To develop viable strategies for preventing or curbing youth violence, we need to understand the scope of the problem and its co-occurrence with other behaviors that concern the public health and policy communities.1-3 Several studies provide estimates of general delinquency among youth,4-6 but research focusing only on violent behavior is sparse and is often based on special populations: youth in the criminal justice system,7 gangs,8 inner-city youth,6'9 high school students,'0"' or White middleclass males.'2 Because violent delinquents may differ from nonviolent delinquents in important ways,'3"4 we need to explicitly disentangle violence from other types of delinquent behavior and to assess its prevalence in more general populations.We also need to understand the degree to which violence and other problems are linked. Prior research indicates that general delinquency may occur as only one of a constellation of problem behaviors,2'4 5 but there is little agreement about how many youth "specialize" in a single high-risk behavior vs multiple problem behaviors. Dryfoos5 has suggested that as many as 25% of the nation's adolescents participate in multiple problem behaviors, including serious delinquency, school failure, substance use, and early sexual activity. In contrast, Elliot et al.4 estimated that less than 1% of 15-to 21 -year-olds were seriously delinquent in 1980, used two or more substances simultaneously, and suffered from mental health problems. And although most studies agree that more males than females are delinquent, substantial disagreement persists over the magnitude and nature of this difference.l'l8Our study examines the prevalence and behavioral context of multiple types of violent behavior in a sample of high school seniors and dropouts originally drawn from 30 middle and junior high schools in California and Oregon.Using liberal and stringent definitions of violence, we explore the correlation between violent behavior and other public health and criminal justice problems. We also examine gender differences in prevalence rates and estimate the extent to which youth engage in multiple problem behaviors.A key contribution of this study is that it overcomes the underreporting bias associated with prevalence rates that are derived from school-based samples of adolescents. I 19 Not only does our sample include a substantial proportion of school dropouts, but we have compensated for any remaining sample attrition by developing weights that allow us to represent the original 7th-grade cohort in the 30 schools. Hence, our estimates have been adjusted for nonresponse due to absenteeism, moving, dropping out of school, or refusal to respond to the survey.
Methods Data SourceWe use a longitudinal database of more than 4500 high school seniors and dropouts (17-to 18-year-old adolescents) from urban, suburban, and rural communities in California and Oregon. The 30 middle schools they originally attended were chosen to represent a broad spectrum of communities, socioeconomic status, and racial and ethnic c...
