Methods: This previously described, NIMH-funded study surveyed a high-risk school-based sample of youth in two large metropolitan areas, using a risk sampling protocol based on school records and teacher reports (Nurius et al, 2008). An ethnically diverse, random sample (N=850) completed nurse-delivered surveys with established measurement scales across multiple sociodemographic, psychological, and interpersonal domains. All constructs demonstrated good to excellent psychometric properties. Victimization was assessed across five domains (2 witnessing and 3 direct victimization forms). Perpetration was also assessed multidimensionally, including fighting, destruction of property, and physically or emotionally hurting others.
Results: Hierarchical multiple regression tested for the unique and cumulative contributions made by three sets of predictors: social disadvantage, psychological and interpersonal risk factors, and prior violent victimization. Each of the three blocks significantly contributed to the R2, and the total model explained 47% of the variance of violence perpetration. The psychological/interpersonal block (e.g., anger, substance use, risk-taking, deviant peer affiliation) contributed strongly (R2Δ=0.38). Family conflict, school instability, and social isolation demonstrated strong bivariate associations, but were nonsignificant when controlling for other risk factors in the final model. The strongest uniquely contributing factors (net of all other predictors) were unmanaged anger and violent victimization history. Violent victimization provided significant increase in the R2, demonstrating its importance above and beyond other disadvantage and risk factors. At the item level, 4 out of the 5 forms of violence exposure made significant unique contributions, the strongest predictor being physical abuse or injury. Sex differences emerged on mean levels of specific perpetration forms. Analysis in early adulthood confirmed sustained higher levels for adolescent perpetrators on all risk factors excepting depression.
Conclusions and Implications: These findings support a multi-risk perspective that includes social disadvantage and psychological and interpersonal risk factors in early intervention and prevention efforts, particularly regarding processes of cumulative stress and support. Key findings speak to prevention potential through targeting mutable contributors (individual/interpersonal factors) as well as promotive factors to ameliorate the effects of factors such as social disadvantage and victimization histories in altering perpetration etiology. Victimization remains one of the strongest unique contributors to violence perpetration, highlighting the need for early intervention to disrupt the victimization-perpetration link. Similarly, unmanaged anger is a particularly important target for intervention for at-risk youth (Blake et al, 2007). We discuss the methodological benefits of multidimensional risk and programmatic implications.