The present dissertation explored the influence of community capacity on behavior problems among adolescents. This study used 1990 census data and the National School Success Profile data set, which comprised a nationally representative sample of 6th- through 12th-grade students (N = 2,099) nested within 93 communities. The study used a contextual effects measurement approach and multilevel logistic regression to examine reports on four dependent variables (drug use, drinking, smoking, and sexual behavior). The study neither proved nor disproved study hypotheses. The present study highlights the need for complex contextual effects models. It suggests the need for conceptual frameworks that include both mediators and moderators such as caregiver support and community peer behavior problems. It also highlights the nuances associated with measuring dependent variables and establishing the structure of random effects in hierarchical generalized linear models. Finally, the study suggests that community interventions should extend beyond community capacity to include adolescents' caregivers and peers.