Identifying Change Levers Related to Program Effectiveness: Evidence from a Large Meta-analysis of Juvenile Delinquency Programs

• In Event: Research on the Effectiveness of Delinquency Prevention Programs

Abstract

Most delinquency programs work through a mediating mechanism, rather than by directly targeting delinquent behavior. There is no need to teach youth how not to engage in delinquent behavior; it is the absence of this behavior that is desired. Most of the research within the “what works” tradition has focused on the effectiveness of specific programs or program types. The aim of this work is to examine mediators or change mechanisms and the degree to which general categories of such mediators are effective levers for altering delinquent behavior. This work will rely on a large meta-analysis of juvenile delinquency programs and examine the extent to which program induced change on a mediator (i.e., effect size on an intermediate outcome) is related to reduced delinquency (i.e., effect size for delinquency). Potential mediators include school performance, social skills, psychological outcomes (e.g., impulse control), vocational skills, etc. A program may be effective at changing a mediator but that change may or may not be related to changes in delinquency. We will present preliminary work that identifies those mediators that when changed also result in less delinquency.

Authors

• David B. Wilson, George Mason University

• Mark W. Lipsey, Vanderbilt University