Evaluation of a Multi-Component School-Level and Family-Based Youth Violence Prevention Program in High Risk Communities

Introduction: There is a growing consensus of the need for comprehensive efforts to reduce youth violence. Such efforts need to include complementary components that focus on individual-level and on broader contextual and environmental factors. Ideally comprehensive prevention efforts should combine universal strategies that focus on all individuals regardless of risk and selective intervention strategies for those at higher risk. Finally, there is recognition of the need to tailor interventions to address contextual factors that vary across communities. This paper describes a 5-year project to evaluate the impact of a comprehensive violence prevention strategy among 10-24 year old youth that included: (a) school-based universal and selective intervention components developed by tailoring aspects of the Olweus Bully Prevention Program to increase its relevance for middle schools serving a primarily African American urban middle school student population; and (b) selective family interventions targeted at youth with high truancy rates.

Methods: Participants were students at middle schools in three communities in Richmond, VA with high rates of crime and poverty. The impact of the intervention was evaluated by a multiple baseline design in which the order and timing of intervention activities was randomized across communities. The frequency of physical, non-physical, and relational violence and victimization in each community was assessed by self-report and teacher ratings of a random sample of 1,745 adolescents. Data were obtained every 3 months through all 5 years of the project to determine the extent to which initiation of the intervention in each community impacted community levels of violence and victimization.

Results: Multilevel modeling examined the extent to which initiation of the intervention in each community altered the trajectory of community-level rates of violence and victimization. Initiation of the intervention was associated with reductions in teacher ratings of adolescents’ frequency of aggression. Intervention effects emerged in different years of implementation for different forms of aggression. Reductions in nonphysical and relational aggression were evident during the second year of implementation, and reductions in physical aggression during the third year of implementation. These effects were generally consistent across gender, grades, and schools. Similar effects were not found on student reports of their behavior or measures of overall school climate.

Conclusions: The positive outcomes for teacher ratings, but not for adolescent reports, suggest the intervention’s effects may have been limited to the school context. Variation in when effects emerged across outcomes suggest that changes in physical aggression may require more sustained intervention efforts.