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.
RECOMMENDATIONS 327 8 Recommendations To make progress in the understanding and control of violent behavior, we call for a balanced program of efforts with short-term and long-term payoffs: (1) problem-solving initiatives of pragmatic, focused, methodologically sound collaborative efforts by policy makers, evaluation researchers, and basic researchers; (2) modifying and expanding national and local violence measurement systems for diagnosing particular violence problems and measuring the effects of interventions designed to solve them; (3) programs of relatively small-scale research projects in areas that have been largely neglected by federal violence research sponsors; and (4) the multicommunity research program described in Chapter 3, which is intended to expand society's capacities to understand and to modify community-, individual-, and biological-level processes that influence individuals' potentials for violent behavior. The problem-solving initiatives and research programs in neglected areas can fairly quickly make incremental contributions to the understanding and control of violent behavior. The improvement of violence measurement systems and the multicommunity research program, while requiring longer initial investment periods, will lay the groundwork for better diagnosis and understanding
RECOMMENDATIONS 328 of violence and for the design of more effective intervention programs. PROBLEM-SOLVING INITIATIVES Available evidence suggests an abundance of promising preventive strategies. As explained in Chapter 7, implementation of these strategies through specific interventions should be an iterative process: diagnosis and intervention design, evaluation involving outcome measures at multiple levels of observation, refining the intervention in light of evaluation results, and replications of the evaluation. Over time, repeated iterations of these steps can expand society's capacity to control violence, and simultaneously contribute knowledge about the processes that cause it. This process should be used in a series of efforts focused on specific components of violenceâeach of which, although significant, accounts for only a small fraction of all violent events. We call this strategy problem solving in violence. Six areas seem especially promising for this problem-solving approach. In the next 15 years, sustained collaboration by policy makers, disciplinary researchers, and evaluation researchers in these initiatives could make major cumulative contributions to better control and understanding of specific violence problems. Recommendation 1: We recommend that sustained problem-solving initiatives be undertaken in six specific areas for which systematic intervention design, evaluation, and replication could contribute to the understanding and control of violence: (a) intervening in the biological and psychosocial development of individuals' potentials for violent behavior, with special attention to preventing brain damage associated with low birthweight and childhood head trauma, cognitive-behavioral techniques for preventing aggressive and violent behavior and inculcating prosocial behavior, and the learning of attitudes that discourage violent sexual behavior; (b) modifying places, routine activities, and situations that promote violence, with special attention to commercial robberies, high-risk situations for sexual violence, and violent events in prisons and schools; (c) maximizing the violence reduction effects of police interventions in illegal markets, using systematic tests and evaluations to discover which disruption tactics for the illegal
RECOMMENDATIONS 329 drug and firearm markets have the greatest violence reduction effects; (d) modifying the roles of commoditiesâincluding firearms, alcohol, and other psychoactive drugsâin inhibiting or promoting violent events or their consequences, with special attention to reducing weapon lethality through public education and technological strategies; ascertaining patterns of firearms acquisition and use by criminals and juveniles; ascertaining and modifying the pharmacological, developmental, and situational processes through which alcohol promotes violent behavior; pharmacologically managing aggressive behavior during opiate withdrawal; ascertaining whether smoking cocaine promotes violence through special pharmacological effects; and reducing drug market violence by reducing demand for illegal psychoactive drugs; (e) intervening to reduce the potentials for violence in bias crimes, gang activities, and community transitions; and (f) implementing a comprehensive initiative to reduce partner assault, including risk assessment; experimentation with arrest, less expensive criminal justice interventions, public awareness campaigns, batterers' counseling programs, alcohol abuse treatment for perpetrators, and family services; and further analyses of the relationships between women's shelter availability and assault and homicide rates. We explain these initiatives more fully in the sections that follow. Development of Violence Potentials Five specific elements of biological and psychosocial development discussed in Chapter 3 warrant particular attention in problem-solving initiatives: â¢ Brain dysfunctions that interfere with language processing or cognition: What is their importance as a risk factor for aggression? How effectively can they be reduced by interventions to prevent substance abuse by pregnant women and to reduce children's exposure to environmental toxins including lead? What is their long-term effect on violent behavior? â¢ Cognitive-behavioral preventive interventions: What is the comparative effectiveness of the following approaches to reducing childhood aggression in different subpopulations: parent training, school-based antibullying programs, social skills training, cognitive-behavioral
RECOMMENDATIONS 330 interventions that stress the undesirability of aggression and teach nonviolent conflict resolution, and promotion of television programs that encourage prosocial, nonviolent behavior and that appeal to children of diverse cultural backgrounds? â¢ Prevention of school failure: What are the comparative and cumulative effects of preschool educational enrichment and early-grade interventions, including tutoring by peers or trained high-school students, on early-grade school failure rates? On childhood aggression? On later adolescent and adult violence? â¢ Development of violent sexual behavior: How important are the following as risk factors: exposure to abnormally high testosterone levels during fetal development, childhood sexual abuse victimization, the learning of tolerant attitudes toward violent acts against women, the development of sexual preferences for violent stimuli, and chronic alcohol use? How do sexual arousal patterns differ between samples of known violent sex offenders and other samples? Are these measures related to any specific neurological, endocrine, or genetic markers? â¢ Systematic evaluations of preventive and therapeutic interventions for sexual violence: development and testing of early preventive strategies and education about sexual violence; and behavioral therapies such as relapse prevention therapy, assertiveness therapy, and anger management therapy. Modifying Places and Situations Even successful developmental interventions will take several years to show effects, and they will prevent only violent acts that have developmental roots. Interventions with a situational perspective should be considered as well in formulating preventive strategies. For a few people, biological interventions may prevent violent behavior in situations that provoke anxiety but not violent behavior by most individuals. One promising initiative for reducing this problem is the development of anxiolytics that act on the serotonergic system. Other situations increase the risk of violent behavior for a broad spectrum of individuals. These risks may be more efficiently reduced by modifying characteristics of places and encounters. Based on the discussion in Chapter 3, the following appear to be among the most promising of these: â¢ Places, routine activities, and encounters: What are the specific risk factors for violent events in high-risk places and encounters,
RECOMMENDATIONS 331 and how can they be effectively modified? By how much are situational risks of violent behavior and victimization increased by alcohol consumption? â¢ Violence in commercial robberies: What interventions can reduce the numbers of persons who control valuables in exposed locations at night in the course of their employment? Where such situations cannot be eliminated, what modifiable factors can reduce the chances and consequences of violent victimization? â¢ Situational prevention of sexual violence: How can the harms of sexual violence be most effectively reduced by separating offenders and victims (e.g., by institutionalizing offenders and sheltering victims of partner assaults), by modifying high-risk situations, and by repairing the physical and psychological consequences of victimization? How should prevention of events involving acquaintances and intimates differ from prevention of events involving strangers? â¢ Violence between custodians and wards: What is the incidence of violent victimizations in schools and prisons? What are the risk factors for violent events in these places? What are the effects of modifying these risk factors? Illegal Markets There are plausible reasons to anticipate violent events in illegal drug markets, yet only fragmentary and conflicting evidence exists on how market disruption by law enforcement agencies affects drug market violence levels. The following problem-solving initiatives should therefore receive high priority: â¢ illegal drug markets (Chapter 4): ascertaining differences in violence 1 patterns for crack houses, "runner-beeper" operations, and open-air drug markets; evaluating the effects on violence of current street-level law enforcement tactics, including possible differential effects in those three types of markets; evaluating the effects of proactive policing on drug- market violence; comparing the effects of centralized and street-level drug law enforcement strategies on violence levels; â¢ illegal firearm markets (Chapter 6): regular "trace studies" of the distribution channels for firearms acquired by juveniles and used in crimes and tests of interventions for disrupting those channels at the wholesale level; evaluations of intervention programs to enforce existing firearms regulations with street-level tactics currently in use for disrupting illegal drug markets (e.g., "buy-bust" operations).
RECOMMENDATIONS 332 Other Violence-Related Aspects of Commodities Beyond their roles in illegal markets, firearms and psychoactive drugs including alcohol influence the chances and consequences of violent events in other ways. Firearms Based on the discussion in Chapter 6, the diagnosis and evaluation components of problem-solving initiatives are important respectively in two relationships between firearms and violence. â¢ Ownership and use: What are the most common patterns of ownership, acquisition, and use of firearmsâespecially for juveniles and criminals, and for handguns and assault weapons (under alternative definitions)? For what purposes are these weapons acquired and carried? How frequently are gun-owning intended crime victims able to deploy their guns in self-defense and prevent injury or death? â¢ Reduction of injury and lethality: How do public education and technological interventions intended to reduce firearms injuries and deaths affect the incidence and consequences of violent firearm use? Alcohol and Other Psychoactive Drugs Three issues that pertain to the roles of both alcohol and other psychoactive drugs in violent events should receive high priority in problem-solving initiatives: â¢ Continue developing and refining interventions to prevent adolescents from becoming abusers of alcohol and other psychoactive drugs, and measure the effects on subsequent violent behavior (Chapter 4). â¢ Use patterns (Chapter 4): What are the most common patterns of psychoactive drug use (including alcohol, and emphasizing use of multiple drugs)? As benchmarks for analyzing causal relationships, what are the profiles of alcohol and other psychoactive drug use by time of day and day of week, for representative samples of different ethnic and socioeconomic subpopulations? What patterns of useâmix of drugs, duration of use, method of administrationâare associated with what patterns of violent behavior? How do these risk relationships differ across categories of persons? â¢ Indirect relationships between psychoactive drugs and violence (Chapters 4 and 5): How is the incidence of violence related
RECOMMENDATIONS 333 to the use of alcohol or illegal psychoactive drugs, through arguments over debts and family arguments over money, time spent away from home, etc.? Because alcohol and other psychoactive drugs have somewhat different relationships to violence, certain problem-solving initiatives apply to only one or the other. Alcohol Problem-solving initiatives at the biological, psychosocial, microsocial, and macrosocial levels may all be useful in reducing alcohol-related violence: â¢ Mount a program to develop a drug that blocks the aggression-promoting effects of alcohol through various modes of action, for example, on the GABAA/benzodiazepine receptor complex (Chapter 4). â¢ Develop and evaluate special alcohol-and-violence counseling programs for youth with diagnosed antisocial personality disorder who abuse both alcohol and other psychoactive drugs, whose parents abuse alcohol, and whose behavior brings them under juvenile court or social service agency jurisdiction (Chapter 4). â¢ For situations that involve drinking alcohol, identify risk factors for violent events and evaluate the violence prevention effects of public information campaigns analogous to "if you drink, don't drive" that encourage people to modify those risk factors (Chapter 4). â¢ Test the effectiveness of alcohol excise tax increases as a means of reducing violent behavior by adolescent males (Chapter 4). Other Psychoactive Drugs For most people, the pharmacological effects of illegal psychoactive drugs do not generally promote violent behavior. Two recommended initiatives are concerned with special circumstances in which such effects may occur: â¢ Test the effectiveness of clonidine and other drugs in managing human aggressive behavior during withdrawal from heroin (Chapter 4). â¢ Special pharmacological effects of crack (Chapter 4): Does the direct and rapid access of smoked cocaine to the brain promote violent behavior even though powdered cocaine use generally does not? Illegal psychoactive drugs are related to violence largely through violent crimes committed in the course of purchasing or distribution
RECOMMENDATIONS 334 in illegal markets. Two recommended initiatives are concerned with reducing this violence by reducing demand for illegal drugs: â¢ Develop pharmacological interventions for reducing users' craving for psychoactive drugs by blocking dopamine and norepinephrine receptor subtypes (Chapter 3). â¢ For incarcerated psychoactive drug users who are convicted of violent crimes or weapons violations, evaluate the violence-control effects of programs that combine in-prison detoxification and treatment with postrelease urinalysis and community-based relapse prevention follow- up, using the models of methadone maintenance for heroin users and therapeutic communities such as the "Stay 'N Out" program for users of cocaine and other illegal psychoactive drugs (Chapter 4). Initiatives at the Social Level Problem-solving initiatives should be mounted to deal with specific social- level violence problems. â¢ Violent bias crimes (Chapter 3): Conduct comparison and cross- validation of incidence estimates obtained by police and by community-based organizations; research on the individual- and community-level effects of violent bias crimes; comparison of three strategies to reduce levels and consequences of violent bias crimes: criminal sanctions, identifying risk factors for victimization and educating members of target groups about how to modify them, and interventions to repair the psychological consequences of violent bias crimes victimization. â¢ Gang-related violence (Chapter 3): Identify risk factors that differentiate violent gangs from other gangs and develop and test interventions to modify those risk factors in violent gangs. â¢ Community transitions (Chapter 3): Identify the violence risk factors in community transitions (especially economic decline and neighborhood gentrification), and develop and test interventions for modifying those risk factors during transition periods. A Comprehensive Initiative to Reduce Partner Assault With the exceptions of arrest in misdemeanor cases and programs for batterers (both of which have generated mixed favorable effects), few of these interventions have been systematically evaluated.
RECOMMENDATIONS 335 Perhaps more important, there has been no systematic investigation of their comparative effectiveness or of how they might be used togetherâin part because these interventions are offered by different public authorities and/or community-based organizations. This leads us to recommend a comprehensive initiative against spouse assault, with the following components: â¢ risk assessment: better estimates in local surveys and the National Crime Survey of the incidence and prevalence of all types of family violence, including attention to cohabiting families; special attention to the incidence of spouse abuse, especially to repeat victimization; tabulations of more information on types of family violence within legal/statistical categories (e.g., disaggregating specific family relationships, capturing sexual assaults on minors); â¢ case control or randomized experimental studies that follow two groups of families over several years: one group would be exposed to a public awareness and education campaign; another offered an infusion of services including shelter, parenting, vocational assistance, assertiveness training, and day care; these treatments could be provided in communities with different law enforcement policies; â¢ a jurisdictional comparative analysis of shelter availability with assault and homicide rates; â¢ evaluation of batterers' programs through experiments using randomized assignment; and â¢ evaluation of law enforcement and criminal justice interventions including arrest as well as less expensive interventions: (a) restraining orders, (b) police warnings in cases without probable cause, (c) police training coupled with intense public education campaigns, and (d) sanctions varying from mandated courses to work programs. IMPROVING VIOLENCE MEASUREMENT SYSTEMS Many questions of fundamental policy and scientific importance cannot be answered today, and emerging violence problems are sometimes slow to be discovered, because of four basic limitations of the systems for gathering information on violence. First, existing systems record only a small slice of all violent behaviors. Second, even for the behaviors covered, they do not provide the data needed to calculate basic conditional probabilities of a violence threat given certain circumstances; of a violent event given a particular kind of threat; or of death, injury, psychological, and financial consequences given the attributes of a violent event.
RECOMMENDATIONS 336 Third, even if otherwise suitable data are available for social units such as geographic or jurisdictional areas, those units are often too large or heterogeneous to permit precise estimates of empirical relationships or intervention effects. Fourth, the information system categories often mask behavioral diversity: multiple reasons for a homicide, for example, or the chain of events that preceded a firearm injury. Recommendation 2: The panel recommends that high priority be placed on modifying and expanding relevant statistical information systems to provide the following: (a) counts and descriptions of violent events that are receiving considerable public attention but are poorly counted by existing measurement systems. These include but are not limited to intrafamily violence; personal victimizations in commercial and organizational robberies; violent bias crimes; and violent events in schools, jails, and prisons; (b) more comprehensive recording of sexual violence, including incidents involving intimates, incidents of homicide and wounding in which the sexual component may be masked, and more complete descriptions of recorded events; (c) baseline measurements of conditions and situations that are thought to affect the probability of a violent event (e.g., potentially relevant neurological disorders, arguments between intoxicated husbands and wives, drug transactions, employees handling cash at night in vulnerable locations); (d) information on the treatment of violence victims in emergency departments, hospitals, and long-term care facilities; links to data on precipitating violent events; and development of these data as a major measurement system; (e) information on long- and short-term psychological and financial consequences of violent victimization and links to data on violent events; (f) measurements of violence patterns and trends for small geographic and jurisdictional areas, as baselines for measuring preventive intervention effects; and (g) information system modifications to record more detailed attributes of violent events and their participants, in order to facilitate more precise studies of risk factors for violence and evaluations of preventive interventions to reduce it.
RECOMMENDATIONS 337 RESEARCH IN NEGLECTED AREAS Seven research areas have been largely starved of resources for decades while applicable theory, measurement, and methodology have advanced. As a result, substantial and rapid progress can be expected from relatively small-scale research in the neglected areas. Recommendation 3: We call for new research programs specifically concerned with the following areas: (a) nonlaboratory research on the instrumental effects of weapons on the lethality of assaults, robberies, and suicide attempts (Chapter 6); (b) integrated studies of demographic, situational, and spatial risk factors for violent events and violent deaths (Chapter 3); (c) comparative studies of how developmental processes in ethnically and socioeconomically diverse communities alter the probabilities of developmental sequences that promote or inhibit violent behavior (Chapter 3); (d) systematic searches for neurobiologic markers for persons with elevated potentials for violent behavior (Chapter 3); (e) systematic searches for medications that reduce violent behavior without the debilitating side effects of âchemical restraint" (Chapter 3); (f) integrated studies of the macrosocial, psychosocial, and neurobiologic causes of sexual and other violence among strangers, intimate partners, and family members (Chapters 3 and 5); and (g) studies of violent behavior by custodians against wards (Chapter 3). These recommendations are explained more fully in the following pages. Instrumental Effects of Weapons We still lack much important basic knowledge about how the choice of weapons influences the lethality of robberies and assaults in actual practice, even though the Centers for Disease Control and the National Institute of Justice now support research on the subject. Specific issues needing investigation include the lethal effects of weapons in actual use; the effects of street-level and wholesale-level
RECOMMENDATIONS 338 police tactics in enforcing firearms regulations; sources of guns used in crimes; sources of guns illegally carried by minors; organization of the guns-for-drugs trade; the circumstances and frequency of firearm deployment for self-defense and the effect of deployment on victimization consequences; disaggregated measures of weapon availability, ownership patterns and motivation, legitimate uses, and acquisition patterns; and the roles in crime of automatic and semiautomatic rifles and handguns, especially by juveniles. Demographic, Situational, and Spatial Factors Many violent events arise from the intersection of demographic, situational, and spatial elements. Much of the relevant research focuses on only one of those levels of analysis at a time. Demographic patterns of violent offending and victimization are well documented. A start was made nearly a decade ago in understanding the situational dynamics of encounters that include violent behavior (Felson and Steadman, 1983). And progress has been made in describing distributions of violent and other crimes across places operationalized as addresses or telephone numbers (Sherman et al., 1989). What is needed is integrated research on how all three kinds of elementsâdemographic, situational, and spatialâconverge to influence individuals' exposure to the risks of victimization in robbery, of sexual and nonsexual assaults, and of death in those events. Comparative Studies Similarly, even though racial and socioeconomic differences in violence rates are well documented and evidence supports a number of explanations, many plausible theories remain untested. Analyses of macrosocial influences on violence, including community-level differences in learning environments and access to prenatal and pediatric care, require comparative research on individuals in multiple communities that vary in their compositionâvariations in ethnic and socioeconomic status mix seem especially important. Similarly, we need to understand how cross-national variation in violence rates is related to other national characteristics.
RECOMMENDATIONS 339 Biological Markers for Violent Behavior Although biological markers for an elevated propensity for violent behavior are currently lacking, findings point to several sites in the nervous system where they may eventually be discovered: â¢ abnormal functioning of the limbic system and temporal lobe of the brain, especially the hypothalamus; â¢ unusual neural discharge patterns in the temporal lobe; and â¢ abnormal activity or metabolism of amines, steroids, or peptides that act as neurotransmitters or modulators, especially dopamine; norepinephrine; serotonin; gamma-aminobutyric acid (GABA); monoamine oxidase (MAO); steroid hormones, particularly testosterone; and glucocorticoids. A common approach to searching for such markers has been psychophysiological research, which explores correlations between aggressive or violent behavior and certain peripheral measures of neurological activity such as heart rate, skin conductance, and evoked potentials. As explained in Chapter 3, these measures are often associated with general arousal levels rather than specific to violent behavior, this approach is likely to identify risk factors specific to violent behavior only through repeated measures linked to records of behavior over time. More research using the following approaches is needed: observing behavior after pharmacologically blocking or triggering the specific neurobiologic process being tested, and using implanted devices to measure and transmit records of neurophysiological activity by animals during their aggressive and other behaviors in seminatural group settings. If neurobiologic markers for violent behavior exist, they are more likely to be found if researchers: â¢ make greater use of general population samples rather than of samples selected on the traditional basis of behavioral or neurological aberrations; â¢ measure neurochemical activity directly in the brain rather than measuring peripheral levels in hair, saliva, urine, blood, spinal fluid, etc.; â¢ design studies to distinguish between neurobiologic causes and effects of violent behavior; and â¢ search for markers in terms of responses to chemical, environmental, or social challenges rather than resting levels.
RECOMMENDATIONS 340 Preventive Pharmacological Interventions Neurobiologic research has already produced drugs that have proven useful as primary or adjunct therapies for managing certain violent human behaviors in specific populations. These include antiandrogen drugs, beta-blockers, and dopamine receptor antagonists. Recent findings suggest that two additional promising leads should be followed: use of antianxiety drugs that operate on the GABAA/benzodiazepine receptor complex to inhibit alcohol-related aggression; and use of anxiolytics that act on the serotonergic system to reduce certain aggressive behaviors. Sexual Violence 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
RECOMMENDATIONS 341 practicable,2 to assess the effectiveness of these interventions in preventing the recurrence of sexually violent acts. Improved classification and measurement are essential for making 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 those who commit violent acts against intimates that do not routinely lead to institutionalization as sex offenders. Violence Involving Custodians and Wards Among community institutions, we surveyed violence in only twoâprisons and schools. We found a dearth of information on the frequencies, causes, and special consequences of violent events in those settings. We believe that theories and data are similarly lacking on violence in other settings populated by custodians and wards, such as mental hospital staff and patients, caregivers and children, and nursing home staff and patients. One type of violence in this categoryâviolent events involving police and other citizensâhas recently become increasingly visible. Knowledge of situational- and individual-level risk factors in all these settings could be particularly helpful in determining whether changes in personnel selection, training, or organizational accountability can affect levels of violence by custodians. There is also a need for research on the consequences of such incidents, both for the victims and for the surrounding community. MULTICOMMUNITY LONGITUDINAL STUDIES There is a need now to lay the basic research groundwork for the next generation of preventive interventions. Recommendation 4: The panel calls for a new, multicommunity program of developmental studies of aggressive, violent, and antisocial behaviors, intended to improve both causal understanding and preventive interventions at the biological, individual, and social levels. As described in more detail in Chapter 3, this multicommunity study would include initial assessments, follow-ups, and randomized experiments for two cohorts in each communityâa birth
RECOMMENDATIONS 342 cohort and a cohort of 8-year-olds. Here, we first explain more clearly the scientific and policy considerations that call for a longitudinal study of violent and aggressive behaviors. We then discuss the problem of maintaining the long- term support needed for the program. Scientific Considerations The methodological advantages of longitudinal panel designs over cross- section approaches for studying the development of delinquent and criminal behavior have been explained in detail elsewhere (Blumstein et al., 1986; Farrington et al., 1986; Tonry et al., 1991). Briefly, they are the ability to observe developmental sequences within individuals (e.g., from low fear response, through aggression, through risk-seeking behaviors, through substance abuse, through violent behavior); the ability to measure the effects of life events that interrupt those sequences; enhanced ability to infer cause-effect relationships by controlling statistically for extraneous variables and establishing time sequences; better information about the accuracy of predictions based on earlier events; enhanced ability to record the forming and breaking of subjects' affiliations with family members, gangs, and other relevant groups; and enhanced ability to record stability or fluctuation over time in individuals' aggressive or violent behavior. In principle, there has long been no reason why studies combining aggregate and individual-level data could not simultaneously study how the structures and cultures of large and small social units, as well as the dynamics of microsocial encounters, influence individuals' potentials for violent behavior. Useful starts have been made in identifying social-level correlates of both criminal career dimensions and violent victimization risk. But knowledge of correlates is not sufficient for understanding processes. And studies of multilevel processes have rarely been carried out because of the substantial burdens they impose on data collection and on researcher's capacities to competently span the theories and methods of the different disciplines concerned with the different levels. We are calling for a program that would begin to remedy these deficiencies. Compared with other programs of longitudinal studies, this one would be distinguished by the combination of: (1) specific emphasis on the relationships between aggressive and violent behavior, including social-level influences on those behaviors;
RECOMMENDATIONS 343 (2) a multicommunity design to facilitate more extensive study of cultural and biosocial influences, both on developmental sequences and on intervention effects; (3) neurobiologic measurements that are as specific for relevant hypothesized processes in the brain as is ethically and technically feasible; (4) designs that facilitate analyses of protective and aggravating conditions and factors in families, peer groups, schools, and communities; (5) randomized tests of interventions that, on the basis of causal understanding at the social, psychological, and biological levels, show promise of fostering the development of prosocial behavior and inhibiting the development of potentials for violent behavior; (6) oversampling of high-risk categories and special efforts to minimize attrition by study subjects in those categories; and (7) cross-validation of official record and self-report versions of violent events. Policy Considerations In the face of visible public concern over violence and its consequences, some may view any investment in basic research, such as a longitudinal study, as impractical or not cost-effective. They might opt for spending the same resources on new prison beds. Although criminals sanction will doubtless remain a major violence control tactic, they are not necessarily the most cost-effective approaches to reducing violence levels. Several risk factors for violence (e.g., poor cognitive capacity, transient neurobiologic conditions, alcohol intoxication) actually inhibit persons from weighing punishment as a possible deterrent to violent behavior. Strategies of selectively longer incarceration terms for convicted persons who have demonstrated elevated potentials for violent behavior will not reduce crime and violence levels without substantial additional increases in prison populations. As explained in Chapter 7, recent experience suggests that if violent crime has been deterred by the threat of incarceration, then it may also have been promoted by other criminogenic factors, some of which arise in the course of individual development. And most fundamentally, criminal sanctions cannot be mobilized until after a violent crime has occurred. As explained in Chapter 3, some developmentalists argue that a small high- risk group can be identified in early childhood and that early preventive intervention strategies can be devised to
RECOMMENDATIONS 344 reduce the portion of this group that grows up to commit violent acts. Some research has already been done along these lines. But much of this work has generally focused on broader behavioral categoriesâconduct disorder, aggression, or delinquency, for example. Risk factors for broad categories of behavior tend to overpredict violent behavior, and the implied intervention strategies may be ineffective in preventing rarer forms of violent behavior. A more specifically focused effort is needed to improve understanding and predictive capability both for a predisposition to aggressive behavior, which is usually observable by age 8, and for the aggressive children whose behavior later becomes violent. Maintaining Support for the Program In calling for a longitudinal study of aggressive and violent behaviors, we are calling for a substantial long-term resource commitment that offers only a reasonable chance of breakthroughs in the understanding and control of violence. The size of the commitment looks smaller in the context of $10 billion (as of 1989) in ongoing prison construction, or of $10 billion per year for space research. Nevertheless, the program will require a commitment of new funds, and the full potential of returns from the investment will not be realized unless funding continues throughout the entire data collection and analysis periodâ10 to 11 years for the birth cohort and 12 to 13 years for the 8-year-old cohort. Ironically, generating and maintaining support for this program may be more difficult than otherwise because relatively few people behave violently. Based on previous longitudinal researchers' experience, Tonry et al. (1991) estimate that about 9 percent of a birth cohort will be arrested for a violent crime before the eighteenth birthday, and that about 36 percent of an early-teen cohort followed for 10 years will report to interviewers that they committed a serious violent act. Given that proportion, a conservative benchmark data collection cost of $500 per year per sample subject would imply a 10-year cost approaching $14,000 per subject who commits a violent offenseâa figure that many would consider too high to warrant this project solely as a means of studying violent behavior. That cost can, of course, be reduced by oversampling from high-risk populations. Moreover, the study will prove cost-effective if the accounting scheme is comprehensive enough, however, because children at greatest risk of aggressive and violent behavior are also at high risk for other problems. Menard et al. report that,
RECOMMENDATIONS 345 of subjects in the National Youth Survey, not only did 36 percent report at least one serious violent offense, but also 87 percent reported general delinquency, 42 percent reported committing index offenses, 94 percent reported illicit alcohol use, 37 percent reported polydrug use, and 24 percent reported mental health problems. In short, even though this study would be the first with a primary focus on violent behavior, it should produce important spin-offs in related areas that are core concerns of agencies in the Office of Justice Programs (a unit of the Department of Justice) and in the Alcohol, Drug Abuse, and Mental Health Administration (a unit of the Department of Health and Human Services). All these areas fall within the basic research mandate of the National Science Foundation. Together, these agencies provide some three-fourths of all federal support for research on violence. FEDERAL SUPPORT FOR VIOLENCE RESEARCH These recommendations will require commitments from federal agency sponsors of research on violence. To get a better picture of exactly who they are, the panel conducted a census of all federal agencies thought to be supporting research and program evaluation on violence as of 1989. The panel was able to account for federal expenditures of $18,080,000 in fiscal 1989. To this amount should be added intramural violence research by 24 full-time equivalent federal employees; with these evaluated at $70,000 each per year,3 this produces a fiscal 1989 total of $20,231,000 classified by responding agencies as violence research.4 To some this size of research expenditures may seem large; however, considered in human terms, it amounts to only $3.41 per 1988 violent victimization. This is a tiny fraction of the estimates by Cohen et al. (Volume 4) of the cost per violent event: $54,100 per rape, $19,200 per robbery, and $16,500 per aggravated assault. As a research topic, violence receives far less support than certain other threats to life. Expenditures on violence research amount to about $31 per year of potential life lost (YPLL) due to violence by age 65. Figure 8-1 compares this amount with research expenditures per YPLL for cancer, heart disease, and AIDS.5 This basis of comparison is, if anything, conservative because it fails to compensate for external social costs that may be associated with violent deaths but not with deaths from disease. Such costs include the deterioration of the quality of life and the loss
RECOMMENDATIONS 346 of legitimate economic activity in communities afflicted with high violence levels, and adverse effects that chronic exposure to neighborhood violence might have on children's education and social development. Figure 8-1 Research expenditures per year of potential life lost for selected causes of death. On this conservative basis of comparison, an increase from the current $20.2 million to $500 million would be needed to bring federal support for violence research into line with federal support for cancer research. Even an increase of that magnitude would bring annual costs for violence research to less than 8 percent of the annual National Institutes of Health budget. It would increase the cost of the societal response to violenceâprimarily criminal justice system costs âby less than 3 percent (see response cost estimates by Cohen et al. in Volume 4). Although an increase of this magnitude would be a proportionate response to violence in America, the panel recognizes that such an increase is unlikely in the current federal fiscal climate. Fortunately, 10 percent of that increase would be sufficient to initiate implementation of the recommendations in this report. Features of the Current Structure Reflecting the diversity of violent behavior itself, the current structure for federally supported research involves many agencies
RECOMMENDATIONS 347 and fosters tremendous diversity in the topics supported. In all, 22 agencies or units of agencies sponsor more than 60 subcategories of violence research. The topics cover most of the important aspects of the understanding and control of violent behavior. The agency respondents were asked to characterize the frequencies with which their violence research programs support various disciplines. While there may well be scholars in all the relevant disciplines whose efforts fail to receive the support they deserve, virtually every relevant discipline has a potential home somewhere among the federal violence research programs. Moreover, inspection of the individual agencies' responses suggests that in programs large enough to support several awards each year, most agencies support a reasonable number of different disciplines. Respondents were asked to characterize the shares of their resources devoted to each of 15 methodologies as zero, minor, or major. Virtually every useful research methodology known to the panel reportedly received a âmajor" share of the resources of at least one sponsoring agency in fiscal 1989. In that year, however, only three programs devoted major shares of resources to program evaluation. The panel is especially concerned about two structural problems with federal support for violence research: (1) the ebb and flow of support for violence researchâa condition that stems from the position of violence as incidental to most sponsoring agencies' mandates; and (2) the dearth of long-term programmatic objectives and support mechanisms, especially for interdisciplinary violence research. Violence as an Incidental For most of the large sponsors of violence research, violence is not a central concern. Rather, violence is considered either in its relationship to some other social problem (e.g., mental illness, drug abuse, alcohol abuse) or as part of some larger mission (e.g., controlling crime, reducing injury). Among the large sponsors of violence research in 1989, for example, violence received only 2 percent of National Institute on Drug Abuse research funds, 1 percent of National Institute on Alcoholism and Alcohol Abuse support, and 17 percent of National Institute of Justice funds. A desirable effect of this position of "violence as an incidental" is the diversity it stimulates. However, it also creates unfortunate distorting effects. First, when agencies that are focused primarily on related problems
RECOMMENDATIONS 348 such as substance abuse are major sponsors of violence research, funding levels for violence research rise and fall with the tide of funding in the related areas. Were interest in those areas to decline during the 1990s, funding for violence research under the current structure could be expected to fall regardless of either the social costs of violence or the expected social and scientific payoffs. Second, when individual researchers apply for support for violence research from a program concerned with another field, award decisions tend to be based on investigators' understanding of and attention to the program's primary focusânot their expertise in violent behavior. This raises the possibility that meritorious proposals in terms of contributing to the understanding and control of violence will be overlooked. Over time, proposers' anticipation of this trend may lead to the "slanting" of proposals to overemphasize the related field regardless of its importance as a cause of violence. Third, over time, violence research proposals come to be judged in terms of classification systems developed for the related fields. For example, in programs governed by psychiatric classification systems such as DSM-III-R, violent behavior is usually grouped with antisocial behavior because "repeated fights or assaults" is one diagnostic indicator for antisocial behavior. Unfortunately, because it is only one of eight such indicators, even an outstanding violence proposal may be rejected as unlikely to make a major contribution to the understanding of antisocial behavior. Barriers to Sustained Interdisciplinary Commitments Certain "big science" projects offer the prospect of contributions to the understanding and control not only of violence and its effects on individual victims and communities, but also of such related problems as drug and alcohol abuse and mental illness. Occasional large-scale programs are essential for significantly advancing the understanding and control of violence. Investigations of its causes should sometimes cross traditional academic boundaries, comparative evaluations of preventive interventions should sometimes span the traditional intervention modes of several public agencies, and both kinds of studies sometimes require investigators to follow large samples of individuals for long periods of time. Unfortunately, interdisciplinary studies that require long-term investments of large-scale financial support currently have no natural home among federal violence research sponsors. When recommendations
RECOMMENDATIONS 349 for longitudinal studies of the development of potentials for crime, delinquency, and prosocial behavior surfaced during the 1980s, no single agency routinely committed sufficient resources for sufficiently long periods, and no agency was in a natural position to organize a consortium of sponsors. To its credit, the Office of Juvenile Justice and Delinquency Prevention (OJJDP) undertook the Program on Causes and Correlates of Delinquency, which involved longitudinal studies of delinquent development in three sites and a commitment to five years of financial support. The National Science Foundation and other sponsors provided short-term funds to reduce the risk of a hiatus during which the research teams might be dismantled, respondent attrition might increase discontinuously, and the future cooperation of local schools and other data providers might be jeopardized. At this writing, pending reauthorization legislation earmarks funds to continue the program for one additional year, but long-term support is in jeopardy. In an innovative institutional arrangement, the National Institute of Justice has joined with the health program of the John D. and Catherine T. MacArthur Foundation to launch the Program on Human Development and Criminal Behavior. Creating the consortium and planning required four years of funding. At this writing funding commitments are ensured for the design and pretest phases only. Long-term funding for data collection is still being negotiated. The experiences of these programs illustrate both the need for a stable organizational framework for underwriting long-term comprehensive research enterprises and the perils of launching such enterprises without long-term commitments. Recent Promising Developments Since the panel's survey of fiscal 1989 support, a number of developments have strengthened the federal enterprise in research on violence. Although we have not systematically canvased all respondents, the developments of which we have become aware raise the prospects for positive complementary lines of evolution in violence research and evaluation and point to several programs that could usefully be continued, restored, coordinated, or expanded. National Institute of Justice The National Institute of Justice (NIJ), the research arm of the Department of Justice, operated a diverse violence research program
RECOMMENDATIONS 350 through fiscal 1990, which covered such topics as child abuse, partner assault, bias crimes, youth criminality and victimization, homicide and drugs, weapons, and psychiatric aspects of violence. A 1990 reorganization eliminated violence and most other stand-alone research programs, but violence is an announced program priority for fiscal 1992, and fiscal 1991 funds were awarded for research on such violence topics as police use of excessive force, serial rape, assaults on correctional officers, and less-than-lethal-force weapons. At the 1991 annual meetings of the American Society of Criminology, the institute's director announced the formation of new divisions of research and of evaluation, and plans to form external scientific advisory groups for all program priority areas, including violence. In addition, NIJ is continuing cosponsorship, with the MacArthur Foundation, of the Program on Human Development and Criminal Behavior, a longitudinal study of the development of prosocial, antisocial, and criminal behavior, including an emphasis on aggression and violent behavior. National Science Foundation The National Science Foundation (NSF) sponsors basic research through a set of programs defined in terms of academic disciplines. Although no program solicited research proposals specifically on violence, the programs within the Directorate for Biological, Behavioral, and Social Sciences collectively span the range of disciplines that are relevant to violent behavior: law and social science, sociology, anthropology, neural mechanisms of behavior, animal behavior, and behavioral neuroendocrinology. Each program maintains an "open window" for proposals and bases its awards on peer-reviewed comparisons of scientific merit and importance within its discipline. Because the various NSF program areas operate largely independently, the overall size and composition of its violence research portfolio is determined by a series of award decisions based on scientific merit assessed from specific disciplinary perspectives, rather than by any coordinated assessment of scientific priorities within the field of violence. A proposal for a coordinated initiative on violence research was considered several years ago but was not pursued. Program evaluation or other policy analysis lies outside the NSF mission, except insofar as it has direct scientific relevance. In late 1991, NSF announced the creation of separate directorates for the biological sciences and for the social, economic, and
RECOMMENDATIONS 351 psychological sciences. The hope has been expressed that, within the latter directorate, research on violent and nonviolent crime will have greater visibility (American Society of Criminology, 1991). It is also important that separation of the directorates does not hinder support of research that reflects an integration of social and neurobiologic perspectives on violent behavior. Centers for Disease Control Through its Center for Environmental Health and Injury Control (CEHIC), the Centers for Disease Control (CDC) has begun to sponsor initiatives intended to measure and reduce the public health consequences of violence: primarily intentional injuries and their aggregate consequences for the health care system. As of fiscal 1989, the reference year for the panel's survey, the program was emphasizing the following components: improving the measurement of intentional injuries and deaths, using epidemiological techniques to identify risk factors, establishing measurable objectives for reducing intentional injuries, designing and evaluating strategies for preventing intentional injuries by modifying risk factors, and mobilizing a range of professions, government agencies, and community-based organizations to pursue those strategies. Since the survey, CDC/CEHIC has completed the development of measurable objectives for reducing interpersonal intentional injury in Healthy People 2000 (U.S. Department of Health and Human Services, 1991). At this writing, it is in the process of drafting a position paper "Prevention of Violence and Violence Injuries," which sets out strategies for preventing intentional injuries and calls for evaluating interventions that pursue those strategies. National Institute of Mental Health At the time of the panel's survey, research on aggressive and violent behavior occurred in three branches of the National Institute of Mental Health: Behavioral Sciences, Neurosciences, and Antisocial and Violent Behavior. Of the three, the latter was the largest, and it sponsored basic research and program evaluations on family violence, on violent behavior by the mentally ill, and on the psychological consequences of violent victimization. At various times, that program supported longitudinal research, with its need for long-term support. During the 1980s, the branch rarely funded research on social-level causes of violent behavior. During 1990, the unit was renamed the Violence and Traumatic
RECOMMENDATIONS 352 Stress Research Branch, and its annual budget was nearly doubled, to $15.1 million. The reorganized program is divided into four research areas, which relate to (1) the causes, prevention, and treatment of perpetrators of interpersonal violence, with special note of family violence, violence against minorities, and homicide; (2) the victims of violence, especially the psychological consequences of victimization; (3) mentally disordered offenders and the violent mentally ill; and (4) victims of disaster, combat, community violence, and terrorism. In fiscal 1991, the program funded 90 research grants and 10 research training grants in these areas. Facilitating the Needed Work The problem facing policy makers is how to preserve diversity while achieving the following objectives: (1) facilitating more basic research on individual animal and human subjects and on organizations, situations, communities, and societies, specifically designed to add to scientific understanding of violent behavior and violent events; (2) integrating basic research more fully with the design, evaluation, and refinement of interventions to prevent the occurrence or recurrence of violence or to reduce and repair the physical, psychological, and other harms of violence to individual victims and community institutions; (3) fostering more interdisciplinary basic research into prosocial as well as aggressive and violent behavior, and evaluating a broader range of violence prevention interventions; (4) facilitating long-term commitments to research and evaluation efforts that could require a decade or more; (5) augmenting intellectual capacity through programs of predoctoral and postdoctoral fellowships; (6) calling attention to violence problems that are receiving insufficient, or insufficiently comprehensive, attention, and stimulating or providing support needed to fill such emerging gaps; and (7) maintaining and disseminating informationâabout ongoing research projects, new promising leads, and newly available findings and dataâto all the diverse communities concerned with violence.
RECOMMENDATIONS 353 NOTES 1To avoid repetitiveness, we are here using the term violence as shorthand to refer to a set of measures, including rates of violent events, of injuries due to violence, and of deaths due to violence. 2 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, 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 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). 3 The figure assumes an average professional salary of $50,000, plus a 40 percent allowance for fringe benefits and administrative costs. 4 Not included in these research expenditures are the expenditures for collection and production of violence statistics. These would include prorated shares of the costs for victimizations in the 1988 National Crime Survey (NCS, $7,000,000), for crimes reported to the police in the 1989 Uniform Crime Reports (UCR, $4,500,000), and for deaths in the 1989 national mortality statistics of the United States ($6,800,000). Prorating each program's costs proportionally to the fraction of incidents reported that would be classified as violent under the panel's definition produces a total estimate of $1,753,000 for violence statisticsâ$1,160,000 in NCS costs, $520,000 in UCR costs, and $73,000 in national mortality statistics costs. 5 Research expenditures for cancer; for heart, lung, and blood disease; and for AIDS are the 1989 actual budget authority figures reported in Public Health Service (1990), respectively, for the National Cancer Institute, for the National Heart, Lung, and Blood Institute (both excluding AIDS), and for AIDS research. Unlike our estimates for violence research, these figures include agency program administration costs; however, they exclude research expenditures elsewhere in the National Institutes of Health that are pertinent to their respective foci. References American Society of Criminology 1991 McCord testifies before NSF. The Criminologist 16(6):1,3. Blumstein, A., J. Cohen, J.A. Roth, and C. Visher, eds. 1986 Criminal Careers and "Career Criminals." Washington, D.C.: National Academy Press.
RECOMMENDATIONS 354 Farrington, David P., Lloyd E. Ohlin, and James Q. Wilson 1986 Understanding and Controlling Crime: Toward a New Research Strategy. New York: Springer-Verlag. Felson, R.B., and H.J. Steadman 1983 Situational factors in disputes leading to criminal violence. Criminology 21(1, February):59-74. 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. Public Health Service 1990 Justification of Appropriation Estimates for Committee on Appropriations, Fiscal Year 1992, Volume III. Washington, D.C.: Public Health Service. Sherman, L.W., P.R. Gartin, and M.E. Buerger 1989 Hot spots of predatory crime: Routine activities and the criminology of place. Criminology 27:27-55. Tonry, M., L.E. Ohlin, and D.P. Farrington 1991 Human Development and Criminal Behavior: New Ways of Advancing Knowledge. New York: Springer-Verlag. U.S. Department of Health and Human Services 1991 Healthy People 2000. Washington, D.C.: U.S. Government Printing Office.