Mental Health
Estimating causal effects of screen time and social media use on adolescent wellbeing: an instrumental variable analysis across 54 countries Yi Yang* Yi Yang Yang Yang Yang Unit for Lifelong Health and Ageing, UCL, UK; Department of Psychiatry, University of Oxford, Oxford, UK
Background: The relationship between adolescent screen time, social media use, and wellbeing is debated, partly because most studies use observational data and methods prone to reverse causation and confounding. Yet, policy reforms and public health interventions targeting adolescents’ digital media use are being implemented often ahead of definitive causal evidence, underscoring the urgent need for robust, causally informed research.
Methods: We apply an instrumental variable approach to estimate causal effects of digital use on wellbeing, using data from 615,133 15-year-olds in 54 countries. To address endogeneity, country- and year-linked instrumental variables were used: International Telecommunication Union (ITU) internet usage for screen time and DataReportal social media penetration for social media use. Causal effects were estimated with two separate instrumental variable models (Two-Stage Least Squares), controlling for individual demographic and socioeconomic covariates plus country-level GDP per capita and urbanization. A Three-Stage Least Squares framework including both exposures and their interaction was also applied. Extensive robustness checks assessed instrument validity and quantified uncertainty.
Results: Higher screen time and more frequent social media use were causally linked to lower life satisfaction, with consistent effects across subgroups defined by usage levels and gender. Among adolescents whose digital use was influenced by the instruments, screen time and social media use were associated with approximately 1- and 0.8-point reductions in wellbeing, respectively. Robustness tests indicated strong instrument relevance and offered supporting evidence for the plausibility of IV assumptions.
Conclusion: Leveraging cross-national variation in digital infrastructure, this study provides timely and robust causal evidence to inform public health and policy responses to adolescents’ digital media use.