While peer mentor programs for students with disabilities in higher education are increasing in prevalence, the research examining the effectiveness of these programs remains limited. In this study, the effect of one college peer mentoring program on academic performance at a small, private four-year university was examined. The sample was drawn from all students registered with the Office of Disability Services (n = 287), some of whom participated in a peer mentoring intervention as well as a comparison group comprised of non-participants. In light of the observational nature of the data, propensity score weighting and matching were used to adjust for possible confounding variables and to explore robustness to different methodological approaches. Logistic and linear regression methods were used to examine the effect of peer mentoring on academic probation status and grade point average (GPA), respectively, while incorporating propensity score-based adjustments. There were no significant differences between the intervention and comparison groups for either outcome; however, intervention group students had a statistically significantly higher number of accommodations available to them. The study illustrates that extant data, when used in conjunction with appropriate statistical adjustments, is a viable alternative to randomized studies. Implications for higher education researchers and practitioners are discussed, including suggestions to collect various types of non-academic data (e.g., satisfaction, well-being, self-determination surveys) as well as examine structural factors of the program (e.g., mentor-mentee relationships, mentor training) in order to better understand the possible benefits of peer mentor programs and the need for collaborative partnerships between disability services, student affairs, researchers, and institutional research staff.