Beyond the “Desperation Visit": A Causal Evaluation of Peer Tutoring Methods and Student Success

Evaluating academic support is frequently complicated by student selection bias. To provide an actionable framework for student retention, this study compares the causal impact of two uniformly trained peer tutoring methods offered concurrently within identical courses ( n  = 6,731). To isolate the true effect of the interventions, data were analyzed using an Average Treatment Effect in the Overlap Population propensity score-weighted framework—a rigorous statistical technique that balances preexisting student characteristics to mimic a randomized trial. Analyzing dosages across one-, three-, and six-sessions revealed stark differences in outcomes. Small group tutoring failed to improve grades, with isolated sessions indicating a negative “desperation visit” effect. Conversely, Supplemental Instruction-adjacent tutoring significantly improved grades after just three sessions ( p  < .001), remaining highly robust to unmeasured confounding (E-value = 34.86). Furthermore, sustained participation in both methods neutralized severe academic deficits. Ultimately, this study provides vital quantitative evidence that structural design dictates tutoring efficacy, guiding institutional resource allocation.