This study investigated the impact of a school-based mentoring program on at-risk students' office referrals, unexcused absences, and school attitude. Students who participated in the mentoring program had statistically significant reductions in office referrals and statistically significant improvements in school attitude. Based on an analysis of mentor interview responses and log entries, mentors were divided into "viewed positively" mentors and "questioned-impact" mentors. Viewed positively mentors reported fewer office referrals, met more consistently with mentees, reported more relaxed mentoring sessions, and shared food and played games more often with their mentees than "questioned-impact" mentors. The results of the study are discussed in terms of the available research on school-based mentoring. In addition, suggestions for future research are provided.