An analysis of more than 183,000 American adults surveyed over an 11-year span indicates that higher daily social media use is consistently linked to an elevated likelihood of reporting depression. Published in npj Mental Health Research, the study utilized machine learning models to evaluate demographic and media habits from 2014 through 2024. Researchers found that social media use ranked among the strongest predictors of self-reported depression, with its contribution to predicted risk rising once use exceeded about two and a half hours per day. These findings suggest that the connection between heavy social media engagement and depression is not limited to teenagers and has remained stable across a decade of evolving digital trends.
The research team, including Martin Block of Northwestern University and colleagues from the University of Cincinnati, examined data from the Media Behavior and Influence Study conducted annually by Prosper Analytics. The analytic sample included 183,300 adults aged 18 to 70, with yearly samples balanced to match U.S. Census benchmarks for age and gender. Participants reported their health conditions and estimated the likelihood of engaging in various media activities during each hour of a typical day. From these ratings, authors calculated average daily minutes spent on social media, internet browsing, television, gaming, email, and news consumption.
Using logistic regression and balanced random forest machine learning models, the researchers assessed how strongly different types of media use related to the odds of reporting depression while holding demographic factors constant. Across the 11-year period, both daily social media time and the share of adults reporting depression climbed steadily. The proportion of respondents reporting depression rose from 15.1 percent in 2014 to 22.0 percent in 2024. After adjusting for demographics, every additional minute of daily social media use was associated with a 0.26 percent increase in the odds of reporting depression, translating to roughly 17 percent higher odds for each additional hour spent on social media per day.
While personal income emerged as the strongest overall influence on predicted depression, social media use consistently ranked among the top predictors from 2019 through 2024, rising from ninth place in 2014. The machine learning models showed that below roughly 150 minutes per day, time spent on social media had a nearly neutral effect on classification, but beyond two and a half hours, additional time increasingly pushed predicted depression risk upward. The authors describe this 150-minute mark as a hypothesis for future studies rather than an established safe limit. Comparable positive links were also found for video games, internet browsing, television, and email, though news consumption showed a weak relationship in only two of the survey waves.
The study highlights that while social media use is a modifiable daily habit, high importance rankings may reflect how much a behavior varies across a population rather than proving that changing the behavior will reduce depression. Subgroup analyses revealed notable demographic patterns, with self-reported depression prevalence ranging between 24 percent and 40 percent for women aged 18 to 25 compared to 15 percent and 25 percent for men in the same age group. The machine learning models performed better among adults aged 45 and older than among the youngest age group.
Researchers caution that because the study relied on repeated cross-sectional surveys sampling new groups each year, it cannot establish whether heavy social media use causes depression. It is possible that adults already experiencing depressive symptoms spend more time online seeking distraction or connection. The relationship could also be bidirectional, with social media exposure contributing to distress while depressive feelings encourage more time online. Additionally, depression was assessed via a single self-reported question rather than clinical interviews, and the sample came from a commercial survey panel rather than a random national sample. Future longitudinal and experimental studies are needed to track individuals over time and determine if keeping daily social media use below specific thresholds offers meaningful mental health benefits.
Sources: PsyPost.