The test of time in school-based mentoring: Role of duration and re-matching on academic outcomes

final3Grossman, J.B., Chan, C.S., Schwartz, S.E.O. & Rhodes, J.E. (2012). The test of time in school-based mentoring: The role of relationship duration and re-matching on academic outcomes. American Journal of Community Psycholog, 49, 43-53

summarized by Stella Kanchewa, M.A., UMass Boston clinical psychology doctoral student

Problem: Youth mentoring programs face the challenge of creating matches that will endure long enough for youth to derive many of the benefits that quality relationships with mentors can generate. Research (e.g., Grossman & Rhodes, 2002) indicates that youth may benefit most from matches that are at least one year in length, yet most school-based mentoring programs generally average relatively short matches, most of which terminate at the end of the school year. How might match length and potential re-matching affect youth outcomes? This study explores several hypotheses within the context of school-based mentoring programs.

Methodology: The current study included a diverse sample of 1,139 adolescents, average age 11.2 years, who were part of a national evaluation of Big Brother Big Sisters’ school-based mentoring program. Students completed surveys at the beginning of the school year before they were randomly assigned to receive mentoring or to a waiting list (in order to serve as a control group). Students completed surveys again at the end of the school year. The student’s teacher(s) also completed surveys at each time point.


After accounting for selection bias, academic impacts were found only amongst those youth with intact matches—teachers rated the achievement of youth in intact matches almost half a point higher than they rated those without the mentoring intervention. Youth who experienced premature match terminations and who were not re-matched showed no significant differences from the controls. Re-matched youth fared worse, however, performing 1.6 points lower than the controls.  This is an intriguing finding, which, if replicated, has important implications for mentoring programs. A number of processes may be at play. First, there is the possibility that the rapid re-matching provided insufficient time for the youth to resolve and make sense of the difficulties or disappointments inherent in the first match. As with other relationship losses, a period of taking stock may be beneficial, rather than quickly launching into a new relationship in which previous problematic behavioral patterns may be repeated. Moreover, the presence of a new mentor may draw attention to the first loss when youth may prefer to quietly withdraw from the program.

A range of factors associated with relationship integrity were identified. In particular, youth who had endured a greater number of life stressors prior to being matched were at greater risk for early match termination. This general pattern of findings is consistent with previous research (Grossman & Rhodes, 2002; Schwartz et al., in press) and suggests that the challenges associated with mentoring youth who have been exposed to relatively high levels of stress are likely to be substantial, potentially overwhelming mentors’ capacity or willingness to help. Case managers should work closely with such dyads to move them beyond the initial, challenging stages of the relationship.  We also found that youth who, according to teachers’ report, tended to overreact to rejection or criticism were slightly more likely to maintain their matches for the entire school year. Perhaps these youth, who tend to be hypersensitive to rejection cues, made a greater investment in the relationship in order to avoid rejection.

Two mentor characteristics were related with the likelihood that a match terminated—being a college student and having prior experience being a mentor. Matches with college student mentors were more likely to terminate early. This finding is perhaps best explained by the unpredictable schedules and transitory nature of college students, which can undermine continuity. Moreover, volunteers who had prior experience being a mentor were more likely to be in long-lasting relationships. This is understandable, as they are likely to have well-defined and realistic expectations about what the experience will entail. The finding is consistent with results of a meta-analysis of mentoring (DuBois et al., 2002), which demonstrated stronger effects among volunteers who had previous experience in helping roles or professions. These findings point to the potential benefits of recruiting volunteers from among programs’ alumni, as well as seeking out volunteers whose backgrounds include prior experience and success in helping roles.

Finally, at the program level, we found evidence that matches were less likely to endure if they were situated in programs that focused more heavily on school work. Most academically focused programs tended to structure meeting times, and matches primarily met in the same place; however, some non-academically focused programs also had structured time and/or matches that met all together. When we tried to disentangle why academically focused programs had more early terminations, we found that having structure was not associated with greater risk of early termination, but that having matches that met at the same time and location was associated with early termination. Such matches may lack the flexibility and responsiveness that is vital to mentoring relationships.

Conclusion: Taken together, the findings are consistent with previous studies, which have shown that mentoring relationships can vary considerably in their effectiveness, depending on the match length.However, at least for school-based mentoring, it is not just the number of weeks that is important, but also the integrity of the match.  The patterns of impacts for intact matches, when compared to broken matches, might also help to explain the relatively disappointing small effects from large-scale, random assignment evaluations of youth mentoring (Bernstein et al., 2009; Herrera et al., 2007). When impacts from all of the matches are combined, positive outcomes can be masked by the neutral, and even negative, outcomes associated with early terminating and reconfigured matches.