The association between social media use and mental health symptoms in middle adolescence: A counterfactual analysis

There has been considerable recent debate surrounding the effects of social media use and adolescents’ mental health. A key source of controversy is whether observed associations might be confounded by ‘third variables’ (which influence both social media use and mental health). To address this, we used counterfactual analysis to account for measured confounding. This approach provides estimates of causal effects by comparing observed outcomes with those that would be expected under an alternative exposure condition. We employed it to evaluate the link between social media use (including messaging and visiting social networking websites) at ages 11 and 14 and later mental health issues (including emotional symptoms, self-harm, and/or suicide attempts) at ages 14 and 17. Data was from the UK Millennium Cohort Study (MCS), a nationally representative sample. The study was conducted with the input of a young persons’ advisory group, who informed our prioritization of the study research question, helped interpret findings, and identify study limitations. Inverse probability of treatment weighting (IPTW) analyses indicated no significant effects of social media use frequency at age 11 (defined as use on most days) or time spent at age 14 (≥2 hours per weekday) on emotional problems or self-harm at age 14 or 17, respectively (analytic n=3,036-4,419). The only exception was significant association between higher time spent on social media at age 14 and lifetime suicide attempts at age 17. These findings highlight the importance of considering potential confounders when examining social media use effects. Given that frequency and time-based measures are unlikely to capture the complexity of this association, we recommend future research also apply similar approaches utilizing more nuanced measures. • Study used counterfactual analysis to account for confounding variables • Data came from UK’s nationally representative Millennium Cohort Study • Found little evidence linking social media use to later mental health issues • Nuanced measures are needed to capture the complexity of social media effects