ABSTRACT
Introduction The purpose of this paper is to explore the potential for predictive analytics and machine learning to improve the performance of juvenile justice risk assessment instruments. The authors will present findings from a case study designed to improve the performance of the risk assessment instrument used by the Florida Department of Juvenile Justice; the Positive Achievement Change Tool (PACT). The study was designed to examine how to improve the accuracy of the PACT and, at the same time, minimize the differential treatment of youth of color in the dispositional process. There is great concern about the differential treatment of minorities in the juvenile justice system and, in particular, the rate of incarceration of minorities in comparison to whites. Accordingly, the researchers were interested to see if the PACT’s performance could be improved while making the risk assessment instrument’s capability fairer and more equitable.
