In September 2021, a Wall Street Journal investigation published a series of internal Facebook documents that became known as the Facebook Files. Among the leaked slides was one that stopped the news cycle cold: research conducted by Facebook's own scientists had found that Instagram worsened body image issues in one in three teenage girls. The slide asked, with an almost clinical detachment: "We make body image issues worse for one in three teen girls." Another internal presentation noted that 13.5 percent of teen girls said Instagram made thoughts of suicide and self-injury worse. These were not findings from academic critics or activist researchers. They were the conclusions of the company's own internal research team, conclusions that the company had declined to disclose publicly.
The Frances Haugen documents arrived into a public conversation that had been growing for years. Since at least 2017, when Jean Twenge published iGen — an analysis of generational survey data tracking the mental health of American teenagers — a credible scientific argument had been building that the smartphone-social media combination was connected to a measurable deterioration in adolescent psychological wellbeing. The charts in Twenge's book showed steep inflections in rates of depression, loneliness, anxiety, and suicide ideation that began around 2012, the year smartphone ownership crossed 50 percent among American teens, and that continued rising through the end of her dataset. The timing was not proof; it was a pattern that demanded explanation.
But the science was — and remains — more complicated than the headline version suggests. On the same question, a 2019 analysis by Amy Orben and Andrew Przybylski found that the effect size of social media use on adolescent wellbeing was comparable to wearing glasses or eating potatoes: statistically detectable but modest. Two serious researchers looking at overlapping evidence reached conclusions that seemed irreconcilable. Understanding why requires looking at what the research actually measured, how the studies were designed, and what each approach can and cannot tell us.
"We're in the middle of an epidemic of mental illness among adolescents, and smartphones and social media are at the center of it." — Jonathan Haidt, The Anxious Generation (2024)
Key Definitions
Passive social media use: Scrolling through feeds, viewing content, and observing others without direct interaction. Consistently associated with worse wellbeing outcomes than active use across multiple studies.
Active social media use: Posting, commenting, messaging, and creating content — forms of engagement that more closely resemble reciprocal social interaction and show weaker or neutral associations with mental health.
Upward social comparison: Evaluating oneself unfavorably by comparing to others who appear to be doing better, more attractive, more successful, or happier. The visual, curated nature of image-sharing platforms structurally promotes upward comparison.
Variable reward schedule: A reinforcement pattern in which rewards arrive unpredictably, shown by B.F. Skinner to produce the most persistent and compulsive behavior. The unpredictability of social validation (likes, comments, follower counts) on social platforms creates this schedule.
Displacement hypothesis: The theory that social media harms mental health not through direct psychological mechanisms but by displacing time that would otherwise be spent on activities with established wellbeing benefits — face-to-face interaction, sleep, physical activity, and unstructured play.
Jean Twenge, iGen, and the 2012 Inflection
Jean Twenge is a social psychologist at San Diego State University who has spent her career tracking generational differences in psychological survey data, particularly data from the Monitoring the Future survey — an annual nationally representative study of American high school students that has run continuously since 1975 — and the Youth Risk Behavior Survey. When Twenge updated her analyses for her 2017 book iGen, she found something that had not been visible in earlier editions of the data: around 2011 and 2012, a broad range of mental health and wellbeing indicators began moving in the same direction for American teenagers.
Rates of loneliness, depressive episodes, and feeling left out increased. Rates of in-person social interaction, sufficient sleep, and self-reported happiness declined. The rates of actual suicide — not just ideation — began increasing in teenage girls after 2007 and accelerated after 2012. Teen boys showed deterioration too, but the female adolescent data was especially striking. The timing aligned not perfectly but plausibly with the spread of smartphones and the rise of Instagram (launched in 2010, acquired by Facebook in 2012) and the shift of social media from desktop-based platforms that teenagers used in the evening to mobile platforms they carried everywhere, including to bed.
Twenge's interpretation was clear: she identified smartphones and social media as the most plausible causal candidates for the inflection. Not the only cause — she acknowledged that other factors, including economic stress following the 2008 financial crisis and shifting academic pressures, were concurrent. But the smartphone-social media shift was the factor most tightly correlated with the timing of the changes and most specifically associated with the activities that increased most (social comparison on image platforms) and decreased most (face-to-face interaction and sleep).
Jonathan Haidt, a social psychologist at New York University, has been the most prominent scientific advocate for Twenge's interpretation. His 2024 book The Anxious Generation synthesized the accumulated evidence across multiple countries — UK, Canada, Australia, and Nordic nations all show similar inflections at similar times — and proposed a four-mechanism model for harm: social deprivation (displacement of real-world friendship), sleep deprivation (phone use in bedrooms), attention fragmentation (constant notification interruption), and exposure to algorithmically amplified harmful content.
The Orben-Przybylski Challenge
Amy Orben, a researcher then at the University of Oxford, and Andrew Przybylski published a paper in Nature Human Behaviour in 2019 that challenged what they called "the popular narrative" around social media and wellbeing. Their analysis drew on three large-scale datasets — the Millennium Cohort Study (UK), the Understanding Society survey (UK), and the U.S. YRBS — with a combined sample of over 355,000 adolescents.
Their central finding was that the association between social media use and wellbeing, while statistically significant in large samples, was extremely small in magnitude. The effect size (beta = -0.05 in one dataset) was smaller than the association between wearing glasses and wellbeing or between regularly eating potatoes and wellbeing. They titled a commentary version of the argument the "Goldilocks hypothesis" — suggesting that both very low and very high social media use were associated with somewhat lower wellbeing, with moderate use showing the least harm.
Orben and Przybylski's critique raised legitimate methodological concerns. Large correlational datasets cannot establish causation. Screen time is typically self-reported and measured with poor precision. The aggregation of all social media uses into a single "screen time" variable obscures meaningful differences between platforms and modes of use. And effect sizes in social psychology are typically small; small does not mean unimportant, but it does mean that single-factor causal narratives are almost certainly oversimplifications.
Haidt and colleagues have argued in response that Orben and Przybylski's analyses are themselves limited by their averaging approach — that small mean effects can mask large effects in specific subgroups (teenage girls, heavy users, vulnerable individuals) and specific platforms (Instagram vs. YouTube, for instance). The methodological debate is ongoing, but the disagreement has been productive: it has pushed researchers toward more fine-grained analyses that distinguish by gender, platform, use type, and developmental stage.
The Facebook Files and What Internal Research Found
The documents leaked by Frances Haugen in 2021 provided a rare window into what a platform with access to its own behavioral data and internal research capacity had discovered about its effects on users. The findings were notable both for what they showed and for what the company had chosen to do with them.
Beyond the body image data, internal Facebook research showed that heavy Instagram use was associated with negative outcomes on multiple wellbeing dimensions for teenage girls. Presentations from internal teams described the platform as producing social comparison that users found harmful, and acknowledged that the algorithmic feed had been optimized for engagement in ways that prioritized emotionally activating content — including content that made users feel bad. One internal document noted that Instagram was "particularly harmful" to approximately 32 percent of teen girls who were already struggling with body image, and that the platform's own suggestions led them to more harmful content.
Haugen's testimony to the U.S. Senate in October 2021 included the claim that Facebook's internal research showed its products were causing harm to teenage users, that the company was aware of this harm, and that it had declined to make changes that would reduce engagement in order to protect revenue. The internal documents supported this narrative in specific areas, though the full picture from the files was more complex than the most dramatic summaries.
The significance of the internal research is not that it resolved the academic debate — internal corporate research has its own methodological limitations and motivations — but that it showed that one of the platforms at the center of the controversy believed its own product was causing harm to a significant subset of its youngest users.
Upward Social Comparison and the Instagram Effect
Whatever the overall effect size debate, the mechanism by which social media most plausibly harms wellbeing for certain users is well-established in psychological research: social comparison, specifically upward social comparison.
Leon Festinger's social comparison theory, originally published in 1954, proposed that people have a fundamental drive to evaluate their own opinions and abilities by comparing them to others. Upward comparison — comparing oneself to those who appear superior — can be motivating under some conditions but reliably produces negative affect and reduced self-evaluation under others. The conditions under which upward comparison is most harmful are those where the comparison is on a dimension the person values highly, where the target appears to be a peer rather than an aspirational distant figure, and where the comparison feels involuntary rather than chosen.
Instagram's design combines all three of these conditions. Its visual format makes appearance a salient comparison dimension. Its social network structure populates feeds with apparent peers. And its algorithm delivers content continuously, making comparison exposure feel involuntary. Kate Fardouly and Lenny Vartanian's 2015 review of experimental and correlational literature on social media and body image found consistent evidence that exposure to idealized images on social platforms reduced women's body satisfaction, with the effects strongest for women who were already body-image conscious.
A study by Vogel, Rose, Roberts, and Eckles published in Personality and Social Psychology Bulletin in 2014 found that exposure to upward comparison targets on a social network reduced self-evaluations of attractiveness and intelligence, while exposure to downward comparison targets raised them — a direct experimental demonstration of the mechanism in a social media-like context. The practical implication is that a feed populated primarily by people presenting idealized versions of their lives is not a neutral information stream; it is a systematic generator of upward comparison that most users are not consciously equipped to discount.
Passive vs. Active Use: Why What You Do Matters More Than How Much
One of the more actionable findings in the social media and mental health literature is the consistent distinction between passive and active use. Ethan Kross, a psychologist at the University of Michigan whose research program has focused on the effects of social media on emotion, and his colleagues found in a 2013 experience-sampling study that Facebook use predicted declines in moment-to-moment wellbeing and life satisfaction over time — but that this effect was driven primarily by passive consumption rather than active communication.
The distinction makes psychological sense. Passive scrolling — consuming others' content without interaction — provides the social comparison exposure discussed above, without the reciprocal validation that direct interaction provides. Active use — commenting, messaging, sharing personal content, coordinating real-world activities — more closely replicates the structure of normal social interaction and can genuinely maintain social connections. Studies by Verduyn and colleagues have found that active Facebook use was not associated with declines in wellbeing, while passive use consistently was.
Melissa Hunt's 2018 randomized controlled trial at the University of Pennsylvania assigned 143 undergraduates to either continue normal social media use or limit use to 10 minutes per platform per day (Facebook, Instagram, Snapchat) for three weeks. The limited-use group showed significantly lower depression and loneliness scores at the end of the trial, even though many of them had not initially reported high levels of social media use. The effect was most pronounced for participants who entered the study with higher baseline depression scores. This was among the first randomized experimental evidence that reducing use produces measurable mental health improvements — not merely that use and mental health are correlated.
Gender Differences and Developmental Vulnerability
The gender difference in social media's mental health associations is one of the most consistent findings in the literature and one of the least easily explained away as confound. Across Twenge's survey analyses, Haidt's international data, and multiple independent studies, girls and young women show substantially stronger negative associations between social media use and wellbeing than boys and young men.
Several explanations have been proposed. Girls' social media use skews more toward image-sharing platforms (Instagram) that emphasize appearance-based comparison, while boys' use skews more toward gaming and YouTube. Girls' social networks in adolescence place higher emphasis on social status and peer evaluation, making them more sensitive to social comparison and cyberbullying. Appearance-based self-evaluation is a stronger predictor of mental health in girls than boys across studies, and visual social media amplifies this domain specifically. Haidt has also argued that the timing of exposure matters: adolescence is a critical window for identity formation and social-emotional development, and heavy social media use may interrupt that development in ways that are more consequential in girls.
Boys are not unaffected — rates of gaming disorder, violent content exposure, and certain depressive indicators have risen — but the platform-use patterns and the magnitude of associations differ. This means that uniform recommendations about social media use will likely need to be gender-differentiated and platform-differentiated to be effective.
Displacement: What Social Media Is Replacing
The displacement hypothesis proposes that social media's most consequential effect on mental health may not be through its direct psychological content but through the time it displaces from activities with established positive effects on wellbeing. Face-to-face social interaction, physical activity, unstructured outdoor play, and sufficient sleep are among the most robustly supported predictors of adolescent psychological health. All four have declined among teenagers in the smartphone era.
Haidt's analysis of Monitoring the Future data found that teenagers in 2017 were sleeping less, spending less time with friends in person, and engaging in less physical activity than teenagers in 2012 — with the greatest declines among the heaviest phone users. If even a portion of the mental health deterioration is attributable to what phones are replacing rather than what they are providing, the intervention implication is the same: reducing phone use allows those beneficial activities to reclaim their place.
Research by Twenge and colleagues comparing weekday versus weekend social media use patterns found that in-person social activity was reliably associated with higher wellbeing, while electronic communication was associated with lower wellbeing — and that the association held even after controlling for how much time students spent in total on each activity.
Practical Takeaways
Use the 30-minute target. Hunt's randomized trial found significant mental health benefits at 30 minutes of total daily social media use. Most people cannot accurately estimate their own use — check your phone's screen time data before forming an opinion.
Shift from passive to active use. Replace scrolling with direct messaging, commenting, and genuine connection. Leave the feed and go to the people.
Remove social apps from your bedroom. Sleep displacement is among the most direct harms in the literature. Charging the phone outside the bedroom is the single simplest structural intervention with the clearest expected benefit.
Turn off all non-emergency notifications. Notifications pull you back into passive consumption at moments of your weakest resistance. Notifications for social apps serve the platform's engagement goals, not yours.
Be intentional about who you follow. Unfollow accounts that reliably produce negative comparison. Follow accounts associated with skill, learning, or genuine connection rather than aspiration and appearance.
Consider the developmental stakes for adolescents. The evidence for harm is strongest and most consistent for teenage girls. Parental norms around phone use — shared family use times, delayed smartphone ownership, keeping devices out of bedrooms — have support from both research and policy bodies.
References
- Twenge, J. M. (2017). iGen: Why Today's Super-Connected Kids Are Growing Up Less Rebellious, More Tolerant, Less Happy — and Completely Unprepared for Adulthood. Atria Books.
- Haidt, J. (2024). The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness. Penguin Press.
- Orben, A., & Przybylski, A. K. (2019). The association between adolescent well-being and digital technology use. Nature Human Behaviour, 3, 173-182.
- Hunt, M. G., Marx, R., Lipson, C., & Young, J. (2018). No more FOMO: Limiting social media decreases loneliness and depression. Journal of Social and Clinical Psychology, 37(10), 751-768.
- Kross, E., Verduyn, P., Demiralp, E., Park, J., Lee, D. S., Lin, N., & Ybarra, O. (2013). Facebook use predicts declines in subjective well-being in young adults. PLOS ONE, 8(8), e69841.
- Fardouly, J., & Vartanian, L. R. (2015). Negative comparisons about one's appearance mediate the relationship between Facebook usage and body image concerns. Body Image, 12, 82-88.
- Vogel, E. A., Rose, J. P., Roberts, L. R., & Eckles, K. (2014). Social comparison, social media, and self-evaluation. Personality and Social Psychology Bulletin, 40(2), 206-222.
- Verduyn, P., Lee, D. S., Park, J., Shablack, H., Orvell, A., Bayer, J., & Kross, E. (2015). Passive Facebook usage undermines affective well-being: Experimental and longitudinal evidence. Journal of Experimental Psychology: General, 144(2), 480-488.
- Haugen, F. (2021). Whistleblower Disclosure: Facebook Internal Research Documents. U.S. Senate Committee on Commerce, Science, and Transportation.
- Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117-140.
- Twenge, J. M., Martin, G. N., & Spitzberg, B. H. (2019). Trends in U.S. adolescents' media use, 1976–2016: The rise of digital media, decline of TV, and the (near) demise of print. Psychology of Popular Media Culture, 8(4), 329-345.
- Haidt, J., & Rausch, Z. (2023). Social media and mental health: A collaborative review. Unpublished manuscript. NYU Stern School of Business.
Related reading: Why We Get Addicted to Our Phones: The Neuroscience of Smartphone Compulsion | Why People Procrastinate: The Psychology Behind Putting Things Off
Frequently Asked Questions
Does social media cause depression and anxiety?
The research evidence is contested and more complex than most media coverage suggests. Large correlational studies by Jean Twenge and Jonathan Haidt show substantial associations between social media use and depression in adolescents, particularly girls. But Amy Orben and Andrew Przybylski's 2019 analysis of three large datasets found effect sizes comparable to wearing glasses or eating potatoes — statistically significant but modest. The most defensible current conclusion is that heavy social media use is associated with worse mental health outcomes for a subset of users — primarily adolescent girls engaging in heavy passive use and social comparison — but is not a universal cause of mental illness.
What does the research actually say about teens and social media?
Jean Twenge's analysis of the Monitoring the Future survey, published in her 2017 book iGen, documented a sharp inflection in adolescent mental health around 2012 — the year smartphone ownership passed 50 percent among American teens. Rates of depression, anxiety, loneliness, and suicide ideation rose steeply after that point, particularly in girls. Jonathan Haidt and colleagues have since built on this work to argue that the smartphone-social media combination is a primary driver of the adolescent mental health crisis. Critics note that correlation does not establish causation and that other factors — academic pressure, economic inequality, climate anxiety — changed simultaneously.
How does Instagram specifically affect body image?
The internal Facebook research leaked by Frances Haugen in 2021 — the so-called Facebook Files — included slides showing that Instagram's own researchers found the platform worsened body image issues in one in three teenage girls. The mechanism is upward social comparison: Instagram's visual format is structurally optimized for presenting idealized, filtered, and curated self-images, creating a stream of upward comparison targets that research by Vogel and colleagues shows reliably reduces self-evaluation.
What's the difference between passive and active social media use?
Ethan Kross and colleagues at the University of Michigan, along with several independent research groups, have found that passive social media use — scrolling through feeds without interacting — is more consistently associated with negative affect and reduced wellbeing than active use (commenting, messaging, creating content). The mechanism for passive use appears to involve social comparison and feelings of exclusion; active use more closely resembles normal social interaction and can be affirming. Melissa Hunt's 2018 randomized trial found that limiting use to 10 minutes per platform per day reduced depression and loneliness significantly over three weeks.
Why is social media designed to be addictive?
BJ Fogg's persuasive technology model, developed at Stanford, and Nir Eyal's Hook Model describe the engineering of variable reward schedules into social platforms. Likes, comments, and notifications function as intermittent reinforcement — the same mechanism that makes slot machines compelling. The unpredictability of social validation (will this post get many likes?) generates dopaminergic anticipation that drives repeated checking behavior. Former insiders including Sean Parker, the first president of Facebook, have acknowledged that this exploitation of psychological vulnerability was conscious and intentional.
How much social media use is too much?
Hunt's 2018 randomized trial found significant mental health benefits from reducing use to 30 minutes per day across all platforms. Orben and Przybylski's 'Goldilocks' analysis suggested that very low and very high use were both associated with lower wellbeing, with moderate use showing the least harm. For adolescents, Haidt recommends no social media before age 16 based on the developmental vulnerability evidence. Current average daily use among American adults is approximately 2-3 hours, with teenagers averaging closer to 5 hours.
What practical steps reduce the mental health impact?
Randomized trial evidence supports: limiting daily use to under 30 minutes per platform, turning off all non-emergency notifications, keeping phones out of bedrooms at night, replacing passive scrolling with active interaction, and periodic full detoxes. Structural approaches — using app timers, removing social apps from the phone's home screen, leaving the phone in another room during meals and social interactions — are more effective than relying on willpower alone, because they change the environment rather than the person.