Application of Research Design and Methods for Optimizing Prevention Science: Growth in Depression among Adolescents

This abstract was presented at the 2018 Society for Prevention Research Annual Meeting which was held May 29 – June 1, 2018 in Washington, DC, US.

Mijin Kim

University of Houston; Alexandria Posada, University of Houston; Wiesner Margit, University of Houston

Introduction: Research on adolescent depression has been guided by a deficit-based model focusing on risk factors (e.g., Graber & Sontag, 2009); in contrast, proponents of positive youth development (e.g., Masten, 2014) emphasize strengths that potentially reduce mental health symptoms and promote thriving. Recently, scholars have considered the complementary nature of these frameworks (e.g., Arbeit et el., 2014). The current study investigated longitudinal associations of developmental assets to growth in depression among adolescents in Germany after accounting for psychosocial protective and risk factors. 

Methods: The participants came from the control group of a longitudinal study. The participants consisted of 722 fifth grade students (Mage= 10.54 yrs. at Wave 1; 56% female). Depression (8 items; α= .83-.90) and developmental assets (personal and social assets, 13 items each; α= .80-.87) were assessed at waves 3-6. Seven psychosocial risk factors (e.g., gender, temperament, yielding to peer pressure) were assessed at baseline. Latent growth curve analysis with time-invariant predictors (i.e., baseline psychosocial protective and risk factors) and a latent time-varying covariate (i.e., developmental assets) were conducted to model growth in depression. Analyses were performed in Mplus version 7.3 using robust maximum likelihood estimation, full information maximum likelihood, and type=complex. 

Results: The fit of the conditional linear LGM model with seven Wave 1 psychosocial predictors and the time-varying covariate developmental assets to the data was acceptable, MLRχ2(89,N= 722) = 207.48, p= .000, CFI = .960, RMSEA= .043 [90% CI: .035, .051], SRMR= .035. Being female (β=.18,p<.01) and reporting a higher degree of yielding to peer pressure (β=.12,p<.05) at Wave 1 were significantly linked with higher initial levels of depression at Wave 3. Higher general activity levels were significantly associated with smaller increases in depression over time (β=-.14,p<.05), whereas female gender was associated with steeper linear increases in depression over time (β=.19, p<.05). Higher levels of developmental assets were consistently related to lower levels of depression at each time points (β=-.27 to -.40; all p<.01), even after adjusting for the latent trajectory of growth in depression over time. 

Conclusion: The findings emphasize that prevention efforts to adolescents’ depression need to integrate both protective and risk factors in the individual level such as gender, activity level, and personal developmental assets, as well as those in the interpersonal and social level such as yielding to peer pressure and social developmental assets.