Society for Social Work and Research

Background and significance: Researchers have noted that there are relatively few studies comparing adolescent sexual offenders with delinquent youth (van Wijk et al., 2006), with mixed findings in terms of the similarities and differences between groups. While there are risk factors common to both groups (e.g., victimization, executive functioning deficits, low school attachment, and family problems), and youthful sexual abusers have been found to commit a great number of nonsexual crimes (Brown & Burton, 2010), a recent meta-analysis found that the delinquency explanation was insufficient to parsimoniously explain adolescent sexual offending. Variables such as family criminality, substance abuse, poor childhood attachment, social incompetence, antisocial attitudes, and exposure to nonsexual violence outside the family were unsupported as risk factors (there were no group differences; Seto & Lalumerie, 2010). However, adolescent sexual offenders have been found to have a higher exposure to violence in the family, greater victimization rates and posttraumatic stress symptomatology, and higher mean scores on personality measures than delinquent youth (Burton, Duty, Leibowitz, 2011; Leibowitz, Laser, Burton, 2011). Exploring discrepancies in the literature, our recent research compared adolescent sexual abusers and delinquent youth on variables that were hypothesized to contribute to sexual and nonsexual crimes.

Symposium theme: In this symposium, we explore the associations between risk and protective factors for sexual and nonsexual offending among a diverse sample of male adolescents (n=502) in court-ordered residential treatment. A battery of self-report measures used widely in juvenile justice settings were administered to the youth in a group format. Advanced statistical models were used: logistic regression, path analysis and Structural Equation Modeling (SEM) to test theoretical models that attempt to explain causal relationships among constructs related to sexual offending. All three papers in this symposium address findings that are consistent with previous research, and findings that were unexpected and necessitate further investigation.

In the first paper, logistic regression models were used to test whether family problems (e.g., mental health and family criminality) predicted sexual and nonsexual offenses. Findings included that the family problems scale score was significant in predicting sexual offender status, but the delinquency scale score was nonsignificant. In the second paper, path analysis was used to explore pathways to sexual aggression and executive functioning was found to have a mediating effect on the relationship between trauma and sexual offending. In the third paper, the importance of identifying protective factors in this population were highlighted, and three structural equation models (SEM) were used to investigate whether school bonding, communication problems, and educational difficulties were related to delinquent acts in both sexual and nonsexual offending youth. School bonding was found to have the largest effect in reducing delinquency among nonsexual offenders.

Understanding the differences between sexually abusive and delinquent youth can help identify the risk factors for recidivism, promote protective factors, and has important implications for social work practice, policy-related decisions, and the assessment and treatment of these adolescents.