Early childhood internalizing problems and maternal depression as predictors of social media-related depressive symptoms in adolescence: a 14-year longitudinal study from Chile

Abstract

To examine associations between social media use and depressive symptoms in Chilean adolescents, test whether early childhood internalizing problems moderate this association, and decompose the intergenerational association between maternal depression and adolescent depressive symptoms via social media use. Longitudinal analysis of the ELPI, a nationally representative Chilean cohort (N = 10,004) followed from infancy (2010) through adolescence (2024). OLS regression and path analysis with PHQ-4 (Cronbach’s alpha = 0.81) as the primary outcome, social media hours/day as exposure, and early childhood CBCL internalizing and maternal CESD-10 as longitudinal predictors. Mediation decomposition used 5,000 percentile-based bootstrap replications. Among 7,129 adolescents with complete data (49.1% female), 54% used social media >3 hours/day. Each SD increase in social media use was associated with 0.33 higher PHQ-4 points (p < .001). Early CBCL predicted adolescent depression (beta = 0.11, p = .002), but the interaction was null (beta = −0.011, p = .720; post-hoc power = 6.5%). Maternal depression predicted adolescent depression directly (beta = 0.041, p < .001) and indirectly through social media (indirect = 0.0059, 95% CI [0.0027, 0.0094]; 0.016 raw PHQ-4 points; 12.6% of total association). Social media use is associated with depressive symptoms in Chilean adolescents. Early childhood internalizing problems and maternal depression are independent risk factors. A modest proportion of the maternal depression-adolescent depression association is statistically channeled through social media use, though causal interpretation requires longitudinal mediation data.