Scrolling into substance use: Social media use frequency and substance use among U.S. high school students

Adolescent substance use imposes lasting costs on education, mental health, and lifetime earnings, with early initiation strongly predicting adult dependence. Despite legal prohibition, approximately 21% of U.S. high school students report current alcohol use, 18% report current nicotine vaping, and 16% report current cannabis use. Social media plausibly contributes to these patterns, yet prior studies rely on broad screen time measures that conflate passive consumption with compulsive checking. We analyzed 2023 Youth Risk Behavior Survey (YRBS) data, the first standalone measure of social media frequency in the YRBS series, from 10,027 U.S. high school students. We estimated survey weighted Linear Probability Models with state fixed effects, adjusting for demographics, mental health, electronic bullying, sleep, and physical activity. Students checking social media more than once per hour show a 19 %age point (pp) higher probability of current alcohol use, a 16 pp higher probability of current nicotine vaping, and a 13 pp higher probability of current cannabis use relative to nonusers, following a dose response gradient. Applied to 17.1 million U.S. high school students, this associates with approximately 1 million additional students drinking alcohol, 870,000 additionally nicotine vaping, and 681,000 additionally using cannabis among the 5.4 million students checking social media more than once per hour. Associations concentrate in alcohol, cannabis, and nicotine vaping, which are prominently normalized online. These findings inform legislative debates over engagement maximizing design features targeting adolescents and support extending content restrictions on alcohol and cannabis influencer marketing.

Adolescent substance use imposes lasting costs on educational attainment, mental health, and lifetime earnings, with initiation during adolescence strongly predicting adult dependence. Despite legal prohibition, approximately 21% of U.S. high school students report current alcohol use, 18% report current nicotine vaping, and 16% report current cannabis use. Social media plausibly contributes to these patterns, yet most prior studies rely on composite screen time measures that conflate passive consumption with compulsive social media checking, limiting inference about frequency-specific risk.

We analyzed 2023 Youth Risk Behavior Survey (YRBS) data, the first standalone measure of social media frequency in the YRBS series, comprising 10,027 U.S. high school students. We estimated survey-weighted Linear Probability Models with state fixed effects, adjusting for sociodemographic characteristics, mental health status, electronic bullying, sleep, and physical activity.

Students checking social media more than once per hour show 19 %age point (pp) higher probability of current alcohol use, a 16 pp higher probability of current nicotine vaping, and a 13 pp higher probability of current cannabis use relative to students reporting no use of social media, each following a dose-response gradient. Applied to 17.1 million U.S. high school students, these estimates associate with approximately 1 million additional students currently drinking alcohol, 870,000 additionally nicotine vaping, and 681,000 additionally using cannabis among the 5.4 million students checking social media more than once per hour.

Associations between social media use frequency and substance use concentrate in alcohol, cannabis, and nicotine vaping, the three substances most prominently normalized in platform content, while lower for cigarettes which are subject to platform content bans and strong social stigma. These findings inform ongoing legislative debates over engagement-maximizing platform design features targeting adolescents and support extending content restrictions on alcohol and cannabis influencer marketing equivalent to those already applied to tobacco.