Social Media Use and Adolescent Mental Health: A Meta-Meta-Analysis Followed by a Quantitative Synthesis of Longitudinal Evidence

Abstract Social media use has been widely debated as a contributor to adolescent mental health problems, yet evidence regarding the strength and interpretation of this association remains contested. We conducted a meta-meta-analysis of quantitative reviews, followed by a complementary meta-analysis of longitudinal evidence extracted from the included reviews to evaluate associations between social media use and mental health outcomes among adolescents aged 10–19 years. Searches combined nine databases alongside backward and forward citation tracking, last updated in March 2026. Seven eligible meta-analyses were identified from 8,780 records, comprising 11 review-level effect sizes spanning depression, anxiety, and broader internalising symptoms. Social media use was associated with poorer mental health outcomes ( SMD = 0.21, 95% CI [0.16, 0.27], p < .001), although heterogeneity was substantial. Despite high quality scores (Mean = 9.57 out of a possible 11), only 3 out of 7 meta-analyses were pre-registered, highlighting the lack of research transparency. To examine the temporal interpretation of the evidence represented within the included reviews, we additionally synthesised 33 longitudinal studies ( N ES = 96) identified within the seven included reviews. A three-level random-effects model of these 96 effect sizes yielded a Cohen’s d of 0.24 (95% CI [0.13, 0.36], p < .001). In terms of forward direction, greater social media use was prospectively associated with poorer subsequent mental health ( d = 0.25, 95% CI [0.12, 0.37], p < .001), whereas poorer mental health was also prospectively associated with greater later social media use ( d = 0.18, 95% CI [0.10, 0.27], p < .001). It is worth noting that both directions are statistically significant, although reverse-direction estimates were smaller in magnitude. Moderator analyses reported by the included meta-analyses were inconsistent, whereas our longitudinal meta-regression found no evidence that mean age or gender composition systematically moderated associations. Overall, findings support a small but consistent association between adolescent social media use and poorer mental health, although longitudinal evidence suggests a more complex and heterogeneous relationship than simple one-way accounts would imply. Future meta-analyses should strengthen conceptual clarity and improve methodological transparency.