In May 2023, U.S. Surgeon General Dr. Vivek Murthy issued a formal advisory on social media and youth mental health, calling it "a profound risk of harm." The advisory synthesized existing research and surveillance data on adolescent social media use.

The report found that teens who use social media for more than 3 hours per day face double the risk of depression and anxiety symptoms compared to those who use it less. Up to 95% of youth aged 13-17 report using at least one social media platform, with average daily use of 3.5 hours. Nearly 40% of children ages 8-12 use social media despite minimum age requirements of 13.

The Surgeon General concluded that "we cannot conclude it is sufficiently safe for children and adolescents" and called on technology companies, policymakers, and parents to take immediate action. This was only the third advisory Dr. Murthy had issued, signaling the seriousness of the concern.

Published in Clinical Psychological Science in 2017, Jean Twenge and colleagues analyzed two nationally representative surveys of U.S. adolescents in grades 8 through 12, totaling 506,820 respondents, alongside national suicide statistics for ages 13-18.

Depressive symptoms and suicide-related outcomes increased significantly after 2010, coinciding with the widespread adoption of smartphones and social media. Among teens spending 5 or more hours per day on electronic devices, 48% reported at least one suicide-related outcome (seriously considering suicide, making a plan, or attempting suicide), compared to 28% of those spending less than 1 hour per day.

The increases were more pronounced among females. The researchers found that screen activities (social media, texting, gaming) were consistently associated with higher depression risk, while non-screen activities (sports, in-person socializing, homework, religious services) were associated with lower risk. The timing of the increase matched smartphone adoption more closely than any other proposed explanation.

Published in Psychiatric Research and Clinical Practice in 2020, Jean Twenge and Thomas Joiner analyzed data from the Monitoring the Future survey and other national tracking studies to document trends in adolescent mental health from 2009 through 2019.

The prevalence of major depressive episodes among adolescents aged 12-17 increased from 8.1% in 2009 to 15.8% in 2019, nearly doubling in a decade. Emergency room visits for self-harm among 10-to-14-year-old girls tripled between 2009 and 2015. Suicide rates among 15-to-19-year-olds increased by 57% from 2007 to 2017.

The increases began in 2011-2012, the same period when smartphone ownership among teens surpassed 50% and Instagram, Snapchat, and other image-based social media platforms gained mass adoption. The authors argued that the timing, consistency across multiple data sources, and the specificity of the increase (concentrated among girls, who use social media more) point to social media as a primary driver.

Published in the American Economic Review in 2022, economists Luca Braghieri, Ro''ee Levy, and Alexey Makarin exploited Facebook''s staggered introduction across U.S. colleges between 2004 and 2006 as a natural experiment, using a generalized difference-in-differences research design.

They found that the rollout of Facebook at a college campus caused a significant increase in symptoms of poor mental health, particularly depression. Students who were most susceptible to mental illness prior to Facebook''s arrival showed increased utilization of mental healthcare services after it became available. The primary mechanism identified was unfavorable social comparison.

Natural experiments are considered stronger evidence for causation than correlational studies because the "treatment" (Facebook arriving at campus) was determined by factors unrelated to student mental health. This design addressed the key objection that depressed people simply choose to use social media more, because the timing of Facebook''s arrival was driven by Mark Zuckerberg''s expansion schedule, not by student demand or pre-existing mental health trends.

Published in eClinicalMedicine (a Lancet journal) in 2019, researchers analyzed data from 10,904 fourteen-year-olds in the UK Millennium Cohort Study, a nationally representative birth cohort.

Compared to 1-3 hours of daily social media use, spending 3-5 hours was associated with a 26% increase in depressive symptom scores for girls (21% for boys). Spending 5 or more hours was associated with a 50% increase for girls (35% for boys). The dose-response relationship was stronger and more consistent in girls.

The researchers identified specific pathways through which social media use led to depression: online harassment, poor sleep quality, low self-esteem, and poor body image. These mediating factors explained a substantial portion of the association, suggesting that the content and social dynamics of social media platforms, rather than screen time alone, drive the mental health effects.

Published in JAMA Psychiatry in 2019, researchers analyzed three waves (2013-2016) of data from 6,595 adolescents in the Population Assessment of Tobacco and Health (PATH) study, a nationally representative U.S. longitudinal cohort.

At each assessment wave, teens spending 3 or more hours per day on social media had the highest risk of developing internalizing problems (depression and anxiety). Any amount of social media use was associated with greater risk of internalizing problems compared to no use, suggesting there is no safe threshold below which social media carries zero risk.

The longitudinal design, following the same teens over three years, provides stronger evidence than cross-sectional snapshots because it shows that social media use at one time point predicts later mental health problems. The nationally representative sample also means results can be generalized to the broader U.S. adolescent population.

Published in PLOS ONE in 2013, University of Michigan researchers conducted an experience-sampling study in which 82 young adults were text-messaged five times per day for two weeks, each time reporting their current feelings, Facebook use since the last message, and direct social interactions.

The more participants used Facebook between one sampling point and the next, the worse they felt at the subsequent measurement. Over the full two-week period, increased Facebook use predicted a decline in life satisfaction. In contrast, direct (in-person or phone) social interactions predicted feeling better over time.

The experience-sampling method captures real-time fluctuations in mood and behavior, avoiding the recall bias that affects traditional surveys. The within-person design (comparing each person to themselves at different time points) also controls for stable individual differences like personality, making the finding that Facebook use precedes mood decline harder to explain as mere correlation.

Published in the American Economic Review in 2020, Stanford and NYU economists conducted a large-scale randomized experiment paying 2,743 Facebook users to deactivate their accounts for four weeks before the 2018 U.S. midterm election.

Participants who deactivated Facebook showed increased subjective well-being (measured by both survey questions and a validated happiness index), reduced political polarization, and increased offline social activities. After the experiment ended and they could return to Facebook, the deactivation group reduced their subsequent Facebook use by an average of 23% compared to the control group, and their reported valuation of Facebook declined.

The reduction in post-experiment use suggests that some Facebook use is habitual rather than satisfying - people use it more than they would choose to if they experienced life without it. This is one of the strongest causal designs in the literature because random assignment eliminates the possibility that pre-existing differences between users and non-users explain the results.

The Centers for Disease Control and Prevention released results from the 2021 Youth Risk Behavior Survey (YRBS) in February 2023, documenting a decade of deteriorating adolescent mental health based on nationally representative data from U.S. high school students.

Among teen girls, 57% reported persistent feelings of sadness or hopelessness (lasting 2 or more weeks) in 2021, up from 36% in 2011, representing a 58% increase over ten years. 30% of teen girls seriously considered attempting suicide in 2021, up from 19% in 2011. Among boys, persistent sadness or hopelessness rose from 21% to 29% over the same period.

While the YRBS does not directly measure social media use, the acceleration in mental health problems beginning around 2012 closely tracks the period when social media and smartphone use became near-universal among American teens. The CDC identified social media as one of several factors warranting urgent attention in its accompanying analysis.

Published in the Journal of Abnormal Child Psychology in 2015, researchers at the University of North Carolina studied 619 adolescents (57% female, mean age 14.6) at two time points to examine whether specific social media behaviors, rather than mere time spent, predict depression.

Technology-based social comparison (comparing one''s appearance, popularity, or life to others seen online) and feedback-seeking (posting content to get likes and comments as validation) were associated with depressive symptoms above and beyond the effects of overall technology use frequency and prior depressive symptoms. The association was particularly strong among girls and adolescents low in peer popularity.

This study is important because it identifies the mechanism: it is not screen time per se that drives depression, but specific social comparison behaviors that social media platforms are designed to encourage. Features like likes, follower counts, curated photo feeds, and algorithmically ranked content all amplify the types of social comparison that predict depression.

Published in the American Journal of Epidemiology in 2017, researchers combined three waves of survey data (2013-2015) from 5,208 subjects in the nationally representative Gallup Panel Social Network Study with objective measures of Facebook activity (likes, posts, link clicks) obtained directly from Facebook.

Increased Facebook use was significantly associated with decreased self-reported mental health, physical health, and life satisfaction. Critically, the negative associations of Facebook use were comparable to or greater in magnitude than the positive impact of real-world offline social interactions. This means that for every unit of social interaction moved from offline to Facebook, the net effect on well-being was negative.

The use of objective Facebook data (rather than self-reported screen time) addressed a common criticism that people poorly estimate their own social media use. The longitudinal design with lagged variables also strengthened causal inference by showing that Facebook use at one time point predicted worse outcomes at the next.