We conducted secondary analyses of program evaluation data involving six high school (HS) mentoring programs. All programs used one-on-one matches but differed in a) how mentors were recruited (voluntary vs. mandated due to class requirement), b) how mentees were referred (individual student referral vs. whole class participation), and c) the goals of mentoring (e.g., improved peer relationships, academic achievement). Variability in program practices suggested three program types or clusters: voluntary mentor/individual student referral/limited goals (Vol/Ind/Ltd), voluntary mentor/whole class referral/peer relationship goals (Vol/Whl/Per), and mandatory mentor/whole class referral/academic achievement goals (Mnd/Whl/Acd). Child, mentor, and teacher data collected pre- and post-mentoring were available for 253 mentor-mentee matches. Results revealed significant differences for mentees across program types: Children matched with HS mentors in Vol/Ind/Ltd and Mnd/Whl/Acd programs had similar outcomes and fared significantly better than mentees in Vol/Whl/Per programs. For mentors, those in mandatory mentoring programs (Mnd/Whl/Acd) reported significantly lower levels of academic, community, and diversity engagement compared to HS mentors in voluntary programs. Discussed are the implications these findings have for program practices involving HS mentors.