ACADEMIC IMPACTS: A CRITICAL COMPARISON OF TUTORING AND SCHOOL-BASED MENTORING

Abstract

The United States education system is not meeting the needs of today’s youth, and this is especially the case for minority and low socio-economic status youth. Although there are a number of policy proposals that could improve youth academic performance, this thesis compares the efficacy of tutoring and mentoring as academic interventions. For mentoring, we specifically examine school-based mentoring programs. Previous research has found mixed results for the impact of both tutoring and mentoring on academic outcomes. For both tutoring and mentoring, some research has found that these programs can improve academic outcomes, while other research has found that these programs have no impact. Although the findings on tutoring and mentoring appear to be similar, no research has directly compared the academic efficacy of these two kinds of programs. Thus, district administrators seeking to improve student performance have no knowledge of which program would be better to implement. We attempt to fill this gap by examining a sample of 741 students, ranging from fourth to ninth grade, who were recommended to receive mentoring. The students were randomly assigned to receive or not receive mentoring; however, all students could also opt to receive tutoring. Thus, the students received either only tutoring, only mentoring, both tutoring and mentoring, or neither tutoring nor mentoring. We use regression analysis to compare teacher-reported student academic performance of the four groups. However, low-achieving students were more likely to receive tutoring. We attempt to control for this by using a number of covariates that theoretically correlate with whether a student chose to receive tutoring and with a student’s academic outcomes. We also use an instrumental variable regression technique to control for this. Finally, we compare academic and non-academic programs in a different way, using regression analysis to compare the impact of academically focused mentoring to that of other mentoring. We find that mentoring is a moderately effective academic intervention. Although we do not find that tutoring over a school year has an academic impact, we do find that those who received tutoring performed on average significantly better if they had also received tutoring during the year before the study. One explanation for this is that tutoring can be a moderately effective academic intervention over a longer period. In fact, our instrumental variable regression analysis finds that the academic impact of tutoring is statistically equivalent to the academic impact of mentoring. Additionally, we find that academically focused mentoring has a statistically equivalent academic impact to other mentoring, suggesting that academic programs perform equally as well as non-academic programs, even at improving academic performance. In general, these findings suggest that tutoring and mentoring can both be modestly effective academic interventions. By examining several previous studies, we also find that the costs of mentoring and tutoring programs are roughly similar and that, for both, more expensive programs do not necessarily yield stronger academic outcomes. Given these cost structures, our findings imply that whether tutoring or mentoring is more suitable depends on the specific circumstances. In particular, the lower-cost program with the most available volunteers is likely to be more appropriate. Because tutoring and mentoring serve different types of students, the appropriate program also depends on which students policymakers seek to target.