Background: Some prevention programs negatively affect participants. Previous research indicates that peers can cause these negative effects. However, little is known about which students may be most vulnerable to negative peer effects in prevention interventions.

Purpose: This study tests the effect of participation in an after-school program (ASP) on student outcomes of peer delinquency, problem behavior and antisocial attitudes and beliefs for students of differing pre-test levels of risk for those outcomes. Drawing on social learning theory, this study examines whether low- and moderate- risk students in the intervention are more likely to acquire delinquent behaviors and beliefs in the ASP than their already-delinquent counterparts.

Participants: 447 middle school students attending underperforming schools in Baltimore County, Maryland.

Intervention: The data are drawn from an experimental evaluation of an after school program which operated in five middle schools in Baltimore County during the 2006-2007 academic year. The overall evaluation of the program found null effects on the wide range of measured outcomes (including academic achievement and delinquency). I explore whether the lack of beneficial program effects is partially attributable to negative effects among low and moderate risk participants who absorbed negative beliefs and behaviors from high-risk peers in the ASP.

Research Design: Randomized, controlled field trial.

Findings: Results indicate that low- and moderate- risk youth are not more likely to experience negative outcomes than high-risk youths. On the contrary, low-risk participants are less likely to experience negative effects than high-risk participants. Students who began the program with elevated negative peer influences grew in this characteristic if they often participated in the ASP but declined in negative peer influences if they less often attended the program. Implications for universal prevention are discussed.