Juvenile justice systems routinely utilize risk assessment instruments (RAIs) to guide the provision of services and custody decisions to optimize youth treatment and system resources. While research has evaluated the validity of RAIs in predicting recidivism risk for justice-involved youth, it has yet to assess the feasibility of using RAIs in a prevention context. The early age and lack of formal delinquency history among this population pose considerable challenges to RAI effectiveness. Drawing data from the Florida Department of Juvenile Justice, analyses explored a sample of 11,472 Prevention Assessment Tool (PAT) assessments administered to youth aged 10 to 16 who had no prior system contact. Off-the-shelf scoring and optimized scoring developed through multivariate and machine learning techniques were employed to validate a prevention risk assessment instrument that predicts the risk of juvenile justice contact until age 18. The results indicate exemplary RAI performance among all optimized statistical techniques across various validation metrics. While performance was consistent across demographic subgroups, there were indications that predictive validity varied across age groups. These findings support the use of RAIs in early intervention and prevention services.
Read full abstract