Influence of Risk Assessment Instruments on Racial/Ethnic Disparities in the Sentencing of Juvenile Offenders

Over the past two decades, there has been a growing trend in the juvenile justice system of administering risk assessment instruments to juveniles and using the results to classify offenders into risk categories and determine their sentence. Some scholars and policy-makers tout risk assessment instruments as capable of reducing racial and ethnic disparities-- a problem affecting the juvenile justice system since its inception. Institutional discrimination, however, would predict the opposite outcome: risk assessment instruments would inadvertently create or maintain existing racial hierarchies, since they could contain wording that would disadvantage racial minorities. Relying on a sample of 26,681 offenders in Florida's Juvenile Justice System between 2006 and 2008, this dissertation examines the role of Florida's risk assessment instrument, the Positive Achievement Change Tool, in mediating the relationship between race/ethnicity and disposition decisions. Although studies have examined race, risk assessments and recidivism, no study has examined the relationship between race, risk assessment and disposition decisions. Findings reveal that the PACT appears to statistically reduce the direct effects of race on the likelihood of receiving a commitment (the less harsh sentence) compared to a community decision (the harsher sentence), although blacks nevertheless remained 12 percent more likely than whites to receive the harsher penalty net of controls. Results revealed no Latino/a-white disposition disparities. Analyses of the criminal and social history scores (the two constituent elements of the PACT) as mediators of the relationship between race and disposition decisions revealed that the addition of the criminal history score reduced the direct effect of being black compared to white on receiving the harsher disposition, but that adding the social history score did the opposite--it was associated with blacks' higher odds of receiving the harsher disposition. Because initial analyses again revealed no Latino/a-white disposition disparities, the criminal and social history scores played no role in mediating disparities between these groups (since there were no significant differences in disposition decisions). Overall, the results are consistent with institutional discrimination, and the conclusion discusses the findings in light of theory, notes limitations, and offers suggestions for future research.