This work assessed the efficacy of a middle-school-based mentoring program designed to increase student use of self-regulated learning (SRL) strategies, self-efficacy for and the perceived usefulness of SRL as well as mathematics and language achievement. A longitudinal cluster randomized trial study design obtained evidence that found differential effects of a school-based mentoring program. Specifically, the performance of 94 seventh grade students naturally nested within four classrooms was measured at baseline and after 3, 6, and 9 months. Two classrooms were each randomly assigned to treatment or control conditions. First, the results indicated that participation in the mentoring program led to significant improvements with regard to all the dependent variables after the 9-month intervention, and significant effects had been observed at 6 months for some variables. Second, the program appears to play a more important role for SRL variables compared with academic variables. Third, the effect sizes were small, small-medium, or medium depending on academic mentoring, the type of variable used to assess the efficacy of the program, or the level of analysis considered, respectively. The effect size of this intervention was equal to or greater than those reported in prior studies. In conclusion, our findings underline the importance of academic mentoring programs that practice SRL strategies and emphasize the relevance of using study designs that provide both cross-sectional and longitudinal data. (Contains 7 tables and 1 figure.)