High-Risk Youth, High Impact: When Targeted Mentoring Works Best

/in , , /by

Miller-Chagnon, R. L., Krause, J. T., Moran, M. J., Haddock, S. A., Zimmerman, T. S., Zhou, H., & Weiler, L. M. (2025). Environmental and individual risk as moderators of a site-based mentoring program for adolescents exposed to adversities. Journal of Community Psychology, 54, e70075. https://doi.org/10.1002/jcop.70075

Introduction

Youth mentoring is widely used to support adolescents exposed to adversity, yet evidence is mixed on whether higher-risk youth benefit more—or disengage more quickly. Miller-Chagnon and colleagues (2025) address this gap by testing whether individual risks (e.g., behavioral challenges, mental health concerns, academic difficulties) and environmental risks (e.g., poverty, family stress, peer difficulties) shape change in mental health and behavior during a targeted, highly supportive site-based mentoring program.

Methods

Participants were 676 youth ages 10–19 and 658 caregivers enrolled in Campus Connections, a 12-week university-based program running four evenings per week (2015–2018). Youth were matched with trained college mentors, supported by weekly instruction and supervision, and had access to marriage and family therapy graduate trainees for “just-in-time” interventions. Risk was measured via a 32-item caregiver checklist. Outcomes included youth-reported depression, anxiety, and delinquency, plus caregiver-reported internalizing symptoms, emotional problems, peer problems, and conduct problems. Analyses used GEEs with Gamma and zero-inflated negative binomial models to test risk-by-time interactions.

Results

Overall, youth showed decreases in depression, anxiety, delinquency, internalizing symptoms, and emotional problems over time. Importantly, greater risk predicted greater improvement: youth with more mental health risk had the largest reductions in internalizing and emotional problems. Youth with more behavioral risk showed increased likelihood of reporting zero delinquency over time. Higher peer-risk youth showed the largest reductions in peer problems (IRR = 0.95). Economic, family, and academic risks did not significantly moderate change.

Discussion

Findings suggest that targeted, well-supported mentoring may be especially protective for adolescents with the greatest behavioral and emotional vulnerabilities. However, the lack of a control group limits causal claims and raises the possibility of regression-to-the-mean. Authors recommend future randomized trials with long-term follow-up and exploration of mechanisms such as emotion regulation and social support.

Implications for Mentoring Programs

Programs may achieve stronger impact when they (a) prioritize youth with elevated risk, (b) provide intensive mentor training and ongoing supervision, and (c) embed clinical “rapid response” supports. Mentoring evaluations should also examine outcomes by risk profile to better understand who benefits most, and why.

Read the full paper here.