The public health approach to the study and prevention of interpersonal violence was given formal recognition in 1984 when Surgeon General C. Everett Koop stated: "Violence is every bit as much a public health issue for me and my successors in this century as smallpox, tuberculosis, and syphilis were for my predecessors in the last century." As the injury and death toll from violent behavior have become increasingly evident, multidisciplinary scholarship in the study of violence has emerged and expanded at an unprecedented pace.
The most widely accepted definition of violence—sometimes termed "intentional interpersonal injury"—is: "behavior by persons against persons that intentionally threatens, attempts, or actually inflicts physical harm" (Reiss and Roth, 1993). The closely related terms "aggression" and "antisocial behavior" are generally applied to lesser forms of violence and include, but are not limited to, behaviors that are intended to inflict psychological harm as well as physical harm.
The public health approach to the study and prevention of violence entails a four-step process:(1) data collection of violence-related problems, assets, and resources; (2) assessment of the possible causes of violence through risk-factor identification; (3) the establishment and evaluation of violence prevention strategies; and (4) the dissemination and implementation of effective strategies. Public health, then, is inherently a research-driven and prevention-oriented science. This approach complements and overlaps with the narrower focus of criminology, which is primarily concerned with forms of violence that constitute crimes and with policies and practices that deter and punish perpetrators.
VIOLENT VICTIMIZATION
Epidemiological data on violence are derived from three primary sources: (1) hospital, emergency medical service, and medical examiner records;(2) police reports and arrest records (and other agency records, such as child protective services for reports of child abuse); and (3) self-report surveys and interviews. In addition, specialized studies that address the particular dynamics and contexts of violence have proven to be important to the understanding and prevention of violence.
The most complete and accurate violencerelated datasets are those on homicide victims. In the United States, the overall homicide victimization rate has fluctuated during the twentieth century from fewer than two homicides per 100,000 in 1900 to a high of nearly eleven homicides per 100,000 in 1980. In 1998, 17,893 individuals were murdered in the United States, which translates into an average daily death toll of forty-nine people. The worldwide 1998 homicide rate was 12.5 per 100,000, significantly higher than the U.S. homicide rate of 6.2 per 100,000. Nevertheless, data from the 1980s reveal that among the forty-one most developed countries, the United States has the third highest homicide rate.
Because violence is not evenly distributed throughout the population, these overall homicide rates provide only a partial picture of homicide's toll. Most notably, homicide victimization in the United States is most prevalent among youth. In 1998, homicide was the second leading cause of death among fifteen-to twenty-four-year-olds. Racial disparities in homicide rates are also disturbingly high. During the second half of the twentieth century, African Americans were murdered at five to eleven times the rate of their white counterparts. Gender differences are equally as dramatic, with males murdered at approximately ten times the rate of females. Finally, the risk of homicide is higher in urban than nonurban areas as well as within neighborhoods characterized by concentrated poverty. Neighborhood social disorganization also increases the probability of violence victimization as well as perpetration.
In comparative studies conducted in the 1990s, the homicide victimization rates in the United States, particularly among children and adolescents, were shown to be several times higher than those in any other industrialized country. In fact, the homicide rate for children under sixteen years old in the United States was five times higher than the corresponding homicide rate for the next twenty-five richest countries combined. The reasons for these elevated homicide rates in the United States are not fully understood; however, probable causes include easier access to firearms, more common and severe patterns of income disparities, and higher levels of racial and ethnic diversity in conjunction with racist and xenophobic attitudes and behaviors.
The question of mechanism, or the means by which people are murdered or injured, is another critical piece of information with respect to our understanding and prevention of violence. The examination of mechanism was particularly helpful in understanding the tremendous increase in homicide victimization rates of adolescents in the United States from 1987 to 1993, and the subsequent downturn through 1998. When the data are disaggregated by mechanism, a clear picture emerges: These trends over time can be accounted for by changes in the number and proportion of youth murdered with a firearm (see Figure 1). The changes in gun use during this period are generally attributed to three major factors: the crack epidemic—which had the effect of destabilizing local drug trafficking markets, rendering them more volatile and violent—and the subsequent petering out of this epidemic; changes in economic opportunity; and changes in policing policy for gun violations.
Most assaultive behavior, however, does not result in death. In 1997 more than 1.75 million people in the United States were treated for assaultive injuries in emergency departments, and more than 10 million individuals aged twelve and over reported that they had been victims of violent crimes. These and other data reveal that young people, African Americans, and males are disproportionately victimized by nonlethal forms of violence, though these disparities are less pronounced than for homicide victimization.
VIOLENCE PERPETRATION AND ASSOCIATED RISK FACTORS
The number and characteristics of individuals who commit murder cannot be precisely determined because of limitations in law-enforcement reporting systems and because identifying information about perpetrators are only available for cases in which an arrest is made or the perpetrator is otherwise identified. Still, some reasonably sound information about adolescents who murder is available: About nine in ten are male, more than half are African American, approximately half act alone, most kill individuals who are close in age and of the same ethnic background, and most use a firearm. The peak or modal age among homicide perpetrators occurs in the late teens and early twenties.
Since the rampage shooting at Columbine High School in the spring of 1999, much concern about violence at schools has been aired, and fears that such events could happen anywhere have emerged. Contrary to these perceptions, however, the number and rate of youth-initiated school violence—both lethal and nonlethal—generally decreased, or at worst remained relatively stable, during the 1990s. Like the homicide victimization rates, the overall juvenile homicide perpetration rate, as well as the aggregate juvenile offending rate for serious forms of violence, rose precipitously in the late 1980s, peaked during the early to mid-1990s, and then decreased through the beginning of the twenty-first century. School-based homicides constitute only 1 percent or fewer of all homicides committed by young people, and schools remain one of the safest environments for children and adolescents.
Equally important to estimating the scope of violence perpetration among youth are efforts to identify risk factors—the characteristics that when present increase the probability that a young person will subsequently engage in violent acts. There are five important aspects of risk factors. First, risk factors tend to be additive—the more risk factors that are present, the more elevated the risk of violence. A single risk factor generally has low predictive power. Even among those children and adolescents with multiple risk factors, few will become violent. Second, risk factors occur, and need to be addressed, at multiple levels, including individual, family, peer group, school, and neighborhood or community levels. Third, different risk factors pertain to different points in the lifespan, with family-level factors playing a greater role for younger children, and peer group and neighborhood factors playing a greater role for older children. Fourth, some risk factors are specific to certain types of violent behavior (e.g., risk factors for sexual violence may be quite different than those for robbery). And fifth, the severity of riskfactor exposure is likely to increase or decrease risk proportionately (e.g., extreme and chronic child abuse is likely to have a more profound effect than lesser forms of child maltreatment).
Several literature reviews have been undertaken on risk factors that increase the probability that children and young teens will subsequently engage in violent behavior. These reviews have sorted out risk factors into two categories: risk factors during the childhood years and risk factors
Figure 1
during the early adolescent years. Risk factors during infancy, and even perinatally, have also been identified, (e.g., child abuse and neglect). This entire body of research, however, is relatively new and far from exhaustive. Therefore, some factors that may in reality increase subsequent risk for violence perpetration may not have been identified in the extant literature because they have been inadequately researched or because of their complexity—the potency of a risk factor may be significantly affected by specific contextualized circumstances (e.g., bystander support), neighborhood norms, and personal history. Similarly, one factor may only become a risk factor, or may become a more potent risk factor, when it occurs in tandem with another factor.
During childhood, the two most powerful predictors of subsequent violence perpetration are substance use and delinquency. Additional, less potent risk factors include aggressive behavior; family violence; inconsistent, overly lax, and harsh disciplinary practices; association with antisocial peers; and poor attitudes toward schooling. Media violence has been shown to increase aggression in the short term, but such exposure has not been linked directly to violent adolescent behavior. Conversely, attempts to reduce violence through media advocacy (e.g., the "Squash It" campaign) have not been shown to reduce rates of violence significantly.
During the early adolescent years, three major and interrelated risk factors have been identified: weak associational ties with nondelinquent peers; strong associational ties with antisocial and delinquent peers; and gang membership. Gang membership, in particular, appears to fulfill important psychological needs with regard to peer acceptance and belonging, as well as the need for enhanced social status, particularly for unpopular youth and for those youth who feel socially powerless. Because gangs serve these fundamental needs, efforts to dissuade young people from joining youth gangs is a more efficient strategy than trying to entice them out of the gang after they have joined, particularly since gangs typically promise to provide valued incentives such as money, power and status, excitement, and, for males, promises of sexual "favors." On the other hand, to ignore current gang members, or rely exclusively on punitive law enforcement efforts, is an inefficient and ineffective violence reduction strategy. Community-based outreach efforts in association with community policing operations are required. Such efforts need to address the psychological, interpersonal, and economic needs of gang members; they should be based upon multiple sources of information about local gang activity; and they should include collaborative efforts involving the police, schools, social service agencies, former gang members, and grassroots organizations.
Additional risk factors during the early adolescent years include antisocial behavior, attending a school in which gangs are prevalent, having been a victim of a violent crime, and residing in a high-crime neighborhood and/or in neighborhoods that have high levels of social disorganization.
While quantitative risk factor analyses are important, qualitative studies based on in-depth interviews, focus groups, and intensive field studies of particular groups of youth provide insights into the dynamics underlying risk-factor analyses and point to additional factors, or combinations of factors, that may be fruitful to study. These studies are important given the generally weak overall predictive power yielded from risk-factor analyses. Examples of such studies include James Garbarino's 1999 study of children and adolescents who have committed violent crimes, Elijah Anderson's 1999 study of the impact of street and cultural norms in an impoverished African-American section of Philadelphia, John Devine's extensive 1996 field studies of school violence in New York City, and Felix Padilla's in-depth 1992 study of the dynamics and culture of a Latino gang in Chicago. These richly textured studies, and others like them, capture the complex and tragic nature of acts of violence. They also provide insights about the psychological logic and developmental history of those who commit violent acts, reminding us that even the most vicious forms of violence can ultimately be understood, though not justified, as uniquely human responses to a volatile mix of difficult circumstances and experiences combined with specific personality and character dynamics.
PREVENTION STRATEGIES
Four major interrelated approaches to the prevention of violence have been articulated: (1) the inculcation or enhancement of protective factors (factors that reduce the probability of violence perpetration among individuals exposed to known risk factors) and/or a corresponding reduction in the number or severity of risk factors, (2) the adoption of self-contained violence prevention programs, (3) the specification of generic strategies (e.g., social skills training) derived by grouping effective and promising programs according to the approach they adopt and the specific program characteristics they utilize, and (4) the elucidation of framing principles that guide the establishment and implementation of programs.
The use of mechanical and electronic surveillance devices (e.g., metal detectors), and the establishment of laws, law enforcement policies, judicial processing, and incarcerative practices remain primarily in the domain of criminology and need to be better integrated with public health approaches. One successful example of this kind of comprehensive and integrated approach was established in Boston. This strategy involved several agencies and programs working together to reduce gun and gang-related violence. The police, probation officers, and courts addressed surveillance, interdiction, and enforcement; legislators passed tougher penalties for gun-related violence; researchers conducted analyses of gun violations; and social workers and religious leaders counseled at-risk youth in the use of nonviolent conflict resolution techniques and offered employment opportunities and program activities. Other approaches to violence prevention, such as changes in public policies,(e.g., foster care policies, school reform, and employment and housing strategies), have received only passing attention within the public health field, with the notable exception of the significant attention paid to firearm policies.
The study of protective factors has been spurred by the long-standing observation that some children who are exposed to several known risk factors do not become violent or otherwise seriously impaired. The task, then, is to identify common characteristics or circumstances that buffer these resilient children from the ill effects of exposure to known risk factors. The scientific study of protective factors, however, is in its infancy and the evidence from this small body of literature is suggestive rather than conclusive.
The most well-documented protective factor is maintaining conventional values, including the rejection of aggressive or violent behavior as an appropriate means to resolve conflict. This characteristic is associated with the peer-level protective factor of associating with peers who hold prosocial values. At the family level, a warm and supportive relationship with one's parents or guardians and engagement in familial bonding activities have been associated with reduced levels of aggression.
As children move into the more high-risk adolescent years, family factors alone do not continue to exert a powerful protective effect. The innoculative effects of protective factors appear to require developmentally appropriate exposures at each stage of development. At the school level, commitment to school has been identified as a protective factor. Finally, because neighborhood and societal change are so difficult to study in controlled studies, and also so challenging to address, protective factors at these levels have not been identified.
The development and implementation of self-contained violence prevention programs has a long-standing history. The introduction of scientific methods to assess the effectiveness of such programs, however, only commenced in the 1980s, with the number and rigor of such evaluations accelerating rapidly during the 1990s. Still, scientific evaluations are very costly and only a small proportion of programs now in use at schools and in communities have been rigorously evaluated.
The programs that have been evaluated are generally highly structured, implemented by professionals, and developed at academic institutions. While this body of research has revealed that some programs do indeed reduce rates of aggression and violence (and that some programs clearly do not work), it is inaccurate to assume that programs that have not been evaluated do not work, or conversely, that they are effective.
It is also inaccurate to conclude that programs that have been shown to be effective will work equally well in all settings and contexts. Very little is known about whether, or how, programs need to be adapted from one setting to another. Some programs may not work equally well for males and females, some may work well in urban but not rural settings, and some programs may work in one cultural context but not another. Some programs are appropriate for all children or youth within a designated age range (universal, or primary, prevention), some are appropriate for children and youth exhibiting or possessing known risk factors (selective, or secondary, prevention), and some programs are appropriate for youth who have already engaged in violence or serious delinquent behavior (indicated, or tertiary, prevention).
In the late 1990s and early 2000s, several compendia of effective, promising, and ineffective violence prevention programs were issued. These include reports by the Surgeon General, the Centers for Disease Control and Prevention, the Center for the Study and Prevention of Violence, the Office of Justice Programs, the National Research Council, the Violence Institute of New Jersey, M. W. Lipsey and D. B. Wilson (1998), and M. B. Greene (1998). The major strategies that have been shown to be effective, along with brief descriptions of illustrative programs, are summarized below; however, readers interested in a full explication of such strategies, along with detailed descriptions of effective and promising programs, are urged to consult sources listed in the bibliography.
The most widely adopted violence prevention strategy emphasizes social skills training to resolve conflict without resorting to aggressive or violent tactics. Social skills training programs generally utilize structured and interactive curricula (e.g., role playing) and are usually classroom based. One example of an effective social skills training programs is Promoting Alternative Thinking Strategies, or PATHS. This program is designed for children from kindergarten through fifth grade and focuses on five specific skills: emotional literacy, self-control, social competence, positive peer relations, and interpersonal problem solving. School-based sessions are taught for approximately thirty minutes each, and the program developers recommend that these lessons should be taught three times per week.
A second overall strategy focuses on parent training and family dynamics. This approach is both educational and therapeutic and based on the theory that a caring, supportive, and stable family life will provide the initial grounding to deter children from subsequently engaging in aggressive, delinquent, or violent behavior. Most commonly, programs are designed to work with parents of young children and are focused on parental decision making, communication, monitoring and sanctioning strategies, and on educating parents about child development. Several family-based strategies have been shown to be highly effective in reducing aggressive and/or violent behavior.
Home visitation, in which therapeutic guidance is provided to parents in their residence, has gained much recognition in recent years. One of the most effective home visitation programs is the Nurse Home Visitation Program, in which a trained nurse visits the home setting during the latter stages of pregnancy through the point at which the child reaches age two. Long-term follow-up studies indicate that the adolescent children of program participants had significantly fewer arrests than control-group adolescents. In addition, two family-oriented programs for adolescents who have exhibited violent and delinquent behavior have also been shown to be effective: Functional Family Therapy and Multisystemic Therapy. Both programs provide intensive family and individual therapy, as well as guidance to parents in addressing practical and everyday problems, and both have effected significant reductions in subsequent delinquent and violent behavior.
A defining feature of a third approach to violence prevention is the central role played by young people in the program's operation and implementation. Four principles underlie such programs: (1) young people understand their own peer culture and what kinds of program components are feasible; (2) young people provide a typically untapped human resource; (3) program norms are more readily diffused through the network of involved youth; and (4) the involvement by young people in implementing such programs provides an alternative for antisocial, violent, and delinquent behavior. The most popular of this class of programs is school-based peer mediation, in which a trained student mediates a dispute between two other students with the goal of establishing a mutually agreed-upon peaceful solution. Other types of programs engage young people in community organizing or advocacy activities. While the small number of peer-operated programs that have been rigorously evaluated has not shown significant reductions in violent or delinquent behavior, the theoretical promise of these programs, the fact that many types of youth-led programs have not been evaluated, and the inherent complexity in evaluating such programs suggest that a decision to forgo or eliminate such programs is premature. Nevertheless, sound policy also suggests that programs should be discontinued if they continuously fail to demonstrate their effectiveness.
Another class of programs utilizes psycho-educational strategies to reduce the likelihood of engagement in violent behavior. The most wellknown type of program within this class of programs is mentoring. While not all mentoring programs are effective, the Big Brothers Big Sisters program model has been rigorously evaluated and shown to be an effective violence prevention strategy. Stand-alone individual counseling, however, is considered an ineffective violence prevention strategy.
Another type of program involves counseling and supportive services for youth who have been exposed to violence, either as victims or as witnesses—both of which are risk factors for subsequent perpetration. In one such program, the Child Development Community Policing Project, police officers receive training in child development and the dynamics of psychological trauma and work together with mental health clinicians—who receive training in police practices and culture—in identifying and responding to children who have been exposed to violence. This program illustrates the potential value of integrating clinical and law enforcement approaches.
Finally, some programs are hybrids, either combining two or more of the approaches outlined above or not fitting neatly into any of the four approaches. One "hybrid" is Olweus's Bullying Prevention Program. This program has several key features, including skills-based classroom training, parent involvement, policy development, "hot spot" analysis, and counseling. Evaluations of this program suggest that it is effective in reducing levels of bullying and harassment. Indeed, multicomponent programs are generally viewed as preferable, particularly for high-risk youth.
Public health efforts to address gun-related violence also do not fit neatly into any of the approaches outlined above. Strategies to reduce gun violence include the promotion of laws and policies that reduce access to guns (some evidence of effectiveness); the adoption of mechanical and electronic means to make guns safer, such as trigger locks and personalized guns (the consistency and quality of such devises are variable and none has been adequately evaluated); educating children in safe gun practices (ineffective); gun buybacks (ineffective); and public information campaigns (no evaluations have been conducted).
As indicated above, an alternative way to approach violence prevention programming is by establishing a set of framing principles that inform their development. While this cannot be done without examining what is known from evaluation studies and from risk and protective factor analyses, it is too early in the evolution of such studies to simply extract these principles from the programs that have been subject to rigorous evaluation and proven to be effective or promising. Some of the principles listed below, therefore, owe more to findings in other areas of public health than they do to the violence prevention field per se. Some principles have been described in earlier parts of this article (e.g., that no single program or approach works equally in all settings and circumstances). What follows is a brief though not exhaustive list of such principles.
The first principle, known as local ownership, suggests that programs will be most successfully operated if the residents in the targeted neighborhood and the specific group of individuals for whom the program is designed to help are centrally involved in the planning, operation, and administration of the program. A second principle multidisciplinarianism, suggests that insights, methods, and approaches from multiple disciplines are needed in developing and implementing violence prevention programs. A third principle, collaboration, suggests that no single agency or group can successfully operate a program in isolation: Violence prevention programs are inherently neighborhood-based and require the engagement of multiple stakeholders.
A fourth principle suggests that a strength-based focus should be emphasized—focusing exclusively on deficits without drawing upon the strengths and interests of the individuals the program is designed to help and the resources available in the community will reduce the probability of success. A fifth principle suggests that committed leadership is necessary for the successful planning and implementation of violence prevention programs. Similarly, staff development is also critical: An untrained, unsupported, and unsupervised staff simply will not succeed in program implementation. Staff also need to be temperamentally suited to the populations with which they work.
Program accessibility is also critically important: If a program is sited in an undesirable location (turf issues are very important for young people), is sited in a difficult-to-get-to location, or is physically unwelcoming or uninviting, then the program will simply not attract participants. Specificity is also important: Programs need to set specific and measurable objectives, otherwise they tend to flounder and evaluation is rendered unfeasible. A final principle is local fit; A program's design and objectives should be derived from a thorough and multipronged assessment of the nature and extent of the violence-related problems in the neighborhood in which the program will be implemented. Additionally, new programs need to fit well into the context of existing programs and strategies.
Perhaps it is fitting to end with a quote from Surgeon General David Satcher, taken from his preface to the Surgeon General's report on youth violence: "As a Nation, we possess knowledge and have translated that knowledge into programs that are unequivocally effective in preventing much serious youth violence."
Alexander, J. F. (1998). Functional Family Therapy. Boulder, CO: Institute for Behavioral Science.
Anderson, E. (1999). Code of the Street: Decency, Violence, and the Moral Life of the Inner City. New York: Norton.
Arredondo, S.; Aultman-Bettridge, T.; Johnson, T.; Williams, K.; and Ninneman, L. (1998). A Study of YouthHandgun Violence. Boulder, CO: Center for the Study and Prevention of Violence.
Berkowitz, L. (1993). Aggression: Its Causes, Consequences and Control. New York: McGraw-Hill.
Blum, R. W.; Beuhring, T.; and Rinehart, P. M. (2000). Protecting Teens: Beyond Race, Income and Family Structure. Minneapolis, MN: Center for Adolescent Health, University of Minnesota.
Blumstein, A., and Wallman, J., eds. (2000). The Crime Drop in America. New York: Cambridge University Press.
Centers for Disease Control and Prevention (1997). "Rates of Homicide, Suicide, and Firearm-Related Death among Children—26 Industrialized Countries." Morbidity and Mortality Weekly Report 46(5):101–105.
Devine, J. (1996). Maximum Security: The Culture of Violence in Inner-City Schools. Chicago: University of Chicago Press.
Elliott, D. S., ed. (2001). Youth Violence: A Report of the Surgeon General. Atlanta, GA: Office of the Surgeon General.
Elliot, D. S., and Tolan, P. H. (1999). "Youth Violence, Prevention, Intervention, and Social Policy." In Youth Violence: Prevention, Intervention, and Social Policy, eds. D. J. Flannery and C. R. Huff. Washington, DC: American Psychiatric Press.
Fagan, J., and Wilkinson, D. L. (1998). "Social Contexts and Functions of Adolescent Violence." In Violence in American Schools, eds. D. S. Elliot, B. A. Hamburg, and K. R. Williams. New York: Cambridge University Press.
Federal Bureau of Investigation (1993). Age-Specific Arrest Rates and Race-Specific Arrest Rates for Selected Offenses: 1965–1992. Washington, DC: Author.
Garbarino, J. (1999). Lost Boys: Why Our Sons Turn Violent and How We Can Save Them. New York: Free Press.
Gilligan, J. (1996). Violence: Reflections on a National Epidemic. New York: Random House.
Greene, M. B. (1996). "Youth and Violence: Trends, Principles and Programmatic Intervention." In Minefields in Their Hearts: A Mental Health Handbook for Children in War and Communal Violence, eds. R. J. Apfel and B. Simon. New Haven: Yale University Press.
—— (1998). "Youth Violence in the City: The Role of Educational Interventions." Health Education and Behavior 25(2):175–193.
Henggeler, S. W. (1997). Treating Serious Antisocial Behavior in Youth: The MST Approach. Washington, DC:U.S. Department of Justice.
Herrenkohl, T. I.; Magiun, E.; Hill, K. G.; Hawkins, J. D.; Abbott, R. D.; and Catalano, R. F. (2000). "Developmental Risk Factors for Youth Violence." Journal of Adolescent Health 26:176–186.
Howell, J. C. (2000). Youth Gang Programs and Strategies. Washington, DC: Department of Justice.
Krug, E. (1999). Injury: A Leading Cause of the Global Burden of Disease. Geneva: World Health Organization.
Lipsey, M. W., and Derson, J. H. (1998). "Predictors of Violent or Serious Delinquency in Adolescence and Early Adulthood: A Synthesis of Longitudinal Research." In Serious & Violent Juvenile Offenders, eds. R. Loeber and D. P. Farrington. Thousand Oaks, CA: Sage Publications.
Lipsey, M. W., and Wilson, D. B. (1998). "Effective Intervention for Serious Juvenile Offenders: A Synthesis of Research." In Serious and Violent Juvenile Offenders, eds. R. Loeber and D. P. Farrington. Thousand Oaks, CA: Sage Publications.
Marans, S., and Berkman, M. (1997). Child Development— Community Policing: Partnership in a Climate of Violence. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.
McCord, J.; Widom, C. S.; and Crowell, N. A., eds. (2001). Juvenile Crime, Juvenile Justice. Washington, DC: National Academy Press.
Mercy, J. A.; Rosenberg, M. L.; Powell, K. E.; Broome, C. V.; and Roper, W. L. (1993). "Public Health Policy for Preventing Violence." Health Affairs 12(4):7–25.
Milne, J. S., and Hargarten, S. W. (1999). "Handgun Safety Features: A Review for Physicians." Journal of Trauma Injury 47(1):145–150.
Murphy, S. L. (2000). Deaths: Final Data for 1998. Atlanta, GA: Centers for Disease Control and Prevention.
Nourjah, P. (1999). National Hospital Ambulatory Medical Care Survey: 1997 Emergency Department Summary. Atlanta, GA: Centers for Disease Control and Prevention.
Office of Juvenile Justice and Violence Prevention (1999). Promising Strategies to Reduce Gun Violence. Washington, DC: Author.
Olds, D. (1998). Prenatal and Infancy Home Visitation by Nurses. Boulder, CO: Institute for Behavioral Science.
Olweus, D. (1993). Bullying at School. Malden, MA: Blackwell Publishers.
Padilla, F. (1992). The Gang as an American Enterprise. New Brunswick, NJ: Rutgers University Press.
Petersilla, J. (1990). "Conditions that Permit Intensive Supervision Programs to Survive." Crime and Delinquency 36:126–145.
Rand, M. (1998). Criminal Victimization 1997. Washington, DC: Bureau of Justice Statistics.
Reiss, A. J., and Roth, J. A., eds. (1993). Understanding and Preventing Violence. Washington, DC: National Academy Press.
Rennison, C. M. (2000). Criminal Victimization 1999: Changes 1998–99 with Trends 1993–99. Washington, DC: U.S. Department of Justice.
Sampson, R.; Raudenbush, S. W.; and Earls, F. (1997). "Neighborhoods and Violent Crime: A Multilevel Study of Collective Efficacy." Science 277:918–924.
Sherman, L. W.; Gottfredson, D.; MacKenzie, D.; Eck, J.; Reuter, P.; and Bushway, S. (1997). Preventing Crime: What Works, What Doesn't, What's Promising. Washington, DC: Office of Justice Programs.
Snyder, H. N., and Sickmund, M. (1999). Juvenile Offenders and Victims: 1999 National Report. Pittsburgh, PA: National Center for Juvenile Justice.
Steinberg, L. (2000). "Youth Violence: Do Parents and Families Make a Difference?" National Institute of Justice Journal 243:30–38.
Thorton, T. N.; Craft, C. A.; Dahlberg, L. L.; Lynch, B. S.; and Baer, K. (2000). Best Practices of Youth Violence: A Sourcebook for Community Action. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control.
Tierney, J. P.; Grossman, J. B.; and Resch, N. L. (1995). Making a Difference: An Impact Study of Big Brothers Big Sisters. Philadelphia, PA: Public/Private Ventures.
U.S. Department of Education (1999). 1999 Annual Report on School Safety. Washington, DC: Author.
Violence Institute of New Jersey (2001). Source Book of Drug and Violence Prevention Programs for Children and Adolescents.
Wintemute, G. J. (1999). "The Future of Firearm Violence Prevention." Journal of the American Medical Association 282(5):475–478.
