Study Sample

The quantitative results of a large number of studies were synthesized using meta-analysis procedures. The 66 studies examined here were drawn from Lipsey and Derzon's bibliography (1998) and supplemented by research reports provided by OJJDP Study Group members and analyses of the Seattle Social Development Project longitudinal data set. The studies selected for this review met the following six criteria:

• Subjects were juveniles living in their community (i.e., they were not incarcerated) when they were first assessed.
  
  
• Subjects were not chosen for having committed prior criminal or violent offenses.
  
  
• Studies measured interpersonal physical violence or acts resulting in physical injury or threat of physical injury to another person, excluding suicidal behavior.
  
  
• Studies identified a modifiable indicator of a meaningful predictor or risk factor. Studies of interactions between multiple risk factors were excluded, as were discussions of race and gender, as predictors of violence.
  
  
• The study design was longitudinal, with results based on prospective or retrospective data so that exposure to risk factors preceded violence.
  
  
• Individual subjects served as the unit of analysis for both independent and dependent variables.
  

Predictors of Youth Violence	Juvenile Justice Bulletin • April 2000