Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
PERSPECTIVES ON VIOLENCE 101 3 Perspectives on Violence In this chapter, we look behind patterns and trends to ask what explains them and how they might be altered. One starting point emphasizes that human behavior is shaped in part by long-term developmental processes through which children learn what events to anticipate, how to respond, and what the outcome will be. Children who learn that violent behavior "works" in families, on television, in violent pornographic literature, among peers, and in the community, it is argued, may have a greater potential to behave violently as adults. This oversimplified psychosocial explanation ignores the fact that even if two individuals could be exposed to identical sequences of experiences as they developed, their potentials for adult violent behavior would differ because their nervous systems process information, recall experiences, and react to events differently. While heredity and the prenatal environment help shape an individual's nervous system by the time of birth, that system is constantly reshaped by experiences, by injuries, by temporary external influences such as stress or psychoactive chemicals, by hormonal activity, and even by violent behavior itself. Even taken together, developmental and neurological explanations are incomplete. The mix of experiences during developmentâtheir violence content, the rewards and punishments for violent and nonviolent behavior, for exampleâ is not the same in every family, school, or community. Quite apart from individuals' development, different communities and groups may reward
PERSPECTIVES ON VIOLENCE 102 or punish violence quite differently depending on its purpose, on the victim's ethnic status or sexual preference, and on his or her preexisting relationship(s) to the perpetrator(s). Places differ in terms of how frequently chance encounters offer the rewards or provocations that elicit violent behavior, how easily one can obtain lethal weapons, and how likely a passerby is to interfere in a violent encounter. To close another loop, however, neighborhoods also differ in terms of the risks they present to developing children's nervous systems in such forms as maternal drug use during pregnancy, head injuries to young children, and exposure to environmental toxins. In short, research strongly suggests that violence arises from interactions among individuals' psychosocial development, their neurological and hormonal differences, and social processes. Consequently, we have no basis for considering any of these "levels of explanation" more fundamental than the others. Because existing studies rarely consider more than one of these levels simultaneously, very little is known about the relevant interactions. Therefore, we found it most convenient to discuss possible explanations and promising preventive interventions level by level, in the order used above. First, we consider psychosocial explanations of violence, especially how early experiences and learning affect individuals' potentials for violent behavior, including sexual violence. Second, we look at violent behavior from a biological perspective, discussing underlying neuroanatomical organization; neurological activityâphysiological, chemical, and hormonal; and the role of genetic transmission. Recognizing that both individual development and violent behavior occur at both levels simultaneously, we assess the prospects for reducing violence through various interventions. We then turn to the larger social context of violence. Attempts to explain ethnic and socioeconomic patterns of violence bring us to consider the influences of family and social disorganization, opportunity structure, and community culture. A discussion of gang-related violence illustrates how these influences interact. We then look at specific circumstances in which violent events occur, focusing on places in the community and how they vary in violence rates. We consider preventive interventions, growing out of work in both criminology and public health, that are suggested by a focus on places and situations. We conclude the discussion of research findings by briefly considering violence in two special places: prisons and schools.
PERSPECTIVES ON VIOLENCE 103 Recognizing how little is known about the relevant interactions between psychosocial, biomedical, and social influences on violence, we conclude this chapter by recommending a major longitudinal study to explore them. As a near- term means of simultaneously increasing understanding and improving violence control capacity, we recommend a series of violence problem-solving initiatives âprograms that exploit both the scientific and the policy potential of rigorous evaluationsâin Chapters 7 and 8. VIOLENCE AND THE INDIVIDUAL Psychosocial Perspectives The research underlying psychosocial perspectives is concerned with the learning of aggressive or nonaggressive behaviors as responses to frustration or to ways of achieving desired goals, and the perceptual and cognitive processes through which individuals interpret their interactions with others and choose their responses. As we note in Chapter 1, behavior described as aggression includes several violent behaviors. There is substantial stability in the potential for aggressive behavior from infancy through adulthood. A number of longitudinal studies show that children who show "aggressive" behavior at around age 8 are more likely than others to exhibit delinquent, criminal, or violent behavior in adolescence and adulthood (Olweus, 1979; Farrington, 1989, 1991). This continuity exists despite substantial changes in environment that usually occur in the late teenage years, as children leave schools, their parents' homes, and sometimes their communities (Olweus, 1984). This behavioral continuity through environmental changes suggests strongly that individual differences associated with childhood aggressive behavior are also risk factors for adult violent behavior, although with weaker predictive power. Part of the stability and overlap in risk factors is explained by overlap between common definitions of aggression and our definition of violent behavior. For example, 5 of the 13 DSM-III-R (American Psychiatric Association, 1987) diagnostic criteria for conduct disorder are violent behaviors under our definition; other commonly used measures of aggression include fighting, and threats of violence, which also meet our standard for violence. But researchers measure aggression using a variety of criteria that do not fit our definition of violent behavior, such as interactions in preschool play groups, laboratory measures such as "hostility machine" use, vandalism,
PERSPECTIVES ON VIOLENCE 104 cruelty to animals, and verbal abuse of other children, for example (see Brain, Volume 2). Among children who exhibit aggressive behavior in childhood, the factors that distinguish the small number who go on to commit violent acts as adolescents or adults from the larger number who do not are not well understood. Identifying the distinguishing factors and explaining the processes through which they cause violent behavior remain fundamental questions for future research. Because aggressive behavior is easier to study than violent behavior, much of what we know comes from the psychological literature on aggression as it relates to violence. Social Learning of Aggression and Violence One explanation of aggressive and violent behaviors invokes the concepts of frustration, social learning, and information processing. More than 50 years ago Dollard et al. (1939) speculated that aggression was the consequence of frustration. Modern views emphasize that aggressive or violent behaviors are learned responses to frustration (Eron, 1990), that they can also be learned as instruments for achieving goals, and that the learning is facilitated by available models of aggressive or violent behavior. In this framework, a child learns aggressive or violent behavior by observing or participating in situations in which violence occurs (e.g., instances of frustration, anger, opportunities for instrumental uses of aggression), by experiencing emotional states that are associated with violence, and by observing or experiencing sanctions (rewards and punishments) related to violent events. From early childhood, such occasions and their violent or nonviolent outcomes may be observed in the family, among peers, elsewhere in the neighborhood, or through the mass media. If such social learning inculcates aggression or violence as an appropriate response to frustration, anger, or opportunities, then the behavior may later be activated by cues such as frustration over school failure or loss of a competition, anger at another's behavior, or recognition of an opportunity to gain money by robbery, social status by fighting, or compliance with one's wishes by threats of violence. Methods of Empirical Testing Aggressive and violent behaviors in the community are usually analyzed using either cross-sectional data, comparing members of a sample at a single age or observation time, or longitudinal data,
PERSPECTIVES ON VIOLENCE 105 following a sample of individuals over time. In either case, data are collected on factors that are hypothesized to be relevant and on sample members' behavior, as measured by their self-reports, by officially recorded events such as arrests, or by the reports of third parties such as family members, peers, or teachers. The merits of cross-sectional and longitudinal approaches, and of self- reports and official records, have been debated extensively (Weis, 1986; Farrington et al., 1986; Gottfredson and Hirschi, 1987; Blumstein et al., 1988). Because tests of many hypotheses require accurate recording of sequences of events and behaviors, longitudinal studies are usually the strongest approach. Where the hypothesized relevant factors can be manipulated (as is often the case), randomized experiments with longitudinal follow-ups are the method of choice. In any case, multiple outcome measuresâbased on self-reports, official records, and third-party reportsâshould be obtained and utilized. In the sections that follow we discuss findings from research on aggressive and violent behavior, on violent sexual behavior, and on preventive interventions. Aggressive and Violent Behaviors Antecedents As explained in Appendix A, researchers have identified many correlates and antecedents of aggressive childhood behavior that are presumed to reflect psychosocial influences: â¢ in infancy: pregnancy and birth complications, low birthweight, and an uninhibited, fearless temperament; â¢ in the preschool years: fearless behavior, hyperactivity-impulsivity- attention deficit, restless behavior, and poor concentration; â¢ in the early school years: daring and risk-taking behavior, poor ability to defer gratification, low IQ, low empathy, and abnormally frequent viewing of violence on television; â¢ early family experiences: harsh and erratic discipline, lack of parental nurturance, physical abuse and neglect, poor supervision, and early separation of children from parents; â¢ early school experiences: school failure and interactions involving bullying and peer rejection; and â¢ factors associated with large low-income families, poor housing, criminal behavior by parents and siblings, and living in high-crime neighborhoods. These are often presumed also to be risk factors for violent
PERSPECTIVES ON VIOLENCE 106 behavior. However, because the behavioral outcomes that are studiedâ aggression, delinquency, or crime, for exampleâare so broad, we cannot say with any precision how they are related to violent behavior. The nature of the relationships that have been found, however, suggest that any causal relationships to violent behavior are likely to involve complex interactions across levels of explanation. For example, as Denno (1990:10-11) explains, research prior to hers found that perinatal complications have a potentially strong link to intellectual deficits in poor environments but not in advantaged families, while measures of intellectual functioning have been empirically linked to delinquency or crime even when family socioeconomic characteristics are statistically controlled. Denno's own study (1990:78-81) links perinatal conditions to mental retardation at age 13-14 and mental retardation to the seriousness of delinquencyâfor females but not males. Many risk factors that emerge during the preschool and early school years âfearlessness, attention deficit, inability to defer gratification, restlessness, and especially early school failureâare among the most robust predictors of childhood aggressive behavior. However, analyses of longitudinal data indicate that while IQ measured at age 8 appears to predict aggression at age 30, this relationship disappears when it is controlled for measures of aggression at age 8. Violent events on television are one route through which social learning of aggressive behavior may occur. Some previous research reviews (e.g., Heath et al., 1989; Huesmann and Miller, in press) and a meta-analysis prepared for the panel (Comstock and Paik, 1990) conclude that frequent exposure to television violence is associated with more aggressive behavior by children and adults. However, there is debate about the underlying causal relationship. The correlations may reflect the joint effect of greater exposure to television violence and a heightened potential for violent behavior, both resulting from poor parental supervision. It may also be that children with a high potential for violent behavior select violent material to watch. There is also debate about the permanence of any direct effect, about the importance of whether the television violence is punished or rewarded, and about the behavioral effects of frustration that might be activated by compelling changes in children's television viewing patterns (see debate between Friedrich-Cofer and Huston, 1986; Freedman, 1986). Moreover, the effects of high levels of neighborhood violence may reduce the effects of viewing of television violence on children's potentials for violent behavior; we are aware of no research on this issue.
PERSPECTIVES ON VIOLENCE 107 Violent offenders as measured by self-reports and official records of crimes are more likely than other adults to have experienced poor parental childrearing methods, poor supervision, physical abuse, neglect, and separations from their parents when they were children (Farrington, 1991). These experiences are commonly reported by violent sex offenders as well (Langevin et al., 1985). School bullying is of interest both as a childhood aggressive behavior and as a precursor of adult violent behavior. Children who are rated âaggressive" by their peers at age 8 have significantly higher average levels of adult aggression, and the underlying causal relationship may involve peer rejection. Childhood aggressiveness has been cited as a common reason for peer rejection (Huesmann and Eron, 1986). However, conflicting findings exist (Cairns et al., 1988), and the role(s) of peer rejection in developing adult aggression is uncertain. Sorting out these relationships is important in choosing whether to try to prevent bullying, peer rejection, or both as approaches to preventing adult aggression. Some factors appear to be protective, in the sense that they reduce the chance of aggressive behavior in individuals who would otherwise be at high risk. One protective factor, a high-income family, was already suggested as reducing the risk-increasing effects of perinatal complications. Other protective factors suggested by previous research include a shy temperament, high IQ, being firstborn, having a stable family, having affectionate caregivers, having parents who regularly attend religious services, and coming from small families characterized by low discord. As with adverse risk factors, however, these relationships have been observed with respect to aggressive but not violent behavior. Preventive Interventions We reviewed evaluations of a number of preventive interventions that seek to alter the psychosocial risk factors listed above (see Appendix A). In careful evaluations, several interventions have shown some success in reducing children's aggressive behavior, but longer-term follow- ups are needed to see if they also prevent violent behavior at later ages. Some of the most promising programs are global, using multiple social learning and behavioral interventions simultaneously to emphasize several of the following: the undesirability of aggression; nonaggressive methods of solving problems; improved social skills; and improved television viewing habits (Guerra, 1990; Tremblay et al., 1991; Pepler et al., 1991). Other promising interventions have a single focus: social skills (Jones and Offord, 1989); or improved television
PERSPECTIVES ON VIOLENCE 108 viewing habits (Huesmann et al., 1983; Eron and Huesmann, 1984; Singer and Singer, 1981). Several communities, including Oakland, California, and San Antonio, Texas, report success with a program for reducing early-grade school failure, one of the risk factors for aggression. The program involves daily tutoring of kindergartners by specially trained high school seniors and seems ripe for a carefully controlled evaluation. Other behavioral approaches that warrant additional testing include parent training in consistent discipline (see review by Kazdin, 1985), combined parent and teacher training (Hawkins et al., 1991), preschool enrichment programs of the Head Start type (Berrueta-Clement et al., 1984), and school-based programs to combat bullying and reduce peer rejection (Olweus, 1991). One preventive approach developed by the public health community was motivated in part by findings that most violent events are preceded by escalation from verbal conflict through insults and threats, and that the likelihood of escalation is increased by the presence of bystanders who do not attempt to mediate (see, e.g., Luckenbill, 1977; Steadman, 1982; Felson and Steadman, 1983). Using culturally specific role plays and other devices, the Boston Violence Prevention Curriculum was developed for use in tenth-grade health classes, to teach methods of interrupting escalation to violence. The curriculum emphasizes the undesirability of violent behavior, nonviolent responses to provocation, and mediation skills (Prothrow-Stith et al., 1987). A 10-site evaluation suggested some positive effects on knowledge and attitudes but little success in changing behavior (Spiro and DeJong, 1991). The lack of success may reflect the previously noted stability of aggressive behavior after middle childhood and the need to reinforce principles taught in school with experiences elsewhere in the community. A version of the program involving other community organizations is still being evaluated. Depending on the results of that evaluation, serious consideration should be given to developing and testing a revised version of the program, for use with children at around the third-grade level. In general, research suggests that preventive interventions are more likely to be successful if they involve parents, peers, teachers, and significant others in the community, and if the intervention is adapted to the cultural norms of the target age, ethnic, and socioeconomic category. Interventions are likely to be more effective if they begin earlyâperhaps as early as the preschool yearsâand if they are based on clear theoretical models of aggression. Evaluations should test multiple interventions simultaneously, to
PERSPECTIVES ON VIOLENCE 109 compare their effectiveness in different populations, particularly with children of different ethnic statuses, socioeconomic statuses, and ages. Follow-up periods must be long enough to ascertain whether interventions in childhood prevent violent behavior as adults. Violent Sexual Behavior By violent sexual behavior we mean a threat or actual use of physical force that either coerces another person to submit to sexual behavior or leads to sexual excitement or release in the perpetrator. This definition includes a wide range of behaviors, and there is profound disagreement over whether the perpetrator of a violent sexual act is motivated by the power achieved through coercion, by the sexual excitement, or both. Designations of violent sexual behavior as acceptable, criminal, deviant, or the product of mental illness have varied over historical periods and across cultures. The designation has also hinged on circumstances. Where it is clear that informed consent could not have been givenâacts against strangers or against children, for exampleâthe acts are usually assumed to be criminal or deviant. For events involving intimates, the designation depends on whether both participants freely gave consentâa matter that may sometimes be ambiguous in one or both of their minds and, for some acts, may be exceedingly difficult for a third party to ascertain. In studying sexual violence, the methods used to study other types of violence are especially difficult to implement. Because we do not yet know enough to identify categories of children at high risk for violent sexual behavior, it is difficult to design efficient samples for longitudinal study or for follow-ups intended to test the effectiveness of early preventive interventions. People may distort self-reports of violent sexual behavior because of embarrassment over discussing sexual behavior generally; however, persons with violent sexual preferences, by definition, are sexually aroused by contemplating or discussing both real and imagined violent events. Reports by victims and witnesses to sexual violence are subject to distortion because of emotional trauma over the events. In arrest reports, sexual motivation is sometimes disguised in certain cases of assault, homicide, and even burglary. All these kinds of reports are subject to error when the distinction between a violent and a nonviolent sexual event hinges on the issue of consent, which entails ambiguities and uncertainties. A valuable behavioral measurement of violent sexual preference
PERSPECTIVES ON VIOLENCE 110 is phallometric measurement of arousal to violent sexual stimuli. However, sexual preference is not equivalent to sexual acting out. Moreover, while the accuracy of phallometric testing has been estimated at 85 to 90 percent in distinguishing between heterosexual and homosexual preferences, and 70 to 75 percent in distinguishing between preferences for children and adults, accuracy in distinguishing between preferences for violent and nonviolent sexual stimuli are considerably poorer (Murphy et al., 1986). Research to date on violent sex offenders points to some of the same psychosocial and biological factors discussed earlier in this chapter with respect to aggressive and violent behavior generally: poor parental supervision; unstable family relationships; being abused or neglected as a child; abnormal functioning of sex hormones; and neurophysiological abnormalities in the temporal lobe of the brain. As discussed more fully in Chapter 4, perpetrators of sexual and nonsexual violence both share a tendency to have histories of alcohol abuse. Indeed, the criminal records of known violent sex offenders often contain notations of other crimes such as assaults, thefts, breaking and entering, and alcohol-related charges. Because of the methodological problems mentioned above, the relationships of these factors and processes to sexual violence are generally less well understood than their relationships to violence generally. However, available findings are sufficient to rule out several simple causal relationships that have occasionally been suggested for violent sex offending: epilepsy, mental retardation, and some XYY or XXY chromosomal syndromes. In the paragraphs that follow we first discuss factors and processes that appear to operate differently with respect to sexual violence than to nonsexual violence, and then discuss prospects for treating violent sex offenders and preventing or mitigating incidents of sexual violence. Deviant Sexual Preferences and Sex-Role Socialization Learning may be involved in one or both of two ways in the development of violent sexual behavior, which is usually studied using different samples and methods from those of other violent behavior. First, violent sexual preferences may be learned through experiences around the time of puberty, in which sexual arousal and release give the reward for fantasies about violent acts. Second, it may also be that, in the process of sex-role socialization, some children learn attitudes that facilitate violent sexual behavior toward women. Such attitudes include the general expectation that males are expected to behave aggressively while females behave passively and submissively (Baron and Straus, 1989; Cherry, 1983;
PERSPECTIVES ON VIOLENCE 111 Russell, 1975; Weis and Borges, 1977), as well as beliefs that certain specific circumstances justify violent behavior against women (Burt, 1980; Check and Malamuth, 1983; and Kikuchi, 1988). The available evidence is inconclusive on the role of sexual preferences as a cause of violent sexual behavior. As one would expect, phallometric tests of men in treatment have found rapists more likely than other men to respond to stimuli depicting rape, and child molesters respond more than others to stimuli depicting children. However, treatment populations may underrepresent men who sexually assault nonstrangers. Moreover, not even all sex offenders in treatment demonstrate deviant arousal patterns (Prentky, 1990), and there are no reliable estimates of the prevalence of men who are sexually aroused by violent fantasies or stimuli but never act on them. The evidence is similarly inconclusive on the importance of sex-role socialization as a cause of sexual violence. Some surveys suggest widespread learning, by the early teenage years, of beliefs that are conducive to rape.1 These findings are disturbing; however, their causal implications for violent sexual behavior are unclear. There is no available evidence on whether such attitudes are more prevalent among men who engage in violent sexual behavior than among men who do not. The theory of sex-role socialization does not attempt to explain violent sexual behavior against children, which is not justified by any widespread social belief system. Pornography Some suggest that pornography encourages violent sexual behavior either by facilitating development of a deviant sexual preference for violence or by helping to inculcate beliefs about male-female sex roles that are conducive to violence (Dworkin, 1979; Morgan, 1980]. Although the panel did not review the research on links between pornography and violence, a number of other scientific panels have investigated pornography and sexual violence, including the 1970 U.S. Commission on Obscenity and Pornography, the 1985 Special Committee on Pornography and Prostitution in Canada, and the 1986 U.S. Attorney General's Commission on Pornography. Scientific literature reviews prepared for these groups have generally concluded that, despite theoretical causal links, demonstrated empirical links between pornography and sex crimes in general are weak or absent. Studies of individual violent sex offenders have found no link between their offenses and their use of pornography; if anything, they do not appear to use pornography as much as the average male.
PERSPECTIVES ON VIOLENCE 112 Researchers have explored the possibility that the use of pornography with violent sexual or nonsexual content may interact with other individual characteristics to elevate the risk of engaging in sexual violence (Donnerstein et al., 1987; Prentky, 1990). In several studies, pornography depicting either sexual or nonsexual violence against women aroused identified rapists more than nonrapists (Quinsey et al., 1984, cited in Prentky, 1990). However, both categories of subjects were also aroused by pornography that depicted consenting sexual activity. This suggests that any effect of pornography on the potential to rape may involve an interaction between violent content and some other characteristic(s) peculiar to rapists. In short, we cannot rule out the possibility of a causal role for violent pornography. Research on this issue should be a high priority. Sexual Abuse in Childhood Sexual abuse in childhood apparently has a very conditional causal relationship to subsequent violent sexual behavior, including child molestation (Prentky, 1990). Childhood sexual victimization is disproportionately common among child molesters, prostitutes, drug addicts, and criminals in general (Hanson and Slater, 1988). Yet prospective studies demonstrate that most victims of sexual abuse in childhood do not become child molesters (Finkelhor, 1984, cited in Prentky, 1990). Research suggests that sexual abuse in childhood plays a conditional role in a causal chain that leads to subsequent violence, depending on the precise nature of the event, the child's emotional reaction to it at the time, the emotional importance and trust that the victimized child had placed in the preexisting relationship to the molester, and the surrounding subcultural norms that reward or condemn sexual and other violent behavior (Prentky, 1990; Kaufman and Zigler, 1987; Browne and Finkelhor, 1986). Mental Illness and Personality Disorders The relationships between violent sexual behavior and both mental illness and personality disorder are in question. Researchers consistently find that, although psychoses and major mental illnesses are rare among sex offenders, all diagnoses occasionally appear (Prentky, 1990). For perpetrators of sexual homicides, researchers such as Revitch (1965) and Dietz (1986) disagree on the prevalences of psychotic illnesses with hallucinations and/or delusions. Sex killers may be more prone to psychosis than are other individuals, although no particular diagnosis such as schizophrenia or psychopathy has been
PERSPECTIVES ON VIOLENCE 113 identified as characteristic. Although the term psychopath has sometimes been used as a label for those who commit sexual homicides, the term has been applied to a variety of different conditions over the years, and there is no scientific basis for considering psychopathy a specific adjunct of sexual homicide or of general sexual violence. A number of studies have attempted to link specific personality characteristics to violent sexual behavior. Both lack of masculinity and extreme aggressiveness have been suggested as predictive characteristics. Current research, however, suggests that sex offenders who have been labeled as unassertive may be no different from offenders in general and that the label may reflect social class differences between prisoner and examiner rather than any inherent characteristic of sex offenders (Segal and Marshall, 1985). Antisocial personality is diagnosed in samples of sex offenders between 40 and 80 percent of the time (Prentky, 1990), and virtually all other psychopathologies co-occur at times. The personality profiles that are most common among sex offenders are also most common for general prison samples (see Megargee et al., 1988, cited in Chaiken et al., Volume 4). Treatment Interventions Historically, the most common treatment for sex offenders has involved behavior therapy, and evaluations are largely limited to case studies in institutions. The majority of the reports available to this day are on homosexual men convicted of child molestation. The most common method of treatment has employed electric shock, paired with pictures of children, to reduce sexual arousal to children generally. The results have been mixed. In some states, the use of these techniques has been banned. Initially, the claims of success were greater than they are today, but it is clear that behavioral techniques have an important role to play in the treatment of sex offenders. Other treatments, such as assertiveness training, anger management, life-skills training, and relapse prevention, are used more at present. Anger management appears to be an appropriate technique for dealing with both nonsexually and sexually violent individuals. The technique can help a person who is generally aggressive to deal with social problems. However, it has not been demonstrated that, in fact, such techniques can alter a long-term pattern of sexually aggressive behavior. More fundamental issues have been raised about whether sex offenders are helped by any form of treatment. Clinicians and
PERSPECTIVES ON VIOLENCE 114 researchers are becoming increasingly aware that sex offenders are resistant to change. A primary factor in treatment plans now is to motivate the sex offender to change. Simpkins et al. (1989), for example, found that the more open the offender was about his sexual behavior, the more progress he made in treatment. However, substantial numbers of sex offenders deny that they have any sexual problem or resist treatment by noncompliance or dropping out. Preventive Interventions Strategies for preventing sexual violence against acquaintances include educating boys and men about roles and responsibilities toward women, and teaching females about the dangers and how to protect themselves. However, these interventions have not been evaluated definitively for their potential to prevent sexually violent behavior. High priority should be given to conducting such evaluations. Because the roles of sexual preferences and sex-role socialization as causes of violent potentials are still uncertain, and methods of changing them are still in development, we do not expect psychosocial interventions to reduce the incidence of sexual violence significantly in the short term. While more effective preventive interventions are being developed and tested, interim strategies for reducing the harm from sexual violence involve situational approaches to preventing events and tertiary prevention, which is intended to repair the consequences to the victim. Situational approaches to prevention of sexual violence against strangers could operate by modifying placesâfor example, by installing emergency telephones and lighting, or by providing escort services. Alternatively, they could operate by altering routine activities. Examples of possible interventions for adult women include education to modify routine activities such as hitchhiking and walking alone in lonely areas, or to conceal status as single in public places. For reducing violence against child victims, training parents in how to choose caregivers and how to "street-proof" children and providing adult escorts all show promise. In this area, however, we lack systematic evaluations of effectiveness. Situational interventions for preventing nonsexual violence are discussed later in this chapter. Yet another approach to intervention is to repair the consequences of sexual violence, which can include numerous physical and emotional symptoms. The most common forms of treatment for the psychological trauma that victims of sexual violence experience are crisis intervention counseling and psychotherapy/ support groups. Other specific interventions to reduce fear and anxiety
PERSPECTIVES ON VIOLENCE 115 include systematic desensitization; stress inoculation training, combining education, relaxation training, and teaching of other coping skills; helping clients to identify cognitive distortions and maladaptive thought patterns and replace them with adaptive responses; and reducing the sexual anxieties and problems that the victim often experiences after a rape. Although much of the research on these treatment strategies reports improvement in the victims, it is flawed by unclear criteria for inclusion, lack of control/comparison groups, and weak conceptualizations of treatment strategies with respect to the symptoms they target. Foa and her colleagues are currently conducting a large-scale outcome study, integrating various modalities discussed above, to identify the most effective forms of treatment. Biological Perspectives All human behavior, including aggression and violence, is the outcome of complex processes in the brain. Violent behaviors may result from relatively permanent conditions or from temporary states. Relatively permanent conditions may result from genetic instructions, from events during fetal or pubertal development, from perinatal accidents, or from brain trauma. Relevant temporary states may be brought on either by some purely internal activity (e.g., brain seizures) or through responses to external stressors, stimuli that produce sexual arousal, ingestion of alcohol or other psychoactive substance, or some other external stimulus. The possibility exists that some violent behavior may be prevented by modifying the precursors of the relevant neurological conditions and states. Biological research on aggressive and violent behavior has given particular attention to the following in recent years: (1) genetic influences; (2) functioning of steroid hormones such as testosterone and glucocorticoids, especially their action on steroid receptors in the brain; (3) functioning of neurotransmitters, particularly dopamine, norepinephrine, serotonin, acetylcholine, and gamma-aminobutyric acid (GABA); (4) opioid and other neuropeptides; (5) neuroanatomical abnormalities of certain brain morphologies; (6) neurophysiological (i.e., brain wave) abnormalities, particularly in the temporal lobe of the brain;
PERSPECTIVES ON VIOLENCE 116 (7) brain dysfunctions that interfere with language processing or cognition; and (8) hypoglycemia and diet. As we discuss in the following pages, correlations have been found between aggressive or violent animal or human behaviors and conditions or states in all these categories. However, interpreting these correlations is difficult. The interactions are complicated, precise measurement of brain activity is difficult, and violent behavior is rare, especially in the animal subjects used in much of the relevant research. The generalizability of experimental findings from other animal species to humans is not always straightforward, both because neurochemical functions vary across species2 and because homologies are hard to draw between specific aggressive and predatory behaviors by animals and violent behaviors by humans. Most of the evidence concerning human subjects consists of correlations rather than experimental results. Causal interpretations are therefore tenuous. Not only are hormonal, neurotransmitter, and neurophysiological processes involved in violent and nonviolent behavior; these functions can be changed as consequences of violent behavior. Given these difficulties, it should not be surprising that no patterns precise enough to be considered reliable biological markers for violent behavior have yet been identified. Researchers have, however, identified some particularly promising areas for further research that may eventually lead to nonintrusive diagnostic and prognostic indices of individuals' potentials for aggression and perhaps to preventive interventions without unacceptable side effects. Many of these findings are summarized in the following pages. Pharmacological links between violent behavior and the use of alcohol and other psychoactive drugs are discussed in Chapter 4. More complete treatments and supporting citations can be found in Volume 2 of the panel's report. Existing findings leave open many questions that are important from scientific or policy perspectives, or both. Recommendations for research to fill these gaps appear in Chapter 8. Genetic Influences Correlations in studies of humans fairly consistently demonstrate genetic influences on individual potentials for antisocial behavior and juvenile delinquency (Bohman et al., 1982; Christiansen, 1977; Cloninger and Gottesman, 1987; Mednick et al., 1984). The
PERSPECTIVES ON VIOLENCE 117 role of genetic mechanisms in criminal and violent behavior has been examined in three studies of twins and adoptees, using data from Scandinavian countries (Bohman et al., 1982; Cloninger and Gottesman, 1987; Mednick et al., 1984). These studies suggest at most a weak role for genetic processes in influencing potentials for violent behavior: the correlations and concordances of behavior in two of the three studies are consistent with a positive genetic effect, but are statistically insignificant. The different findings with respect to the two categories of behavior raise the possibility that some genetic process that is relevant to violence may be concealed because violent behavior is rare and/or etiologically heterogeneous. In any event, the implications of these findings to violence patterns in the United States are uncertain because the measures of heritability derived from relatively homogeneous Scandinavian populations are not readily extrapolated to U.S. society. By analogy, some evidence suggests that individuals may differ in their genetic predispositions to at least some forms of alcoholism (Cloninger et al., 1978; Plomin, 1989). That variation can be expected to have a greater effect on the distribution of alcoholism rates when alcohol is more readily available and inexpensive, and where alcoholism prevention programs are less successful, for example. Similarly, the effects of any variation in genetic predisposition to aggressive or violent behavior can be expected to depend on such factors as families' responses to aggressive behavior by their developing children, the availability of weapons, and the financial, social, and legal punishments and rewards for violent behavior. Well-documented principles of behavior genetics have at least three implications for any genetic process that may be found in the future to affect potentials for violent behavior. First, any such process would have to involve many genes and substantial environmental variation, rather than any simple chromosomal syndrome that might be useful as a marker or risk factor. Second, because human gene pools change over periods of time measured in generations, and because gene pools within any country are so diverse, genetic processes alone cannot explain either short-run fluctuations in violence rates over time or variation in violence rates across countries. Third, in studies of psychological and social influences on violent behavior, the designs should collect the data on parents and siblings that are needed to control for confounding genetic influences, and greater use should be made of samples of twins, sibling pairs, and adoptees to develop a more precise understanding of the relevant genetic processes.
PERSPECTIVES ON VIOLENCE 118 Over the past 40 years, occasional case studies of sexually deviant individuals and families have raised the possibility that some simple XXY or XYY chromosomal syndrome may transmit a potential for sexually violent behavior (Baker and Stoller, 1968; Baker et al., 1970; Bartlett, 1968; Pasqualini et al., 1957; see also reviews by Kessler and Moos, 1970; and by Owen, 1972). Case studies provide only weak tests of this hypothesis, and the panel is aware of no study using modern behavior genetic designs that has tested for genetic influences on sexual violence. In view of the general difficulties of establishing genetic relationships to rare behaviors that are also subject to environmental influences, however, it would be premature to rule out the possibility that some complex interaction involving multiple genes and life experiences may account for some instances of sexual violence. In rodents, the successful establishment of aggressive and nonaggressive strains of a species, as well as comparisons across strains, demonstrates significant heritability of aggression (Carey, Volume 2). However, animal studies also show that tendencies toward aggression can be modified by experience, contextual cues, and the social environment. Hormonal Mechanisms Because testosterone is the principal androgenic hormone (i.e., the principal hormone in producing masculine characteristics), it has been more frequently studied in relation to sexual violence than to other violent behaviors. Studies find high prevalences of elevated testosterone levels in clinical samples of violent sex offenders, and there is some suggestion that this correlation increases with the violence of the crime. However, the correlation is probably confounded by alcohol abuse, which modifies testosterone levels in complex ways and is also associated with violent behavior. Animal and human research suggests that prenatal and perinatal exposure to abnormal levels of androgenic hormones permanently alters the relationship between aggressive behavior and steroid hormones (see Brain, Volume 2, Tables 1, 3, and 4, for supporting citations). In female mice and rhesus monkeys, greater adult aggression has resulted from injections of testosterone during pregnancy or at the time of birth. Follow-ups of human children find that girls who were accidentally exposed to inappropriate androgenic hormones during fetal development display an unusually high long-term tendency toward aggression, while boys prenatally
PERSPECTIVES ON VIOLENCE 119 exposed to antiandrogenic steroids show decreased aggressiveness. It is not clear, however, whether the correlations in humans reflect direct hormonal influences on behavior or some psychosocial interaction involving peers' reactions to abnormal genital development caused by the abnormal hormone exposure. Some correlations have been cited as suggesting that testosterone may promote immediate aggressive responses to provocative stimuli. However, these patterns are weaker in nonhuman primates than in lower animals, and weaker still in humans. Moreover, studies of humans provide fairly strong evidence of reverse causal relationships: that being subject to aggression or being defeated in competition decreases human androgen levels, and that winning a competition increases them (see, e.g., Brain, Volume 2). Two kinds of studies offer evidence on how manipulations of hormone levels affect violent behavior. First, antiandrogen drugs show some promise as adjunct therapy for violent sex offenders, but the effect seems to depend more on reducing the sex drive than on dampening the violent behavior. Second, studies of the effects of administering synthetic derivatives of steroids (i.e., anabolic steroids) suggest that, despite occasional anecdotes concerning athletes in training, injections of anabolic steroids exert at most a weak influence accounting for little variation in violent behavior. Recent studies suggest that the functioning of two other sex hormonesâ dehydroepiandrosterone sulfate (DHEAS) and luteinizing hormoneâmay be associated with unusual sexual behavior, including sexual violence.3 These include both basal studies, which compare hormone levels in peripheral blood samples, and challenge tests, in which a hormone is injected to stimulate the brain and reactions are measured. Animal studies suggest that maximal information is extracted when one has both kinds of tests, because the interactions between resting levels and reactions to challenges are not yet well understood. The relationships between violent behavior and these sex hormones need more systematic investigation. Neurotransmitters and Receptors Of at least 50 known neurotransmitters, 4 have been studied most thoroughly in relation to aggressive behavior (primarily in animals, but occasionally in humans): dopamine, norepinephrine, serotonin, and GABA. Findings from these studies are reviewed in Volume 2, and the following discussion is based on those sources.
PERSPECTIVES ON VIOLENCE 120 Among other neurological functions, dopamine activates neural processes for pleasure and reward; cocaine, amphetamine, and other psychoactive drugs stimulate these processes (see Chapter 4 for further discussion of drug effects on dopamine receptors). Although evidence from animal studies points to large natural changes in brain dopamine systems during and following aggressive and defensive behavior, there is no evidence on whether similar responses occur in humans. It is possible that the brain dopamine activity in animals is one of the rewarding aspects of aggression, but specific markers for dopamine involvement in any specific type of aggression are not available. Antipsychotic drugs, especially those that antagonize the D2 subtype of dopamine receptor are frequently used to quell acute violent outbursts or as long-term âchemical restraint" for violence-prone persons confined in institutions. However, these drugs alter behaviors other than aggression and, as with cocaine and amphetamine, chronic use of these drugs produces an array of neurological problems (see Miczek et al., Volume 2, Chapter 3). Norepinephrine transmission and its contribution to the "flight or fight" syndrome has been well documented for decades. Researchers have recently identified specific brain regions in which this activity takes place. Noradrenergic activity, however, occurs in conjunction with high levels of arousal generally, and no specific noradrenergic marker in any brain region has been identified for violent or aggressive behavior. In the past dozen years the most significant clinical use of noradrenergic drugs has involved those that inhibit the activity of beta-type adrenergic receptors (beta-blockers). These drugs are used to manage violent behavior by hospitalized retarded, schizophrenic, and autistic patients. However, their aggression-reducing effect may occur through their actions on certain subtypes of serotonin receptors, and their behavioral effects are not specific to aggression. Clonidine, a drug that targets alpha-type adrenergic receptors, has been used to manage irritability, aggression, and defensive acts during withdrawal from addiction to alcohol, nicotine, and opiates. Serotonin is the neurotransmitter that has been most intensively studied in animal and human research on violent behavior. Marked changes in serotonin synthesis, release, and metabolism have been observed in individuals of many nonhuman animal species that have repeatedly engaged in violent or aggressive behavior. But the substantial variation in serotonin functions across species makes extrapolations from one species to another, including humans, problematic.
PERSPECTIVES ON VIOLENCE 121 Since the late 1970s, several researchers using human subjects have reported inverse correlations between serotonin concentrations in blood or spinal fluid and various measures of aggressive, impulsive, or suicidal behavior. Recent studies of Finnish males by Virkkunen, Linnoila, and associates (Virkkunen et al., 1989; Linnoila et al., 1989) find that a complex interaction among serotonin concentration, alcoholism, responsivity to glucose challenge (a test for hypoglycemia), and aspects of monoamine metabolism may be related to these behaviors. More generally, a number of correlative studies reviewed by Ellis (1991) suggest that low monoamine oxidase (MAO) blood concentrations may indicate elevated potentials for suicide, sensation seeking, impulsiveness, alcoholism, and criminality. More consistent replications and a better understanding of the relationship between measures in blood, spinal fluid, and relevant brain regions are needed before these factors can be accepted as biological markers for violent behavior. Certain drugs that act on the serotonergic system have been found in animal trials to reduce aggressive behavior and to decrease anxiety in humans. Current clinical research with novel drugs acting on serotonin receptor subtypes promises to reveal specific antiaggressive effects in humans with fewer side effects. Beta- blockers and lithium may achieve their antiaggressive effects through action on these serotonin receptor subtypes. GABA was long thought to act as a neurotransmitter that inhibited aggressive behavior in animals. The latest research finds that GABA activity has excitatory as well as inhibitory actions in many discrete brain regions, but no pattern provides a marker for aggressive or violent behavior in animals or humans. The GABAA receptors are of particular clinical interest as the site of action for benzodiazepines, one of the most important antianxiety drugs used to reduce aggression and to sedate very agitated individuals. Evidence from animal and human studies documents the calming effects of these drugs for most persons. However, under certain conditions, these drugs occasionally lead to violent outbursts called "paradoxical rage." GABAA receptors are also one of the targets of alcohol action in the brain, and these receptors represent a promising site for pharmacological interventions to reduce alcohol-related violent behavior. Neuroanatomical Abnormalities Experiments involving electrical stimulation of specific brain regions in cats have established two distinct types of aggressive behavior, predation and affective defensive aggression (i.e., "rage"),
PERSPECTIVES ON VIOLENCE 122 which involve different circuits of neural activity. The limbic system and perhaps secondarily the cerebral cortex appear to modulate both types of behavior (see Mirsky and Siegel, Volume 2). These findings raise the possibility that a similar typology may apply to human aggression, perhaps involving the same neural pathways. However, concern for human subjects prevents researchers from using the techniques needed to assess how applicable neurological models of feline or other animal aggression are in classifying or explaining human aggression. Available evidence from small samples of humans also points tentatively to structures in the limbic system, especially the amygdala and medial hypothalamus, as brain sites associated with certain violent behaviors. However, this evidence has been seriously questioned, in part because of the general difficulty of studying the limbic system directly. Perhaps more importantly, virtually all the human subjects were observed in the course of treatment for diagnosed brain abnormalities or for highly unusual histories of chronic unprovoked violent behavior. There is no sound basis for generalizing from either of these atypical groups to statements about brain regions that might be involved in violent behavior by persons in the general population. Neurophysiological Abnormalities Several studies suggest rare episodes of violence either before or between seizures associated with temporal lobe epilepsy (Mirsky and Siegel, Volume 2). However, the question remains unanswered as to whether patients with seizure disorders have greater or smaller potentials for violence than the general population. New, less intrusive brain imaging techniques may facilitate research that would provide better understanding of the brain structures involved in aggressive and violent behavior in the general population. A link between abnormal, not necessarily violent, sexual behavior and brain disorders was suggested by Kolarsky et al. (1967), who found that among epileptics, unusual sexual behavior occurred more commonly when neurophysiological abnormalities appeared in the temporal lobe than in other sites. These findings were later supported by some studies of animals and, most recently, by studies using advanced tomographic techniques that found a high prevalence of temporal lobe impairment among humans who engage in sexually anomalous behaviors (Langevin, 1990). Preliminary research suggests that men with sadistic sexual preferences
PERSPECTIVES ON VIOLENCE 123 show temporal lobe anomalies more often than do other offenders. These correlations cannot be said to show either that temporal lobe abnormalities cause violent sexual behavior or that epilepsy is a risk factor for violent or nonviolent sex offending. Other Brain Dysfunctions Several kinds of indirect evidence suggest that abnormalities of brain functioning increase the risks of violent behavior, although no specific neurophysiological marker has yet been identified (see Mirsky and Siegel, Volume 2, on which the following discussion is based). First, a large number of studies indicate that neuropsychological deficits in memory, attention, and language/verbal skillsâwhich sometimes follow from limbic system damageâ are common in children who exhibit violent or aggressive behavior (see, e.g., Mungas, 1988; Miller, 1987; Piacentini, 1987; Lewis et al., 1988). While this correlation could reflect some direct relationship between limbic system damage and aggressive behavior, it is more likely to reflect less direct results of distorted social interactions with peers resulting from impaired communication skills, or to arise from frustration over the inability to compete successfully with peers in cognitive tasks. Second, certain peripheral measures of nervous system activity such as heart rate and skin conductance have been found to differ on average between samples of normal controls and samples of criminals, delinquents, conduct-disordered children, and psychopaths (Kagan, 1989; Raine and Venables, 1988a; Raine et al., 1990a; Siddle et al., 1973; Wadsworth, 1976). Third, abnormal patterns of brain electrical activitiesâslow waves and the P300 component of event-related brain potentialsâhave been found in samples of incarcerated criminals and psychopaths (see, e.g., Williams, 1969; Hare, 1978; Raine and Venables, 1988b; Raine et al., 1990b). The appropriate interpretation of these patterns is not entirely clear because neither the samples nor the environments reflect the community. The findings, however, raise questions that can be more clearly resolved through research that uses new technology for brain scans, especially magnetic resonance imaging (MRI) and positron emission tomography (PET). Such research is expensive and subject to human subjects requirements. From the standpoint of designing violence control interventions, perhaps the most useful information about brain dysfunctions and violent behavior is the extent to which they arise from social-environmental conditions. Some individuals' dysfunctions may
PERSPECTIVES ON VIOLENCE 124 well be a result rather than a cause of aggressive behavior: they may originate in head injuries inflicted by others in retaliation. However, research reviewed by Mirsky and Siegel (Volume 2, Table 13) points to the following environmental correlates of cognitive deficits, hyperactivity, and other precursors of violent behavior: exposure to environmental lead (e.g., in air contaminated by leaded fuels, from paint in older houses, from plumbing systems) and maternal use of opiates, cocaine, alcohol, and tobacco during pregnancy. In addition to other likely benefits for children's health, interventions to reduce substance abuse by pregnant women and to reduce small children's environmental lead exposure should be considered potential long-range preventive interventions for violence. However, health education is frequently insufficient by itself to alter expectant mothers' behavior during pregnancy (Institute of Medicine, 1985). Hypoglycemia and Diet Several kinds of studies suggest possible relationships between aggressive behavior and sugar intake and metabolism, but the interpretation of findings is problematic (see more detailed reviews in Kanarek, Volume 2; Venables and Raine, 1987). For example, some studies find that interventions to reduce sugar consumption reduce antisocial behavior by institutionalized juvenile offenders, and some studies find correlations between sugar intake and hyperactivity in young children; however, the findings are inconsistent. Moreover, outside controlled settings, children's sugar intake may be related to family characteristics such as socioeconomic status and parental supervision, which are also correlated with aggressive behavior (see Kanarek, Volume 2). Similarly, hypoglycemia (i.e., low blood glucose levels and associated behavioral symptoms) is relatively common in violent offenders (Virkkunen, 1986), and the peak time of day for assaults in prison coincides with the peak time for acute symptoms of hypoglycemia. Although these findings suggest that some relationship exists, they should not be interpreted as evidence of a direct causal link, because at least two alternative explanations cannot be ruled out. First, alcohol use increases susceptibility to hypoglycemia and is also related to the occurrence of violent events. Second, hypoglycemia has also been theoretically linked to both low heart rate and electroencephalogram (EEG) slowing, which are common among violent offenders. Occasional claims have surfaced that various food additives (e.g.,
PERSPECTIVES ON VIOLENCE 125 artificial colors or preservatives) are associated with hyperactivity or other precursors of violent behavior (see, e.g., Feingold, 1975). Although open trials in family settings tend to support these claims, carefully controlled double-blind experiments do not. Such relationships may exist, but they appear to affect at most a small fraction of preschool children. Occasional studies have found correlations between violent behavior or nondisease mortality (including intentional injury) and cholesterol intake or levels. These findings are mixed and therefore inconclusive, and the studies have generally failed to control for other behavioral changes that may have accompanied changes in cholesterol intake (see Kanarek, Volume 2). Some studies of rodents, cats, and humans suggest that nicotine temporarily reduces aggressive behavior. Of course, the potential use of nicotine to control such behavior should not be considered independently of the enormous health risks of tobacco smoking (Miczek et al., Volume 2). Developmental and Biomedical Evaluation and Research Needs We have presented a number of findings about the biological and psychosocial development of individual potentials for violent behavior. A number of questions that still remain unanswered either have been mentioned or are implicit in the discussion. We recapitulate them here. Aggression and Violence Our discussion of the psychosocial and biological development of potentials for violent behavior points to three promising intervention strategies that could begin at least in early childhood, and in one case before birth: (1) prevention of brain injuries, substance abuse by pregnant women, exposure to lead, and other prenatal, perinatal, and postnatal events linked to brain dysfunctions that increase individuals' potential for aggression; (2) cognitive-behavioral preventive interventions including parent training, school-based antibullying programs, social skills training, and interventions that stress the undesirability of aggression, teach nonviolent conflict resolution, and promote viewing of television programs that emphasize prosocial behavior; and (3) prevention of school failure through preschool educational
PERSPECTIVES ON VIOLENCE 126 enrichment and through kindergarten tutoring by specially trained high school students. These strategies are complementary rather than mutually exclusive: each can be expected to prevent some children from developing potentials for violent behavior, but none will be universally effective. Long-term evaluation programs are needed to learn how to adapt these intervention modes to specific subpopulations of children and to specific types of aggressive and violent behavior. Sexual Violence Very little is known about how potentials for sexual violence develop, how violent sex offenders differ from the general male population in terms of either sexual preferences or socialization toward women, or how the occurrence and recurrence of violent sexual behavior can be prevented. We place particular priority on four sets of research questions: (1) What are the roles of genetic-environmental interactions, sexual abuse in childhood, the learning of tolerant attitudes toward rape, and chronic alcohol abuse in the development of individual potentials for violent sexual behavior? (2) What is the role of violent deviant sexual preferences in causing violent sexual acts? How do these preferences differ between samples of known violent sex offenders and other samples? What events produce these violent preferences? What role, if any, does violent pornography play in their development? Are there specific neurological, endocrine, or genetic markers either for them or for an elevated potential to act on them? What subpopulations, if any, would benefit from hormone therapy? (3) What preventive and educational strategies show promise of reducing sexual violence involving intimates, acquaintances, or strangers? (4) Several psychosocial interventionsârelapse prevention therapy, assertiveness therapy, and anger management therapyâshow some promise of changing some individuals' violent deviant sexual preferences, especially when combined with pharmacological interventions that reduce the intensity of the sex drive. Systematic evaluation is needed, involving randomized experiments where practicable,4 to assess the effectiveness of these interventions in preventing the recurrence of sexually violent acts. Improved classification and measurement are essential for making
PERSPECTIVES ON VIOLENCE 127 progress on these issues, but they present special problems in the context of sexual violence. Classification systems should be refined and improved to facilitate both better developmental understanding and more effective treatment. In addition, special priority should be placed on improved classification of violent sexual acts against intimates that do not routinely lead to institutionalization. Advancing Biomedical Understanding For many people, attention to genetic, neurophysiological, neurochemical, and neuroendocrine processes that underlie violent behavior raises concerns about the potential for ethically unacceptable future preventive interventions. The specters of eugenics, preemptive incarceration based on individuals' biomedical profiles, and maintaining classroom discipline with drugs that dull children's creativity are all occasionally raised. Actually, as we have explained in previous pages, better biomedical understanding of violence has added violence prevention to the list of justifications for rather benign interventions with families, such as teaching pregnant women to avoid drug use and new mothers to protect their children from head injuries, and interventions in communities, such as reducing children's exposure to lead. A primary dilemma to accumulating better biomedical understanding is that less intrusive and more broadly accepted biomedical research methods tend to provide information that is more remote from the causative brain mechanisms. For example, analyzing urine or hair samples imposes negligible burdens on subjects, but the results confound relevant neurochemical processes with unrelated events that occurred long before or long after an episode of violent behavior. Sampling bloodâa somewhat more intrusive procedureâeliminates some confounding influences, but abnormal blood concentrations of serotonin metabolites, for example, may reflect serotonin production, metabolism, or excretion effects resulting from violent behavior together with any other life events rather than any causal effect specific for violence. Spinal tapsâan uncomfortable procedure that involves some riskâproduce measures that mirror ongoing brain activity somewhat better, but they are unrelated to the initiation or termination of violent acts and cannot distinguish the brain region in which the activity occurred. Modern brain imaging techniques locate brain activity somewhat more precisely but demand several hours of subjects' time, involve the injection of radioactively labeled tracers
PERSPECTIVES ON VIOLENCE 128 into the bloodstream, and sometimes trigger claustrophobic episodes in some people. Decisions to undertake specific research projects should balance the burden on subjects against the value of the information that may be obtained and the likelihood of success in obtaining that information. Clearly, for obtaining any given piece of information, less invasive procedures are preferred over more invasive ones. Among research projects that impose equal burdens on subjects, those that are intended to obtain more specific information about neurological or neuroendocrine functioning and behavior, or that study responses to demands and challenges, should be preferred over studies of more peripheral processes or of resting levels of chemicals or neurological activity. Specific challenge tests are needed to distinguish among current biomedical theories of aggressive and violent behavior, such as the serotonin deficit, imbalance in the autonomic nervous system, and a physiological basis of temperamentâall of which have partial support, at least, from animal studies. As an important source of information about the basic genetic, neurobiologic, developmental, and environmental factors that determine aggressive and violent behavior, proper use of animal models is a means of reducing research burdens on human subjects. The systematic study of animal models for adaptive, predatory, and pathological forms of aggressive behavior represents the primary means to gain insights into all the relevant neurobiologic mechanisms, including neurotransmitter systems and neurophysiological processes. In turn, these insights are a basis for exploring and designing effective interventions. Extrapolations from animal models of aggression and violent behavior to the human condition need to take species differences into account. The most promising findings are based on work with several animal species, at least one of which ought to be a primate species. Particularly profitable lines of inquiry for animal models are those that focus on the individual organism, its genetic predispositions, its development, and its interactions with social units and environmental events. The critical ethical dilemma of experimental research on aggression and violent behavior is the trade-off between experimental preparations with high realism and face validity versus the demand for protection of research subjects against stress and pain. It is unacceptable for experimental biomedical research with humans to achieve high face validity in studies of violent behavior. Therefore, it is necessary to develop and implement animal models,
PERSPECTIVES ON VIOLENCE 129 often as a means of building an empirical basis for subsequent research that places acceptable but nonnegligible burdens on human subjects. Each study of animal or human subjects needs to be designed to inflict the least possible stress and harm on the subjects without invalidating the information to be obtained. VIOLENCE AND SOCIAL PROCESSES Individuals, including those who behave violently, live and act in collective social units such as families, gangs, markets, voluntary organizations, and communities. People who commit crimes are influenced by a host of factors beyond their personal and family characteristics (Gottfredson and Taylor, 1986; Robins, 1978). How do social processes affect an individual's potential for violent behavior? In this section we draw on a diversity of research to discuss factors and processes in communities, neighborhoods, and other social units that are related to violent crime. We explore the links between poverty and violence, and the possible roles of culture in poor communities. The section closes with a discussion of recent research on youth gangs and violence, as an illustration of how collective social units both shape and are shaped by the communities of which they are a part. We know a great deal more about the extent and nature of violence and related behavior within the individual and social levels of explanation than we do about relationships that cross those levels. Most studies focus on a single level of explanation, or on rates of violence for categories of individuals or areal units, rather than linking them. Although the lack of adequate data is responsible for this state of affairs, adequate theory to guide data collection and analysis is equally at fault. For these reasons, we must draw on a broad range of studies and on the most persuasive empirical findings and theoretical arguments concerning the manner in which different levels of explanation relate to one another. How Do Ethnic Status and Poverty Affect Violence? Chapter 2 explains that blacks, Hispanics, and American Indians are at greater risk than whites for becoming victims of violent crime. However, a look behind this simple pattern reveals a more complex picture. The panel looked for evidence on violence rates for the smallest, most homogeneous areas for which data are availableâcensus tracts. 5 Four studies were found in which the dependent variable
PERSPECTIVES ON VIOLENCE 130 was the homicide victimization rate and the independent variables were measures of the racial or ethnic composition and of the economic status of the census tract. The studies covered three cities in Ohio (Muscat, 1988), New Orleans (Lowry et al., 1988), Atlanta (Centerwall's, 1984 reanalysis of Munford et al., 1976), and Boston. All are epidemiological studies of recent vintage. Although these studies were not designed to answer the question we pose, are of uneven quality, and focus on different types of homicide, they are nevertheless the best evidence available. In all four studies, there is an interaction between ethnic status and socioeconomic status. At low socioeconomic levels, blacks have much higher risks of becoming homicide victims than do whites. At higher socioeconomic levels, the difference between blacks and whites disappears and even reverses in one of the studies. Although the details of each study vary somewhat, the results are generally the same. The relationship is illustrated in Figures 3-1 and 3-2, which depict the New Orleans and Atlanta findings. Figure 3-1 Mean annual homicide rate, New Orleans (1979, 1982, 1985, 1986).
PERSPECTIVES ON VIOLENCE 131 Figure 3-2 Intraracial domestic homicide rates, Atlanta, 1971-1972. SOURCES: Population dataâ1970 U.S. census; homicide dataâAtlanta Police Department and Medical Examiner Offices, Fulton and DeKalb counties, Georgia. Links Between Poverty and Violence The interaction described above suggests that socioeconomic status, as measured using some indicator of poverty, is a useful starting point for understanding and controlling violence. More than 50 years ago, the classic work of Shaw and McKay (1942) on the ecology of crime and delinquency led to the conclusion that three structural factorsâlow economic status, ethnic heterogeneity, and residential mobilityâresulted in the disruption of community cohesion and organization, which in turn contributed to variations in crime and delinquency among communities. This conclusion was buttressed by the fact that high rates of delinquency persisted in communities characterized by these factors over many years despite high population turnover in the communities, which changed the ethnic and racial character of their residents. Researchers since then who have studied community-level variations have found associations between poverty and high rates of official
PERSPECTIVES ON VIOLENCE 132 delinquency (Gordon, 1967), but few studies focused specifically on violent crime. Those that did continued to find that homicides were disproportionately concentrated in areas of poverty (Bullock, 1955; Bensing and Schroeder, 1960; Beasley and Antunes, 1974; Mladenka and Hill, 1976). This pattern has held up regardless of which ethnic group occupied the poor areas. Three of these studies examined violent crime in Houston (Bullock; Beasley and Antunes; Mladenka and Hill). Like Shaw and McKay, each reported high correlations between violent crime rates and measures of poverty. Areas in Houston with high rates of violent crime were also characterized by high population density and a high proportion of black residents. Recent research finds correlations between violence rates and many community characteristics that are distinct from, but related to, socioeconomic status: concentrations of poverty, residential mobility and population turnover, family disruption, high density in housing and population, features of local social organization (e.g., low density of friends and acquaintances, few social resources, weak intergenerational ties in families and communities, weak control of street-corner peer groups, and low organizational participation in community life), and opportunities associated with violence (such as gun density, drug distribution networks, and the location of nonhousehold activities conducive to violence) (summarized by Sampson and Lauritsen, Volume 3). But finding statistical relationships between community characteristics and violence still does not explain how they are related to poverty and, in turn, how they increase violence. And it does not explain why race differences in violent crime rates tend to disappear when poverty is included as an explanation. In the next section we briefly review available findings that link violence rates not only to poverty levels in a community, but also to specific dimensions of poverty. Dimensions of Poverty One reason it is so difficult to disentangle the effects of race/ethnic status from community influences is because poor whites and poor blacks live in areas that are very differentâeconomically and otherwiseâmaking simple comparisons impossible. In 1980, for example, in the five largest U.S. cities, 85 percent of poor blacks lived in poverty areas, compared with only 30 percent of poor whites; nearly 40 percent of poor blacks lived in areas characterized by extreme poverty, compared with 7 percent of poor whites.
PERSPECTIVES ON VIOLENCE 133 Another source of difficulty is that observed relationships between particular communities and violence, by race, necessarily reflect differences between communities that are notâand probably cannotâbe accurately measured, statistically controlled, or otherwise fully taken into account. Compared with whites, the âecological niches" in which poor blacks live are disadvantaged in ways that defy easy measurement, such as job quality, access to high-status social networks, and exposure to conventional role models (Wilson, 1987:58-60). Despite these difficulties, in broad and summary terms the research finds that community influences often combine with poverty and with one another to produce high rates of violent crime. Although some studies report that the racial composition of communities has a direct effect on violence, virtually all studies report that the effects of race are also mediated by other factors, such as family structure and community change (Sampson and Lauritsen, Volume 3). When these factors are controlled statistically, the effects of race and ethnic status diminish or disappear altogether. This argument, it should be emphasized, is not theoretically limited to black-white comparisons. More data and research have been devoted to such comparisons, in part because blacks are the largest ethnic/racial minority in the United States. To the extent that the community-level correlates of violence are found to characterize other ethnic and racial groups, the expectation is that they too will experience high rates of violence. Empirical evidence is lacking, in part because, as we note in Chapter 2, the relevant statistical systems have only recently begun to record other ethnic designations. Community Characteristics Researchers have identified a number of community characteristics that are related to violence rates. For example, in studies of neighborhood rates of violent crime, measures of the density of multiunit housing, residential mobility, and the prevalence of disrupted family structures generally accounted for more variation than did measures of poverty and income inequality (Sampson, 1983, 1985, 1987). Another study, of 57 neighborhoods in three cities, found a similar link between levels of violence and residential mobility in poor, but not in affluent, neighborhoods (Smith and Jarjoura, 1988:40, cited in Sampson and Lauritsen, Volume 3). High population density was associated not only with high rates of violence but also with high rates of poverty, teenagers in the population, single-parent households,
PERSPECTIVES ON VIOLENCE 134 and nonwhite populationsâall commonly found correlates of violence. The interactions among these factors make it difficult to estimate their independent effects using data on a cross section of neighborhoods at a single point in time. However, in a study of changes in violence rates from 1970 to 1980 in 277 Baltimore neighborhoods, Taylor and Covington (1988) found sharp increases not only in areas that experienced increasing concentrations of people in poverty but also in neighborhoods in which gentrification improved housing characteristics and altered family composition. The suggestion that neighborhood change per se is associated with increased violence levels is consistent with the theory that disturbances in social organization of any kind are a factor in increasing neighborhood violence levels. In a study of changes in homicide rates in metropolitan areas and states from 1960 to 1980, Land et al. (1990) report similar findings for poverty, income inequality, and family disorganization. Although it is difficult to isolate specifically how these factors affect individual behavior, including violence, the National Research Council's Committee on National Urban Policy marshaled impressive evidence that in recent decades poverty and its correlates have become increasingly concentrated in inner-city neighborhoods (Lynn and McGeary, 1990; McGeary and Lynn, 1988). This concentration has occurred through changes at the regional levelâ for example, migration from the South to certain northern cities, especially New York, Chicago, Philadelphia, Newark, and Detroitâand through changes in the structure of economic and other opportunities across neighborhoods within large- and medium-sized metropolitan areas (Wilson, 1991:2; Jargowsky and Bane, 1990:44-48). Social and Economic Structure and Organization Several studies demonstrate the special vulnerability of urban minorities to fundamental economic changes in some cities and to population changes within specific communities (Wilson, 1987; Hagedorn, 1988; Sullivan, 1989; Anderson, 1990). Primary among these, Wilson argues, are "structural economic changes, such as the shift from goods-producing to service-producing industries, the increasing polarization of the labor market into low-wage and high-wage sectors, technological innovations, and the relocation of manufacturing industries out of the central cities" (1987:39). The "concentration effects" of ghetto poverty have been a central focus of William Julius Wilson's (1987) work. One effect of
PERSPECTIVES ON VIOLENCE 135 the increased concentration, Wilson argues, has been the social isolation of inner cities from mainstream institutions. As the black middle class grew during the 1980s, many middle- and working-class black families found it possible to move out of the ghetto into areas with more desirable housing and other neighborhood amenities. In turn, their leaving removed an important "social buffer" that might have deflected "the full impact of the â¦ prolonged and increasing joblessness that plagued inner-city neighborhoods in the 1970s and early 1980s" (Wilson, 1990:75). Community residents left behind in areas of extreme poverty have experienced high levels of family disruption, poor access to prenatal and child care, and low infant birthweightâfactors that earlier in this chapter we described as developmental precursors of aggressive and violent behavior. Recent research confirms their association with violence. These same communities and families also experience high levels of family disruption via divorce, desertion, and female-headed families; the breakdown of community networks of informal social control (Wallace and Wallace, 1990; Sampson, 1985; Messner and Tardiff, 1986; Smith and Jarjoura, 1988; all reviewed in Sampson and Lauritsen, Volume 3); and changes in routine activities that increase their risks of violent victimization. The breakdown of social organization is exemplified in Furstenberg's (1990) comparison of families in two poor Philadelphia communities. One is a black housing project in which parents characteristically practiced an individualistic style of parental management. Mistrust among neighbors was high, and "mothers complained of their inability to distinguish neighborhood youth from strangers, and they bemoaned the alienated sense in which local parents ignored the open misbehavior of youth in the community." Despite enormous expenditures of time and energy devoted to their children, "parents seemed unable to band together to solve common problems." The other community was white and poor, but was characterized by strong social networks among families, âeven those with poor parenting skills" (p. 24). Here, the "high degree of observability keeps youth in check. The task of parents inside the home is reinforced by the support rendered by other parents outside the home" (p. 25). Sampson concludes that parenting is much more than the "individualistic process that contemporary society makes it out to be" (p. 26). The link between levels of social organization and violence is also illustrated in a comparison of military and nonmilitary communities. For black males, the homicide death rate for soldiers
PERSPECTIVES ON VIOLENCE 136 was only 9 percent of the civilian rate for the same age category; the difference between military and nonmilitary whites was somewhat smaller (Rothberg et al., 1990). While this striking difference in the risk of homicide could be due to soldiers' higher levels of education, the researchers who conducted the study emphasize the control of violence by community regulation of lifestyle, noting the importance of both formal and informal regulations in controlling people's behavior. The diminution of adults' ability to supervise children in their families and neighborhoods exemplifies a breakdown of what Coleman (1990) has called social capitalâthe quality of intergenerational relationships. Research is needed to test the lack of social capital as a causal factor in violence. Community Culture Poverty, high turnover in population, and other features of socially disorganized communities affect community culture by impeding communication, erecting barriers to common values, and making effective collective action to achieve such values virtually impossible (Skogan, 1990). Under such circumstances in an isolated community, widely accepted cultural values become "unviable," and an oppositional culture may begin to shape what are considered to be appropriate standards of behavior (Anderson, 1978). In Wilson's words: "As the basic institutions declined, the social organization of inner-city neighborhoods (sense of community, positive neighborhood identification, and explicit norms and sanctions against aberrant behavior) likewise declined. This process magnified the effects of living in highly concentrated urban poverty areasâeffects that are manifested in ghetto-specific culture and behavior" (1987:138). A vivid picture is provided by Elijah Anderson's ethnographic study of adjoining community areas in Philadelphia: Northton, a black community formerly of mixed socioeconomic status, now increasingly occupied by the ghetto poor, and The Village, a gentrifying mixed-race community. Anderson (1991) analyzed the complex oppositional cultureâwidely shared in some poor black communitiesâthat has resulted from the unique historical circumstances, structural economic changes, and institutional failures experienced by the ghetto poor he has observed. The street culture requires skills utterly unlike those required in the service economy. While workers employed in the old manufacturing economy could act tough and use coarse language, these behaviors limit opportunities for employment in the service economy.
PERSPECTIVES ON VIOLENCE 137 Drugs exacerbate the problems. "The drug economy is in many ways a parallel, or a parody, of the service economy (with an element of glamour thrown in). Rival drug dealers claim particular street corners to sell their wares. Corners are literally bought and sold, and they belong to the one who has the power to claim the space for the time being; such claims may result in territorial disputes that are sometimes settled by violence. â¦ For many young men the drug economy is an employment agency superimposed on the existing gang network. Young men who 'grew up' in the gang, but now are without clear opportunities, easily become involved; they fit themselves into its structure, manning its drug houses and selling drugs on street corners. With the money from these 'jobs,' they support themselves and help their families, at times buying fancy items to impress others" (Anderson, 1990:244). The language of the street legitimizes and enhances involvement in the underground economy. Young people speak of crime as "going to work" and "getting paid" (see also Sullivan, 1989). Groups of young black males have been observed to chant âa beat, a beat, a beat" as they build themselves up to assault and robbery. In another example of social capital deterioration, older people, even "old heads" who once sought to guide the young along conventional, law-abiding lines, feel threatened, let down, and compromised, and many are ambivalent about the criminal behavior of the young, as the following field observation suggests (Anderson, 1990:243): Outside the barbershop on a corner in Northton, old heads â¦ steal a closer look at what "must be" a drug dealer. Self-conscious, they try to hide their interest, paying a kind of civil inattention to the show. Others, including a group of young boys nearby, have come over for a look at the "bad ride," but for them the driver is really the main attraction. He is a dealer. That is the way they make sense of the charismatic, close-cropped young black man literally "dripping with gold," dressed in his full urban regalia, including midnight blue Adidas sweat suit and new white sneakers. Rosetinted glasses thinly shade his eyes. In a show of style, he profiles; lightly stepping into his Lincoln Town Car, he exhibits his cool. He sinks into the richly upholstered front seat, leans on the armrest, and turns up the stereo. His performance is all the more remarkable because of his youth (about twenty-one), which contrasts sharply with his expensive possessions. Local people are intrigued by this figure. Ambivalent, they recoil at his image, but they envy him just the same. The old heads feel especially conflicted; most definitely they condemn his carâwhich "he did not
PERSPECTIVES ON VIOLENCE 138 have to work for." But secretly, some even admire him. He's "getting over," making it, though not in the right way, the way they were taught. Among many poor blacks in the ghetto, ambivalence easily becomes alienation. A commonly voiced theme is "the Plan," encapsulating a host of issues and images that negatively affect the ghetto poor, including unemployment, police brutality, drugs, AIDS, incarceration, gentrification, and competing ethnic group businesses (Anderson, 1990). The Plan is discussed as a conspiracy engaged in by the wider community to commit genocide on the black community. It is systematically perpetrated by white officials and caretakers and, perhaps inadvertently, by some better-off blacks. Black-on-black crime is viewed as particularly strong proof of the conspiracy, because so little is done to curb it and to protect ghetto citizens. Incidents of unpunished police brutality in the black community are also taken as evidence. Institutions that most affect the life chances of the ghetto poor are profoundly distrusted, especially law enforcement, the business community, schools, public welfare, and city authorities. Even those who do not espouse the genocide theory are affected by the siege mentality it engenders. They come to feel that, at the very least, the wider society intentionally disrespects blacks. Mutual suspicion and concern with respect pervade the ghetto poor community. Under such circumstances, social order becomes precarious. Because of concern for being "dissed" (disrespected), respect is easily violated. Because status problems are mixed with extreme resource limitations, people, especially young people, exaggerate the importance of symbols, often with life-threatening consequences. These consequences are exacerbated by the widespread belief that authorities view black life as cheap, hardly worth their attention. This view is reinforced when black-on-white crime receives more attention by authorities and by the media than does black-on-black crime. The result is that people feel thrown back on their own limited resources. They arm themselves, take offense, and resist in ways that contribute to the cycle of violence. Conclusions Structural characteristics, such as the availability of legitimate and illegitimate job opportunities, and cultural characteristics, such as beliefs and values that are transmitted in families and neighborhoods, interact powerfully with individual and social factors
PERSPECTIVES ON VIOLENCE 139 to produce behavioral outcomes, among individuals and communities. The complexity of these relationships and their theoretical linkages defy simple characterization. We concur with the assessment of Jencks and Mayer (1990), who conclude: "We badly need better studies of neighborhoods' impact on teenage crime. We especially need studies that focus on the effect of very poor neighborhoods, including large public housing projects. We also need studies that follow families as they move in and out of very poor neighborhoods and examine how such moves affect teenagers' behavior" (p. 162). We also need studies of the effects of crime on communities, including large public housing projects, and of community effects of population turnover in neighborhoods. As should be clear from the preceding discussion, communities vary in their ability to control violence, in the opportunities they present for crime and violent confrontations, and in the extent to which their local cultures encourage or discourage violence. Youth Gangs And Violence: An Example Of Social Processes At Work Frederic M. Thrasher, the first social scientist to study gangs systematically, defined the gang as "an interstitial group originally formed spontaneously, and then integrated through conflict" (Thrasher, 1927:57). He discovered that gangs were to be found principally in the "interstitial areas," that is, the city's slums. From observation he concluded that the "ganging process is a continuous flux and flow, and there is little permanence in most of the groupsâ (p. 31). He went on to describe a variety of types of gangs, gang characteristics, and characteristics of individual members, and gang activities. Since Thrasher's classic work, many definitions of gangs have been suggested. Most of these include some reference to the delinquent or other forms of socially disruptive behavior that are the objects of concern (Klein, 1971; Klein and Maxson, 1989; Hagedorn, 1988). Several classifications of gangs have been suggested (Cloward and Ohlin, 1960; Miller, 1975, 1980; Taylor, 1990a). All definitions of gangs include (or imply) common elements: gangs are groups whose members meet together with some regularity, over time, on the basis of group-defined criteria of membership and group-defined organizational characteristics. For our purposes, then, gangs are self-determining, non-adult- sponsored groups that demonstrate continuity over time. When gang members become involved with others in the commission
PERSPECTIVES ON VIOLENCE 140 of violent acts, to the extent that these acts develop out of gang interaction or are related to gang membership, the behavior of such groups may be regarded as gang-related, even though they may involve few members of a gang. Defining gangs and gang-related in this way avoids the circularity of including in the definition the behavior that is to be explained. The definition excludes small groups that are organized only for the commission of crime, such as drug selling crews (Williams, 1989). How prevalent are youth gangs? Again the definitional issue makes the question difficult to answer. In his report to the Violence Commission, Klein (1969) estimated that in the Los Angeles census tract with the most gang members "only 6 percent of the 10- to 17-year-old age bracket were gang affiliated" (p. 1445). This estimate may be low by our more inclusive definition, and it may not reflect current levels of gang involvement in some cities. Gangs appear to have proliferated since Klein's research, particularly in certain areas of large cities, and to have spread to cities that previously had no "gang problem." Spergel et al. (1989) placed 1987 gang membership in Los Angeles city and county at about 70,000, and recent police estimates approach 100,000. We know that many gangs are extremely violent, yet violent behavior among even the most violent gangs is relatively rare. When violent episodes occur within or between gangs, or when gangs attack others or destroy property, some members typically do not participate. Moore's (1987) observation that levels of lethal violence among gangs in the 1950s and 1970s varied "from one clique to another within the same gang" (p. 218) seems to be equally applicable to contemporary gangs and to lesser forms of violence as well. Explaining Violent Gang Behavior The explanations of violent gang behavior can be organized roughly parallel to the individual-level and social-level explanations of violent behavior that we used earlier in this chapter. In the following pages, we first consider possible explanations in terms of individual gang members and then consider explanations in terms of the functions of violent behavior for the gang as a social unit. Finally, we then discuss explanations in terms of cultural and structural forces in the communities that gangs inhabit. Violent Behavior by Individual Gang Members In terms of individuals, there is no reason to believe that learning processes differ
PERSPECTIVES ON VIOLENCE 141 between members of gangs and others. However, violent behavior is more prevalent among gang members than others of the same age and gender, and it often differs in other ways. Klein and Maxson (1989) suggest that homicides by gang members are not as closely related to intimate personal relationships as other homicides are, and that gang membership is often implicated in both homicide perpetration and victimization. Moreover, fighting between gangs has been observed to occur over extended periods (Schwartz, 1987) and to victimize persons who are not gang members, often inadvertently (Jankowski, 1991). A study of high school students and dropouts in three cities found that higher percentages of gang members than others reported participation in delinquent behaviors (Fagan, 1990). And, despite the much higher involvement of males than females in law violation, "prevalence rates for female gang members exceed the rates for nongang males for all 12 behavior categories" (p. 12). Violent offenses among the behaviors studied included felony assault, minor assault, and robbery. Of equal importance is that the frequency with which the violent offenders committed such offenses differed little by either gender or gang membership. Together with the universally observed variability of violent behavior among gang members, this finding strongly suggests that similar processes produce frequent and persistent violent behavior among gang members and others alike. Gangs exert important influences on violent behavior and on the communities in which they live. However, more general processes, at the individual and social levels of explanation, must account for individual and community violenceâ especially its frequency and persistence. Field research with gangs is one key to understanding the nature of these processes (Hagedorn, 1988; Jankowski, 1991; Moore, 1987; Vigil, 1988; Padilla, 1990). As noted by Short (1991), Jankowski emphasizes the "defiant individualist character" of members of the gangs he observed, which includes such attributes as intense competitiveness, mistrust or wariness in dealings with others, self- reliance, and a survivalist orientation toward a hostile world (see also MacLeod, 1987; Hagedorn, 1988). One of the chief organizational problems for gangs is to reconcile the conflicting needs and demands of the defiantly individualist members of the group. The Functions of Violence for the Gang Studies of young people in many cities and different time periods, using a variety of research methods, find that fighting with age peers, particularly
PERSPECTIVES ON VIOLENCE 142 among young males, is more characteristic of individuals and groups with lower socioeconomic status than among higher-status youth. Where gangs are most common, socialization into violence begins early in life. And intragang fighting is more common than is either intergang fighting or violent behavior directed outside the gang (Jankowski, 1991; Miller et al., 1961; Hagedorn, 1988). Sullivan (1989) found that fighting with age peers, often with serious consequences, occurred at an early age among white, Hispanic, and black youth gangs in Brooklyn. The experience gained in this way often was later applied to the "systematic pursuit of income" (p. 109). Most fighting among boys in their early and middle teens is about status (MacLeod, 1987; Horowitz, 1983; Short and Strodtbeck, 1965). For some, however, the scarcity of resources, the symbolic significance of some types of property, and the lack of access to legitimate means of acquiring it translate into violence at an early age. Fighting within and between gangs, as well as attacks on others, often serves group purposes, such as demonstrating personal qualities that are highly valued by the gang or reinforcing group solidarity by disciplining individual members (Jankowski, 1991). Violent behavior by individual gang members is also heavily influenced by group values and the perceived requirements of group membership or status within the group. Gang conflict often occurs when a gang believes that its status or reputation, its turf, or its resources (e.g., its share of a drug market) are threatened by another gang. A member who feels his status within the gang threatened may react aggressively. The response may be direct, as in an attack on a threatening person, or indirect, as when a gang leader acts to reestablish his position by acting violently toward someone outside the gang in response to a perceived threat to his leadership (Short, 1990, 1991). The interpersonal and group dynamics of such behavior are not well understood. A small research literature suggests that it is neither irrational nor entirely predictable solely on the basis of individual characteristics. Group processes such as those noted above are implicated in much violent behavior by gang members (Jankowski, 1991; Short, 1990; Farrington et al., 1981; Short and Strodtbeck, 1965). At this level of explanation, interpersonal relationships and group processes often interact to produce threats, insults, and conflicts that escalate to produce violent exchanges, even to the extreme of "victim- precipitated" homicides (Wolfgang, 1957; Schwartz, 1987).
PERSPECTIVES ON VIOLENCE 143 Community Influences on Gang Violence Social forces that influence young people range from their families and communities to the global economy. Local neighborhoods and communities are especially important as the social settings that are most immediately experienced (Schwartz, 1987; Reiss and Tonry, 1986). To gang members, even the meaning of crime is shaped both by the neighborhood in which they spend their time almost entirely unsupervised by adults, and by the consumerist culture promoted in the media (Sullivan, 1989:249). Status threats are likely to be played out with violence within the gang milieu. Opportunities for violent behavior are greater in this context, especially for young people who have been socialized into violence or who lack the social skills to achieve status through nonviolent means. Adding to this context competition over drug markets produces an extremely volatile mix, which often results in violence. Many of Thrasher's observations from the 1920s apply equally well to contemporary gangs. Most youth gangs are found in inner cities, in areas of poverty, physical deterioration, and institutional breakdown. Because racial and ethnic minorities tend to be concentrated in such areas, manyâbut not allâ gangs are comprised of minority youths. These gangs are therefore subject to the cultural and structural forces that affect minority communities. Cultural Influences Besides the cultures promoted in the media, some of the major influences on youth gangs in this country are immigration to the United States and the position of ethnic and racial minorities in this country and in local communities. Ethnic antagonisms are often a source of gang violence (Jankowski, 1991; Sullivan, 1989; Vigil, 1987; MacLeod, 1987; Moore et al., 1978; Short and Strodtbeck, 1965) and some ethnic traditions encourage violence (Chin, 1990; Vigil, 1988). The ethnic character of violence and of gangs reflects the ethnic composition of inner cities, as evidenced by the emphasis among Hispanic groups on honor (above references; Schwartz, 1987; Horowitz, 1983) and the violence associated with traditional tongs and triads of Chinese gangs (Chin, 1990). Other ethnic variations in gangs, and in violence, relate to the status of ethnic groups in U.S. society, local communities, and social settings as well as to ethnic traditions. Much non-Hispanic white gang violence continues to be directed against minority youth, especially blacks (Jankowski, 1991; Sullivan, 1989), and at times against Hispanics and others (Moore, 1987). Jankowski observes that in largely Irish Boston communities a tradition of adult social and athletic clubs is an important
PERSPECTIVES ON VIOLENCE 144 factor in community acceptance of youth gangs. And in many Chicano communities of East Los Angeles, street gangs have become "quasi- institutionalized" since the 1930s and 1940s (Moore et al., 1978; Moore, 1987, 1991; Vigil, 1987). Structural Influences Recent field studies of gangs report that many young men remain active in gangs rather than growing out of them, as was the case in the past. While marriage, family responsibilities, and jobs once were the normal route out of the gang, the scarcity of good, stable job opportunities now makes that progress elusive. Much recent research documents the decline in manufacturing jobs available in inner cities, the rise of a service economy that offers few opportunities for advancement, and an increasingly segmented labor force in which whites dominate higher-paying jobs while minorities are consigned to low-wage (often part-time) work, welfare, and the illegal economy. As noted earlier in this chapter, as more affluent and stable minority populations have moved out of poverty areas, the communities left behind have lost their most effective role models and institutional leaders (Anderson, 1990; Lemann, 1991). The effects of these developments on gangs and gang violence are not clear. Some field research reports that the presence of older members in gangs has led to more rational pursuit of economic gain or to their transformation into groups organized solely for criminal purposes. This is particularly the case when gangs become heavily involved in drug distribution, in which case violence is likely to become more instrumental in character (Taylor, 1990a, b; Chin, 1990; Williams, 1989). As noted above, the availability of attractive, conventional role models is reported to have declined. Many "old heads"ârespectable and respected middle- and working-class adults who made it a point to advise young people as to acceptable conductâhave left these communities. Those who remain often find themselves ignored, disparaged, or threatened. Younger, flashier, and (at least temporarily) successful drug dealers now appeal to even younger youth (Anderson, 1990). In Hispanic communities in East Los Angeles, "good" jobs that left the area as a result of economic shifts have been replaced by unstable, low-wage, and unsheltered jobs, much as has occurred in other cities. Immigrants from Mexico compete with gang members and others for these jobs. Older gang members who have not found stable conventional jobs, many of whom have been in and out of prison and whose family and community ties tend to be
PERSPECTIVES ON VIOLENCE 145 weak, hang around with younger gang members. The result, Moore reports, is that the gang is perpetuated as an agent of "street socialization" rather than becoming more integrated into conventional community life (Moore, 1991; see also Short, 1991). Conclusion The importance of local community cultures and structures of economic opportunity for understanding and controlling violent behavior can hardly be overestimated. Although important clues exist, the processes that result in individual violence, gang violence, and violence in large gatherings of mainly young people are a high priority for future research. Among the clues, the interaction of social factors and intergenerational relationships seems most promising. When communities are weakened by demographic and economic shifts that concentrate poverty and destabilize their institutions, conventional processes of socialization and control fail. When intergenerational relationships break down or are distorted by such developments, the likelihood that gangs will flourish and compete with one another, often with deadly consequences, is enhanced. We do not consider in this report a number of important questions about youth gangs, on which there is substantial research and controversy. Among these questions are how and why gangs form, why youths join gangs, how gang presence and membership affect youths' development, how gangs differ from other youth groups, what nonviolent activities gangs engage in, and how to suppress gang activity. Distinctions between traditional youth gangs and drug distribution organizations, and the roles of both in drug market violence, are discussed in Chapter 4. THE CIRCUMSTANCES OF VIOLENCE To this point in this chapter, we have discussed individual and social processes that may contribute to a person's potential for violent behavior. A promising approach to the understanding and control of violenceâa perspective with roots in both criminology and public healthâis to focus on the places where violence occurs. The incidence of violent events varies widely in spaceâby city, neighborhood, and specific address. For example, although 97.8 percent of all Minneapolis addresses generated zero robbery calls in 1986, 8 addresses generated more than 20 calls each (Sherman et al., 1989).
PERSPECTIVES ON VIOLENCE 146 Individual- and aggregate-level violence risk measures have different implications for violence reduction strategies. Once a place with high individual risk (i.e., many violent victimizations per person at risk or, ideally, per person- hour at risk) is identified, private citizens have an incentive to reduce their own risk by avoiding the place, and business proprietors in the place have an incentive to "harden" it with guard dogs, private security, or various physical changes (Cook, 1986). But because of high daily traffic, some places like shopping centers with relatively low individual risks may have high aggregate risks (i.e., they account for a disproportionate share of violent victimizations in an area per year). In such places, the social benefits of violence risk reduction may exceed the total of all private benefits, and government interventions to reduce risks (e.g., by increasing police patrols or by making the place physically less conducive to violence) are sometimes warranted. Planning intervention strategies oriented toward location begins with an understanding of the risk factors in places in the general community. We first consider what is known about how the probabilities and consequences of violent events in a place are related to the routine activities of people who inhabit it. We then consider risk factors for violence in various physical environments. We then consider violence in two special placesâprisons and schools. Routine Activities And High-Risk Encounters According to "lifestyle" and "routine activities" theories, people's daily activities affect violent victimization rates by determining how often potential victims come together with potential perpetrators in places where no one is likely to intervene (Hindelang et al., 1978; Cohen and Felson, 1979). The evidence from American data for routine activities theory is largely indirect; "nighttime activity outside the home" is the most specific behavior that is statistically associated with higher rates of violent victimization. Most of the supportive evidence concerns personal characteristics that are presumed to be associated with activities that increase the chances of victimization: being a teenager or young adult, male, single, unemployed, and low-incomeâthe last effect is especially strong for robbery and negligible for rape. These studies find blacks more likely than whites to be robbed, but no more likely to be assaulted. Analyses of Canadian and British surveys, which record more details about respondents' activities, find higher risks of violent
PERSPECTIVES ON VIOLENCE 147 victimization correlated specifically with spending more nights per week out alone, greater frequency of drinking in pubs, and use of public transportation. We have no way of knowing whether these are also risk factors in the United States, which differs significantly in terms of the use of private transportation, access to neighborhood bars and other sources of alcohol and drugs, and racial composition. Although these findings tempt one to suggest that limiting one's outside activities will also limit exposure to violent victimization, especially in bars or on public transportation alone at night, the evidence suggests that a more complex chain of events may limit the effectiveness of such a strategy in neighborhoods most vulnerable to violence. Some communities present higher risks than others because of the community-level characteristics discussed earlier (e.g., social disorganization, concentration of poverty) rather than any individual's characteristics or routine activities (Sampson and Wooldredge, 1987). Residents of high-risk neighborhoods do rearrange their routine activities because they fear victimization, according to hosts of studies reviewed by Cook (1986) and Warr (Volume 4), yet their efforts fail to offset the higher risk associated with their neighborhoods (Miethe and Meier, 1990). The evidence is too fragmented, and the crime types analyzed to date are too broad, to conclude that community context and individual precautions interact as described here to determine violent victimization risks. Further work is needed to answer the practical question of what specific activities people should avoid, in which locations and at what times, to reduce their risk of violent victimization. The Physical Environment Research has identified only a few physical characteristics of residential and public places as unambiguously associated with the risk of violence (Sampson and Lauritsen, Volume 3; Taylor and Gottfredson, 1986). Public areas near apartment units, located in geographic areas with high rates of divorce and population turnover, are particularly dangerous (Sampson and Wooldredge, 1987). Such areas concentrate large numbers of adolescents, who are unsupervised while their single parents are at work. Because of the high population turnover, potential monitors are unlikely to know these children, and potential victims tend to be unknown to their attackers. Apartment housing areas, especially those on city blocks that are large in area and population, with high population
PERSPECTIVES ON VIOLENCE 148 density, and near public housing projects and high schools are also associated with high risks of violence (Beasley and Antunes, 1974; Mladenka and Hill, 1976; Roncek, 1981; Roncek and Faggiani, 1986; Sampson, 1983; Schuerman and Kobrin, 1986; Smith and Jarjoura, 1988). One particularly dangerous place is the open-air drug market. Although there is no systematic geographic research on the spatial characteristics of drug markets (Harries, 1990), attributes of "desirable" drug markets have been identified (Garreau, 1988): (1) easy on-off: a narrow one-way access road for automobile traffic, down which dealers can monitor arriving customers or marked police vehicles without being "blind-sided"; (2) wide-openness: multiple escape routes (e.g., through hallways, breezeways, or woods) behind sellers on the street; and (3) barricades: obstacles to police cars and officers on foot weighed down by equipment, which are relatively surmountable by teenage boys. During the 1970s, under the rubric of "defensible space," prescriptions for changing the physical environment were offered to reduce crime and violence (Newman, 1973a,b, 1975). Recommendations included physical barriers to transform large housing projects into clusters of smaller ones, design innovations to facilitate surveillance, intermingling residential areas with safe areas, and use of garden-style rather than high-rise apartment designs. A demonstration project in Hartford, Connecticut, experimented with physical changes designed to create symbolic entrances to the neighborhood and to close off side streets to through traffic (Gardiner, 1978). Evaluations of these techniques indicate some success in reducing crime and especially fear of crime, at least in the short run (Newman and Franck, 1980, 1982; Fowler, 1981; Fowler and Mangione, 1981). But the desired effects failed to materialize in one multiethnic housing project (Merry, 1981a,b) and proved to be short-lived in the Hartford demonstration. The reason for the inconsistent and temporary effects appears to be that crime and violence arise from interactions between the social environment and the physical environment, which cannot be controlled entirely through manipulations of the physical environment. The communal feeling of territoriality and mutual protection, on which the defensible space concept rests, apparently failed to materialize in the multiethnic housing project because of lingering anonymity and mistrust across ethnic groups. In Hartford,
PERSPECTIVES ON VIOLENCE 149 citizens' participation in neighborhood patrols and maintenance of the symbolic entrances fell off after an initial reduction in the fear of crime. To intruders bent on crime, this deterioration may well have signaled a decline in local guardianship and concern and therefore in the risk of detection by responsible monitors. These findings suggest that architectural innovations help cohesive communities to make dangerous places less so, but are insufficient in themselves to create long-term safety. Architectural change is one of a number of surveillance techniques for crime prevention (Clarke, 1983). Another techniqueâreported to have limited success âis to increase traffic by responsible monitors through dangerous places. Another approach is called âtarget hardening," with such devices as alarms, dogs, locks, etc. Clarke concludes that, although these have been successful in reducing property crimes such as thefts and burglaries, they have no clear-cut effects on violence risks. A third type of prevention is called "environmental management" and includes techniques designed to make targets for instrumental violence less accessible or less attractive to would-be perpetrators. Exact-change requirements for passengers on New York City buses, for example, caused a substantial reduction in bus driver robberies (Chaiken et al., 1974) and airport screening procedures brought about a substantial reduction in skyjackings (Landes, 1978). The most elaborate approach to situational interventions comes from the public health perspective. One example of a public health framework for controlling intentional (i.e., violent) injuries appears in Table 3-1. The framework addresses (1) primary preventionâprevention of the violent event, (2) secondary preventionâreduction of the harm caused by the event, and (3) tertiary preventionârepair of the harm. It draws attention to strategies that involve not only altering the behavior of perpetrators and victims but modifying hazards in the environment, such as weapons, architectural barriers, and valuables in the possession of a vulnerable robbery target. Examples of primary prevention tactics include locating automatic teller machines in well-lit, high-traffic areas; limiting cash on hand in retail stores and posting signs to that effect; installing plexiglass shields for lone employees who must handle cash in public places, especially at night; and staggering school hours so that older children are in class while vulnerable younger children are walking to and from school. Secondary prevention tactics are
PERSPECTIVES ON VIOLENCE 150 TABLE 3-1 Public Health Injury Control Framework PRIMARY PREVENTION 1. Prevent creation of the hazard. â¢ Avoid placing automatic teller machines in high-risk locations. â¢ Restrict the sale of handguns. 2. Reduce the amount of hazard brought into being. â¢ Limit cash on hand at convenience stores to less than $30. â¢ Restrict the size of magazines in semiautomatic rifles. 3. Prevent the release of a hazard that already exists. â¢ Equip police officers with firearms that have "user-specific" safeties. â¢ Increase scene visibility for passerby or employee surveillance of high-risk locations. 4. Modify the rate of release of a hazard from its source. â¢ Ban magnum charges in handgun ammunition to decrease muzzle velocities and wounding potential. â¢ Ban automatic weapons and semiautomatic assault weapons. 5. Separate by time or space the hazard from the people to be protected. â¢ Restrict knives and firearms from high-risk locations through use of metal detectors and other forms of weapons screening. â¢ Release younger children from school 30 minutes before older children and stagger recess times. 6. Physically separate the hazard from people to be protected by barriers. â¢ Install bullet- and knife-proof shields in cabs. â¢ Install bullet-proof booths for clerks in all-night gas stations and high-risk convenience stores. SECONDARY PREVENTION 7. Modify surfaces and basic structures to minimize injury. â¢ Ban hollow point and explosive bullets to decrease injury severity. â¢ Substitute plastic containers for glass containers in bars and at large sporting events. 8. Make the people to be protected more resistant to damage. â¢ Equip all officers involved in law enforcement with bullet-proof vests. â¢ Equip special high-risk police units with combat-style head protection and body armor. TERTIARY PREVENTION 9. Begin to counter damage already done. â¢ Adopt the enhanced 911 system on a universal basis. â¢ Provide all paramedics and emergency medical technicians with special training in the management of penetrating injuries. 10. Stabilize, repair, and rehabilitate injured people. â¢ Support a national system of trauma centers. â¢ Promote a multidisciplinary approach to the treatment and rehabilitation of victims of violence, including psychological as well as physical care.
PERSPECTIVES ON VIOLENCE 151 aimed at reducing the lethality of accessible weapons through such means as requiring the use of plastic rather than glass beverage containers in crowds; reducing access to firearms; providing bullet-resistant clothing to police and private security officers in vulnerable circumstances; and reducing available firepower through restrictions on muzzle velocity and hollow-point ammunition. Tertiary prevention tactics include enhanced emergency telephone service (which automatically displays to the dispatcher the address of the calling telephone); training emergency medical technicians on penetrating injuries; building trauma centers; and providing psychological and physical rehabilitation to victims of violence. Convenience stores are a focus of private efforts to reduce violence. According to the National Institute on Occupational Safety and Health (NIOSH), retail trade is the highest-risk category for female occupational homicides (Bell, 1991). Most of them occur in the course of robberies, although customers' demands to violate age and time restrictions on beer and wine sales also trigger occasional high-risk encounters. Among the possible risk factors for occupational homicide identified by NIOSH (1992) are interaction with the public, the exchange of money, and evening employment as possible risk factors. Local passenger transportation and retail food sales are especially high-risk occupations. In Gainesville, Florida, over a five-year period, 96 percent of all convenience stores were reportedly robbed, compared with 36 percent of the fast food restaurants, 21 percent of the gas stations, and 16 percent of the liquor stores (Clifton, 1987). A number of risk factors for convenience store robbery have been suggested, for example: (1) location on a low-traffic street near limited commercial activity (Duffala, 1976); (2) a small parking lot, longer operating hours, not offering gasoline service, and being located near socially disorganized neighborhoods (D'Allessio and Stolzenberg, 1990); and (3) operating with only one clerk in the store (Jeffery et al., 1987). Interventions that have been implemented to prevent convenience store robberies include requirements to operate with two clerks, and a "robbery prevention package" adopted by the National Association of Convenience Stores. The package includes reducing the amount of cash available and posting signs to that effect; clearing the windows to make the cash register and safe visible to passersby; increasing lighting, inside and out; training employees in robbery prevention techniques; moving cash registers to more visible positions; and eliminating escape routes. As discussed in Chapter 7, evidence on the effectiveness of these interventions is inconclusive.
PERSPECTIVES ON VIOLENCE 152 Violence in Special Places Violence at a given level or severity is not equally problematic in all places. Social concern is greater in places where it impairs social functions that are especially important, or in places where the victims are legally required to be. Unfortunately, knowledge from research lags behind social concern in two such special places: prisons and schools. Prison Violence A fundamental question about prison violence is the extent to which inmates' violent behavior is imported from their lives in the community, learned in prison as a standard means of social interaction ("prisonization"), or triggered by prison conditions, such as frustration over the deprivation of accustomed privileges. Studies using institution-level data find some evidence that the deprivations associated with extraordinary crowding, long parole referral dates, and few visitors are associated with more assaults by inmates on one another and on staff. They also find some evidence of importation: prisons with greater numbers of inmates who behaved violently on the street experience higher levels of prison violence (Ellis et al., 1974; Gaes and McGuire, 1985). We could find no strong empirical evidence that prisonization is a common cause of interpersonal prison violence. However, several ethnographic studies suggest that incarcerated street gang members import gang activities, including illegal markets, into the prison setting and tailor them to the special conditions there (Jacobs, 1977; Fleisher, 1989). Illegal markets in prison, like those in the community, operate under codes of conduct (enforced by violence) in the absence of legal systems for protecting franchise territories, resolving disputes, enforcing agreements, and recruiting employees (Fong, 1987; Moore et al., 1978; Toch, 1977). Studies also suggest that both gang violence and sexual violence in prisons seem to maintain the racial and ethnic affinities and rivalries for social dominance that characterize the inmates' home communities (Lockwood, 1980). We discussed gangs earlier in this chapter; Chapter 4 has a further discussion of illegal markets. Social Control and Prison Violence In recent decades, two legal processes âthe proliferation of determinate and mandatory minimum sentences and litigation over prison reformâhave transformed the nature of social control mechanisms available to prison administrators.
PERSPECTIVES ON VIOLENCE 153 A vital policy question is whether these processes have aggravated or inhibited prison violence. In both cases, existing research provides no clear answers, and the ongoing nature of the legal processes both demands and facilitates studies of how they are affecting prison violence. One could speculate that the predictability of determinate sentences would reduce prison violence by reducing frustration caused by uncertainty over release dates. Alternatively, one could speculate that a determinate or mandatory minimum sentence, by ruling out the chance of reducing time served through good behavior, removes a restraint on violent behavior. In an analysis of California prison experience in the mid-1970s with new mandatory minimum sentences (Forst and Brady, 1983) the results were mixed: an increase in inmates' assaults on one another and on staff, with and without weapons, but a decrease in prison homicides. The rubric of prison reform covers a host of mechanisms available for social control of prisons that have been changed recently in many places by litigation, social activism, and new management structures (Barak-Glantz, 1981; Engel and Rothman, 1983; Flynn, 1976; Park, 1976; Rubel and Goldsmith, 1980; Toby, 1983a). Perhaps the most thoroughly studied reforms occurred in the Texas prison system, where the court decision in Ruiz v. Estelle (S.D. Texas, 1980) required changes to redress excessive crowding, excessive disciplinary force, the designation of selected inmates empowered to discipline other inmates, vague rules and arbitrary enforcement, inadequate health care, and inadequate fire and safety standards. The mandated reforms were implemented in 1983, and the two in-depth studies document a clear pattern of decreased violence by staff and increased violence by inmates during the following two years (Marquart and Crouch, 1985; and Ekland-Olson 1986). Inexperienced staff were hired to fulfill some requirements, residual inmate resentment from the prereform period continued, inmates' urges to test the system in the wake of their new rights could have been anticipated, and the abrupt realignment of rights and responsibilities necessarily placed inmates and staff in unfamiliar social interactions. Research on Texas prison experience since 1985 could help to distinguish permanent from temporary effects. A more encouraging view of the effects of innovation in correctional facilities is the experience with direct supervision in jails (Nelson, 1988). Direct supervision involves both design and management: replacement of standard cell rows or pods under remote
PERSPECTIVES ON VIOLENCE 154 surveillance with "living units" of approximately 50 cells clustered around a day room, with 24-hour management by officers located inside the unit. Direct supervision has usually been introduced as a means of personnel cost reduction rather than as a court-ordered reform; Nelson reports several examples of reduced inmate violence. Influences on Prison Violence Situational influences on prison violence include crowding, the facility's conduciveness to monitoring, and the nature of interpersonal interactions. Individual influences include inmates' psychopathology and their concept of personal space. Two distinct concepts of crowding are in use: (1) spatial density, the number of inmates per unit of area, and (2) social density, the number of inmates per partitioned unit. Increases in both spatial and social density (the latter measured by the use of dormitories) are associated with increases in inmate assaults on one another and on staff (Gaes, 1985). These measures have larger effects on crowding than do the proportion of young inmates or the rate of inmate turnover in the facilityâtwo measures that are often linked to increased violence. However, none of the studies that Gaes reviewed have both facility-level and individual-level measures of crowding, so it is not clear whether inmates who live in less crowded circumstances are less likely than others to be perpetrators or victims of violence. Also, as documented by Ekland-Olson (1986), any violence reduction effect from a decrease in crowding may be offset at least temporarily by a breakdown of social control, as in the Texas case. The location of a violent event can influence prison violenceâespecially the extent to which it can be monitored. One study finds, as expected, that assaults in prisons frequently occur in areas that are difficult to monitor (Atlas, 1983). However, many assaults also occur in well-supervised areas, such as detention and high-security areas. These findings are difficult to interpret: they may suggest that assaults are spontaneous, therefore not easily deterrable by monitoring. The rates may also be affected by variations in time spent in the well-monitored areas, or by the violence potentials of the populations assigned to the high-security areas. Although sexual assaults in prisons are widely assumed to be commonplace, we could find little published systematic research on the subject (Prentky, 1990). Several unpublished studies report highly variable victimization rates, depending on the prison sampled,
PERSPECTIVES ON VIOLENCE 155 the definition used, and the reference period. This work suggests no particular risk factors. Two psychological characteristics are specifically associated with violent behavior in prison. One is psychopathology; Toch (1989) notes cases of patterned violence, in which inmates regularly repeat a specific pattern of violent behavior. The other individual risk factor for fights and other violent behavior in prison is an exaggerated sense of social distanceâtheir desired physical distance from others, which may be repeatedly violated in densely populated prisons (Walkey and Gilmour, 1984). School Violence Although 5.3 percent of all 1990 nonfatal violent victimizations of persons aged 12 and over reportedly occurred in school buildings, and another 4.5 percent on school property (Bureau of Justice Statistics, 1992), relatively few studies of school violence have been carried out. Until recentlyâas violence in urban schools has emerged as an issue in its own rightâthe focus in schools was on disruptive behavior, a category that ranges from talking in class to vandalism to violent behavior. As a result, far less is known about school violence than about prison violence, much of that knowledge simply confirms common sense expectations, and most of it is based on old data. A special victimization survey sponsored by the National Institute of Education (NIE, 1978) reported far more school robberies than the National Crime Survey, a discrepancy that should be more fully explored and explained, since it could be due to errors in either survey. Violence rates in secondary schools are highest in school districts marked by higher crime rates and more street fighting gangsâan indication that the school reflects the community, and evidence that is consistent with the importation of violent behaviors into the school, by both students and intruders. According to an analysis of National Crime Survey data by McDermott (1979), the majority of violent victimizations of students and teachers were committed by strangers: more than 90 percent of reported rapes, 80 percent of robberies, and 60 percent of aggravated and simple assaults. While some of these could have been committed by fellow students who were previously unknown to the victims, some may also have been committed by intruders (Toby, 1983b). The NIE study also reports higher rates of student violence in schools in which students perceive signs of ineffective social control: undisciplined classrooms, lax or arbitrary enforcement of
PERSPECTIVES ON VIOLENCE 156 school rules, and a weak principal. It is not clear whether the ineffective control gives a kind of permission that encourages violent behavior, or whether high violence levels create fear among authorities that undermines their will to impose discipline. Despite expectations that mandatory desegregation efforts might increase school violence either because of student frustration or because cross- area busing would increase anonymity, the NIE survey found no strong effect. Finally, the NIE survey provides aggregate-level evidence that violence levels are related to students' attachments to the values that schools are intended to promulgate. In secondary schools, violence rates increased with the percentages of students who did not aspire to good grades, who did not view their curricula as relevant, and who did not believe that their school experience could positively influence their lives. A MULTICOMMUNITY LONGITUDINAL RESEARCH PROGRAM Earlier in this chapter, we surveyed knowledge about aggressive and violent behavior from three perspectives: psychosocial development, biomedical, and social. While each perspective has contributed findings with scientific and policy implications, these findings are pieces in a mosaic that is yet to be assembled. We cannot yet answer the question, How do psychosocial, biomedical, and social processes operate and interact to explain violence patterns in the United States today? More specifically, many risk factors for aggression were listed earlier in this chapter, yet we do not know either which ones actually cause aggression or how they are related to subsequent violent behavior. Promising causal theories (e.g., shy versus uninhibited temperaments, social learning of aggression, serotonin deficits, neurologically based cognitive and attention deficits, attachment theory) that are discussed more fully in Appendix A and Miczek et al. (Volume 2, Chapter 6) have yet to be compared against one another in a common framework. Knowledge of the specific developmental sequences that link perinatal and preschool events, childhood aggression, and adult violence is still only approximate, fragmentary, and segregated from community and ethnic contexts. Better understanding of these sequences could improve the timing of interventionsâperhaps in age ranges in which behavior is relatively malleable, or just before key developmental transitions that separate periods of relative behavioral stability.
PERSPECTIVES ON VIOLENCE 157 As discussed earlier, there is a need to resolve conflicting findings about the interactions between bullying, peer relations, aggression, and violence. Because developmental studies rarely examine community influences or compare sequences across communities or ethnic categories, we do not know the extents to which the well-documented continuity in aggression reflects continuity in individuals versus the environment. Childrearing practices may have different effects in different communities or cultures, and protective factors at the community and family levels compensate for adverse circumstances at either level. Even without this integrated perspective on violence, researchers and practitioners have designed interventions that succeeded in reducing aggressive or violent behaviorâby small amounts, in selected samples, and in certain communities. As we recommend in Chapter 8, systematic programs to refine and evaluate these interventions should be launched immediately. Over the long term, however, designing more comprehensive violence prevention strategies requires better answers to the basic question posed above. With a 12- to 15-year perspective in mind, therefore, we call for a multicommunity longitudinal research program to investigate the psychosocial, biomedical, and social processes that explain: (a) why some children exhibit patterns of aggressive behavior at early ages while most do not and (b) why only a minority of aggressive children go on to commit violent acts as adults. In principle, there has long been no reason why investigators, by combining community- and individual-level data, could not study how social structures and cultures interact with individuals' psychosocial development to influence their potentials for violent behavior. But such studies have rarely been carried out because of the substantial burdens they impose on data collection and on scholars' capacities to span competently the range of relevant theories and methods. Longitudinal studies pose methodological and human subject problems not ordinarily encountered in experiments or cross-section surveys. These will require special attention in their design and implementation. Methodological and measurement problems arise because of the length of time subjects are studied. Among these problems are selective attrition of subjects and locating subjects who move or change their names. Complex analytical issues arise in separating community context and change variables from individual-level variables. Likewise, human subject issues are more complex. Some of the
PERSPECTIVES ON VIOLENCE 158 measures, particularly those relating to biomedical measures, may be viewed as particularly intrusive and raise questions about the conditions for informed consent. Similarly, there are issues of whose consent is required. For younger cohorts, parental consent is required, but is the consent of only one parent required when parental information pertains to both and at what age is the consent of the child to be obtained? There likewise are special issues of protecting confidential information, especially those relating to subpoena or when information is secured on physical and sexual abuse of subjects when there are legal reporting requirements. Despite these difficulties, relevant longitudinal studies have occasionally been carried out with selected small populations; these are reviewed by Farrington, Tonry et al. (1991), and Appendix A. However, even taken together they fail to answer the basic question posed above for a variety of reasons. Many collected data too infrequently to describe event sequences in detail (Olweus, 1980; Magnusson and Bergman, 1988; McCord and McCord, 1956). Developmental data in several of the most important studies were collected several decades ago (e.g., McCord and McCord, 1956; Wolfgang et al., 1972; Huesmann and Eron, 1984) or in foreign countries (e.g., Moffitt and Silva, 1988; Farrington and West, 1990; Olweus, 1980; Kolvin et al., 1988; Magnusson and Bergman, 1988), so their findings may not apply to violence in today's urban American communities. Several recent longitudinal studies (Elliott et al., 1989; Huizinga et al., 1991; Loeber et al., 1991; and Thornberry et al., 1991) collected data at sufficiently close intervals in American communities; however, they do not focus on aggression and violence, and they do not cover the prenatal/birth- to-adult ages. Perhaps the greatest difference between the program we recommend and previous longitudinal studies is that our recommended program would provide an unparalleled opportunity to examine the relationships between biomedical variables and violent behavior. As explained earlier in this chapter, modern research using animal and human subjects suggests links between violent behavior and variables ranging from exposure to alcohol and drugs in the uterus, to serotonin deficits, to neuroendocrine functioning, and to head injuries, for example. It is unlikely that any of these links constitutes a causal relationship in isolation. Longitudinal designs offer the opportunity to detect individual differences in these variables early in life, to examine their stability and change over time, and to ascertain how variation in them is related to variation in behavior. For behaviors such as aggression and violence, which are fundamentally influenced by development, tracking
PERSPECTIVES ON VIOLENCE 159 individuals over time represents an essential scientific strategy for discovering such relationships. Once discovered, the relationships may suggest ways, within strict ethical guidelines, to enhance the effectiveness of humane, beneficial, and minimally intrusive violence prevention strategies. The logic behind the program we recommend follows from the discussions in this chapter of the development of aggression and violence, of relevant biological and psychosocial risk factors, and of promising preventive interventions. The program would involve studies in multiple communities that vary in their demographic and socioeconomic composition, their urban status, and their neighborhood and community organization (see this chapter for examples), to investigate how variations and change in these factors alter the probabilities of various developmental sequences. In each community, the proposed program calls for study of multiple cohorts from prenatal to adult, each selected to overrepresent high-risk categories. The major components are similar for each study: a comprehensive initial assessment of biological, temperamental, cognitive, and behavioral characteristics in members of each cohort; longitudinal follow-ups of âhigh-risk" and "normative" subsamples of each cohort, with at least annual data collection from multiple sources, to assess how the children's life courses and future behaviors are related to the characteristics and key life transitions measured; measures of community environments and changes in them; and randomized experiments to measure the effects of one or more promising preventive interventions on other high-risk and normative samples in each cohort. In 10 years, plus time for analysis of the data, this program would lay the groundwork for new approaches to violence prevention by enhancing our understanding of (1) biological and behavioral characteristics of infants that increase their risk of growing up to commit violent acts; (2) protective factorsâ conditions, events, and processes in families and communities that prevent the development of high potentials for violent behavior in high-risk individuals; and (3) preventive interventions grounded in knowledge of protective factors. The three components are described in the following pages. The Initial Assessment In each study the initial assessment would collect a substantial amount of data, which could be used not only for analysis but also, if feasible, to subdivide each sample in each community into
PERSPECTIVES ON VIOLENCE 160 high-risk and normative subsamples. The specific criteria for defining the subsamples of cohort subjects should be developed as part of study design but should include behavioral indicators of temperament and cognition, and biomedical indicators. To distinguish between "inhibited" and "uninhibited" temperament, batteries of behavioral measures can be applied as early as the age of 4 months. In the behavioral battery, the infant could be exposed to unfamiliar visual, auditory, and olfactory stimuli, and various changes are taken as measures of autonomic responsivity. Levels of motor activity and irritability would be quantified from videotapes. This battery would permit the detection of uninhibited infants who showed relatively low levels of motor activity and irritability in response to cognitive testing. In later cohorts, these measures could be augmented with parent, teacher, peer, and significant other ratings of aggressive behavior. Cognitive capacity can be measured by about the age of 4 years, using tests of the knowledge of letters and of numbers, and short-term memory for words. The biomedical measures should be carefully selected according to the principles described earlier in this chapter. Under these principles, one category of data would include items that can be collected in simple interviewsâfamily drug and alcohol use and aggressive behavior, alcohol and drug exposure in utero, exposure to heavy metals, and some head injuries, for example. These variables appear to be linked biomedically to the potential for violent behavior and, under strict confidentiality, the data can be collected without burden on study subjects. Not only could analyses of these simple biomedical measures provide promising leads, but also their inclusion in statistical models would improve estimates of how psychosocial influences and interventions affect potentials for violent behavior. In a second category, certain noninvasive, low-cost, and reasonably accurate physiological measuresâwhile not specific enough to identify the brain activity that links a violent act to a specific stimulusâcan be used in a longitudinal design to establish predictive relationships. For example, saliva measures of gonadal and adrenal hormones offer reasonably accurate indicators of endocrine activity. In a longitudinal design, saliva measures could be used to trace the onset and progression of puberty. When linked to longitudinal data on peer interactions and behavior, the information could be used to estimate predictive relationships for violent behavior. Once established, such relationships might be used to identify categories of youth for whom social skills training
PERSPECTIVES ON VIOLENCE 161 or some other intervention might be especially effective in reducing chances of violent behavior. Similar cases can be made for measuring evoked brain potentials, heart rate and heart rate variability, and galvanic skin responses because noninvasive procedures exist. While these are not very specific probes into cortical and autonomic system functioning, they place negligible risks and burdens on subjects. They can contribute to the testing of hypotheses about the arousability of the autonomic nervous system, lateralization of brain functions, and the presence of seizure activity. As with endocrine activity, analyses of repeated measures linked to records of behavior over time may reveal useful predictive relationships. A third category of biomedical measurement techniques imposes nonnegligible burdens or risks on subjects. Among these, spinal taps and brain imaging techniques at the current state of image resolution measure brain activity too imprecisely and are too nonspecific to violent behavior, to warrant their inclusion in our proposed program. In contrast, pharmacological or endocrinological challenge tests offer the prospect of information about sensory thresholds and physical mobilization in stressful situations. Because neurological markers for violence potentials are likely to exist in such responses rather than in resting levels of chemicals or neurological activity, the challenge test methodology should be considered for inclusion in the program we recommend. However, to guard against problematic, potentially irreversible, side effects in the study subjects, only challenge tests that have been carefully explored and calibrated should be considered in this program. Among those that can be safely conducted, only those that test specific hypotheses about biomedical causes (e.g., a serotonin deficit, a neurologically based temperament) or that promise new preventive or therapeutic interventions should be included. Longitudinal Follow-up In each cohort, the longitudinal follow-up would involve annual contacts. Annual measurements in early cohorts should include the demographic characteristics, delinquent and criminal behavior, and drug and alcohol abuse of all household members; measures of the subjects' physical growth, minor physical anomalies, injuries, and illnesses; measures of subjects' development, including neuromotor performance, IQ, perception, communication skills, and temperament.
PERSPECTIVES ON VIOLENCE 162 The annual measures should also attempt to capture family interaction patterns, such as family degree of organization, stability, bonding between children and parents, stimulation by caregivers, parental discipline and conflict management, and measures of physical, social, religious, and moral functioning in the family. They should also record important life events; characteristics of the community, neighborhood, day-care facilities; and, after school entrance, behavioral observations of teachers and peers. And they should record behaviors including aggression and its precursors, as well as interactions with peers, family members, and other caregivers. For each of the preadolescent cohorts, the final assessment should emphasize outcome measures including parent, peer, and teacher reports of the child's aggressive behavior, laboratory assessments of the child's behavior with same-sex peers, and behavioral measures of autonomic responsivity. For all later cohorts, we recommend annual collection of self-reports and official records of subjects' violent, antisocial, delinquent, and criminal behavior; behavioral indicators of temperament (e.g., impulsivity, attention deficit, risk taking); indicators of cognitive structure (e.g., hostile bias, perceptual accuracy, and realistic appraisal of behavioral consequences); indicators of cognitive functioning (e.g., intelligence, cognitive ability, academic attainment); indicators of commitment to the normative social structure (e.g., indicators of expected future achievements, of respect for community norms, of alienation from social norms); interactions with parents and peers; characteristics of family structure and family members' behavior; and work history. The study design calls for studying the development of each cohort in different kinds of community environments. Although cohort subjects will initially be selected in different community environments, those environments will change as the subjects develop, and subjects and their families will change community environment. Consequently, it will be essential to track both changes in community environments and the environmental history of each subject. Among the community characteristics to be investigated that have been linked to risk of delinquent, criminal, and other kinds of violent behavior are the peer networks of neighborhoods (e.g., violent gangs); the kinds of illegal activity in the community (e.g., illegal drug markets); the opportunities for legitimate work and occupational mobility; the value and normative structure of the community; and the family structure of communities (e.g., proportion of children in single- parent households).
PERSPECTIVES ON VIOLENCE 163 Analysis should focus on separating community- from individual-level effects. Analyses of data from all cohorts would permit much more precise documentation than is currently available of how temperament affects developmental processes, and of how developmental processes and life events interact to determine the risk of aggressive behavior at all ages to adult status. In addition, the analyses would identify protective factors that prevent aggressive children from developing into violent adults. Randomized Experiments The randomized-experimental component of the proposed research is intended to compare preventive interventions such as the three categories mentioned earlier: prevention of relevant brain dysfunctions, cognitive- behavioral interventions, and prevention of early-grade school failure. We have in mind comparative assessments of intervention "packages," designed to answer two kinds of questions. First, how do the interventions in each package compare in their effectiveness at preventing high-risk children from developing aggressive or violent behaviors? Second, do the interventions provide other benefits to children, including those who are not classified as at high risk for violence? The latter question is important because it usually becomes necessaryâboth to broaden political support and to avoid stigmatizing high-risk childrenâto provide an intervention program to all children in a geographic area rather than to only a subset. To answer both questions, a classification rule would be needed to classify study subjects as high-risk or normative, but the interventions would be tested on both groups. The study should also be designed to measure the effects of community- level interventions both on the cohort samples and on community violence levels. Possible interventions include parent education programs beginning at the prenatal stage, anti-violence campaigns by religious and community-based organizations, and efforts to limit exposure to violence in the media. Outcome measures for evaluating the individual-level interventions should include, of course, measures of aggressive and violent behavior as reported by the study subjects, parents, peers, teachers, social workers, and, for the early cohort, juvenile justice authorities. Also, because demonstrating benefits beyond violence control will hasten public acceptance of programs, outcome measures should include other outcomes where specifically relevant
PERSPECTIVES ON VIOLENCE 164 to the intervention: reading and arithmetic ability, short-term memory, retention in school, educational achievement, television viewing habits, and substance abuse and sexual activity during adolescence. NOTES 1 In a survey of 1,700 Rhode Island boys and girls in the seventh to ninth grades (Kikuchi, 1988), 20 percent answered that a male on a date had a right to sexual intercourse against the women's consent if "he spent a lot of money on her." Another study finds that about 55 percent of college males in a sample reported "some" or "higherâ likelihood of raping someone in the future (Malamuth, 1981). Another reports that a majority of a sample of 598 men believed that rape reports usually result from a women's concealing a pregnancy or seeking reprisal against a man and endorsed other expressions of tolerance toward rape (Burt, 1980). Another study found that scores on an Attitudes Toward Women Scale were similar for incarcerated rapists, incarcerated nonsexual offenders, and nonoffenders with low socioeconomic status. 2There is relative evolutionary constancy for acids, peptides, and steroids, their synthetic and metabolic enzymes, their receptors and messenger systems in mammalian nervous systems. However, there is considerable specialization in the functions of neurotransmitter systems across animal species. The most reliable information comes from studies comparing at least two species, one of which ought to be primate. 3 A study by Bain et al. (1987) found that pedophilic men manifest disturbances in the hypothalamic-pituitary axis and overrespond to injections of hormones that stimulate gonadal function. Bain et al. (1988) found abnormal levels of DHEAS in sexually aggressive men, but an ACTH challenge test was not significant. 4 We are aware of only one randomized experimental evaluation of a treatment program for violent sex offenders that involves follow-up outside an institutional setting. The Sex Offender Treatment and Evaluation Project (SOTEP), mandated and funded by the California state legislature, involved random assignment of volunteers to a relapse prevention program during the last 18-30 months of their prison terms. By 1989, 98 treatment and control subjects had been released to the community, and they are still being followed up. Since the intervention did not shorten participants' incarceration terms and provided treatment that was not otherwise available, the study did not increase danger to the community. Preliminary findings suggest that the relapse prevention program had some success (Miner et al., 1990). 5Using census tracts has limitations: they are not individual level and therefore say nothing about the risks of an individual, black or white, with higher or lower status than his neighbors. But information from census tracts is better than information from states or metropolitan areas because tracts are more homogeneous.
PERSPECTIVES ON VIOLENCE 165 References American Psychiatric Association 1987 Diagnostic and Statistical Manual of Mental Disorders, DSM-III-R (Third Edition, Revised). Washington, D.C.: American Psychiatric Association. Anderson, E. 1978 A Place on the Corner. Chicago: The University of Chicago Press. 1990 Streetwise: Race, Class, and Change in an Urban Community. Chicago: The University of Chicago Press. 1991 Alienation and Crime Among the Ghetto Poor, prepared for the Panel on the Understanding and Control of Violence (unpublished). Atlas, R. 1983 Crime site selection for assaults in four Florida prisons. Prison Journal 63:59-71. Bain, J., R. Langevin, R. Dickey, and M. Ben-Aron 1987 Sex hormones in murderers and assaulters. Behavioral Sciences & the Law 5:95-101. Bain, J., R. Langevin, R. Dickey, S. Hucker, and P. Wright 1988 Hormones in sexually aggressive men: I. Baseline values for eight hormones. II. The ACTH test. Annals of Sex Research 1:63-78. Baker, H.J., and R.J. Stoller 1968 Sexual psychopathology in the hypogonadal male. Archives of General Psychiatry 18:631-633. Baker, D., M. Telfer, C. Richardson, and G. Clark 1970 Chromosome errors in men with antisocial behavior: Comparison of selected men with Klinefelter's syndrome and XYY chromosome pattern. Journal of the American Medical Association 215(5):869-878. Barak-Glantz, I. 1981 Toward a conceptual scheme of prison management styles. Prison Journal 61:42-58. Baron, Larry, and Murray A. Straus 1989 Four Theories of Rape in American Society. New Haven, Conn.: Yale University Press. Bartlett, D.J. 1968 Chromosomes of male patients in a security prison. Nature 219:351-353. Beasley, R.W., and G. Antunes 1974 The etiology of urban crime: An ecological analysis. Criminology 11:439-461. Bell, Catherine A. 1991 Female homicides in United States workplaces, 1980-1985. American Journal of Public Health 81(6):729-732
PERSPECTIVES ON VIOLENCE 166 Bensing, R.C., and O. Schroeder 1960 Homicide in an Urban Community. Springfield, Ill.: Charles C Thomas. Berrueta-Clement, J.R., L.J. Schweinhart, W.S. Barnett, A.S. Epstein, and D.P. Weikart 1984 Changed Lives. Ypsilanti, Mich.: High/Scope. Blumstein, Alfred, Jacqueline Cohen, and David P. Farrington 1988 Longitudinal and criminal career research: Further clarifications. Criminology 26:57-74. Bohman, M., C.R. Cloninger, S. Sigvardsson, and A.L. von Knorring 1982 Predisposition to petty criminality in Swedish adoptees: I. Genetic and environmental heterogeneity. Archives of General Psychiatry 39:1233-1241. Browne, A., and D. Finkelhor 1986 Impact of child sexual abuse: A review of the research. Psychological Bulletin 99:66-77. Bullock, H.S. 1955 Urban homicide in theory and fact. Journal of Criminal Law, Criminology, and Police Science 45:565-575. Bureau of Justice Statistics 1992 Criminal Victimization in the United States, 1990. Washington, D.C.: U.S. Government Printing Office. Burt, M.R. 1980 Cultural myths and support for rape. Journal of Personality and Social Psychology 38:217-230. Cairns, Robert B., Beverley D. Cairns, Holly J. Neckerman, Scott D. Gest, and Jean-Louis GariÃ©py 1988 Social networks and aggressive behavior: Peer support or peer rejection? Developmental Psychology 24(6):815-823. Centerwall, Brandon S. 1984 Race, socioeconomic status, and domestic homicide, Atlanta, 1971-2. American Journal of Public Health 74(8):813-815. Chaiken, J.M., M.W. Lawless, and K.A. Stevenson 1974 The impact of police activity on subway crime. Urban Analysis 3:173-205. Check, J.V.P., and N.M. Malamuth 1983 Sex role stereotyping and reactions to depictions of stranger versus acquaintance rape. Journal of Personality and Social Psychology 45:344-356. Cherry, F. 1983 Gender roles and sexual violence. Pp. 245-260 in E.R. Allgeier and N.B. McCormick, eds., Changing Boundaries: Gender Roles and Sexual Behavior. Palo Alto, Calif.: Mayfield. Chin, K. 1990 Chinese Subculture and Criminality: Non-Traditional Crime Groups in America. Westport, Conn.: Greenwood.
PERSPECTIVES ON VIOLENCE 167 Christiansen, K.O. 1977 A review of studies of criminality among twins. In S.A. Mednick and K.O. Christiansen, eds., Biosocial Bases of Criminal Behavior. New York: Gardiner. Clarke, R.V. 1983 Situational crime prevention: Its theoretical basis and practical scope. Pp. 225-256 in M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research. Vol. 4. Chicago: University of Chicago Press. Clifton, W., Jr. 1987 Convenience Store Robberies in Gainesville, Florida: An Intervention Strategy by the Gainesville Police Department. Gainesville, Florida Police Department, photocopy. Cloninger, C.R., and Gottesman, I.I. 1987 Genetic and environmental factors in antisocial behavioral disorders. Pp. 92-109 in S.A. Mednick, T.E. Moffitt, and S.A. Stack, eds., The Causes of Crime: New Biological Approaches. New York: Cambridge University Press. Cloninger, C.R., K.O. Christiansen, T. Reich, and I.I. Gottesman 1978 Implications of sex differences in the prevalences of antisocial personality, alcoholism, and criminality for familial transmission. Archives of General Psychiatry 35:941-951. Cloward, R.A., and L.E. Ohlin 1960 Delinquency and Opportunity. Glencoe, Ill.: Free Press. Cohen, L., and M. Felson 1979 Social change and crime rate trends: A routine activities approach. American Sociological Review 44:588-607. Coleman, J.S. 1990 Foundations of Social Theory. Cambridge, Mass.: Harvard University Press. Comstock, G., and Paik, H. 1990 The Effects of Television Violence on Aggressive Behavior: A Meta-analysis. Unpublished report to the National Academy of Sciences Panel on the Understanding and Control of Violent Behavior, Washington, D.C. Cook, P.J. 1986 The demand and supply of criminal opportunities. In M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research. Vol. 7. Chicago: University of Chicago Press. D'Allessio, S., and L. Stolzenberg 1990 A crime of convenience: The environment and convenience store robbery. Environment and Behavior 22(2):255-271. Denno, Deborah W. 1990 Biology and Violence: From Birth to Adulthood. Cambridge: Cambridge University Press.
PERSPECTIVES ON VIOLENCE 168 Dietz, P.E. 1986 Mass, serial, and sensational homicides. Bulletin of the New York Academy of Medicine 62:477-491. Dollard, J., L.W. Doob, N.E. Miller, O.H. Mowrer, and R.R. Sears 1939 Frustration and Aggression. New Haven, Conn.: Yale University Press. Donnerstein, E., D. Linz, and S. Penrod 1987 The Question of Pornography. New York: Free Press. Duffala, D.C. 1976 Convenience stores, armed robbery, and the physical environment. American Behavioral Scientist 20:227-246. Dworkin, A. 1979 Pornography: Men Possessing Women. New York: G.P. Putnam's Sons. Ekland-Olson, S. 1986 Crowding, social control, and prison violence: Evidence from the post-Ruiz years in Texas. Law & Society Review 20(3):389-421. Elliott, D.S., D. Huizinga, and S. Menard 1989 Multiple Problem Youth: Delinquency, Substance Use, and Mental Health Problems. New York: Springer-Verlag. Ellis, D. 1991 Monoamine oxidase and criminality: Identifying an apparent biological marker for antisocial behavior. Journal of Research in Crime and Delinquency 28:227-251. Ellis, D., H.G. Grasmick, and B. Gilman 1974 Violence in prisons: A sociological analysis. American Journal of Sociology 80:416-443. Engel, K., and S. Rothman 1983 Prison violence and the paradox of reform. Public Interest 70-73:91-105. Eron, L.D. 1990 Understanding aggression. Presidential address to the World Meeting of ISRA, Banff, Alberta, June 12. Eron, L.D., and L.R. Huesmann 1984 The relation of prosocial behavior to the development of aggression and psychopathology. Aggressive Behavior 10:201-212. Fagan, Jeffrey 1990 Social processes of drug use and delinquency among gang and non-gang youths. In C. Ronald Huff, ed., Gangs in America. Newbury Park, Calif.: Sage Publications. Farrington, D.P. 1989 Early predictors of adolescent aggression and adult violence. Violence and Victims 4:79-100. 1991 Childhood aggression and adult violence: Early precursors and later-life outcomes. Pp. 5-29 in D.J. Pepler and K.H. Rubin,
PERSPECTIVES ON VIOLENCE 169 eds., The Development and Treatment of Childhood Aggression. Hillsdale, N.J. : Erlbaum. Farrington, David P., Leonard Berkowitz, and Donald J. West 1981 Differences between individual and group fights. British Journal of Social Psychology 20:163-171. Farrington, David P., Lloyd E. Ohlin, and James Q. Wilson 1986 Understanding and Controlling Crime: Toward a New Research Strategy. New York: Springer-Verlag. Farrington, D.P., and D.J. West 1990 The Cambridge study in delinquent development: A long-term follow-up study of 411 London males. Pp. 115-138 in H.J. Kerner and G. Kaiser, eds., Criminality: Personality, Behavior, and Life History. Berlin: Springer-Verlag. Feingold, B.F. 1975 Hyperkinesis and learning disabilities linked to artificial food flavors and colors. American Journal of Nursing 75:797-803. Felson, R.B., and H.J. Steadman 1983 Situational factors in disputes leading to criminal violence. Criminology 21(1, February):59-74. Finkelhor, D. 1984 Child Sexual Abuse: New Theory and Research. New York: The Free Press. Fleisher, M. 1989 Warehousing Violence. Newbury Park, Calif.: Sage Publications. Flynn, E.E. 1976 The ecology of prison violence. In A.K. Cohen, G.F. Cole, and R.G. Bailey, eds., Prison Violence. Lexington, Mass.: Lexington Books. Fong, R.S. 1987 A Comparative Study of the Organizational Aspects of Two Texas Prison Gangs: Texas Syndicate and Mexican Mafia. Ph.D. dissertation. Forst, M.L., and J.M. Brady 1983 The effects of determinate sentencing on inmate misconduct in prison. The Prison Journal 63:101-113. Fowler, F.G. 1981 Evaluating a complex crime control experience. In L. Bickman, ed., Applied Social Psychology Annual. Vol. 2. Hillsdale, N.J.: Erlbaum. Fowler, F.G., and T.W. Mangione 1981 An Experimental Effort to Reduce Crime and Fear of Crime in an Urban Residential Neighborhood: Reevaluation of the Hartford Neighborhood Crime Prevention Program. Draft executive summary. Cambridge: Harvard/Massachusetts Institute of Technology, Center for Survey Research.
PERSPECTIVES ON VIOLENCE 170 Freedman, J.L. 1986 Television violence and aggression: A rejoinder. Psychological Bulletin 100:372-378. Friedrich-Cofer, L., and A.C. Huston 1986 Television violence and aggression: The debate continues. Psychological Bulletin 100:364-371. Furstenberg, Frank 1990 How Families Manage Risk and Opportunity in Dangerous Neighborhoods. Paper presented at the annual meeting of the American Sociological Association, Washington, D.C. Gaes, G.G. 1985 The effects of overcrowding in prison. Pp. 95-146 in M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research. Vol. 6. Chicago: University of Chicago Press. Gaes, G., and W.J. McGuire 1985 Prison violence: The contribution of crowding versus other determinants of prison assault rates. Journal of Research in Crime and Delinquency 22:41-65. Gardiner, R.A. 1978 Design for Safe Neighborhoods. Washington, D.C.: U.S. Government Printing Office. Garreau, J. 1988 The invisible hand guides D.C.'s visible menace: Street-corner drug trade provides a model of capitalism, say economists, police. Washington Post, December 12, pp. A1, A14- A15. Gordon, Robert A. 1967 Issues in the ecological study of delinquency. American Sociological Review 32(6):927-944. Gottfredson, Michael, and Travis Hirschi 1987 The methodological adequacy of longitudinal research on crime. Criminology 25:581-614. Gottfredson, S.D., and R.B. Taylor 1986 Person-environment interactions in the prediction of recidivism. In R. Sampson and J. Byrne, eds., The Social Ecology of Crime. New York: Springer-Verlag. Guerra, N. 1990 Social cognitive approaches to the prevention of antisocial behavior in children. Paper presented at the Workshop on Hostility and Sociability, Warsaw, Poland (September). Hagedorn, J. 1988 People and Folks: Gangs, Crime, and the Underclass in a Rustbelt City. Chicago: Lakeview Press. Hanson, R.K., and S. Slater 1988 Sexual victimization in the history of sexual abusers: A review. Annals of Sex Research 1:485-500. Hare, R.D. 1978 Electrodermal and cardiovascular correlates of psychopathy. In
PERSPECTIVES ON VIOLENCE 171 R.D. Hare and D. Schalling, eds., Psychopathic Behavior: Approaches to Research. New York: John Wiley & Sons. Harries, K.D. 1990 Serious Violence: Patterns of Homicide and Assault in America. Springfield, Ill.: Charles C Thomas. Hawkins, J.D., E. Von Cleve, and R.F. Catalano 1991 Reducing early childhood aggression. Journal of the American Academy of Child and Adolescent Psychiatry. Heath, L., L.B. Bresolin, and R.C. Rinaldi 1989 Effects of media violence on children. Archives of General Psychiatry 46:376-379. Hindelang, M., M. Gottfredson, and J. Garofalo 1978 Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, Mass.: Ballinger. Horowitz, Ruth 1983 Honor and the American Dream. New Brunswick, N.J.: Rutgers University Press. Huesmann, L.R., and L.D. Eron 1984 Cognitive processes and the persistence of aggressive behavior. Aggressive Behavior 10:243-251. 1986 Television and the Aggressive Child. New York: Erlbaum. Huesmann, L.R., and L.S. Miller in press Long term effects of repeated exposure to media violence in childhood. In G. Comstock, ed., Public Communication and Behavior. Vol. 3. Orlando, Fla.: Academic Press. Huesmann, L.R., L.D. Eron, R. Klein, P. Brice, and P. Fischer 1983 Mitigating the imitation of aggressive behaviors by changing children's attitudes about media violence. Journal of Personality and Social Psychology 44:899-910. Huizinga, D., F. Esbensen, and A.W. Weiher 1991 Are there multiple paths to delinquency? Denver youth survey. Journal of Criminal Law and Criminology 82(1):83-118. Institute of Medicine 1985 Preventing Low Birthweight. Washington, D.C.: National Academy Press. Jacobs, James B. 1977 Stateville: The Penitentiary in Mass Society. Chicago: University of Chicago Press. Jankowski, Martin Sanchez 1991 Islands in the Street: Gangs and American Urban Society. Berkeley, Calif.: University of California Press. Jargowsky, Paul A., and Mary Jo Bane 1990 Ghetto poverty: Basic questions. Pp. 16-67 in Laurence E. Lynn, Jr., and Michael G.H. McGeary, eds., Inner-City Poverty in the United States. Washington, D.C.: National Academy Press. Jeffery, C.R., R.H. Hunter, and J. Griswold 1987 Crime analysis, computers, and convenience store robberies.
PERSPECTIVES ON VIOLENCE 172 Appendix D to Wayland Clifton, Jr., Convenience store robberies in Gainesville, Florida: An intervention strategy by Gainesville police department. Paper presented at the 1987 meeting of the American Society of Criminology, Montreal, November. Jencks, C., and S.E. Mayer 1990 The social consequences of growing up in a poor neighborhood. Pp. 111-186 in L.E. Lynn, Jr., and M.G.H. McGeary, eds., Inner-City Poverty in the United States. Washington, D.C.: National Academy Press. Jones, M.B., and D.R. Offord 1989 Reduction of antisocial behavior in poor children by non-school skill-development. Journal of Child Psychology and Psychiatry 30:737-750. Kagan, J. 1989 Temperamental contributions to social behavior. American Psychologist 44:668-674. Kaufman, J., and E. Zigler 1987 Do abused children become abusive parents? American Journal of Orthopsychiatry 57:186-192. Kazdin, A.E. 1985 Treatment of Antisocial Behavior in Children and Adolescents. Homewood, Ill.: Dorsey Press. Kessler, S., and R. Moos 1970 The XYY karyotype and criminality: A review. Journal of Psychiatric Research 17:29-34. Kikuchi, J.J. 1988 Rhode Island develops successful intervention program for adolescents. NCASA News (Fall):26-27 Klein, Malcolm 1969 Violence in American juvenile gangs. Pp. 1425-1460 in Donald J. Mulvihill and Melvin M. Tumin, eds., Crimes of Violence. Vol. 13. Staff report submitted to the National Commission on the Causes and Prevention of Violence. Washington, D.C.: U.S. Government Printing Office. 1971 Street Gangs and Street Workers. Englewood Cliffs, N.J.: Prentice-Hall. Klein, Malcolm, and Cheryl Maxson 1989 Street gang violence. Pp. 198-234 in Neil Alan Weiner and Marvin E. Wolfgang, eds., Violent Crime, Violent Criminals. Newbury Park, Calif.: Sage Publications. Kolarsky, A., K. Freund, J. Machek, and O. Polak 1967 Male sexual deviation: Association with early temporal lobe damage. Archives of General Psychiatry 17:735-743. Kolvin, I., E.J.W. Miller, M. Fleeting, and P.A. Kolvin 1988 Social and parenting factors affecting criminal-offence rates: Findings from the Newcastle thousand family study (1947-1980). British Journal of Psychiatry 152:80-90.
PERSPECTIVES ON VIOLENCE 173 Land, Kenneth, P. McCall, and L. Cohen 1990 Structural co-variates of homicide rates: Are there any invariances across time and space? American Journal of Sociology 95:922-963. Landes, W.M. 1978 An economic study of U.S. aircraft hijacking, 1961-1976. Journal of Law and Economics 21:1-32. Langevin, R. 1990 Sexual anomalies and the brain. Pp. 103-113 in W.L. Marshall, D.R. Laws, and H.E. Barbaree, eds., Handbook of Sexual Assault: Issues, Theories, and Treatment of the Offender. New York: Plenum Press. Langevin, R., J. Bain, M.H. Ben-Aron, R. Coulthard, D. Day, L. Handy, G. Heasman, S.J. Hucker, J.E. Purins, V. Roper, A.E. Russon, C.D. Webster, and G. Wortzman 1985 Sexual aggression: Constructing a predictive equation. A controlled pilot study. Pp. 39-76 in R. Langevin, ed., Erotic Preference, Gender Identity, and Aggression in Men: New Research Studies . Hillsdale, N.J.: Lawrence Erlbaum. Lemann, Nicholas 1991 The Promised Land: The Great Black Migration and How It Changed America. New York: Alfred A. Knopf. Lewis, D.O., J.H. Pincus, B. Bard, and E. Richardson 1988 Neuropsychiatric, psycho-educational, and family characteristics of 14 juveniles condemned to death in the United States. American Journal of Psychiatry 145:584-589. Linnoila, M., J. DeJong, and M. Virkkunen 1989 Family history of alcoholism in violent offenders and impulsive fire setters. Archives of General Psychiatry 46:613-616. Lockwood, D. 1980 Prison Sexual Violence. New York: Elsevier. Loeber, R., M. Stouthamer-Loeber, W. Van Kammen, and D.P. Farrington 1991 Initiation, escalation and desistance in juvenile offending and their correlates. Pittsburgh youth study. Journal of Criminal Law and Criminology 82(1):36-82. Lowry, Philip W., Susan Hassig, Robert Gunn, and Joyce Mathison 1988 Homicide victims in New Orleans: Recent trends. American Journal of Epidemiology 128(5):1130-1136. Luckenbill, D.F. 1977 Criminal homicide as a situated transaction. Criminal Homicide (?):176-186. Lynn, Laurence E., Jr., and Michael G.H. McGeary, eds. 1990 Inner-City Poverty in the United States. Washington, D.C.: National Academy Press. MacLeod, Jay 1987 Ain't No Makin' It: Leveled Aspirations in a Low-Income Neighborhood. Boulder, Colo.: Westview Press.
PERSPECTIVES ON VIOLENCE 174 Magnusson, D., and L. Bergman 1988 Individual and variable-based approaches for longitudinal research on early risk factors. In M. Rutter, ed., Studies of Psychosocial Risk: The Power of Longitudinal Data. Cambridge: Cambridge University Press. Malamuth, N.M. 1981 Rape proclivity among males. Journal of Social Issues 37:138-157. Marquart, J.W., and B.M. Crouch 1985 Judicial reform and prisoner control: The impact of Ruiz v. Estelle on a Texas penitentiary. Law & Society Review 19(4):557-586. McCord, W. and J. McCord 1956 Psychopathy and Delinquency. New York: Grune and Stratton. McDermott, M.J. 1979 Criminal Victimization in Urban Schools. Albany, N.Y.: Criminal Justice Research Center. (Analytic Report SD-VAD-8 1979.) McGeary, Michael G.H., and Laurence E. Lynn, Jr., eds. 1988 Urban Change and Poverty. Washington, D.C.: National Academy Press. Mednick, S.A., W.F. Gabrielli, and B. Hutchings 1984 Genetic influences in criminal convictions: Evidence from an adoption cohort. Science 224:891-894. Megargee, E.I., M.J. Bohn, and J.L. Carbonell 1988 A Cross-Validation and Test of the Generality of the MMPI-Based Offender Classification System. Final report of National Institute of Justice grant no. 1596001138A1, Tallahassee, Fla.: Florida State University. Merry, S.E. 1981a Defensible space undefended: Social factors in crime prevention through environmental design. Urban Affairs Quarterly 16:397-422. 1981b Urban Danger: Life in a Neighborhood of Strangers. Philadelphia, Pa.: Temple University Press. Messner, S., and K. Tardiff 1986 Economic inequality and levels of homicide: An analysis of urban neighborhoods. Criminology 24:297-318. Miethe, T., and R. Meier 1990 Opportunity, choice, and criminal victimization. Journal of Research in Crime and Delinquency 27:243-266. Miller, L. 1987 Neuropsychology of the aggressive psychopath: An integrative review. Aggressive Behavior 13:119-140. Miller, W.B. 1975 Violence by Youth Gangs and Youth Groups as a Crime Problem in Major American Cities. Washington, D.C.: U.S. Government Printing Office.
PERSPECTIVES ON VIOLENCE 175 1980 Gangs, groups, and serious youth crime. Pp. 115-138 in D. Shichor and D. Kelly, eds., Critical Issues in Juvenile Delinquency. Lexington, Mass.: D.C. Heath. Miller, Walter B., Hildred Geertz, and Henry S. G. Cutter 1961 Aggression in a boys' street-corner group. Psychiatry XXIV:283-298. Miner, M.H., J.K. Marques, D.M. Day, and C. Nelson 1990 Impact of relapse prevention in treating sex offenders: Preliminary findings. Annals of Sex Research 3:165-185. Mladenka, K., and K. Hill 1976 A reexamination of the etiology of urban crime. Criminology 13:491-506. Moffitt, T.E., and P.A. Silva 1988 Neuropsychological deficit and self-reported delinquency in an unselected birth cohort. Journal of the American Academy of Child and Adolescent Psychiatry 27:233-240. Moore, Joan W. 1987 Variations in violence among Hispanic gangs. Pp. 215-30 in Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez, eds., Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. 1991 Institutionalized Youth Gangs: Why White Fence and El Hoyo Maravilla Change So Slowly. Working Group on the Social and Economic Ecology of Crime and Drugs in Inner Cities, Social Science Research Council, unpublished. Moore, J.W., with Robert Garcia, Carlos Garcia, Luis Cerda, and Frank Valencia 1978 Homeboys: Gangs, Drugs, and Prisons in the Barrios of Los Angeles. Philadelphia, Pa.: Temple University Press. Morgan, R. 1980 Theory and practice: Pornography and rape. In L. Lederer and A. Rich, eds, Take Back the Night: Women on Pornography. New York: William Morrow and Company. Mungas, D. 1988 Psychometric correlates of episodic violent behaviour: A multidimensional neuropsychological approach. British Journal of Psychiatry 152:180-187. Murphy, W.D., M.R. Haynes, S.J. Stalgaitis, and B. Flanagan 1986 Differential sexual responding among four groups of sexual offenders against children. Journal of Psychopathology and Behavioral Assessment 8:339-353. Muscat, Joshua E. 1988 Characteristics of childhood homicide in Ohio, 1974-84. American Journal of Public Health 78(7):822-824. National Institute of Education 1978 Violent SchoolsâSafe Schools: The Safe School Study Report to the Congress. Washington, D.C.: U.S. Government Printing Office.
PERSPECTIVES ON VIOLENCE 176 National Institute for Occupational Safety and Health 1992 Homicide in U.S. workplaces: A strategy for prevention and research. Washington, D.C. Nelson, W.R. 1988 Cost savings in new generation jails: The direct supervision approach. National Institute of Justice Construction Bulletin (July). Newman, O. 1973a Architectural Design for Crime Prevention. Washington, D.C.: U.S. Government Printing Office. 1973b Defensible Space: Crime Prevention Through Urban Design. New York: Macmillan. 1975 Design Guidelines for Creating Defensible Space. Washington, D.C.: U.S. Government Printing Office. Newman, O., and K.A. Franck 1980 Factors Influencing Crime and Instability in Urban Housing Developments. Washington, D.C.: U.S. Government Printing Office. 1982 The effects of building size on personal crime and fear of crime. Population and Environment 5:203-220. Olweus, D. 1979 Stability of aggressive reaction patterns in males: A review. Psychological Bulletin 86:852-875. 1980 Familial and temperamental determinants of aggressive behavior in adolescent boys: A causal analysis. Developmental Psychology 16:644-660. 1984 Development of stable aggressive reaction patterns in males. Pp. 103-137 in R.J. Blanchard and D.C. Blanchard, eds., Advances in the Study of Aggression. Vol. I. Orlando, Fla.: Academic Press. 1991 Bully/victim problems among schoolchildren: Basic facts and effects of a school based intervention program. Pp. 411-448 in D.J. Pepler and K.H. Rubin, eds., The Development and Treatment of Childhood Aggression. Hillsdale, N.J.: Erlbaum. Owen, D.R. 1972 The 47, XYY male: A review. Psychological Bulletin 78(3):209-233. Padilla, Felix 1990 Going to Work: The Entrepreneurial Side of the Gang. Unpublished manuscript. Park, J.W.L. 1976 The organization of prison violence. In A.K. Cohen, G.F. Cole, and R.G. Bailey, eds., Prison Violence. Lexington, Mass.: Lexington Books. Pasqualini, R.Q., G. Vidal, and G.E. Bur 1957 Psychopathology of Klinefelter's syndrome. Lancet 2:164-167.
PERSPECTIVES ON VIOLENCE 177 Pepler, D.J., G. King, and W. Byrd 1991 A social-cognitively based social skills training program for aggressive children. Pp. 361-379 in D.J. Pepler and K.H. Rubin, eds., The Development and Treatment of Childhood Aggression. Hillsdale, N.J.: Erlbaum. Piacentini, J.C. 1987 Language dysfunction and childhood behavior disorders. Advances in Clinical Child Psychology 10:259-285. Plomin, R. 1989 Environment and genes. American Psychologist 44(2):105-111. Prentky, R.A. 1990 Sexual Violence. Paper prepared for the Panel on the Understanding and Control of Violent Behavior, National Research Council. March 1 . Prothrow-Stith, D., H. Spivak, and A. Hausman 1987 The violence prevention project: A public health approach. Science, Technology, and Human Values 12:67-69. Quinsey, V.L., T.C. Chaplin, and D. Upfold 1984 Sexual arousal to nonsexual violence and sadomasochistic themes among rapists and non-sex-offenders. Journal of Consulting and Clinical Psychology 52:651-657. Raine, A., and P.H. Venables 1988a Skin conductance responsivity in psychopaths to orienting, defensive, and consonant-vowel stimuli. Journal of Psychophysiology 2:221-225. 1988b Enhanced P3 evoked potentials and longer P3 recovery times in psychopaths. Psychophysiology 25:30-38. Raine, A., P.H. Venables, and M. Williams 1990a Relationships between CNS and ANS measures of arousal at age 15 and criminality at age 24. Archives of General Psychiatry 47:1003-1007. 1990b Relationships between N1, P300 and CNV recorded at age 15 and criminal behavior at age 24. Psychophysiology 27:567-575. Reiss, Albert J., Jr., and Michael Tonry, eds. 1986 Communities and Crime. Chicago: University of Chicago Press. Revitch, E. 1965 Sex murder and the potential sex murder. Diseases of the Nervous System 26:640-648. Robins, L.N. 1978 The interaction of setting and predisposition in explaining novel behavior: Drug initiations before, in, and after Vietnam. In D.B. Kandel, ed., Longitudinal Research on Drug Use. New York: John Wiley & Sons. Roncek, D. 1981 Dangerous places: Crime and residential environment. Social Forces 60:74-96.
PERSPECTIVES ON VIOLENCE 178 Roncek, D., and D. Faggiani 1986 High schools and crime: A replication. Sociological Quarterly 26:491-505. Rothberg, Joseph M., Paul T. Bartone, Harry C. Holloway, and David H. Marlowe 1990 Life and death in the U.S. Army. Journal of the American Medical Association 264(17):2241-2244. Rubel, R.J., and A.H. Goldsmith 1980 Reflections on the rights of students and the rise of school violence. Pp. 71-79 in K. Baker, and R.J. Rubel, eds., Violence and Crime in the Schools. Lexington, Mass.: Lexington Books. Russell, D.E.H. 1975 The Politics of Rape: The Victim's Perspective. New York: Stein & Day. Sampson, R.J. 1983 Structural density and criminal victimization. Criminology 21:276-293. 1985 Neighborhood and crime: The structural determinants of personal victimization. Journal of Research in Crime and Delinquency 22(1):7-40. 1987 Personal violence by strangers: An extension and test of the opportunity model of predatory victimization. Journal of Criminal Law and Criminology 78:327-356. Sampson, R.J., and Wooldredge 1987 Linking the micro- and macro-level dimensions of lifestyle- routine activity and opportunity models of predatory victimization. Journal of Quantitative Criminology 3:371-393. Schuerman, L., and S. Kobrin 1986 Community careers in crime. Pp. 67-100 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Vol. 8. in Crime and Justice: A Review of Research. Chicago: University of Chicago Press. Schwartz, Gary 1987 Beyond Conformity or Rebellion: Youth and Authority in America. Chicago: University of Chicago Press. S.D. Texas 1980 Ruiz v. Estelle 503 F. Supp. 1265. Segal, Z.V., and W.L. Marshall 1985 Heterosexual social skills in a population of rapists and child molesters. Journal of Consulting and Clinical Psychology 53:55-63. Shaw, C.R., and H.D. McKay 1942 Juvenile Delinquency and Urban Areas. Chicago: University of Chicago Press. Sherman, L.W., P.R. Gartin, and M.E. Buerger 1989 Hot spots of predatory crime: Routine activities and the criminology of place. Criminology 27(1):27-55.
PERSPECTIVES ON VIOLENCE 179 Short, James F. 1990 Gangs, neighborhoods, and youth crime. Criminal Justice Research Bulletin 5 (4):1-11. 1991 Poverty, ethnicity and crime: Change and continuity in U.S. cities. (Review essay.) Journal of Research in Crime and Delinquency 28(4):501-518. Short, James F., and Fred L. Strodtbeck 1965 Group Process and Gang Delinquency. Chicago: University of Chicago Press. Siddle, D.A.T., A.R. Nicol, and R.H. Foggit 1973 Habituation and over-extinction of the GSR component of the orienting response in anti-social adolescents. British Journal of Social and Clinical Psychology 12:303-308. Simpkins, L., W. Ward, S. Bowman, and C.M. Rinck 1989 The Multiphasic Sex Inventory as a predictor of treatment response in child sexual abusers. Annals of Sex Research 2:205-226. Singer, J.L., and D.G. Singer 1981 Television, Imagination, and Aggression: A Study of Pre- Schoolers. Hillsdale, N.J.: Erlbaum. Skogan, Wesley G. 1990 Disorder and Decline: Crime and the Spiral of Decay in American Neighborhoods. New York: Free Press. Smith, D.R., and G.R. Jarjoura 1988 Social structure and criminal victimization. Journal of Research in Crime and Delinquency 25:27-52. Spergel, Irving A., G. David Curry, Ruth Ross, and Ronald Chance, eds. 1989 Survey. National Youth Gang Suppression and Intervention Project. School of Social Service, University of Chicago. Spiro, A., III, and W. DeJong 1991 Preventing Interpersonal Violence Among Teens: Field Tests and Evaluation of the Violence Prevention Curriculum for Adolescents. Final report for National Institute of Justice grant no. 87-IJ-CX-0009. Steadman, Henry J. 1982 A situational approach to violence. International Journal of Law and Psychiatry 5:171-186. Sullivan, M. 1989 Getting Paid: Youth Crime and Work in the Inner City. Ithaca, N.Y.: Cornell University Press. Taylor, Carl S. 1990a Gang imperialism. Pp. 103-115 in C. Ronald Huff, ed., Gangs in America. Newbury Park, Calif.: Sage Publications. 1990b Dangerous Society. East Lansing, Mich.: Michigan State University Press. Taylor, Ralph B., and Jeanette Covington 1988 Neighborhood changes in ecology and violence. Criminology 26(4):553-589.
PERSPECTIVES ON VIOLENCE 180 Taylor and Gottfredson 1986 Environmental design, crime, and prevention: An examination of community dynamics. Pp. 387-416 in A.J. Reiss, Jr., and M. Tonry, eds., Communities and Crime. Vol. 8. Chicago: University of Chicago Press. Thornberry, T.P., A.J. Lizotte, M.D. Krohn, M. Farnworth, and S.J. Jang 1991 Testing interactional theory: An examination of reciprocal causal relationships among family, school, and delinquency. Rochester youth development study. Journal of Criminal Law and Criminology 82(1):3-35. Thrasher, Frederic M. 1927 The Gang: A Study of 1,313 Gangs in Chicago. (Abridged ed., 1963.) Chicago: University of Chicago Press. Toby, J. 1983a Violence in school. Pp. 1-47 in M. Tonry and N. Morris, eds., Crime and Justice: An Annual Review of Research. Vol. 4. Chicago: University of Chicago Press. 1983b Violence in School. Washington, D.C.: National Institute of Justice, Research in Brief. Toch, H. 1977 Police, Prisons, and the Problem of Violence. Chapter 4. Washington, D.C.: U.S. Government Printing Office. 1989 Violence in prisons. Pp. 267-285 in K. Howells and C.R. Hollin, eds., Clinical Approaches to Violence. New York: John Wiley & Sons. Tonry, M., L.E. Ohlin, and D.P. Farrington 1991 Human Development and Criminal Behavior: New Ways of Advancing Knowledge. New York: Springer-Verlag. Tremblay, R.E., J. McCord, H. Boileau, P. Charlebois, C. Cagnon, M. LeBlanc, and S. Larivee 1991 Can disruptive boys be helped to become competent? Psychiatry. Venables, P.H., and A. Raine 1987 Biological theory. Pp. 3-28 in B. McGurk, D. Thornton, and M. Williams, eds., Applying Psychology to Imprisonment: Theory and Practice. London: Her Majesty's Stationery Office. Vigil, James Diego 1987 Street socialization, locura behavior, and violence among Chicano gang members. Pp. 231-241 in Jess F. Kraus, Susan B. Sorenson, and Paul D. Juarez, eds., Research Conference on Violence and Homicide in Hispanic Communities. University of California, Los Angeles. 1988 Barrio Gangs. Austin, Tex.: University of Texas Press. Virkkunen, M. 1986 Reactive hypoglycemia tendency among habitually violent offenders. Nutrition Reviews 44:94-103. Virkkunen, M., J. DeJong, J. Bartko, and M. Linnoila 1989 Psychobiological concomitants of history of suicide attempts
PERSPECTIVES ON VIOLENCE 181 among violent offenders and impulsive fire setters. Archives of General Psychiatry 46:604-606. Wadsworth, M.E.J. 1976 Delinquency, pulse rate and early emotional deprivation. British Journal of Criminology 16:245-256. Walkey, F., and D.R. Gilmour 1984 The relationship between interpersonal distance and violence in imprisoned offenders. Criminal Justice and Behavior 11(3):331-340. Wallace, R., and D. Wallace 1990 Origins of public health collapse in New York City: The dynamics of planned shrinkage, contagious urban decay and social disintegration. Bulletin of the New York Academy of Medicine 66:391-434. Weis, Joseph G. 1986 Issues in the measurement of criminal careers. Pp. 1-51 in A. Blumstein, J. Cohen, J.A. Roth, and C. Visher, eds., Criminal Careers and âCareer Criminals." Vol. II. Washington, D.C.: National Academy Press. Weis, Kurt, and Sandra S. Borges 1977 Victimology and rape: The case of the legitimate victim. Pp. 35-75 in D.R. Nass, ed., The Rape Victim. Dubuque, Iowa: Kendall/Hunt. Williams, D. 1969 Neural factors related to habitual aggression: Consideration of those differences between those habitual aggressives and others who have committed crimes of violence. Brain 92:503-520. Williams, Terry 1989 The Cocaine Kids: The Inside Story of a Teenage Drug Ring. Menlo Park, Calif.: Addison-Wesley. Wilson, William Julius 1987 The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press. 1990 Race-neutral programs and the democratic coalition. The American Prospect 1(Spring):74-81. 1991 Studying inner-city social dislocations: The challenge of public agenda research. American Sociological Review 56(February):1-14. Wolfgang, Marvin E. 1957 Victim-precipitated criminal homicide. Journal of Criminal Law, Criminology, and Police Science 48(I):1-11. Wolfgang, Marvin E., Robert M. Figlio, and Thorsten Sellin 1972 Delinquency in a Birth Cohort. Chicago: University of Chicago Press.