Introduction
Violence is one of the most serious threats to the health of youth in the United States and around the world. In America, young people are disproportionately represented as both victims and perpetrators of violence. National school-based data indicate that violence is prevalent in many schools. The 1993 Youth Risk Behavior Survey indicates that 42 percent of adolescents were in a physical fight during the 12 months preceding the survey and 22 percent carried a weapon during the 30 days preceding the survey (Kann et al., 1995). Another study found that 50 percent of boys and 25 percent of girls reported being physically attacked by someone at school (Centers for Disease Control and Prevention [CDC], 1992). The CDC also reported 105 violent deaths in school in the two-year period from 1992 to 1994. The National Center for Education Statistics (NCES) found that more than half of public schools experienced some crime during the 1996-97 school year, and one in ten schools reported at least one serious violent crime during that year (NCES, 1998). In Minnesota, there were 1119 incidents involving dangerous weapons in the 1996-97 school year (Children, Families and Learning, 1998).

Combating In-School Violence
The National Center for Education Statistics (1998) reports that 78 percent of public school principals indicated having some formal school violence prevention or reduction program. Research has shown that types of interventions are numerous and vary widely. In the following table, we have listed a series of school-based violence prevention interventions.

Best Practices
This is a summary of "best practices" in school-based violence prevention. The information was obtained from reviews of the literature and computer databases, and phone interviews with Minnesota state agency prevention staff and representatives of the Minnesota Principals' Association. School-based violence prevention activities have been targeted at risk and protective factors operating at several levels: individual, family, community, and society (Edari and McManus, 1998; American Psychological Association, 1993; our Table 2). The following review will identify strategies employed at each level, their relation to existing research on risk and protective factors, and what is known regarding their effectiveness in reducing youth violence.

Individual-Level Interventions
At present, much of the policy discussion regarding youth violence has been grounded in the belief that this problem results primarily from the dysfunctional behavior of individual students (Hyman and Perone, 1998). Consequently, many existing interventions have been targeted at individual students with problematic behavior. Because interventions at an individual level are easier to implement, and their effects easier to measure, their value may be exaggerated (Tolan and Guerra, 1998). Webster (1993) has noted the intuitive appeal of school-based conflict resolution programs in spite of insufficient demonstrations of their effectiveness. On the other hand, a comparison of two different middle school violence prevention programs by Durant et al. (1996), found that the curriculum including conflict resolution was more successful in reducing severe physical fights that required medical attention. Because there are so few well-designed evaluations that include behavioral outcomes, the effectiveness of conflict resolution and peer mediation programs has not been determined (Tolan and Guerra, 1998). Incorporating individual-level approaches into broader interventions, ensuring that they are developmentally appropriate, and promoting extension of skills into real-life situations appear to be critical to facilitating the success of interventions at this level.

Family-Level Interventions
Multiple sources have concluded that family involvement is an essential component of promising programs (Dusenbury et al., 1997; Drug Strategies, 1998; Minnesota Department of Education, 1992), while others have classified it as potentially the most promising type of intervention implemented thus far (Wilson-Brewer, 1995; Tolan and Guerra, 1998).

Tolan and Guerra (1998) recommend three approaches that appear to be most successful in decreasing adolescent delinquent behavior: teaching behavioral parent training to decrease negative parenting and coercive and dominating styles of interaction that put children at risk for increased aggression and later delinquency; promoting activities that encourage emotional cohesion and shared values within a family; and aiding families to address external demands and stresses that may contribute to violence and a lack of connectedness.

It is critical to understand that many youths who commit violent acts have been witnessing violence as a standard means of family interaction their entire lives. Attempting to reverse this trajectory through interventions targeted at adolescents will certainly be much more laborious, and probably less effective, than strategies aimed at reducing family violence earlier in children's development. Therefore, in addition to considering how to promote non-violent behavior among teens through targeting the family environment, strategies that can be employed at an early age should also be widely implemented. Home visiting programs by nurses and other trained care givers, starting immediately after birth, have demonstrated effectiveness in reducing child abuse, child injury, and arrests of children up to age 19 (National Institute of Justice, 1998; Olds et al., 1998).

Community-level Interventions
Peer Group
Peer acceptance is important to all of us, but it becomes particularly critical as youths enter early adolescence and progress through high school. Dominant norms are powerful enough that, even if an adolescent opposes violence, perceived group norms that violence is appropriate and even desirable may supercede individual beliefs (Drug Strategies, 1998). Intuitively it would seem that an appropriate response to deal with the possibility of negative peer influence would be to remove the most aggressive students from the classroom. However, these approaches may not be the most effective means of tackling the problem, and may in fact result in perpetuating it. An evaluation of the "St. Louis Experiment" (Feldman, 1992) suggests that integrating peers who have better social skills and behaviors with at-risk youth results in more positive outcomes than separating the two. In this study there was a decrease in antisocial behavior among at-risk youth, while better-socialized youth retained their beliefs and behaviors in the integrated groups.

School Physical Environment
Using metal detectors and other security systems has shown some promise in reducing school-related weapon-carrying, although these approaches appear to have no effect in reducing threats and fighting in schools (CDC, 1993). A school system in New York City implemented a system-wide metal detection program by equipping security staff with hand-held metal detectors and conducting unannounced searches of students at a cost of $300,000 per year per school. (Wilson-Brewer and Spivak, 1994). In addition to confiscating over 2,000 weapons, all weapon-related incidents decreased, attendance increased, and anecdotal evidence indicates that students reported feeling safer at school. Of note, this intervention was implemented concurrently with several other interventions including violence prevention curricula, peer mediation programs, and crisis intervention teams.

In addition to being extremely costly, a fundamental drawback to environmental changes, such as metal detectors, is that it can send a message to children that they are not to be trusted, which threatens their sense of school-connectedness and self-worth. Hence, not only are metal detectors unlikely to reduce violent behavior, they may actually encourage it, by lowering student morale. Schools must decide how to balance the need for immediate security measures with the possible negative implications of such measures (Drug Strategies, 1998). In certain school environments, metal detectors may be necessary, but they will only be effective if combined with actions that demonstrate that the school values the students (Drug Strategies, 1998).

School Culture and Norms
While physical environment changes can be implemented relatively quickly, altering school culture will require long-term efforts. But these efforts are critical to the primary prevention of school violence. Teaching practices, discipline strategies, rules and policies, and school-wide norms regarding violence all impact the attitude and motivation of students.

Appropriate teaching practices can impact school-connectedness in several ways. Teachers are in a critical position to help students participate in and derive meaning from school tasks. Baker (1998) describes the unintentional tendency of teachers to regard high and low achievers differently in the classroom, perpetuating the low school-connectedness of those experiencing less academic success. Teachers must be adequately trained in the most successful methods to promote youth development.

Nationally and in Minnesota, suspension has been widely-employed to address juvenile delinquency. According to a survey conducted by the National Center for Education Statistics (1998), out-of-school suspension lasting five or more days was the most commonly-reported action taken against students for fighting. Costenbader and Markson (1998) conducted research on a group of students who were suspended for aggressive or violent behavior. When suspended, the most common response of students was that they felt "angry" or "happy to get out of the situation." A majority of the students found suspension not at all helpful or of little use. Many students indicated they believed they would be suspended again. Although suspension may be necessary in some cases, it should be employed with the knowledge that it can exacerbate risk factors such as poor academic performance, feeling disconnected from school, stigmatization by peers, and increased exposure to the delinquent subculture (Costenbader and Markson, 1998). Furthermore, suspensions have not been shown to reduce violent or disruptive behavior in school (Wilson-Brewer, 1995).

Neighborhood/Community
Research on risk and protective factors indicates that in addition to the importance of family-connectedness, good relationships with any adult--a teacher, coach, clergy member, neighbor--may serve as protective for youth. Several studies suggest the effectiveness of programs designed to change the roles of at-risk youth in their communities and increase their motivation to behave in a pro-social manner (Tolan and Guerra, 1998).

Society-level Interventions
Efforts that reduce the accessibility of handguns, alcohol and drugs, and media violence exposure in the environment of youth are likely to reduce the frequency and lethality of adolescent violence (Edari and McManus, 1998; Tolan and Guerra, 1998).

Best Bets Based on the Literature
Throughout the literature, there is evidence that some interventions have been examined thoroughly enough to demonstrate effectiveness, ineffectiveness, or promise (Drug Strategies, 1998; National Institute of Justice, 1998; Tolan and Guerra, 1998). Below is a concise overview of these findings:

Views from Minnesota
Minnesota schools face the same kinds of safety concerns as any other urban or rural schools in the country. Is the level of violence increasing or decreasing in Minnesota? And what violence prevention and intervention strategies are Minnesota schools employing to create safe schools?

Minnesota has a system of independent school districts, thereby impeding a statewide violence prevention strategy. Nonetheless, school districts are implementing violence prevention strategies of varying degrees. A sampling of strategies was obtained through surveys conducted with four elementary and four secondary school principals, representing urban, suburban and rural school districts. In addition, a survey was conducted with the Safety and Security Coordinators of the largest urban school district in Minnesota. The quotes interspersed in the next few paragraphs are those of the survey respondents.

All but one school, an urban alternative (drop-in/re-entry) school serving 9-12th grade, reported having very few incidences of violence; in addition, statistics indicate a downward trend. Most incidents, as described by the school principals, involve pushing/shoving, verbal confrontations, acting out in class, and being disrespectful to the teacher or other students. The schools reported very few "real fights," that is fights involving hitting.

All schools take a multi-dimensional approach to violence prevention and intervention, simultaneously implementing different types of intervention as described in Table 1. Without exception, creating a culture of nonviolence was the keystone to their success in creating a safe school. Seven out of eight school principals explained that a culture of nonviolence is created through messages and actions conveying that teachers and staff of the school care about, value, and hold high academic and behavioral expectations for students. Elementary schools, in particular, have school-wide slogans, such as "gentle touch and kind words," or "use wisdom," to remind students of their responsibilities to themselves and others. Most of the principals emphasized the importance of everyone in the school treating students with respect, being consistent with discipline and communicating clearly about consequences, having teachers who really believe that all students can succeed, and creating opportunities for students to be heard if they believe they've been treated unfairly. In the end, "it's the personal relationships that make the difference."

All elementary schools have some variation of a social skills curriculum that teaches decision-making, working with others, and respecting others through words and actions; one third grade is using the curriculum to meet the requirements of the profiles for learning. All schools have an intensive social skills curriculum for students who are repeatedly disruptive. Except for an alternative school, the secondary school principals reported very little in the way of school-wide individual social skill building. All elementary schools surveyed train students to be peer mediators; one secondary school is planning to implement such a program next year. All report success with the method and credit the success, in part, to the staff who model conflict resolution techniques in their relationships with students and with other adults in the school.

Only the secondary schools made environmental/technological changes to improve safety, including hand-held metal detectors used intermittently, magnetic doors, and adding a lunch hour to reduce the number of students in the lunchroom. All schools surveyed monitor their halls as well as the lunchroom. The Safety and Security Coordinator reported having a School Liaison (security staff) in every school. One elementary school needed to replace the playground and elected to enlarge it so that students were not playing on top of one another, thereby reducing the potential for conflicts.

All schools identified a need for additional resources, defined almost exclusively as financial. These resources would be used for (in order of frequency stated):

• Reducing class size and teacher/student ratios;
• A school counselor in every school;
• A second adult to be on school buses;
• Services for parents;
• More health professionals (nurses) for the schools;
• More space (one principal gave up her office to have a classroom to teach reading).

One secondary school principal in a school serving nearly 1600 students said that the issue wasn't money; rather, it was allowing schools to "have the autonomy to make decisions with and for students, staff and teachers."

In addition to school-based efforts, every principal and the Safety and Security Coordinator spoke at length about the importance to address the factors outside of school that create a culture of violence and that schools need to promote messages of peace as vigorously as society promotes messages of violence. Everyone spoke of the importance of reaching parents and helping them develop or improve effective parenting skills. Other factors identified include:

• The pervasiveness of media violence, especially where there is no evidence of pain or of consequences associate with the violence;
• The amount of violence some children witness in their community;
• The high mobility of many students and their families;
• The unwillingness of communities to view raising children as a responsibility of every member of that community.

Finally, everyone surveyed expressed the concern that society believes that schools are solely responsible for preventing youth violence in schools. Rather, as one teacher stated, "It's all of our responsibility." The schools are a great delivery system and they require the support of society at-large to make them successful.

Recommendations
In conclusion, the literature and survey findings support the following recommendations regarding "best practices" in developing school-based violence prevention interventions. Interventions must be based on theoretical models derived from the scientific research on youth violence; interventions that include a combination of violence prevention strategies are likely to be the most effective. In order to further our knowledge about what works and what doesn't work in reducing school violence, all interventions must be carefully evaluated to determine their effects on violence and related behaviors. Given the complexity of the problem of youth violence, initiatives that link school and community agencies and programmers with researchers are likely to promote the development and implementation of successful intervention strategies and rigorous evaluation designs.

Authors' Note: The following are examples of interventions that possess qualities described in our recommendations:

The Second Step Curriculum (Grossman et al, 1997).
The P. A. L. S. Program (Guth, 1995).

American Psychological Association Commission on Youth Violence. 1993. Violence and youth: Psychology's Response. Washington, DC: American Psychological Association.

Baker, Jean. 1998. "Are We Missing the Forest for the Trees? Considering the Social Context of School Violence." Jl of School Psych. 36:1, pp.29-44.

Blum, Robert W. and Rhinehart, P.M. 1997. Reducing the Risk: Connections that Make a Difference in the Lives of Youth. Division of General Pediatrics and Adolescent Health, University of Minnesota.

Bry, B.H. 1982. "Reducing the Incidence of Adolescent Problems through Preventive Intervention: One- and Five-Year Follow-Up." Am Jl of Community Psych. 10:3, pp.265-76.

Centers for Disease Control and Prevention. 1992. Proceedings of the Third National Injury Control Conference. Atlanta, GA: Centers for Disease Control.

Centers for Disease Control and Prevention. 1993. "Violence-Related Attitudes and Behaviors of High School Students-New York City, 1992." MMWR. 42:40, pp. 773-77.

Costenbader, Virginia and Samia Markson. 1998. "School Suspension: A Study with Secondary School Students." Jl of School Psych. 36:1, pp.59-82.

Drug Strategies. 1998. Safe School, Safe Students. Washington, DC: Drug Strategies.

Durant, Robert H. et al. 1996. "Comparison of Two Violence Prevention Curricula for Middle School Adolescents." Jl of Adol Health. 19, pp.111-17.

Dusenbury, Linda et al. 1997. "Nine Critical Elements of Promising Violence Prevention Programs." Jl of School Health. 67:10, pp.409-14.

Edari, Ronald and Patricia McManus. 1998. "Risk and Resiliency Factors for Violence." Violence Among Children and Adolescents. 45:2, pp. 293-305.

Feldman, R.A. 1992. "The St. Louis Experiment: Effective Treatment of Antisocial Youths in Prosocial Peer Groups," in Preventing Antisocial Behavior: Interventions from Birth Through Adolescence. McCord, J. and R. Tremblay, eds., pp. 233-52. New York, NY: Guilford.

Gainer, Patricia S., Daniel W. Webster, and Howard R. Champion. 1993. "A Youth Violence Prevention Program: Description and Preliminary Evaluation." Arch Surg. 128, pp.303-08.

Grossman, David C., et al.1997. "Effectiveness of a Violence Prevention Curriculum Among Children in Elementary School." JAMA. 277, pp. 1605-1611.

Guth, Julie T. 1995. Teacher P. A. L. S.: A Guide to the Peaceful Alternatives and Life Skills Program. Brainerd MN: Mid-Minnesota Women's Center.

Hausman, Alice, Glenn Pierce, and Lebaron Briggs. 1996. "Evaluation of Comprehensive Violence Prevention Education: Effects on Student Behavior." Jl of Adol Health. 19, pp.104-10.

Hyman, Irwin A. and Donna C. Perone. 1998. "Introduction to the Special Theme Section on School Violence." Jl of School Psych. 36:1, pp. 3-5.

Kann, L et al. 1995. "Youth Risk Behavior Surveillance-United States, 1993." MMWR. 44:SS-1, pp.1-56.

Minnesota Department of Children, Families, and Learning. 1998. Dangerous Weapons Incidents in Minnesota Schools, 1995-1996 and 1996-1997 School Years. St. Paul, MN: Minnesota Department of Children, Families, and Learning.

Minnesota Department of Education. 1992. Promising Prevention Strategies: A Look at What Works. MN: Minnesota Department of Education.

National Center for Education Statistics. 1998. Violence and Discipline Problems in U.S. Public Schools: 1996-97. U.S. Department of Education.

National Institute of Justice. 1998. Preventing Crime: What Works, What Doesn't, What's Promising. National Institute of Justice.

Olds, David et al. 1998. "Long-term Effects of Nurse Home Visitation on Children's Criminal and Antisocial Behavior: 15-Year Follow-up of a Randomized Controlled Trial." JAMA. 280:14, pp.1238-244.

Prothrow-Stith, Deborah and Sher Quaday. 1995. Hidden Casualties: The Relationship Between Violence and Learning. Washington, DC: National Health and Education Consortium and National Consortium for African American Children, Inc.

Resnick, Michael D., et al. 1997. "Protecting Adolescents from Harm: Findings from the National Longitudinal Study on Adolescent Health." JAMA. 278:10, pp.823-32.

Tolan, Patrick and Nancy Guerra. 1998. What Works in Reducing Adolescent Violence: An Empirical Review of the Field. Boulder, CO: Center for the Study and Prevention of Violence.

Webster, Daniel W. 1993. "The Unconvincing Case for School-Based Conflict Resolution Programs for Adolescents." Health Affairs. Winter 1993, pp127-41.

Wilson-Brewer, Renee. 1995. "Peer Violence Prevention Programs in Middle and High Schools." Adolescent Medicine: State of the Art Reviews. 6:2, pp.233-50.

Wilson-Brewer, Renee and Howard Spivak. 1994. "Violence Prevention in Schools and Other Community Settings: The Pediatrician as Initiator, Educator, Collaborator, and Advocate." Pediatrics. 94:4, pp.623-30