Despite the expected benefits of organized activity involvement (e.g., sports, clubs), inconsistencies in associations between activity involvement and internalizing and externalizing problems may be explained in part by limitations of measurements and variations between individuals. To address these gaps, a latent variable of organized activity participation was tested as a predictor of internalizing and externalizing problems, and initial child adjustment was tested as a moderator of the outcomes from activity participation. Participants included 431 adolescents (52.2% female; ages 12–13 in seventh grade) from the Child Development Project. Adolescents self-reported activity involvement (seventh grade) and internalizing problems (seventh and ninth grades); mothers reported on adolescents’ externalizing problems (seventh and eighth grade). Structural equation models showed that an activity involvement latent variable predicted lower internalizing problems. The interaction between activity involvement and initial level of externalizing problems predicted externalizing problems. Specifically, higher levels of activity involvement predicted lower levels of externalizing problems at initially lower levels of externalizing problems. However, at higher levels of initial externalizing problems, higher levels of activity involvement predicted higher levels of externalizing problems. The results suggest that activity involvement reduces risk for subsequent internalizing problems but could increase or decrease risk for subsequent externalizing problems depending on initial levels of externalizing problems.