Temporal Associations Between Social Media Use and Depression



Document Type

Journal Article

Publication Date


Publication Title

American Journal of Preventive Medicine





First Page


Last Page





Introduction: Previous studies have demonstrated cross-sectional associations between social media use and depression, but their temporal and directional associations have not been reported. Methods: In 2018, participants aged 18–30 years were recruited in proportion to U.S. Census characteristics, including age, sex, race, education, household income, and geographic region. Participants self-reported social media use on the basis of a list of the top 10 social media networks, which represent >95% of social media use. Depression was assessed using the 9-Item Patient Health Questionnaire. A total of 9 relevant sociodemographic covariates were assessed. All measures were assessed at both baseline and 6-month follow-up. Results: Among 990 participants who were not depressed at baseline, 95 (9.6%) developed depression by follow-up. In multivariable analyses conducted in 2020 that controlled for all covariates and included survey weights, there was a significant linear association (p<0.001) between baseline social media use and the development of depression for each level of social media use. Compared with those in the lowest quartile, participants in the highest quartile of baseline social media use had significantly increased odds of developing depression (AOR=2.77, 95% CI=1.38, 5.56). However, there was no association between the presence of baseline depression and increasing social media use at follow-up (OR=1.04, 95% CI=0.78, 1.38). Results were robust to all sensitivity analyses. Conclusions: In a national sample of young adults, baseline social media use was independently associated with the development of depression by follow-up, but baseline depression was not associated with an increase in social media use at follow-up. This pattern suggests temporal associations between social media use and depression, an important criterion for causality.

Open Access

Green Accepted