Over the 4-year period, attention focused on whether there were enough adults with flexible time and emotional resources to mentor at-risk youth; whether mentoring could be integrated into large-scale youth-serving institutions, specifically juvenile justice agencies; whether a set of practices was available that characterized the adult role in an effective mentoring relationship; what level of training and support activities, services, and costs was required to administer mentoring programs; and whether participation in mentoring programs changed the attitudes, perceptions, and behaviors of at-risk youth and mentors. The six Campus Partners in Learning programs examined during the course of research had a common core but varied widely in how they implemented that core. For example, mentee grade levels ranged from 4th to 9th, with 7th and 8th grades the most common levels. Programs ranged in size from 8 to 24 matches. Mentees were recruited through their school or the public housing project where they resided. Some programs held their meetings primarily on the college campus, while other programs conducted meetings at the mentees' schools. Some programs consisted almost exclusively of individual interaction, others focused on group sessions, and one was evenly divided between individual and group interaction. Most programs were not initially structured to take into account the unique circumstances of college students. Program elements included mentor recruitment, mentor screening, mentor training, mentee orientation, matching, and mentor/mentee interaction. Interviews with 29 pairs of mentors and youth showed that 45 percent formed successful relationships. Mentors in successful relationships had a common approach to their mentees; they allowed the relationships to be youth-driven. Responses to questionnaires administered to 50 mentors and 52 youth indicated that mentors exhibited improved self-esteem, perceived scholastic competence, and satisfaction with their social skills but did not show improvements in communication skills, grade point averages, or the sense that they could change the world. Mentees were exposed to additional social and cultural activities, their sense of control over their lives improved, but there were no behavioral changes or improvements in academic performance. Appendixes contain program site descriptions and information on the research methods. 16 references and 8 tables