This study examined the association between participation in a school-based group-mentoring program (Project Arrive) and academic outcomes in a group of ninth-grade students who had been identified as at high risk of high-school dropout (n = 239). Comparison participants were ninth-grade students with similar levels of risk (n = 980). Using inverse probability of treatment weighting (IPTW) with a propensity score to reduce selection bias, and a partially clustered model to account for nonindependence of data in program students, we found that Project Arrive students earned more credits by the end of 9th grade and 10th grade; and reported increased instructional time by the end of 9th grade, than comparison students.