Early adolescents, in particular, adolescents at-risk are placed on islands of isolation where they are separated from the rest of society. During this important stage of their lives they are offered little to do that is worthwhile or important. By denying young people an immediate role in our society, we prolong their dependence, undermine their self-esteem, and cripple their capacity to care. A sad consequence is that youth often turn their energies toward activities destructive to themselves and the society that ignores them. This study focuses on the effect of engaging at-risk youth in a community service program, where they engage with the elderly in nursing homes, on their self-esteem and academic performance. This study used an experimental design to examine the relationship between the independent variable, a community service program called Magic Me, and the dependent variables of self-esteem, grades, and school attendance. The researcher used a one-way multivariate analysis of covariance to analyze the data collected on 95 students in two middle schools in the city of Baltimore, and one in Baltimore county. This analysis revealed no significant results, between the experimental and control groups, on the dependent variables. However, a repeated measures analysis, which analyzed self-esteem scores over three points in time, revealed a significant difference on this variable in one of the schools. Further examination showed that the community service group in this school was led by the most experienced facilitator. The results of this study suggest that, unless there is a specific academic component(s) within the community service program that it may be unrealistic to expect improvements in academic performance. The study also suggests that there be a relationship between group leader style and experience, and how students perceive themselves after the community service experience.