UC Berkeley

The Positive Youth Justice (PYJ) paradigm is gaining traction in the juvenile justice system. This paradigm shifts focus from youths’ deficits to their strengths, opening new avenues to help young people build prosocial assets and desist from delinquent behavior. However, there is little conceptual or empirical guidance for translating the PYJ approach into real-world interventions. In this study, I focus on one promising construct within the Positive Youth Development framework—prosocial identity (i.e., the extent to which young people view themselves as prosocial)—and empirically test key hypotheses drawn from a conceptual model for intervention (Jian & Skeem, 2023). The central inquiries of the study include how prosocial identity should be defined as a treatment target in the juvenile justice context and under what conditions prosocial identity protects against reoffending. The primary aims are to first examine the basic protective effects of prosocial identity against reoffending in a justice-involved sample. Here, I include prosocial identity at the level of subcomponents—other-oriented (emphasizing benefits to others) and conventional (emphasizing law-abiding and conventional commitment), as well as holistically (prosocial identity as the combination of the two subcomponents). Second, I examine whether three factors—prosocial identity prominence (importance to the self), prosocial identity internal validation (confidence that the self can be achieved), and prosocial identity external validation (e.g., social support from peers and/or caregivers, school connectedness)—may moderate any protective effect of prosocial identity on recidivism. The secondary aim is to examine the generalizability of measures of prosocial identity that were developed with non-referred populations, to justice-involved youth.

I addressed these aims using data on a subsample of 760 court-referred youths. I conducted psychometric analyses and a series of progressively stringent analyses (from regression to causal modeling incorporating machine learning). The primary aims yielded three main findings. (1) Contrary to hypotheses, other-oriented identity, but not conventional identity, predicted recidivism independently. (2) There was no empirical support for the notion that prosocial identity causally protects against recidivism, after rigorously adjusting for covariates. (3) Among the proposed moderators, peer support (a form of external validation) significantly moderated the relation between prosocial identity and recidivism (OR=0.99) in logistic regressions. The predicted probability of recidivism decreased as prosocial identity increased, only for those with moderate or high level of peer support. The moderating effect of peer support remained significant even after rigorously adjusting for all covariates, but not after using causal modeling. For the secondary aim, results indicate that prosocial identity measures developed with non-referred youth can be generalized to justice-involved youth.

These findings provide mixed support for hypotheses derived from the identity-based conceptual model. Research that directly tests aspects of the model is necessary to advance our understanding of whether and how deliberately increasing prosocial identity among justice- involved youth will reduce recidivism.