Date of Award


Document Type

Open Access Dissertation



First Advisor

Dawn K. Wilson


Depression is a common psychiatric problem experienced in adolescence with nearly a quarter of all adolescents experiencing a major depressive episode before adulthood. Previous evidence indicates that physical activity (PA) is a known protective factor for depressive symptoms and major depressive disorder in adolescents and that PA self-efficacy (i.e., self-confidence) improves self-esteem, which in turn reduces depressive symptoms. Furthermore, PA self-efficacy may be more protective against depression for individuals who value or consider PA to be highly important. The proposed study aimed to determine the cross-sectional and longitudinal PA-depressive symptoms relation in adolescents. The study was designed specifically to test a mediation model, wherein PA self-efficacy mediated the cross-sectional and longitudinal PA-depressive symptoms relations, and to test a moderation model, wherein PA valuation moderated the PA self-efficacy-depressive symptoms relation. The participants were a subsample of youth enrolled in the Active by Choice Today trial, a randomized controlled trial that tested the efficacy of an afterschool theory-based PA intervention. Students were recruited for the larger trial through flyers and events at their middle schools. These 409 sixth graders were enrolled in the first two years of the trial. The mean age was 11.4 (SD=0.65), most (72%) were African American, and about half (55%) were females. Participants completed the Children’s Depression Inventory (CDI) and self-report measures of PA self-efficacy and PA valuation. Participants wore accelerometers and seven-day moderate-to-vigorous PA (MVPA) estimates were calculated from the raw PA counts. In both the cross-sectional and longitudinal models, there was no direct effect of MVPA and depressive symptoms (p>0.05). In the cross-sectional model, there was a significant relation between MVPA and PA self-efficacy (a path; B=0.006, SE=0.002, p=0.009) and a significant relation between PA self-efficacy and depressive symptoms (b path; B=-0.262, SE=0.069, p<0.001), partially supporting the proposed mediation model. The hypothesized moderating effect of PA valuation was not supported in the cross-sectional or longitudinal models. A secondary finding of interest was that African Americans reported significantly fewer depressive symptoms than their peers at T1 (t=-2.32, p=0.021) and at T2 (t=-1.99, p=0.047). Overall, these findings suggest that, when objectively measured, MVPA may not be correlated with depressive symptoms in young adolescents. However, the findings supported the notion that domain-specific PA self-efficacy may be related to depressive symptoms. Finally, from a developmental perspective the data suggested that, in early adolescence, African American youth may experience fewer depressive symptoms than their peers. The theoretical and clinical implications of these findings are discussed.