Yes, Social Media Really Is a Cause of the Epidemic of Teenage Mental Illness

Two major problems with a review in Nature

For centuries, adults have worried about whatever “kids these days” are doing. From novels in the 18th century to the bicycle in the 19th and through comic books, rock and roll, marijuana, and violent video games in the 20th century, there are always those who ring alarms, and there are always those who are skeptics of those alarms. So far, the skeptics have been right more often than not, and when they are right, they earn the right to call the alarm ringers “alarmists” who have fomented a groundless moral panic, usually through sensational but rare (or non-existent) horror stories trumpeted by irresponsible media. 

But the skeptics are not always right. I think it is a very good thing that alarms were rung about teen smoking, teen pregnancy, drunk driving, and the exposure of children to sex and violence on TV. The lesson of The Boy Who Cried Wolf is not that after two false alarms we should disconnect the alarm system. In that story, the wolf does eventually come. 

The question before us now, on the topic of teens and social media, is this: Are the skeptics correct that we are going through just one more groundless moral panic over teens and tech in which adults are freaking out while, in fact, the harms are so minimal that they shouldn’t be a cause for worry? Or did the wolf really arrive around 2012, and has been mauling young people ever since via their smartphones and social media accounts? (Of course, there are researchers who reside in the space between these two perspectives.)

Psychologist Candice Odgers has taken the skeptical side for many years now (along with researchers including Amy Orben, Andrew Przybylski, Jeff Hancock, Chris Ferguson, and Aaron Brown), while Jean Twenge and I have been writing as alarm ringers. It has been a normal and productive academic debate. Engagement with each other’s arguments is how science makes progress. Even if we never convince each other, the broader scientific and policy communities tune in to the debate, and eventually, they’ll move one way or the other.

Odgers recently stated the skeptics' case in an essay in Nature titled The Great Rewiring: Is Social Media Really Behind an Epidemic of Teenage Mental Illness? The essay offered a critique of my recent book, The Anxious Generation. Odgers’ primary criticism is that I have mistaken correlation for causation and that “there is no evidence that using these platforms is rewiring children’s brains or driving an epidemic of mental illness.”  She also warns that my ringing of a false alarm “might distract us from effectively responding to the real causes of the current mental-health crisis in young people,” which, she suggests, are social ills such as racism, economic hardship, and the lingering impact of the 2008 Global Financial Crisis and its disparate impact on children in low SES families. 

In this response essay, I’ll present the two main problems I see with the skeptics' approach, as exemplified in Odgers’ review: 

1. Odgers is wrong to say that I have no evidence of causation

2. Odgers’ alternative explanation does not fit the available facts.

1) Odgers is wrong to say that I have no evidence of causation

Odgers’ central claim is that I have mistaken correlation for causation and that I have “no evidence” that social media is a cause, rather than a mere correlate, of the current epidemic of adolescent mental illness. Odgers says that I am just “making up stories by simply looking at trend lines.”

In 2018, when I entered this debate as a co-author of The Coddling of the American Mind, it was true that the great majority of published studies on “digital media” and mental health were correlational, and every social scientist knows that correlational studies often suggest to our intuitive minds a causal pathway that vanishes when experiments using random assignment are published. (Think of the changing guidance on the health effects of fat, carbs, and red wine). But even in 2018 there were a few experimental studies on social media and mental health. For example, college students who were asked to reduce their social media use for three weeks generally experienced mental health benefits compared to the control group. Zach Rausch, Jean Twenge, and I began to collect all the studies we could find in 2019, and we organized them by type: correlational, longitudinal, and experimental. We put all of our work online in Google Docs that are open to other researchers for comment and critique. You can find all of our “collaborative review” documents at AnxiousGeneration.com/reviews

The main document that collects studies on social media is here:
Social Media and Mental Health: A Collaborative Review

In that document, we list dozens of correlational and longitudinal studies. These large datasets provide us with correlation coefficients that tell us how variables in the dataset are related to each other, and they tell us when and for whom those relationships are stronger. Those studies reveal a fairly consistent relationship in which heavy users of social media are at much higher risk of mental illness or poor mental health than everyone else. A widely cited 2018 study of 14-year-olds found that girls who spend five hours or more on social media per day are three times as likely to be depressed as girls who use social media only a little or not at all. For boys, the ratio is lower, closer to two-to-one. A meta-analysis of 26 such studies found that the risk of depression increased by 13% for each hour increase on social media for adolescents (and that increase was even higher for girls).¹ 

In that document, we also list 22 experimental studies, 16 of which found significant evidence of harm (or of benefits from getting off of social media for long enough to get past withdrawal symptoms). To give just a few examples: 

1. Alcott and colleagues (2020) randomly assigned 2743 adults to either deactivate their Facebook accounts for one month or not. This study also found that deactivation significantly improved subjective well-being and that “80% of the treatment group agreed that deactivation was good for them.” The treatment group was also more likely to report using Facebook less and having uninstalled the app from their phones post-experiment.

2. Brailovskaia and colleagues (2022) have done one of the only studies I have seen that incorporate both social media reduction and physical activity increases. They randomly assigned 642 participants to (1) reduce social media by 30 minutes a day for two weeks, (2) increase physical activity by 30 minutes a day for two weeks, (3) follow both instructions, or (4) do nothing. The researchers found the strongest effects within the combined condition (#3). This group reported the largest decreases in depressive symptoms and increases in life satisfaction and subjective happiness compared to other groups. 

3. There are also a number of experiments that have looked at Instagram's unique negative impacts on women, including the finding that it is more harmful to women than is Facebook.

In that document, we also list nine quasi-experiments or natural experiments (as when high-speed internet arrives in different parts of a country at different times), eight of which found evidence of harm to mental health, especially for girls and women. (Odgers cited only the 9^th one, which relies on a dataset that is made up of unreliable estimates that do not detect known rises in mental illness, as Zach showed here.) To give just one example: Arenas-Arroyo and colleagues (2022) looked at the links between the staggered rollout of broadband internet in Spain between 2007-2019 and hospital discharge diagnoses of behavioral and mental health cases of adolescents. They found a significant effect of the arrival of high-speed internet, but only among adolescent girls.

There are also a number of studies showing mental health improvements, increases in physical activity, and reductions in bullying when schools go phone-free—one of the few natural experiments that specifically targets both adolescents and group-level effects. (These group-level effects are illustrated brilliantly here). 

I am not saying that academic debates are settled by counting up the number of studies on each side, but bringing so many studies together in one place gives us an overview of the available evidence, and that overview supports three points about problems with the skeptics’ arguments.

First, if the skeptics were right and the null hypothesis were true (i.e., social media does not cause harm to teen mental health), then the published studies would just reflect random noise² and Type I errors (believing something that is false). In that case, we’d see experimental studies producing a wide range of findings, including many that showed benefits to mental health from using social media (or that showed harm to those who go off of social media for a few weeks). Yet there are hardly any such experimental findings. Most experiments find evidence of negative effects; some find no evidence of such effects, and very few show benefits. Also, if the null hypothesis were true, then we’d find some studies where the effects were larger for boys and some that found larger effects for girls. Yet that’s not what we find. When a sex difference is reported, it almost always shows more harm to girls and women. There is a clear and consistent signal running through the experimental studies (as well as the correlational studies), a signal that is not consistent with the null hypothesis.

Second, and most germane to Odgers’ review: She and the other skeptics are free to critique the hundreds of studies Zach and I cite in our essays at After Babel and in chapters 1, 5, 6, and 7 of The Anxious Generation. They can certainly say that they are not persuaded and that they will not be persuaded until the perfect experiment is done. But they cannot say that I am drawing only on correlational studies, or that I have “no evidence” of causation, or that I do not understand the difference between correlation and causation. I laid out the evidence for causality (not just correlation) between social media use and mental health outcomes for girls and walked the reader through the Google Doc and multiple kinds of evidence in this post in early 2023:

Social Media is a Major Cause of the Mental Illness Epidemic in Teen Girls. Here’s the Evidence.

I presented several conceptual problems with the skeptics’ claims about causality and evidence in this essay: Why Some Researchers Think I’m Wrong About Social Media and Mental Illness. For example, I noted that the skeptics focus on testing one narrow model of causality that treats social media consumption as if it were an individual act, like consuming sugar, and then looks for the size of the dose-response relationship in individuals. But much of my book is about the collective action traps that entire communities of adolescents fall into when they move their social lives onto these platforms, such that it becomes costly to abstain. It is at that point that collective mental health declines most sharply, and the individuals who try to quit find that they are socially isolated. The skeptics do not consider the ways that these network or group-level effects may obscure individual-level effects, and may be much larger than the individual-level effects.

Third, even beyond the published experiments, there is ample evidence of causation that should be relevant to parents, lawyers, and legislators: eyewitness testimony. When you ask members of Gen Z what they think is causing their high rates of mental illness, they often point to social media and particularly Instagram as a major cause.³ (See section 4 of our main collaborative review doc.)  I have been searching for essays by members of Gen Z that defend social media and the phone-based childhood as being good for mental health. I can’t find them, but I find plenty of members of Gen Z who say that social media damaged them or their generation. Also relevant: many members of Gen Z are starting organizations to fight back.

Meta collected some eyewitness testimony accidentally. In that famous internal study brought out by whistleblower Frances Haugen, the researchers found that Instagram is particularly bad for girls. They wrote: “Teens blame Instagram for increases in the rate of anxiety and depression. . . . This reaction was unprompted and consistent across all groups.” Quantitative researchers such as Odgers are free to place less weight on qualitative studies, but they are a kind of evidence in social science research. They are relevant when we try to sort out multiple theories about causation, especially when the platforms will not share data with scientists so the experts with the deepest insights into what social media is doing to teens are the teens themselves. They see it happening. Do they count as “no evidence” since their claims are not peer-reviewed?

2) An Alternative That Does Not Work

The second major problem with Odgers’ review is that she proposes an alternative to my “great rewiring” theory that does not fit the known facts. Odgers claims that the “real causes” of the crisis, from which my book “might distract us from effectively responding,” are longstanding social ills such as “structural discrimination and racism, sexism and sexual abuse, the opioid epidemic, economic hardship and social isolation.” She proposes that the specific timing of the epidemic, beginning around 2012, might be linked to the 2008 Global Financial Crisis, which had lasting effects on “families in the bottom 20% of the income distribution,” who were “also growing up at the time of an opioid crisis, school shootings, and increasing unrest because of racial and sexual discrimination and violence.”

I agree that those things are all bad for human development, but Odgers’ theory cannot explain why rates of anxiety and depression were generally flat in the 2000s and then suddenly shot upward roughly four years after the start of the Global Financial Crisis. Did life in America suddenly get that much worse during President Obama’s second term, as the economy was steadily improving? 

Her theory also cannot explain why adolescent mental health collapsed in similar ways around the same time in Canada, the UK, Australia, and New Zealand, as Zach and I showed in this post:

The Teen Mental Illness Epidemic is International, Part 1: The Anglosphere

Adolescent Self-harm Episodes in Five Anglo Nations

Figure 1. Since 2010, rates of self-harm episodes have increased for adolescents in the Anglosphere countries, especially for girls. For data on all sources and larger versions of the graphs, see Rausch and Haidt (2023). (Data for Canada is limited to Ontario province, which contains nearly 40% of the population of Canada.)

Nor can she explain why it also happened at roughly the same time in the Nordic countries, which lack most of the social pathologies on Odgers’ list, as Zach and I showed in this post:

The Teen Mental Illness Epidemic is International, Part 2: The Nordic Nations

Percent of Nordic Teens with High Psychological Distress  (Ages 11-15)

Figure 2. Percent of Nordic Teens with High Psychological Distress. Data from the Health Behavior in School-Age Children Survey (2002-2018). Graphs and data were organized, analyzed, and created by Thomas Potrebny and Zach Rausch. See 1.1.1 of Nordic Adolescent Mood Disorders since 2010.

Nor can she explain why it also happened in much, though not all, of Western Europe:

The Youth Mental Health Crisis is International Part 4: Europe

Percent of European Teens with High Psychological Distress by Country-Level Individualism (Ages 11-15)

Figure 3. Changes in psychological distress by high vs. low individualism of each country, split by sex. Source: Health Behavior in School-Age Children Survey, 2002-2018 (See Zach’s spreadsheet for data points) and Hofstede Insights.⁴

Nor can she explain why suicide rates for Gen Z girls (but not always boys) are at record levels across the Anglosphere, as we showed in this post:

Suicide Rates Are Up for Gen Z Across the Anglosphere, Especially for Girls

Rises in Suicide Among Adolescent Girls (Ages 10-19)

Figure 4. All datasets and figures can be found in the post. 

If Odgers was correct that the “real causes” of the epidemic are America’s social ills, then we would not find these patterns in so many countries. I just can’t see a causal path by which America’s school shootings, lockdown drills, poverty, or racism caused girls in Australia to suddenly start self-harming or dying by suicide at the same time as so many American girls.

An equally large problem for Odgers’ explanation is that it commits her to the prediction that the increases in mental illness were largest for teens in low SES families. After all, her explanation for why there was a four-year delay between the onset of the GFC and the onset of the mental health crisis was because the effects lasted longer for those “in the bottom 20% of the income distribution,” who “continue to experience harm.” 

Jean Twenge tested Odgers’ explanation by looking to see whether rates of major depressive episodes increased faster for teens in families below the poverty line (shown in red in Figure 5 below) versus those whose families’ incomes were at least double the poverty line (shown in blue). As you can see, there was no difference between the two groups up through 2012 (which is contrary to Odgers’ thesis about the differential impacts of the GFC on mental health), and then a difference opened up after 2012, but in the opposite direction of Odgers’ prediction.

Percent of American Adolescents with Depression by Family Income Level (Ages 12-17)

Figure 5. Percent of U.S. 12- to 17-year-olds experiencing major depressive episodes in the last 12 months, by family income level. Source: Data from the nationally representative National Survey of Drug Use and Health; analysis by Jean M. Twenge for the Generation Tech Substack.⁵ 

Twenge has also addressed 13 other possible explanations for the mental health crisis and shows why they don’t fit the data.

In short: Odgers has pointed to an alternative causal explanation that A) does not fit the timing in the U.S., B) does not fit the social class data in the U.S., and C) does not fit the international scope of the crisis. 

What Now?

Parents, teachers, and legislators cannot wait any longer; they want to do something about the ever-rising levels of anxiety, distraction, and suffering. Whose policy prescriptions should they follow, given the current state of the evidence and the relative risks and costs associated with each path? 

If leaders and change-makers were to embrace Odgers’ causal theory about the “real causes” rather than being “distracted” by mine, then the way forward is to first solve society’s biggest social problems, most of which we have been working on for decades. Perhaps this time we’ll make more progress, and in ten or twenty years, rates of teen mental illness will begin to decline. 

And if Odger’s causal theory turns out to be wrong? We’ll have spent another decade or two locked in the usual culture war battles over spending on social programs that may or may not be effective, and we will have lost another generation to mental illness. 

In contrast, if leaders and change makers were to embrace my account of the “great rewiring of childhood,” in which the phone-based childhood replaced the play-based childhood, what policy implications follow? That we should roll back the phone-based childhood, especially in elementary school and middle school because of the vital importance of protecting kids during early puberty. More specifically, we’d try to implement these four norms as widely as possible: 

1. No smartphones before high school (as a norm, not a law; parents can just give younger kids flip phones, basic phones, or phone watches).

2. No social media before 16 (as a norm, but one that would be much more effective if supported by laws such as the proposed update to COPPA, the Kids Online Safety Act, state-level age-appropriate design codes, and new social media bills like the bipartisan Protecting Kids on Social Media Act, or like the state level bills passed in Utah last year and in Florida last month).

3. Phone-free schools (use phone lockers or Yondr pouches for the whole school day, so that students can pay attention to their teachers and to each other)

4. More independence, free play, and responsibility in the real world.

Note that these four reforms, taken together, cost almost nothing, have strong bipartisan support, and can be implemented all right now, this year, if we agree to act collectively.

And what if it turns out that I am wrong? What if, in reality, the multinational collapse of adolescent mental health in the early 2010s was not caused by the arrival of phone-based childhood; it was just a big coincidence. Will kids be damaged by these four norms? I don’t think so. What irreversible harm will be done to children who spend more time listening to their teachers during class, more time playing and exploring together outdoors, and less time sitting alone hunched over a device?

We are now 12 years into a public health emergency that began around 2012. In The Anxious Generation, I offer a detailed explanation of what caused it (drawing on many academic fields) and a detailed path by which we can reverse it. I know of no plausible alternative explanation, nor have I found anyone offering a realistic alternative pathway out.

We certainly need skeptics to challenge alarm-ringers, who sometimes do ring false alarms. God bless the skeptics. But at a certain point, we need to take action based on the most plausible theory, even if we can’t be 100% certain that we have the correct causal theory. I think that point is now.

P.S., See Zach’s follow-up post addressing claims by skeptical researchers who think the increases in rates of mental illness are exaggerated.  

1. Odgers claims that the correlation is actually reverse correlation: depression causes teens to spend more time on social media. It’s true that some longitudinal studies have found that an increase in depression precedes an increase in time on social media, but other studies have found the reverse, and some have found that the relationship is bidirectional.

2. Yes, there could be a “file drawer problem” if researchers on one side are systematically discouraged from publishing, so the missing “positive” studies are all sitting in file drawers in researchers’ offices. But because findings of benefits would be unusual and newsworthy, I don’t believe that there is a strong or consistent bias against the skeptics. 

3. I focus here on eyewitness testimony from members of Gen Z, but there is ample eyewitness testimony from teachers, and from parents, who see the effects of heavy phone use on classroom behavior, learning, and family life. 

4. High individualistic countries include United Kingdom, Hungary, Netherlands, Italy, Belgium, Denmark, France, Sweden, Ireland, Latvia, and Norway. Low individualistic countries include Austria, Slovakia, Spain, Russia, Greece, Croatia, Bulgaria, Romania, Slovenia, Portugal, Ukraine, and North Macedonia.

Measure for distress: We used a measure of high psychological distress from the HBSC that includes four items that tap different symptoms of psychological distress: feeling low, feeling nervous, feeling irritable, and having sleep difficulties. Respondents rated how often they felt these four symptoms over the last six months (1: Every day, 2: More than once/week; 3, About every week; 4, Every month; 5, Rarely or never). We operationalized high psychological distress as having three or more of the four psychological ailments every day or more than once a week over the last six months. 

Using this measure, we computed the average psychological distress score in each country, for girls and for boys separately, in each of the 5 survey years. This gave us two new datasets (one for girls, one for boys), each with 33 rows (one for each country), which facilitated the various cross-country comparisons that we report.

5. Note: This is a screening study of the population, not a study of diagnoses or treatment. Thus the differences cannot be caused by willingness or ability to seek treatment. 2020 data was excluded as sampling times varied from other years.