Research

How does TikTok use affect news consumption? Leveraging detailed device-level mobile browsing and app use data and a shock to early adoption of TikTok as a novel, plausibly exogenous source of variation, I estimate the causal effect of TikTok use on news consumption, political knowledge and attitudes in the United States. I find that increased TikTok use substantially reduces news consumption related to major news events such as elections, national disasters and major political developments. TikTok use causes decreases in news consumption along both the extensive and intensive margins. It also reduces the length of individuals’ news consumption sessions, suggesting that reduced attention spans contribute to the effect. Survey evidence shows that the effect is not limited to mobile devices and has downstream consequences. Individuals in high-TikTok-use media markets have lower traditional news consumption and political knowledge, and higher partisan polarization in presidential approval after, but not before, TikTok’s proliferation.

Social media may harm mental health, but population-scale causal evidence remains limited. Existing studies mostly rely on self-reported outcomes, small samples, and short time horizons. We link population-scale health insurance claims across the social media era (2006–2020) to geographic variation in Twitter use. Mental-health claims increase 28% more over this period in metropolitan areas with above-median Twitter use. To isolate causality we leverage an exogenous shock to Twitter adoption at the 2007 South-by-Southwest festival. We estimate that a one-percentage-point increase in metropolitan-area adult Twitter use, relative to that year’s national average, led to 3.5 additional annual claims per 100 insured for anxiety, depression, and eating disorders. The increase is consistent from 2011–2020, larger for women and, among the privately insured, adolescents. These results suggest social media may harm mental health directly, through content, and indirectly, by changing social interactions.

How does social media affect the composition of political protests in the United States? Using early adoption of Twitter as a plausibly exogenous source of variation in county-level Twitter penetration, and comprehensive data on protest events, this paper finds that Twitter penetration increases the frequency of protests overall, but also radicalizes them. Twitter disproportionately fuels protests with participation of “extreme” groups — groups that are particularly militant, radical, or hateful — both on the political left and right. Survey data suggest that coordination is not the only mechanism: county-level Twitter penetration also intensifies opinions, plausibly increasing individuals’ willingness to partake in protests involving extreme groups.