In this study, the long-term academic effects of mentoring relationships in higher education are explored, with particular emphasis on students in ethnic categories with a low first-year retention rate. A sample of 339 undergraduate students in a student-faculty mentor program was statistically paired with 339 nonmentored controls on gender, ethnicity, class level, and entering grade point average (GPA). At the end of the one-year mentoring experience, mentored students had a higher GPA, completed more units, and had a higher retention rate. Eleven years after the onset of this study, records revealed that by graduation the GPAs of the mentored students did not differ significantly from those of their controls and there were no differences in graduation rate. However, the mentored students remained on campus to pursue graduate study and teaching credentials at a higher rate. There was no apparent advantage associated with matching students with mentors based on gender, but students matched with mentors of the same ethnicity showed a higher cumulative GPA and graduation rate and also entered graduate study at a higher rate. (Contains 2 tables.)