The program consists of three major components: a mentoring program in which youths are matched with older adults, who provide ongoing support and encouragement in weekly interactions; community service activities designed to promote involvement with and better understanding of the frail elderly; and a school-based life-skills curriculum. Approximately 400 6th-grade students in Springfield, Massachusetts, an ethnically diverse community with a large proportion of public school students living in low-income families, participated in the evaluation over a 3-year period. Students' classes were randomly assigned to one of three experimental conditions: the mentor condition, in which students received the life-skills curriculum, community service activities, and were assigned an elder mentor; the curriculum condition, in which students received only the life-skills curriculum and community service activities; and the control condition, in which students received no intervention. Youths involved in the study completed questionnaires on three occasions: prior to the initiation of program activities, at the conclusion of the program, and again 6 months following the cessation of program activities. Results indicated that mentoring was associated with significantly lower levels of problem behavior and substance use, as well as significantly higher levels of self-confidence, self-control, cooperation, and attachment to both the school and the family. Moreover, students who received mentoring reported significantly lower levels of alcohol use. In contrast, few positive effects of the life-skills curriculum or community service activities were observed. 5 tables and 29 references