Adolescent Developmental Pathways Among Depression, Conduct Problems, and Rejection: Integrative Data Analysis Across Three Samples

• Published in: Journal of clinical child and adolescent psychology 2024-06, p.1
• Main Authors: McClaine, Rachel N, Connell, Arin M, Magee, Kelsey E, Ha, Thao, Westling, Erika, Brown-Iannuzzi, Jazmin, Stormshak, Elizabeth, Shaw, Daniel S
• Format: Article
• Language: English
• ISSN: 1537-4424; 1537-4424
• DOI: 10.1080/15374416.2024.2359063

The current study investigated sex differences in longitudinal associations among youth depression, conduct problems, and peer rejection from ages 11 to 16. We hypothesized that girls would follow the irritable depression model, which posits that depression leads to conduct problems, and that peer rejection would mediate this relationship. We hypothesized that boys would follow the cumulative failure model, which suggests that conduct problems predict future depression, mediated by peer rejection. We used integrative data analysis to combine three datasets, creating an aggregate sample of 2,322 adolescents, 58.4% of an ethnic minority group, and 51.3% boys. Using random-intercept cross-lagged panel modeling with data from ages 11-16, we conducted a nested model comparison. Results indicated that a model which allowed paths to differ by sex demonstrated better model fit than a constrained model. While depression, conduct problems, and peer rejection were relatively stable over time and had correlated random intercepts, there were few crossover paths between these domains for either sex. When the strengths of individual crossover pathways were compared based on sex, only the path from conduct problems at age 13 to depression at age 14 was significantly different, with this path being stronger for girls. These results suggest that stable, between-person effects largely drive relationships between depression, conduct problems, and peer rejection during adolescence, whereas there are few transactional, within-person pathways between these domains. This pattern of findings demonstrates the utility of random intercept cross-lagged panel modeling for disentangling between- and within-person effects.