Document Type


Publication Date


Published In

Journal Of Community Psychology


Schools can benefit from understanding how to use community volunteers to achieve academic goals. A randomized control evaluation, involving 1,139 students from 71 schools, of the school-based mentoring program of Big Brothers Big Sisters of America found modest but statistically significant improvements in the teacher-rated academic performance and self-reported scholastic efficacy of mentored students. The present study explores the causal mechanism behind these effects. We find that a close relationship between mentor and protégé appears key to better academic outcomes. Because relationship closeness is not randomly assigned, we use two-stage least squares and other methods to control for potential selection bias. The role of emotional closeness as a mediator of program effects is evident across mentoring relationships of various lengths and statuses. Students were more likely to feel close to their mentors in programs that included weekly meetings and opportunities for mentor–protégé pairs to interact outside of a large-group setting.


This work is a preprint that is freely available courtesy of Wiley,

Included in

Economics Commons