How Effective Are Mentoring Programs forYouth? A Systematic Assessment of the Evidence

摘要

During the past decade, mentoring has proliferated as an intervention strategy for addressing the needs that young peo-ple have for adult support and guidance throughout their development. Currently, more than 5,000 mentoring programs serve an estimated three million youths in the United States. Funding and growth imperatives continue to fuel the expan-sion of programs as well as the diversification of mentoring approaches and applications. Important questions remain, however, about the effectiveness of these types of interventions and the conditions required to optimize benefits for young people who participate in them. In this article, we use meta-analysis to take stock of the current evidence on the effective-ness of mentoring programs for youth. As a guiding conceptual framework for our analysis, we draw on a developmental model of youth mentoring relationships (Rhodes, 2002, 2005). This model posits an interconnected set of processes (social-emotional, cognitive, identity) through which caring and meaningful relationships with nonparental adults (or older peers) can promote positive developmental trajectories. These processes are presumed to be conditioned by a range of indi-vidual, dyadic, programmatic, and contextual variables. Based on this model and related prior research, we anticipated that we would find evidence for the effectiveness of mentoring as an approach for fostering healthy development among youth. We also expected that effectiveness would vary as a function of differences in both program practices and the characteristics of participating young people and their mentors. The meta-analysis encompassed 73 independent evalua-tions of mentoring programs directed toward children and adolescents published over the past decade (1999-2010). Overall, findings support the effectiveness of mentoring for improving outcomes across behavioral, social, emotional, and academic domains of young people's development. The most common pattern of benefits is for mentored youth to exhibit positive gains on outcome measures while nonmentored youth exhibit declines. It appears then that mentoring as an inter-vention strategy has the capacity to serve both promotion and prevention aims. Programs also show evidence of being able to affect multiple domains of youth functioning simultaneously and to improve selected outcomes of policy interest (e.g., aca-demic achievement test scores). From a developmental stand-point, benefits of participation in mentoring programs are apparent from early childhood to adolescence and thus not confined to a particular stage of development. Similarly, although programs typically have utilized adult volunteers and focused on cultivating one-to-one relationships, those that have engaged older peers as mentors or used group formats show comparable levels of effectiveness. Collectively, these findings point toward the flexibility and broad applicability of mentoring as an approach for supporting positive youth development. Several other aspects of our findings, however, underscore a need for caution. These include a failure of evaluations to assess several key outcomes of policy interest (e.g., juvenile offending, obesity prevention) or to determine whether bene-fits for youth are sustained at later points in their develop-ment. More generally, we find that gains on outcome measures for the typical young person in a mentoring program have been modest (equivalent to a difference of 9 percentile points from scores of nonmentored youth on the same measures). This level of impact is within the range of effects observed for other types of interventions for children and adolescents but fails to reflect discernible improvement over the previous generation of mentoring programs (DuBois, Holloway, Valentine, & Cooper, 2002). Variability in program effectiveness, although less pronounced, also continues to be evident even after accounting for methodological differences in studies. In ana-lyzing this variability,