Does Self-Report of Aggression After First Arrest Predict Future Offending and Do the Forms and Functions of Aggression Matter?

The current study tested whether a self-report measure of aggression (i.e., the Peer Conflict Scale; PCS) would predict later delinquency, after controlling for other risk factors, and tested whether the different forms and functions of aggression contributed independently to this prediction. Self-report of aggression was assessed at the time of first arrest, and both self-report of delinquency and official arrests were assessed at 5 different time points over a 30-month follow-up period in a sample of male adolescent offenders (N = 1,216; Mage = 15.12, SD = 1.29 years) arrested in 3 regions (i.e., western, southern, northeast) of the United States. Aggression predicted both later total and later violent self-reported delinquency (odds ratio [OR] = 1.02, 95% confidence interval [CI: 1.01, 1.02]), even after controlling for youths' self-reported lifetime history of delinquent acts and callous-unemotional (CU) traits (i.e., Inventory of Callous-Unemotional Traits) collected at initial arrest. Further, only self-reported aggression (but not lifetime delinquency and CU traits) contributed independently (OR = 1.02, 95% CI [1.00, 1.03]) to the prediction of arrests for violent offenses. Finally, the predictive utility of aggression was largely accounted for by physical and reactive aggression, with limited incremental prediction provided by relational and proactive aggression. These findings support the potential utility of self-reports of aggression, such as the PCS, when assessing risk for future violence. Findings also suggest that the utility of these self-reports of aggression cannot be solely accounted for by other risk factors often included in typical risk assessment tools. Public Significance Statement This study suggests self-reported history of aggression, especially physical and reactive aggression, assessed immediately following first arrest predicts future violent offending in male adolescents, even when controlling for other known risk factors such as baseline levels of antisocial behavior and callous-unemotional traits. Thus, violence risk assessments designed to detect those youth who are most likely to be violent to determine the optimal placement following arrest would benefit from considering adolescents' reports of past aggression.