Social media use in adolescents with and without mental health conditions

Abstract

Concerns about the relationship between social media use and adolescent mental health are growing, yet few studies focus on adolescents with clinical-level mental health symptoms. This limits our understanding of how social media use varies across mental health profiles. In this Registered Report, we analyse nationally representative UK data (N = 3,340, aged 11-19 years) including diagnostic assessments by clinical raters alongside quantitative and qualitative social media measures. As hypothesized, adolescents with mental health conditions reported spending more time on social media and were less happy about the number of online friends than adolescents without conditions. We also found hypothesized differences in social media use by condition type: adolescents with internalizing conditions reported spending more time on social media, engaging in more social comparison and experiencing greater impact of feedback on mood, alongside lower happiness about the number of online friends and lower honest self-disclosure. In contrast, those with externalizing conditions only reported higher time spent. These findings emphasize the need to consider diverse adolescent mental health profiles in policy and clinical practice.

Fig. 1. Differences in social media use for the three group comparisons.

Top: hypothesis 1 (H1, any mental health condition versus no condition). Middle: hypothesis 2 (H2, internalizing/externalizing versus no condition). Bottom: hypothesis 3 (H3, internalizing versus externalizing condition). Data are presented as mean differences based on Hedges’s g effect size (g) and its corresponding 90% CI. The shaded area indicates the SESOI (g = 0.4, corresponding to d = 0.4). If the 90% CI lies completely within the SESOI, we concluded equivalence, and therefore no meaningful differences. Bolded effect sizes reflect comparisons that supported our hypotheses, while faded effect sizes reflect comparisons that did not support our hypotheses. The starting sample size includes social media users (N = 519 for adolescents with any mental health condition, N = 282 for internalizing conditions, N = 104 for externalizing conditions and N = 2,821 for no mental health condition). However, the exact sample sizes for each comparison are reported in Supplementary Tables 5 and 6 and differ for each dimension of social media use, given that we planned to analyse those separately. Source data.