The goals of this research program were to develop, implement, and evaluate a mentoring program for Latino college freshmen and to identify processes that account for the effects of the program. In two longitudinal studies, at-risk Latino freshmen (mentees) were mentored from fall to spring by upper division or graduate students from psychology and counseling majors and compared in the spring to an equivalent sample of nonmentored students (nonmentees). In both studies, mentees showed improvement in psychosocial factors that underlie academic performance. In the second study, mentees decreased in depression and stress and were less likely than nonmentees be classified as being at risk for poor academic outcomes. The amount of mentor–mentee contact and the quality of the relationship contributed to positive outcomes for mentees. Results suggest that mentors are of value in alleviating psychosocial risk factors. Selecting at-risk students and using experienced peers as mentors make the program cost effective.