violence is both intrinsic to drug selling and, in urban areas, part of a generalized pattern of deviance, then the patterns of violence within drug selling are specific applications of behaviors that also occur in other con- texts. Thus, it would be unwise to conclude that the drug business makes people violent or that people are violent in the context of drug selling but not elsewhere. Drug selling is etiologically related to violence, but only because violence is intriitisic to drug selling. It is more likely that drug selling provides a context that facilitates violence, in which violence is acceptable given the illicit nature of drug selling and the absence of other forms of legal recourse or social control. Nevertheless, crack sellers more often are involved in violence and drug use. The crack market apparently has intensified the social processes that sustain both drug-related and other violence. Crack has evolved in a specif- ic and econoililic social context, in which social and economic transforma- tions have altered the formal and informal controls that previously had 36 shaped violent behaviors. Also, factors unique to crack distribution appar- ently contribute to the increased violence. The expansion of illicit drug sales in New York City has paralleled the decrease in legitimate economic opportunities in this decade. Participation in the informal economy has increased, especially among minorities living in neighborhoods where the demand for goods and services in the informal economy rivals participation in the formal economy (Sassen-Koob 1989). In the volatile crack markets, crack sometimes has become a "currency of the realm," a liquid asset with cash value that has been bartered for sex, food, or other goods (Inciardi, in press; Williams 1989). Sellers or users with large amounts become targets for "take offs" by either other sellers or users wanting the drug. In turn, violence as self-defense is a common theme and an essential element in controlling situations in which large volumes of crack are present (Bourgois 1989). Johnson et al. (in press) suggest that there is a process of self-selection and social selection of violent persons in the crack trade that accounts for higher levels of violence than in previous drug epidemics. These people are used both to maintain internal discipline in drug-selling groups and as combatants in territorial disputes. Harnid (1990) attributes increases in violence asso- ciated with crack to the erosion of formal and informal social controls in neighborhoods whose human, social, and economic capital has been depleted over the past two decades. High rates of residential mobility and declining capital investment have contributed to an ecology of violence in several inner-city areas. The emergence of a volatile crack market perhaps has benefitted from these processes and intensified them. The participation of generally violent offenders in the crack trade, coupled with decreased con- trols and increased clime opportunities in socially fragmented areas, may account for the increased violence in the crack market. lf street-level drug sellers, in general, and crack sellers, in particular, exhibit behaviors that are part of a generalized pattern of deviance, then the charac- terization of crack-selling networks as a new organized crime menace has disillusioned the public as to appropriate social policies. If these new or- ganizations are responsible for drug selling and its attendant violence, then it is difficult to explain the unlimited flow of new people who are selling drugs. Policies that seek sources of conspiratorial decisions to sell drugs risk the danger of reifying the image of drug dealers as cold businessmen and entrepreneurs and rejecting debate on other policies that might address the entry of young men and women into drug selling and a wide range of violent behaviors. If violence and drug selling in the crack market reflect the social and economic disorganization of the neighborhoods where crack selling is concentrated, then policy should reflect sensible thinking about how to strengthen social areas to control crimes, stop the production of violent offenders, and mitigate crime-producing conditions. 37 FOOThOThS 1. Adler points out that, although violence was rare in the drug-selling scenes she observed, it was always in the background as an implied threat in lieu of legal recourse to mediate business disputes. 2. MeGeary and Lynn (1988) comprehensively reviewed the economic re- structuring of American inner cities over the past 20 years. 3. The definition of social control used here is similar to the processes de- scribed by Black (1983) and refers to the processes that people use to respond to deviant or antagonistic behaviors. These may include verbal expressions of disapproval or threats or sanctions that may either punish or incapacitate. Self-help refers to responses to aggression or threat. 4. Belenko et al. (in press) analyzed arrest patterns for crack offenses. 5. This approach to grouping subjects used their relative proximity in a specified dimensional space. The nonhierarchical centroid method was less useful than the hierarchical models as a heuristic tool, as it dis- played neither agglomerative nor divisive linkages (dendograms). How- ever, this weakhess was addressed by running sequential solutions that specified cluster sizes from three to seven. Comparisons of each suc- cessive iteration approximated a divisive hierarchical analysis. This classificatory procedure posed no question of statistical significance in the derivation procedure. The clusters were a heuristic tool that was instructive for partitioning drug sellers into groups for descriptive and analytic purposes. The types should be interpreted cautiously, however, as the procedure is sensitive to shifts in sample composition. 6. Johnson et aL (1985) defined each type of role. These definitions were read aloud to respondents during the interview. 7. Participants in New York City refer to their groups as "crews," "posses," or other terms specific to locales or ethnicity of the members. Such groups are distinct from groups of street~corner youths or youth gangs, in that drug-selling activities provide the rationale for group affiliation. They also may be polyethnic groups, unlike the ethnic or neighborhood affiliations common in youth gangs. Williams (1989) described "crews" in New York, and Klein et al. (in press) described the confluence of drug selling and "traditional" youth gangs in Los Angeles. 8. ANCOVA procedures first considered covariate effects in descending order of their F scores and adjusted the main effects and interactions for effects of covariates. 38 REFERENCES Adler, P.A- ~eling and Dealing: An Ethnogrœiphy of an Upper-Level Llealing and Smuggling Communily. New York: Columbia University Press, 1985. Aldenfelder, M.S., and Blashfield, R.K- Cluster Analysts. Beverly Hills, CA: Sage Publications, 1984. Anglin, M.D. A review of alcohol, drugs, and aggression. J Stud Alcohol 35(5):469A90, 1984. Ball, J.C.; Sehaeffer, J.C.; and Nurco, D.N. The day4o-day criminality of heroin addicts in Ballimore: A study in the continuity of offense rates. Drug Aicohol Depend 12:119-14Z 1933. Belenko, S.; Fagan, J.; and chin, K- Criminal justice respos to crack. J Res Crime Del-œen~t 21(1), in press. Biernacki, P., and Waldorf, D. Snowball sampling: Problems and tech- niques of chain referral sampling. Sociol Meth Res 10:141-163, 1981. Black, D. Crime as social controL Am Sociol Rev 48:3445, 1983. Bourgois, P. In search of Horatio Alger: Culture and ideology in the crack economy. Contemp Drug Probl 16(4):619~50, 1989. Chin, Ko~Lin. Chinese Triad Societies, Tongs, Organized Crime, and Street Gangs in Asia and the United States. Unpublished doctoral dissertation, Department of Sociology, University of Pennsylvania, 1986. Cooper, B.M. Motor City breakdown. Village Voice, December 1, 1987. pp. 23-35. Fagan, J. The social organi:ration of drug use and drug dealing among urban gangs. Criminology 27(4):633-670, 1989. Fagan, J., and Chin, K. Initiation into crack and cocaine: A tale of two epidemics. Contemp Drug Probl 16(4):579-618, 1989. Fagan, J., and Chin, K- Social prosses of initiation into crack use and dealing. J Drug Issues, in press. Farley, J.E Disproportionate Black and Hispanic unemployment in U.S. metropolitan areas. Am J Econ Sociol 46(2):129-1SO, 1987. Farley, R., and Allen, W.R. The Color Line and the Quality of Lifr in America, New York: Russell Sage Foundation, 1987. Fields, A.J. Weedslingers: Young black marijuana dealers. In: Beschner, G., and Friedman, A.S., eds. Teen Drug Use. Lexington, MA: lexing- ton Books, 1985. pp. 85-104. Frank, B.; Hopkins, W.; and Lipton, D.S. Current drug use trends in New York City. Albany: New York State Division of Substance Abuse Ser- vices, June 1987. Goldstein, P.J. The dru~violence nexus: A tri-partite coriceptual frame- work. J Drug Issues 15:493-506, 1985. Goldstein, PJ. Drugs and violent crime. In: Weiner, NA, and Wolfgang, M.E., eds. Pathways to Criminal Violence. Newbury Park, CA: Russell Sage Foundation, 1989. pp. 1648. 39 Goldstein, P.J.; Belluci, P.A.; Spunt, B.J.; and Miller, T. Frequency of co- caine use and violence: A comparison between men and women. New York: Narcotic and Drug Research Inc. Unpublished manuscript. Goldstein, PJ.; Brownstein, H.H.; Ryan, P.; and Belluci, P.A. Crack and homicide in New York City, 1989: A conceptually based event analysis. Con temp Drug Probl, in press. Goldstein, P.J.; Upton, D.S.; Preble, E.; Sobel, I.; Miller, T.; Abbott, W.; Paige, W.; and Soto, F. The marketing of street heroin in New York City. J Drug Issues 14:553-566, 1984. Goldstein, P.J.; Upton, D.S.; Spunt, B.J.; Belluci, P.A.; Miller, T.; Cortez, N.; Khan, M.; and Kale, A. Drug Related Involvement in Violent Epi- sodes. Final Report, grants DA 03182 and DA 04017, National Institute on Drug Abuse. New York: Narcotic and Drug Research Inc., 1987. Ham id, A. The political economy of crack-related violence. Contemp Drug Probl 17(1):31-78, 1990. Hochschild, J.L. Equal opportunity and the estranged poor. Ann Am Acad Political Soc Sci 501:143-155, 1989. Hopkins, W. Street perspectives on drug use. Presented at the Conference on Drug Abuse Prevention, John Jay College of Criminal Justice, New York, NY, April 7, 1989. lanni, F.A.J. Blxk Mafia: Ethnic Succession in Organized Crime. New York: Simon and Schuster, 1974. Inciardi, J.A. Beyond oocaine: Basuco, crack, and other coca products. Contemp Drug Probl 14(3):461-492, 1987. Inciardi, J.A. Sex for crack: A research note. Con temp Drug Probl 16(4)689-702, 1989. Johnson, B.D.; Goldstein, P.J.; Preble, E.; Schmeidler, J.; Upton, D.; Spunt, B.; and Miller, T. Taking Care of Business: The Economics of Crime by Heroin Abusers. lexington, MA: lexington Books, 1985. Johnson, B.D.; Williams, T.; Dei, K.; and Sanabria, H. Drug abuse and the inner city: Impacts of hard drug use and sales on low income communi- ties. In: Wilson, J.Q., and Tonry, M., eds. Crime and Justice: An Annual Review of Research. Vol. 13. Chicago: University of Chicago Press, 1990. pp. 9-67. Kasarda, J.D. Jobs, migration and emerging urban mismatches. In: McGeary, M.G.H., and Lynn, LE., eds. Urban Change and Poverty. Washington, DC: National Academy Press, 1988. pp. 148-198. Kasarda, J.D. Urban industrial transition and the underclass. Ann Am Acad Political Sx Sci 501:26-47, 1989. Klein, M.W. Offense specialization and versatility among juveniles. Br J Criminol 24:185-194, 1984. Klein, M.W.; Maxson, C.; and Cunningham, L. Street gang involvement in rock cocaine selling. Final report to the National Institute of Justice, grant 85-ll-CX-0057. Los Angeles: UCLA Social Science Institute, 1988. 40 Kozel, N.J., and Adams, E.H., eds. Cocaine Use in America: Epidemito- logical and Clinical Perspectives. National Institute on Drug Abuse Research Monograph 61. DHHS Pub. No. (ADM)85-1414. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1985. Massey, D.S., and Eggers, M.L. The ecology of inequality: Minorities and the concentration of poverty, 19701980. Am J Sociol 95(5):1 153-1188, 1990. McBride, D.C. Drugs and violence. In: Inciardi, J.A., ed. The Drug- Crime Connecdo~ Beverly Hills, CA: Sage, 1981. pp. 105-124. McGeary, M.C.H., and Lynn, L.E., eds. Urban Change and Poverti'. Washington, DC: National Academy Press, 1988. Mieczkowski, T. Geeking up and throwing down: Heroin street life in Detroit. Criminology 24(4):645-666, 1986. Mieczkowski, T. Crack distribution in Detroit. Conitemp Drug ProbI, in press. Murphy, S.; Waldorf, D.; and Reinarman, C. Drifting into dealing: Be- coming a cocaine seller. San Francisco: Institute for Scientific Analysis. Unpublished manuscript. Musto, D. An American Disease: Origins of Narcotic ControL 2d ed. New York: Oxford University Press, 1989. Newsweek Kids and Cocaine. March 17, 1986. pp. 58-68. New York Times. Crack: A Disaster of Historic Dimensions, Still Grow- ing. March 28, 1989a. p. A20. New York Times. Report from the Field on an Endless War. March 12, 1989b. Section 4, p. 1. New York Times. Selling Milk, Bread, and Cocaine in New York. March 30, 1989c. p. Al. New York Times. Detroit Drug Empire Showed All the Traits of Big Busi- ness. December 18, 1989d. p. Al. New York Times. Even with TNT, New York City Can't Destroy Drug Bazaars. June 1, 1989e. p. Al. Sampson, R.J. Crime in cities: The effects of formal and informal social control. In: Reiss, A.J., Jr., and Tonry, M., eds. Communities and Crime. Chicago: University of Chicago Press, 1986. pp. 271-310. Sampson, R.J. Urban black violence: The effect of male joblessness and family disruption. Am J Sociol 93(2):34&382, 1987. Sassen-Koob, S. New York City's Informal Economy. In: Portes, A.; Castells, M.; and Benton, L.A., eds. The Informal Economy: Studies in Advanced and Less Developed Countries. Baltimore: Johns Hopkins University Press, 1989. pp. 60-77. Siegel, R.K. Cocaine free base abuse: A new smoking disorder. J Psy- choactive Drugs 14:321-337, 1982. Siegel, R.K. New patterns of cocaine use: Changing doses and routes. In: Kozel, N.J., and Adams, E.H., eds. Cocaine Use in America: Epidemlo- logical and Clinical Perspectives. National Institute on Drug Abuse Research Monograph 61. DHHS Pub. No. (ADM)85-1414. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1985. 41 Siegel, R.K. Cocaine smoking: Nature and extent of coca paste and co- caine freebase abuse. In: Washton, A.M., and Gold, M., eds. Cocaine: A Clinician's Handbook New York: Guilford Press, 1987. pp. 175-191. Spitz, H.I., and Rosecan, J.S. Cocaine reconceptualized: Historical over- view. In: Spitz, H.I., and Rosecan, J.S., eds. Cocaine Abuse: New Directions in Treatment and Research~ New York: Brunner-Mazel, 1987. pp. 5-18. Tienda, M. Neighborhood effects and the formation of the underclass. Pre- sented at the Annual Meeting of the American Sociological Association, San Francisco, CA, August 1989a. Tienda, M. Puerto Ricans and the underclass debate. Ann Am Acad Politi- cal Soc Sci 501:105-119, 1989b. Vigil, J.D. Barrio Gangs. Austin, 'DC: University of Texas Press, 1988. Wacquant, L.D., and Wilson, W.J. The costs of racial and class exclusion in the inner city. Ann Am Acad Political Soc Sci 501:825, 1989. Waldorf, O.D. Final report of the ex-sellers project: An explanatory study of indirect criminal justice pressures on cocaine sellers. National Institute of Justice grant 89-IJ-CX4)O36. Washington, DC: U.S. Department of Justice, 1989. Waldorf, O.D.; Reinarman, C.; and Murphy, S. Cocaine Changes. Phila- delphia: Temple University Press, 1990. Washton, A., and Gold, M. Recent trends in cocaine abuse as seen from the "800-Cocaine Hotline." In: Washton, A.M., and Gold, M., eds. Cocaine: A Clinician's Handbook New York: Guilford Press, 1987. pp. 10-22. Watters, J.K.; Reinarman, C.; Fagan, J. Causality, context and contingency: Relationships between drug abuse and delinquency. Contemporary Drug Problems 12(3):351-373, 1985. Williams, T. The Cocaine Kids. New York: Addison-Wesley, 1989. Zahn, M. Homicide in the Twentieth Century: Trends, types and causes. In: Gurr, T.R., ed. V'wlence in America, Volume I: The History of Crime. Newbury Park, CA: Sage, 1989. pp. 216-234. Zimmer, L. Operation Pressure Point. Occasional Paper of the Center for Research on Crime and Delinquency, New York University School of Law. New York: New York University School of Law, 1987. ACKNOWLEDGMENTS Support for this research was provided by U.S. Department of Justice grant 87-U-CX~()O64, "Changing Patterns of Drug Abuse and Criminality among Crack Users," frorn the National Institute of Justice. The opinions are those of the authors and do not reflect the policies or views of the U.S. Depart- ment of Justice. Bruce Johnson, Ph.D., and Eloise Dunlap, Ph.D., both with Narcotic and Drug Research, Inc., are acknowledged for their critical contributions to the design of this research and the difficult data collection effort. 42 AUThORS Jeffrey Fagan, Ph.D. Associate Professor School of Criminal Justice Rutge~~The State University of New Jersey 15 Washington StIe:t Newark, NJ 07102 K~lin CIlin, Ph.D. Senior Reseatch Analyst New York City Criminal Justice Agency 305 Broadway New York, NY 10007 43 Violence Associated With Acute Cocaine Use in Patients Admitted to a Medical Emergency Department Steven L. Brody INTRODUCTION Cocaine abuse has been an increasing public health concern over the past decade. In the early 198Os, medical attention focused on dramatic cocaine- associated complications, including myocardial infarction, stroke, and sudden death (Cregler and Mark 1986). That attention broadened in the latter half of the 19&)s as social and economic conditions were marked by an explo- sion in interpersonal violence and violent crimes, including "cocaine related" homicides (Johnson et al. 1987; Harruff et al. 1988) and assaults, to include an investigation of the psychopharmacologic effects of cocaine (Gawin and Ellinwood 1988; Dackis and Gold 1988; Johanson and Fischman 1989). Medical, trauma, and psychiatric visits to hospitals continued to rise, and, by 1988, data from the National Institute on Drug Abuse (MDA) Drug Abuse Warning Network (DAWN) indicated that cocaine was the most fre- quently mentioned illicit substance involved in emergency department (ED) visits and in medical examiner reports on drug-related deaths (National Institute on Drug Abuse 1989). Consequently, many researchers are begin- fling to explore one of the most challenging aspects of the cocaine abuse problem-the relation between cocaine use and violenc~a complex issue that includes social, economic, and medical factors. One of the most comprehensive explanations of the relations between drug use and violence is the tripartite scheme developed by Goldstein (1986; Goldstein et al. 1988). In this analysis, one dimension leading to violence is termed "systemic," which is related to drug distribution and trafficking. A second is "economic compulsive," which describes the violence associated with acquisition of money to purchase drugs and includes muggings and property crimes. A third factor is "psychopharmacologic," or violent 44 spite of the growing body of research linking cocaine to violence and aggressive behavior, much remains poorly understood. Data supporting the hypothesis that a direct psychopharmacologic effect of cocaine leads to vio- lent behavior is limited by a paucity of observations of acutely intoxicated users and the problems inherent in performing human studies with a drug known to be highly toxic with often unpredictable effects. The purpose of this chapter is to describe a group of patients who presented to a medical ED (MED) with violent and aggressive behavior associated with acute cocaine use. While largely observational, this information may help to explain one facet of the complex relationship between cocaine use and violence. METHODS Patients described in this study were seen over a 2-year period between August 1986 and August 1988. All patients were seen at the MED of Grady Memorial Hospital, the major provider of indigent care in Atlanta, GA. The MED has over 65,0(X) patient visits per year and is the primary unit for the acute management of drug-abuse-related problems. Other areas within the hospital that provide emergency care include a surgical-trauma area, a gynecology-obstetrics unit, a pediatric emergency department, and a psychiatric crisis clinic. Patients with cocaine-related violent behavior admitted to the MED were pooled from two separate patient data bases that had been used for previous studies. One patient set (A) was a consecutive series of 223 patients who visited the MED with cocaine-related problems over a 6month period be- tween August 1986 and February 1987 (Brody et aL, in press). In this study, the medical records of all patients with the term "cocaine" in the MED record were retrospectively reviewed. Patients were also taken from a second data base (B) that was a nonconsecutive series of 29 patients with cocaine-associated rhabdomyolysis (a clinical and laboratory syndrome re- sulting from skeletal muscle injury and the release of cell contents into the blood) who came to the MED between January 1987 and August 1988. In each of these studies, records were made of demographic information, including patient age and sex, details of drug use, including route of use and frequency; specific medical complaints; physical examination findings; laboratory data, including toxicologic data; management, including acute drug therapy; and patient outcome. Data from these two studies were examined for inclusion in the present study if there was information in the MED record documenting violent or aggressive behavior associated with the ED visit. Criteria for violent or aggressive behavior included evidence of one or more of the following: assault, destruction of property, "combative" or "agitated" behavior, and other "uncooperative" or threatening behavior requiring physical restraint. 46 Each patient had evidence of acute cocaine use documented by at least two of the following three criteria: (1) a history of cocaine use within 12 hours; (2) witnessed behavior, symptoms, or clinical findings consistent with acute cocaine intoxication as described by other studies (Gay 1982); or (3) detection of cocaine metabolites in the urine. RESULTS Patients Thirty-seven patients with violent or aggressive behavior associated with acute cocaine intoxication were identified from the two patient data sets described above. Nineteen patients were identified in data set A, repre- senting 8.1 percent of all visits to the MED for acute and chronic cocaine- associated medical problems over a 6-month period. An additional 18 patients were identified in data set B. Over the 2-year, nonoverlapping period that brackets these two data sets, there was an estimated total of 900 visits to the MED for acute and chronic cocaine-associated medical prob- lems. Therefore, the patients with violent or aggressive behavior associated with acute cocaine use represented at least 4 percent of all cocaine-related visits to the MED. Patients included 31 men and 6 women with a mean age of 28.2 years (range 16 to 46 years). All routes of cocaine use were used; however, route was not specified in five patient charts. Intravenous injection was used by 45 percent of patients, 33 percent smoked cocaine, nasal insuffla- tion ("snorting") was specified by four patients, and one patient ingested cocaine orally. Four patients used multiple routes. Estimates of amounts of drug use and frequency of drug use were extremely variable and not well documented. While some patients described daily use of several grams, many patients described themselves as "occasional" users. Cocaine use was verified by toxicologic testing to quantify cocaine metabolite in patient's urine. Drug testing was not done in five patients. Cocaine metabolites were not detected in the urine of four patients despite a history of acute cocaine use, making the diagnosis of acute cocaine use unclear in these patients. Other drugs of abuse and alcohol were commonly used acutely, in combina- tion with cocaine. Alcohol use was determined by history or was detected in the blood of half the patients, although the blood level was less than 100 mg/dL in all but one patient. Other drugs used with cocaine, as deter- mined by history or toxicologic testing, included berizodiazepines in four patients, opiates in three patients, marijuana in three patients (determined by history only), phencyclidine in two patients, and tricyclic antidepressant in one patient. Five patients used more than two substances of abuse, includ- ing cocaine. Only 19 patients used only cocaine. Two patients used cocaine and marijuana. 47 Behaviors Violent behavior or aggressive behavior was described "in the field" just prior to the MED visit (and was often the reason the patient was brought to the hospital by police or family), occurred during the period that the patient was in the MED, or occurred in both settings. Violent behavior was de- scribed by police in 20 patients (54 percent), by paramedics in 6 patients (2 percent), by friends or family members in 11 patients (30 percent), and was directly observed in the MED by physicians and nurses in 30 patients (81 percent). Behavior was determined to be associated or not associated with a psychot- ic or delirious state. Nonpsychotic behavior was described in 14 cases (38 percent). In these cases, behavior was described as "combative," "un- cooperative," or "agitated." Description of a typical patient follows. A 31-year-old woman was brought to the MED for violent behavior after she was injected with cocaine. In the ambu- lance, she was "nervous" and refused intravenous therapy. In the MED, she was alert but uncooperative, fighting with the staff~ "acting wildly," and repeatedly "leaping off the stretcher." She was restrained but continued to refuse to answer questions. After 2 hours, she was conversant and cooperative. Behavior was described as "delirious," "paranoid," or associated with altered mental status in 23 cases (62 percent). These patients had a behavior pat- tern that was typical of cocaine-induced psychosis or had an altered mental status with disorientation and violent behavior after a seizure or syncopal spell. These patients also shared many of the combative and agitated features of the nondelirium cases. The following patient had such a presentation: A 42-year-old man was brought to the MED by police after threatening to harm his mother after he used cocaine. He was found by police running in the street yelling "people are going to kill me." He was initially disorient- ed, hypertensive, and tachycardic. He was combative and was managed with limb restraints and an intramuscular in- jection of haloperidol. He became lucid within several hours. Seven patients specifically assaulted others (often security guards or police personnel), and property destruction was noted prior to the MED visit in four additional patients. The following is an illustrative case. 48 A 34-year-old man with a long history of cocaine use came to the MED after "testing" intravenous cocaine prior to a large purchase. After injecting the cocaine, he began having hallucinations, chest pain, and shortness of breath. According to friends he then "went crazy" and began to destroy the furniture in the room. Pretending to have a gun, he entered the room next door, destroyed the furni- ture, and passed out. He was alert but anxious in the MED. He was admitted for management of rhabdomyolysis. Many of these violent behaviors were associated with activity involving extreme exertion. Often patients were running down streets, had prolonged struggles with police, or, in one case, climbed a large fence around a high- way after injecting 1 g of cocaine. Attempts by police officers to stop these patients were commonly met with struggles and fighting. Medical Complications In addition to behavioral changes, patients often had serious medical symp- toms or complications. Cardiovascular complaints including chest pain, often associated with dyspnea and diaphoresis or palpitations were noted by seven patients. Hypertension was common. The following was an extreme case. A 25-year-old man with a history of mild hypertension was brought in by police for assault after he smoked "a large amount" of cocaine. He was combative but complained of chest pain in the MED. His blood pressure was 300/210 mm Hg, and he was given intravenous labetelol for control of his blood pressure and admitted to the hospitaL Serious neurologic complications, all previously known to be associated with cocaine use (Lx)wenstein et al. 1987), occurred in 11 patients, including S who developed coma following violent behavior, 4 who had seizures, and 2 who experienced syncope. Violent and aggressive behavior commonly occurred after seizure or syncope as the following case illustrates. A 19-year-old man had a seizure after smoking crack and then began to fight with his friends. Despite being held down, he kicked the paramedics and screamed, "I'm going to kill the [person] who gave me crack." In the MED, he violently fought with the staff. He was fully restrained and given intramuscular haloperidol. Another serious medical complication was rhabdomyolysis, diagnosed in 18 patients, 2 of whom required dialysis for renal failure. This high incidence 49 is due to the bias introduced by the use of the rhabdomyolysis data set (data set B) for patient selection. Respiratory arrest following a period of violent behavior occurred in three patients, all requiring endotracheal intubation and ventilatory support. This potentially fatal complication has been previously described by Wetli and Fishbain (1985) in a description of patients with "excited delirium." laboratory test abnormalities occurred in many of the patients with violent behavior including leukocytosis (white blood cell count greater than 10,000 cells/mm3), elevated serum creatinine (greater than 2.0 mg/dL), and a mild metabolic acidosis. Approximately one-third ?~ patients had a fever (oral temperature greater than 38 0C), and two patients were hyperthermic (tern- perature greater than 40 0C). Minor trauma comprised of multiple lacera- tions or abrasions occurred in 11 patients (30 percent). Therapy Most patients did not receive a specific drug therapy for violent behavior. Seven patients had full resolution of altered mental status and behavioral changes at the time of evaluation in the MED. Extremity (limb) restraints (leather or cloth) were used for 13 patients and required multiple medical staff members for application. Dmg therapy was rarely used. Haloperidol was used in six patients, all of whom required extremity restraints. One patient received intravenous lor- azepam for behavior management, and one patient (previously described) received intravenous labetolol to control severe hypertension. Of the 37 patients, 20 (54 percent) were admitted to the hospital for man- agement of medical complications or for evaluation of persistent abnormal mental status. This included all 20 of the patients identified in data set B and 3 of 17 patients from data set A. Of those not admitted, six were re- leased into police custody, two were transferred to the psychiatry department for further evaluation, and nine were discharged home from the MED. All patients who were admitted were alive at the time of discharge from the MED or the inpatient service. DISCUSSION This study found that patients with acute cocaine intoxication may present with a wide variety of violent and aggressive behavior patterns. Further, observations from this study suggest that patients with cocaine-associated violent or aggressive behavior, seen in the ED of a large inner-city hospital are acutely ill patients who are difficult to manage and have multiple com- plex medical complications as a consequence of cocaine intoxication. These findings are in agreement with previous studies linking the pharmacologic 50 effects of the cocaine with violent behavior. Wetli and Fishbain (1985) were among the first to describe clinical characteristics of a series of acute- ly ill cocaine-intoxicated patients with violent behavior. Several other groups have documented violent behavior occurring in the setting of acute cocaine intoxication. Honer et al. (1987) provide a limited description of 70 patients with acute psychiatric symptoms, of which at least half had some violent behavior; however, details were not provided. Further, Roth et al. (1988) described a large series of acutely intoxicated patients admitted for rhabdomyolysis who were often violent, combative, and agitated. To- gether, these patient observations demonstrate a "proneness to violence," particularly associated with cocaine-induced psychosis, as was described by Post (1975). Additionally, these data support previous observations that vio- lent behavior can be a manifestation of cocaine intoxication in the absence of psychosis (Manschreck et al. 1988). There are several lines of evidence that support a psychopharmacologic basis for cocaine-induced violent behav- ior in humans. Oocaine is a complex pharmacologic agent that acts as a local anesthetic and as a central nervous system (CNS) neurochemical modulator. The major CNS effects of acute cocaine use are increases in the major neuro- transmitters: dopamine, norepinephrine, and serotonin (Gold et al. 1986). These occur because cocaine blocks the intrasynaptic reuptake of these neurotransmitters, resulting in a flood of intrasynaptic neurochemicals and, consequently, increased postsynaptic stimulation by these neurochemicals. Behaviorally, the increased dopamine levels are likely responsible for cocaine-induced euphoria at low levels and dysphoria at higher levels (Gold et al. 1986; Johanson and Fischman 1989). Dopamine is postulated to be the key neurotransmitter responsible for positive reinforcement or drug "craving" (Ritz et aL 1987). Norepinephrine increases levels of alertness and, together with dopamine, results in increased psychomotor activity and seizures. Examples of peripheral effects of increased norepinephrine trans- mission are cardiovascular findings of hypertension, tachycardia, and arrhythmia (Dackis et al. 1989). chronic cocaine use is hypothesized to deplete the neurotransmitter pool of dopamine, norepinephrine, and serotonin, and, therefore, to result in a decrease in neurotransmitter stimulation (Gold et al. 1986). Recent studies in rats chronically treated with cocaine have demonstrated a decrease in brain levels of dopamine metabolites (Wyatt et al. 1988). Evidence that this may occur in humans is supported by data showing that serum prolac- tin, a hormone under tonic dopamine inhibition, is increased in chromc cocaine abusers (Gawin and Kieber 1985a; Mendelson et aL 1988). While there are several potential mechanisms for this, a decreased dopamine effect is the most attractive explanation. Similar neuroendocrine changes have been correlated with aggressive behav- ior and suicide. Fishbein et al. (1989) observed that serum prolactin levels 51 are greater in groups of substance abusers (including noncocaine substances) who have high levels of aggressive behavior, suggesting that this hormone may be a marker for neurochernical changes that increase aggressive behav- ior. Cocaine withdrawal states are also associated with doparnine depletion and high prolactin levels (Mendelson et al. 1988). During this withdrawal period, which can begin within hours of discontinuing cocaine use, users may become irritable and agitated and may be prone to violent behavior (Gawin and KIeber 1986). Fishbein et aL (1989) and others (Brown et al. 1982) have also suggested that modulation of serotonin is important in aggressive behavior. A de- crease in serotonin, which is hypothesized to occur with chronic cocaine use (Gold et aL 1986), has been observed to occur in rats following acute Co- caine injection (Hanson et al. 1987) and may have a role in violent behav- ior. The hypothesis that a fall in the neuroinhibitory effects of serotonin may be related to aggressive behavior is supported, in part, by data on hu- mans, which show that cerebral spinal fluid levels of serotonin rnetabolite are decreased, possibly due to serotonin depletion, in individuals with corn- pulsiveaggressive behavior and those with suicidal behavior (Brown et aL 1982). Studies that investigate changes in neurochernical levels indicate that repeated doses of cocaine ("chronic"), even over 24 hours, result in differ- ent effects than single doses (Johanson and Fischrnan 1989; Hanson et aL 1987). Therefore, it is perhaps misleading to label patients in this study as "acutely" intoxicated, since it is unlikely that a single dose of cocaine was used. Although data concerning the intensity and chronicity of cocaine use were not available in the present study, Gawin and Kieber (1985b) have emphasized that the binge use of several grams of cocaine over several days is not uncommon. In addition, Brower et al. (1988) found that cocaine users with psychotic or violent symptoms used more cocaine over more days than those without symptoms. This chronic and intense use may be a key factor for precipitating violent behavior and deserves further attention in future studies that seek to identify discrete biologic factors that determine violent behavior. Additional evidence that the psychopharmacologic effects of cocaine are linked to violent behavior comes from investigations of the effects of amphetamine, a cocainelike stimulant, on aggressive behavior. Ampheta- mine, like cocaine, increases CNS dopaminergic activity and results in increased drug self-administration; chronic use results in dopamine depletion (Seiden 1985; Gawin and Ellinwood 1988). chronic use also results in a classic drug-induced psychosis, which includes inappropriate aggressive behavior (Seiden 1985; Sato 1986). Ellinwood (1971) described 13 persons who committed homicide while intoxicated by amphetamine. Asnis and Smith (1978) also described patterns of violent behavior in amphetamine users but suggested that personality and environmental factors played 52 important roles. As with human cocaine studies (Johanson and Fischman 1989), only limited human amphetamine behavior studies have been done. Beezley et al. (1987), using volunteer college students and relatively low doses (0.32 mg/kg) of oral dextroamphetamine, failed to show that the drug caused increased aggressive behavior when compared to placebo; however, this model may not be analogous to use patterns in chronic methampheta- mine or other stimulant abusers. Studies that show a lack of relationship between cocaine use and violent behavior are primarily studies of persons arrested for violent crimes (Collins et al. 1988; Kozel and DuPont 1977). These study populations differ from the populations that have linked cocaine with violent behavior in that the latter are often hospitalized patients. It is possible that acutely hospitalized patients described in the present study and by others (Brower et al. 1988; Honer et al. 1987; Jekel et al. 1986) were using higher doses of drug or using the drug more frequently and, despite committing violent crimes, were taken to a hospital for medical management instead of being incarcerated. The possibility that jailed users are a different population from hospitalized users suggests that there is a dos~response factor related to violent behav- ior. This is consistent with Post's (1975) psychiatric description of acute cocaine as a spectrum or "continuum" of clinical syndromes. With lower doses, the patient experiences a feeling of increased power that may be associated with maniclike hyperactivity and a proneness to violence, but without a change in sensorium or mental status. With more severe intoxica- tion, the patient presentation is that of a drug-induced psychosis and is asso- ciated with violent behavior. This also emphasizes another potential differ- ence between studies related to the interpretation of the term "violent behavior." For example, violent behavior associated with psychosis or a delirium state is likely to be viewed as a different behavior than an assault committed while the cocaine abuser is only mildly intoxicated. In the former state, the individual may be termed "a psychiatric patient" and is taken to a hospital, while in the latter, the abuser is a "criminal" and is taken to jail. The violence or aggressive behaviors associated with the psychopharmaco- logic effect of cocaine as described in this study may be multifactorial, a possibility that underlines the potential limitations of this study. Important factors include underlying psychiatric disease, environmental factors, and the effect of concomitant drug use. Underlying psychiatric disease is a com- mon problem among the indigent homeless who frequent the inner-city hospitals. Further, Teplin (this volume) emphasieed that psychosocial personality disorders are more common among drug users. In this context, there may be an increased incidence of violent behavior in the study popu- lation, but since there is not a control, cocaine-nonusing population for com- parison, the question cannot be answered. Even if there is an increase in psychopathology in the study population, cocaine plays an important role. 53 As Post (1975) noted, the patient with underlying psychiatric disease is prone to cocaine-induced behavioral changes. There are several environmental pressures that may be important in causing violent behavior in the described patient population. Subcultural behaviors associated with gangs, crowded living conditions, and adverse relationships with law enforcement officers are likely important factors but are beyond the scope of this discussion. Also, the ED environment itself may contrib- ute to aggressive and violent behavior. long waiting times, crowded condi- tions, and poor staff~atient communication in a high-stress setting have been implicated as a cause for violent behavior (Lavoie et aL 1988). Finally, other intoxicants, coingested with cocaine, may cause violent or aggressive behavior. In the present study, almost half of the patients had detectable, though low, blood alcohol levels (less than 100 mg/dL). Both acute and chronic use of alcohol has been associated with violent behavior (Collins and Schlenger 1988). This may be a confounding factor in many other cocain~violence studies because over 85 percent of cocaine users use alcohol (Roehrich and Gold 1988). Alcohol is rapidly metabolized and use cannot be detected after several hours so that studies, such as Collins et aL (1988) that depend on drug screens, may miss this important substance. Other drugs used by patients in this study that have been previously associ- ated with violent behavior include amphetamine, opiates, and phencyclidine (Collins et aL 1988). The patients in this MED study had a high incidence of cocaine-use-related medical problems, and a high percentage required hospital admission. This particular group of patients may not be representative of all patients with cocaine-associated violent behavior, because half were selected from a data base of patients that were admitted with cocaine-associated rhabdomyolysis, and if the rhabdomyolysis group is excluded, the hospitalization rate is only 18 percent. Nonetheless, the need for hospitalization of the violent cocaine- intoxicated patient should not be discounted. Wetli and Fishbain (1985) emphasized the importance of prompt, aggressive medical care for the vio- lent patient presenting with "excited delirium," after they noted that several of these types of patients died while in police custody following arrests for assault and other crimes. The management of the patient with acute cocaine intoxication and violent behavior does not usually require drug therapy (Brody et al., in press; Derlet and Albertson 1989). The half-life of cocaine is short, less than 1 hour when smoked or used intravenously (Johanson and Fischman 1989), and, as was the case in most of the patients in this study, the acute behav- ioral changes rapidly resolved spontaneously. Benzodiazepines, especially diazepam, have been shown to be the most efficacious agent for the man- agement of acute cocaine intoxication, but a drug from this class was used in only one patient in this study. Animal studies show that diazepam 54 effectively prevents seizure and death, while other drugs, such as the beta adrenergic antagonist propranolol, are ineffective (Catravas and Waters 1981). Haloperidol has been most widely recommended as the drng of choice for the management of acutely psychotic patients (Ellison and Jacobs 1986). Anecdotally, it was effective and without complications when used in these cocaine-intoxicated patients. Sherer et al. (1989) found that pre- treatment with haloperidol decreased the "pleasantness" of the cocaine effect and attenuated the cocaine-mediated hypertension. As a dopamine antago- nist, it may play a beneficial role in the cocaine user with high-dopamine states but theoretically may be less effective in the chronically depleted, bingeing patient. The use of leather or cloth limb and trunk restraints for the control of the combative patient is common and often necessary to protect the staff and the patient from bodily harm. The patient who continues to struggle against restraints may be at risk for other medical complications and therefore should be evaluated for adjunctive drug therapy with haloperidol or benzodiazepine. Above all, an orderly approach to potential and acute medical problems with attention to respiratory, cardiovascular, and neurologic systems, is essential. Those caring for the violent cocaine abuser should be aware that the violent behavior may be short lived but that other serious medical prob- lems may coexist that will not resolve spontaneously, i.e., violent behavior in the cocaine user should be considered a marker for associated medical problems. CONCLUSION The association of acute cocaine intoxication and violent behavior appears to be primarily related to a state of intense cocaine intoxication. Several potential neurotransmitter mechanisms may link cocaine with violent and aggressive behavior. Further animal studies are needed to continue to investigate neurochemical changes that correlate with behavioral changes. Future research in man should include an investigation of acute neurochemi- cal and endocrinologic changes associated with the cocaine-intoxicated patient. These studies must be controlled for environmental factors, poly- drug use, and underlying psychiatric disease. long-term followup of co- caine users may reveal chronic behavioral, neurochemical, and endocrino- logic changes and may be important for future treatment programs. REFERENCES Asnis, S.F., and Smith, R.C. Amphetamine abuse and violence. j Psyche- delic Drugs 10:371-377, 197& 55 Beezley, D.A.; Gantner, A.B.; Bailey, D.S.; and Taylor, S.P. Amphetamines and human physical aggression. j Res Pers 21:52-60, 1987. Brody, S.L.; Slovis, C.M.; and Wrenn, K.D. Cocaine related medical prob- lems: A consecutive series of 233 patients. Am J Med 88:325-331, 1990. Brower, K.J.; Blow, F.C.; and Beresford, T.P. Forms of cocaine and psychiatric symptom. Lancet 1:50, 1988. Brown, G.L.; Goodwin, F.K.; and Bunney, W.E. Human aggression and suicide: Their relationship to neuropsychiatric diagnoses and serotonin metabolism. In: Ho, B.T.; Schoolar, J.C.; and Usdin, E., eds. Serotonin in Biologic Psychiatry. New York: Raven Press, 1982. pp. 287-307. Catravas, J.D., and Waters, I.W. Acute cocaine intoxication in the con- scious dog: Studies on the mechanism of lethality. j Pharmacol E~ Ther 217:350-356, 1981. Collins, J.J.; Powers, L.L.; and Craddock, A. Recent drug use and violent arrest charges in three cities. Center for Social Research and Policy Analysis, Research Triangle Institute, NC. Presented at the American Society of Criminology, Chicago, II, 1988. Collins, J.J., and Schlenger, W.E. Acute and chronic effects of alcohol use on violence. j Stud Alcohol 49:516-521, 1988. Cregler, L.L., and Mark, H. Medical complications of cocaine abuse. N Engl J Med 315:1495-1499, 1986. Dackis, C.A., and Gold, M.S. Psychopharmacology of cocaine. Psychiatr Ann 18:528-530, 1988. Derlet, R.W., and AIbertson, T.E. Emergency department presentation of cocaine intoxication. Ann Emerg Med 18:542-546, 1989. Ellinwood, E.H. Assault and homicide associated with amphetamine abuse. AmjPsychiatry 127:1170-1175, 1971. Ellison, J.M., and Jacobs, D. Emergency psychopharmacology: A review and update. Ann Emerg Med 15:962-968, 1986. Fishbein, D.H.; IOzovsky, D.; and Jaffee, J.H. Impulsivity, aggression, and neuroendocrine responses to serotonergic stimulation in substance abusers. Biol Psychiatry 25:1049-1066, 1989. Gawin, F.H., and Ellinwood, E.H. Cocaine and other stimulants: Actions, abuse and treatment. N Engl J Med 318:1173-1182, 1988. Gawin, F. H., and KIeber, H.D. Neuroendocrine findings in chronic cocaine abusers: A preliminary report. Br J Psychiatry 147:569-573, 1985a. Gawin, F.H., and KIeber, H.D. Cocaine use in a treatment population: Patterns and diagnostic distinctions. In: Kozel, N.J., and Adams, E.H., eds. Cocaine Use in America: Epidemiologic and Clinical Perspectives. National Institute on Drug Abuse Research Monograph 61. DHHS Pub. No. (ADM)85-1414. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1985b. pp. 182-192. Gawin, F.H., and KIeber, H.D. Abstinence symptomatology and psychiatric diagnosis in cocaine abusers. Arch Gen Psychiatry 43:107-113, 1986. Gay, G.R. Clinical management of acute and chronic cocaine poisoning. Ann Emerg Med 11:562-572, 1982. 56 George, F.R. Cocaine produces low dose locomotor depressant effects in mice. Psychopharmacology 99:147-150, 1989. Gold, M.S.; Dackis, C.A.; Pottash, A.L.C.; Extein, L.; and Washton, A. Cocaine update: From bench to bedside. Adv Akohol Subst Abuse 5:35- 60, 1986. Goldstein, P.J. The drugsviolence nexus: A tripartite conceptual frame- work. j Drug Issues 15:493-506, 1986. Goldstein, P.J.; Bellucci, P.A.; Spunt, B.J.; Miller, T.; Cortez, N.; Khan, M.; Durrance, R.; and Vega, A. Female drug related involvement in violent episodes. Narcotic and Drug Research, Inc., Report for the New York State Division of Substance Abuse, March 1988. Hadfield, M.G. Cocaine: Peak time of action on isolation-induced fighting. Neuropharmacology 21:711-713, 1982. Hanson, G.R.; Matsuda, L.A.; and Gibb, J.W. Effects of cocaine on meth- amphetamine-induced neurochemical changes: Characterization of cocaine as a monoamine uptake blocker. j Pharmacol E~ Ther 242:507-513, 1987. Harruff, R.C.; Francisco, J.T.; Elkins, S.K.; Phillips, A.M.; and Fernandez, G.S. Cocaine and homicide in Memphis and Shelby County: An epi- demic of violence. j Forersic Sci 33:1231-1237, 1988. Honer, W.G.; Gerwitz, G.; and Turey, M. Psychosis and violence in cocaine smokers. Lancet 2:451, 1987. Jekel, J.F.; Allen, D.F.; Podlewski, H.; Clarke, N.; Dean-Patterson, S.; and Cartwright, P. Epidemic free-base cocaine abuse: Case study from the Bahamas. Lancet 1:459-462, 1986. Johanson, C.-E., and Fischman, M.W. The pharmacology of cocaine related to its abuse. Pharmacol Rev 41:3-52, 1989. Johnson, T.E.; lebenow, G.C.; Miller, M.; Gonzalez, D.L.; and Carroll, G. Urban murders: On the rise. Newsweek, February 9, 1987. p. 30. Kantak, K.M. Magnesium alters the potency of cocaine and haloperidol on mouse aggression. Psychopharmacology 99:181-188, 1989. Kozel, N.J., and DuPont, R.L. Trends in drug use and crime and their rela- tionship in an arrested population. Am J Drug Alcohol Abuse 4:413-429, 1977. Iavoie, F.W.; Carter, G.L.; Danzl, D.F.; and Berg, R.L. Emergency depart- ment violence in United States teaching hospitals. Ann Emerg Med 17:1227-1233, 1988. Lowenstein, D.H.; Massa, S.M.; Rowbotham, M.C.; Collins, S.D.; McKinney, H.E.; and Simon, R.P. Acute neurologic and psychiatric com- plications associated with cocaine abuse. Am J Med 83:841-846, 1987. Manschreck, T.C.; Laughery, J.A.; Weinstein, C.C.; Allen, D.; Humblestone, B.; Neville, M.; Podlewski, H.; and Mitra, N. Characteristics of freebase cocaine psychosis. Yale J Blol Med 61:115-122, 1988. 57 Mendelson, J.H.; Teoh, S.K.; Lange, U.; Mello, N.K.; Weiss, R.; Skupny, A.; and Ellingboe, J. Anterior pituitary, adrenal, and gonadal hormones during cocaine withdrawal. Am J Psychiatry 145:1094-1098, 1988. Miczek, K.A. Effects of 1-dopa, d-amphetamine and cocaine on intruder- evoked aggression in rats and mice. Prog Neuro-Psychopharmacol 1271- 1277, 1977. National Institute on Drug Abuse. Semiannual Report. Data from the Drug Abuse Warning Network Statistical Series G, Number 22. DHHS Pub. No. (ADM)89-1607. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1989. Post, R.M. Cocaine psychosis: A continuum model. Am J Psychiatry 132:225-231, 1975. Ritz, M.C.; Lamb, R.J.; Goldberg, S.R.; and Kuhar, M.J. Cocaine receptors on dopamine transporters are related to self-administration of cocaine. Science 237:1219-1223, 1987. Roehrich, H., and Gold, M.S. 800-COCAINE: Origin, significance, and findings. Yale J Biol Med 61:149-155, 1988. Roth, D.; Alarcon, F.J.; Fernandez, J.A.; Preston, R.A.; and Bourgoignie, J.J. Acute rhabdomyolysis associated with cocaine intoxication. N Engl J Med 319:673-677,1988. Sato, M. Acute exacerbation of methamphetamine psychosis and lasting dopamine supersensitivity~ clinical survey. Psychopharmacol Bull 22:751-756, 1986. Seiden, L.S. Methamphetamine: Toxicity to dopaminergic neurons. In: Brown, R.F., and Friedman, D.F., eds. Neuroscience Methoas in Drug Abuse Research. National Institute on Drug Abuse Research Monograph 62. DHJIS Pub. No. (ADM)85-1415. Washington, DC: Supt. of Docs., U.S. Govt. Print. Off., 1985. pp. 100-116. Sherer, M.A.; Kumor, K.M.; and Jaffe, J.H. Effects of intravenous cocaine are partially attenuated by haloperidol. Psychiatry Res 27:117-125, 1989. Wetli, C. V., and Fishbain, D.A. Cocaine-induced psychosis and sudden death in recreational cocaine users. j Forensic Sci 30:873-880, 1985. Wyatt, R.J.; Karoum, F.; Suddath, R.; and Fawcett, R. Persistently de- creased brain dopamine levels and cocaine. JAA~ 259:2996, 1988. ACKNOWLEDGMENTh Martha M. Wilber, M.D., Corey M. Slovis, and Keith D. Wrenn, M.D., assisted with data collection and analysis. 58 AUTHOR Steven L. Brody, M.D. Former Director Medical Intensive Care Unit Grady Memorial Hospital and Assistant Professor of Medicine and Emergency Medicine Emory University School of Medicine Atlanta, GA 30303 Current position: Senior Staff Fellow Pulmonary Branch National Heart, Lung, and Blood Institute National Institutes of Health Building 10, Room 6D-03 Bethesda, MD 20892 59 The Operational Styles of Crack Houses in Detroit Tom Mieczkowski INTRODUCTION This chapter describes data from a study that examined the principles of management and organization in typical crack-house operations in Detroit, Mr. These methods involve explicit violence, such as the use of firearms, beatings, and intimidation, as a part of the operational procedure of the crack house. The crack house exhibits implicit violence as well in the nature of the social interactions between clients and sellers and between clients themselves. The chapter also focuses on placing the operational techniques established in crack houses within the larger historical context of drug dealing. Thus, periodically, this chapter will refer to the street sales literature based on heroin as the drug of choice. It is not our intent to uncritically equate heroin selling with crack selling, nor is it within the scope of this chapter to explore all the complex contrasts and similarities between a generic approach to street hustling and the qualifications of that approach for each specific drug type. Based on research over the last decade, however, it is evident that the broad set of dynamics that constitute "hustling" as an aspect of street life, drug use, and drug sales is applicable, in some degree at least, to crack retailing. Furthermore, historic models developed with data based on heroin selling operations should not be a priori excluded as invalid. Because the generalized concept of the distribution of cocaine involves a complex set of actions and actors, the data presented will be limited to describing street-level sales; the manner by which sellers and users effec- tively accomplish exchanges within their own locales; and the utilization of violence to accomplish and facilitate these ends. Descriptions will concen- trate on the "street scene" or lowest end retail activities associated with the network of drug distribution. 60 LITERATURE REVIEW The "street" conception of drug distribution has been developed in the liter- ature for several decades. Studying street sellers of drugs is originally asso- ciated with the work of Edward Preble (Preble and Casey 1969). In recent years, this level of sales has been explored by Redlinger (1975), Caplowitz (1976), Wepner (1977), Johnson et aL (1985), Mieczkowski (1986), Pearson (1987), and a host of other criminologists, anthropologists, and behavioral scientists. Although explicit discussions of crack selling and data derived from research on crack sales activities are quite rare, a small amount of work has been done. Inciardi, for example, has surveyed street people in Miami, FL, and has presented some data about their involvement with crack selling (Inciardi 1986; Inciardi, this volume). Also, the interplay between economic management, drug abuse, street life, and violent behavior has been explored by Goldstein (1981), as well as Nurco et al. (1985). Another re- lated work is Hanson and colleagues' Life with Heroin, which is an elabora- tion on these themes within the heroin subculture (Hanson et al. 1985). The present work continues in this direction by developing descriptions of social behavior of street crack sellers. The "street scene" in drug sales and use refers to a loosely structured social system by which retail consumers of drugs are supplied with low-cost, small dose increments of illicit substances. It is an active, transient, and impro- vised market place that takes on a diverse situational character. In Detroit, the street level of drug sales has three general dimensions. 1. Street Sales. Street sales are the open-air, sidewalk, or roadway sales of small retail quantities of drugs to walk-up or drive-up customers. There is no required prior conspiracy or consultation between buyer and seller. Relatively recent descriptions of this system are in Geberth (1978), Mieczkowski (1986), Hanson et al. (1985), and Hagedorn (1988). this technique frequently represents the least sophisticated method of distribution. 2. Runners and Beepermen. This system involves elements of prior con- sultation or interaction between buyer and seller. The buyer may enter into that interaction directly, or the buyer may utilize an intermediary who may have prior relationships with a seller. Runners act as sales agents for the primary retailer. The term "runner" may also connote an intermediary (touter) who retrieves drugs for a consumer and receives, in terms of reward, a portion of the drugs secured for the end user. A beeperman is a retail seller who distributes by prior telephone consulta- tion with a consumer. The term is used because the prior consultation occurs by telephone and is initiated by contact with a phone pager, or "beeper." Typically, the beeperman may rendezvous at an agreed locale with the consumer, deliver the contraband to a home or office, or re- quire the consumer to come to a particular place to receive the drugs. 61 Beepers, being widely available at low cost, have become increasingly popular mechanisms for drug sales. The mere possession of a beeper, for example, may elevate one's status in street culture differentiating one from a "street seller?' or "corner boy." 3. The Crack House. The crack house represents a third method of retail marketing. Its most distinctive feature is the use of a fixed and secured locale, to which report all manner of customers. It operates in various modes or styles, which will be described in some detail in this chapter. The crack house's relative permanence distinguishes it in comparison to the first two techniques, which are transient methods. METHODOLOGY The data are derived from the Detroit Crack Ethnography Project (DCEP) funded by the Bureau of Justice Assistance (grant number OJP-88-M 39J). The data for this chapter comes from interview transcripts with 100 self- reported dealers and user