In my previous post, we discovered that the teen mental health crisis might not be as dramatic as The Anxious Generation, Jonathan Haidt’s book, claims. Today, we’re diving deeper into the research that’s supposed to prove social media harms our teens and why that research isn’t as convincing as it first appears.
Jonathan Haidt asserts in The Anxious Generation that “dozens of experiments” confirm that social media causes anxiety and depression in teens. But when we examine these studies closely, they reveal flaws that undermine their conclusions.
Take the “No More FOMO” experiment that Haidt cites. Researchers told 143 psychology students they were studying social media use, then asked some to limit their Facebook, Instagram, and Snapchat time. Here’s the catch: The students knew exactly what the study was about. They’d heard countless times that social media is bad for mental health. When researchers asked them to report on their well-being, is there a chance they didn’t know the “right” answer?
Even with this built-in bias toward finding problems, the study barely found anything. No effects on anxiety, self-esteem, or overall well-being. There were small changes in loneliness and depression, and only for students who were already struggling.
Another study Haidt referencesshowed teen girls’ Instagram selfies—some original, some digitally enhanced. The enhanced photos led to lower body satisfaction. But again, every teenage girl has likely heard that perfect Instagram images harm body image. When you tell participants you’re studying “facial preferences” and then show them obviously manipulated photos, you’re practically telegraphing what you want them to say.
Recognizing the obvious issues with controlled experiments, researchers turned to “natural experiments,” studying what happens when broadband internet rolls out to different regions at different times. The logic seems sound: If social media harms mental health, we should see clear declines as internet speed improves.
But the results don’t support this thesis as cleanly as Haidt suggests. In Spain, researchers found effects only in men, not women, despite the study’s abstract claiming effects for “both males and females.”
The Italian study mostly covered a period before widespread social media use, yet still found mental health effects, suggesting other online activities might be responsible.
The English study was the most rigorous, tracking 6,000 children across 3,765 neighborhoods. But the results were puzzling: Broadband was associated with better exam performance from ages 10 to 11 but worse performance at age 16. The largest effect was a tiny 0.6 percent decrease in how children felt about their appearance.
Here’s one of the most revealing critiques in this entire debate. Dr. Amy Orben at Cambridge University found that teens’ digital technology use correlates with their well-being at about the same level as their well-being correlates with eating potatoes.
Both correlations were statistically significant in a dataset of over 60,000 people. Both explained similar tiny amounts of variance in teen well-being—less than 1 percent.
We don’t blame potatoes for teen depression. So why do we blame social media?
This illustrates a crucial problem with big datasets. When you have enough participants, you can find statistically significant correlations between almost anything. The question isn’t whether the correlation exists. It’s whether it matters in the real world.
When we look at the full body of research on social media and teen mental health, three patterns emerge:
The effects are tiny. Even studies that find negative effects usually explain less than 1 percent of how teens feel. That’s like saying a teen feels sad because they didn’t eat breakfast while ignoring their family problems, financial stress, school pressure, and sleep issues.
The effects aren’t consistent. Dr. Orben’s research shows that scientists can get very different results from the same data depending on how they analyze it. Some studies pick “the most negative possible” way to interpret their findings.
Within-person effects are even smaller. Studies that track how changes in one person’s social media use affect their own well-being over time show even smaller effects than studies comparing different people at one point in time.
If social media explains less than 1 percent of teen well-being, what explains the other 99 percent? Research consistently points to several factors that matter far more:
1. Family relationships top the list. In a UK study of teens who showed up at emergency rooms for self-harm, 64 percent cited family relationships as their primary problem. Mental health issues, supposedly driven by social media, ranked fifth.
2. Social connections matter enormously, but it’s about quality, not quantity. Strong friendships can be especially protectivewhen teens aren’t getting support from family.
3. Economic security has massive impacts that dwarf any effects from screen time. Children’s suicide rates are higher in counties with greater poverty concentration.
4. Sleep and physical health create both causes and effects of mental health struggles. While screens can interfere with sleep, so can family stress, feeling unsafe in your neighborhood, and early school start times.
5. Academic pressure shows up in research consistently. A meta-analysis of 52 studies found evidencelinking academic pressure to mental health problems in 48 of them. Consider the case studies of Palo Alto, California, and the anonymized “Poplar Grove” community, where suicide rates are four to five times the national average. These teens had everything Haidt says should protect them from social media’s harms, like tight community bonds, involved parents, shared values. Yet they experienced devastating mental health crises.
6. School environment creates another major impact on teen mental health. Recent data shows troubling increases in the percentage of students who were threatened or injured with a weapon at school, bullied at school, or missed school due to safety concerns.
The research tells us that fighting over phones might be the wrong battle. Instead of focusing on screen time limits, we might ask: Am I a parent my teen feels safe talking to? How are their friendships? Do they seem overwhelmed by school?
This isn’t about ignoring technology’s role. It’s about putting it in proper perspective. When we create blanket policies based on flawed research, we risk overlooking the real factors that affect our teens’ well-being. We also risk harming teens who benefit from digital connections, like LGBTQ+ youth who use social media as a lifeline when their families don’t accept them.
The evidence points to a complex interplay of factors, of which screen time makes up a fairly small part. That means our response needs to be more nuanced than simply taking away phones.
If you or someone you love is contemplating suicide, seek help immediately. For help 24/7 dial 988 for the 988 Suicide & Crisis Lifeline, or reach out to the Crisis Text Line by texting TALK to 741741. To find a therapist near you, visit the Psychology Today Therapy Directory.