Advancing social media data science to examine the causal relationship between social media content and adolescent health-risk behaviours

Social media platforms like TikTok may negatively shape adolescents' health behaviours, contributing to health inequalities. Exposure to alcohol, drug use, and anti-social behaviour (ADA) content on social media correlates with increased offline engagement in these behaviours. However, existing research relies on biased measures (e.g., time-use diaries, self-reports), which fail to capture the complexity of content exposure, limiting causal claims and evidence-based policy. Guided by a Policy Advisory Group - including stakeholders from Ofcom, the Department for Science, Innovation, and Technology - and Youth Advisory Groups, this project aims to address critical policy priorities by: (1) Developing innovative methodologies to accurately measure real-time social media content exposure by using a hybrid human-generative AI approach to identify ADA-content in TikTok data donations (obtained via the TikTok Application Programming Interface) combined with Ecological Momentary Assessment. (2) Assessing the causal effects of ADA-content on offline ADA-behaviours in adolescents (ages 13-24) from the socioeconomically and ethnically diverse Born in Bradford Age of Wonder and SmartHealth Young Adult Studies. Ultimately, the project will integrate public voices for real-world impact, inform public health strategies and legislation, and advance knowledge via a public engagement campaign featuring influential figures (i.e., United Nations, Idris Elba), high-impact publications, and policy briefings.