Big News: The Surgeon General Calls for a Warning Label on Social Media

Dr. Murthy is right. The evidence of widespread harm to adolescents is now strong.

The U.S. Surgeon General, Vivek Murthy, has long been a leader on the youth mental health crisis. He wrote a book in 2020 on loneliness (Together: The Healing Power of Human Connection in a Sometimes Lonely World), and in 2023 he issued a landmark report on loneliness. In 2023, he also published a major report, a Surgeon General’s Advisory reviewing the research on social media and concluding that:

The current body of evidence indicates that while social media may have benefits for some children and adolescents, there are ample indicators that social media can also have a profound risk of harm to the mental health and well-being of children and adolescents.

Yesterday, June 17, Dr. Murthy dropped a bomb: An essay in the New York Times in which he called for government-mandated warning labels on social media, akin to those that a previous Surgeon General called for in 1964, on cigarettes.

In this post, we want to support the Surgeon General by making two points:

1. In public health emergencies, we don’t wait for certainty

2. The empirical evidence backing up the Surgeon General is now very strong

We Don’t Wait for Certainty

Dr. Murthy starts off his essay like this;

One of the most important lessons I learned in medical school was that in an emergency, you don’t have the luxury to wait for perfect information. You assess the available facts, you use your best judgment, and you act quickly. The mental health crisis among young people is an emergency — and social media has emerged as an important contributor. Adolescents who spend more than three hours a day on social media face double the risk of anxiety and depression symptoms, and the average daily use in this age group, as of the summer of 2023, was 4.8 hours. Additionally, nearly half of adolescents say social media makes them feel worse about their bodies.

It is time to require a surgeon general’s warning label on social media platforms, stating that social media is associated with significant mental health harms for adolescents.

How different this is from the mindset used in many parts of the academic research community! When we review articles for academic journals or debate with our colleagues, we insist on very high levels of proof, at least beyond a reasonable doubt. We are very careful about what we admit in a scientific journal. We pride ourselves on our skepticism, and we treat one kind of error (publishing or believing something that turns out to be false) as far more serious than the other kind of error (failing to publish or believe something that turns out to be true). 

This is why the debate with our critics has been going on for five years and will likely continue for many more years before we reach widespread consensus, as happened with public health threats such as tobacco and leaded gas. Most of our critics don’t argue that social media is harmless; their position is that the evidence we have amassed is not strong enough to be confident that social media is harmful. Nearly all now agree that there are associations between hours of usage and levels of depression/anxiety, but some critics argue that the correlations are too small to have caused such a big change in rates of adolescent mental illness. They agree that many experiments show benefits from quitting social media, but also point out some studies that show no benefits and even some that suggest harm from quitting social media (more on this in a moment).

This is how normal academic debates work. Their critiques have forced us to dig deeper and think harder. In purely “academic” science, we have plenty of time to debate the details of measurement, operationalization, and statistical analyses used in each study. 

The public health community is part of the scientific community too, but their time horizons are shorter, and their cost-benefit analyses are very different. Because threats to public health can arise quickly and spread rapidly, they do not have the “luxury to wait for perfect information.” Public health officials must consider the costs of both kinds of errors. 

Suppose that everyone who thinks that social media has harmed their children is wrong. Suppose everyone has simply fallen for another “moral panic,” and so America enacted the steps urged by the surgeon general or the four norms that we recommend in The Anxious Generation. What harms would result from that mistake? 

1. A lot of kids would be forced to wait until high school to get social media (as the Surgeon General urges) 

2. Phones would be locked up during the school day, so kids would spend more time playing and talking together in person—and sometimes even listening to the teacher.

3. Very little money would be wasted, as the four new norms cost nothing to implement beyond the cost of phone lockers.¹

But what if the millions of parents, teachers, principals, psychologists, and members of Gen Z who believe that social media is harming young people turn out to be correct, and yet America continues to do nothing, waiting another decade for the psychological research community to reach wide consensus? In this case, the cost of additional suffering and learning loss is almost incalculable. It would mean another ten years of kids experiencing unprecedently high rates of anxiety, depression, self-harm, and suicide. Another ten years of students with declining educational attainment due to having a massive distraction machine in their pockets, disrupting their focus with notifications every few minutes. It would mean another ten years of rising costs to all 50 states for psychiatric emergency room visits and subsequent hospitalizations as the mental health crisis rolls on. 

The Surgeon General is absolutely right that our top goal here cannot be certainty; it must be protecting children based on the best available evidence.

The Evidence of Widespread Harm Is Now Very Strong

Since 2019, we (Jon and Zach, along with Jean Twenge) have been collecting and organizing all the relevant published studies we can find — posting the titles, links, and abstracts in publicly visible Google docs. We now have dozens of such documents on many aspects of the crisis, from the scale of the youth mental health crisis to the impact of phone-free schools. We take all the studies on all sides, and we lay them out in neatly categorized buckets so that everyone can see for themselves what the nature of the evidence is on all sides. 

Our main document is titled

Social Media and Mental Health: A Collaborative Review

Please check that out for yourself, or take our guided tour of the document in this Substack post:

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

or in this more recent update:

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

Jon described much of this evidence in The Anxious Generation, and we have gone much deeper into it in many substack posts. For the benefit of journalists and general readers interested in the Surgeon General’s recommendation, here is a brief narration of the many kinds of evidence and the points that we believe we have established as we have responded to arguments from other researchers. The links are mostly to Substack posts here at After Babel, where we go into great detail on each point.

1. The adolescent mental health crisis is real

From the 1990s through the mid-2000s, there was little sign of any youth mental health problem in the U.S. in any of the long-running nationally representative datasets. But by 2015, adolescent mental health was a 5 alarm fire, with steeply rising rates of loneliness, anxiety, depression, self-harm, and suicide. 

The trends are not plausibly explainable by changing diagnostic criteria or by a greater willingness of Gen Z to report mental illness as the stigma around it declined. No explanation has been given as to why de-stigmatization proceeded suddenly and rapidly around 2012 onward and affected only the young, and no evidence of such rapid de-stigmatization has been provided. If there has been de-stigmatization, it seems likely it’s been going on for decades. Yet teen depressive symptoms barely budged between 1991 and 2011 and then suddenly shot upward. Behavioral data also show significant rises in ER admissions and hospitalizations for self-harm episodes among adolescent girls, as well as a rise in actual suicides for both boys and girls since 2010.

2. The crisis is international, happening across most of the developed world

We have done an extensive study of 10 nations (Anglo and Nordic) and found rates of depression, anxiety, and other measures of ill-being have been rising since 2010 in all of them. We have also published a study of Europe using the Health Behavior in School-Age Children Survey, which found that high psychological distress is rising across nearly all Western European nations (with the exception of Spain), and many Eastern European nations. (Note that we have worked to address where and why cultural variation exists, including some Eastern European nations and within more religious communities in the United States).

In addition, economist David Blanchflower has found—across a variety of datasets—that happiness used to be U-shaped by age, with middle age being the least happy; but that is not true anymore. Young adults (ages 18-25) are now the least happy. He has found this trend in 82 countries around the world. He argues that the changes began around 2014.

Emergency department visits and hospitalizations for non-fatal self-harm among adolescent girls have been rising in the five Anglo countries. When we look at suicide rates and split the results by sex and age, there are increases in teen girl suicides during this period, including in the U.S., UK, Australia, Canada, and several other nations. Rates of suicide are very high among adolescent girls in New Zealand but have been at those high levels for many years. In fact, suicide rates among Gen Z girls are higher than any previous generation before them that we have data for. (We do acknowledge variation in self-harm and suicide across nations).

[Note that there are a few reports and studies that do not report the same kind of global trends that Blanchflower has found. But there are a number of issues with using these datasets to look at year-to-year trends among young people in specific nations. The three include: The Gallup World Poll, The Global Burden of Disease, and WHO Global Health Estimates.²]

3. There is no viable alternative explanation for the multi-national collapse of adolescent mental health other than the arrival of the “phone-based childhood” across the developed world in the early 2010s. 

In The Anxious Generation and here at After Babel, we do not blame social media for the entire disaster. We argue that the multi-national collapse of youth mental health is primarily driven by the loss of the play-based childhood, culminating in the “great rewiring” of childhood that happened between 2010 and 2015 as the new phone-based childhood became the norm. (The loss of community is a big part of the story too, predating and contributing to the loss of the play-based childhood.) Social media is a big part of the story, especially for girls, so it is the first place to take action, and we applaud the Surgeon General for calling for warning labels.

Some of our critics say we are barking up the wrong tree, and, in fact, all this talk about social media will distract researchers from finding the real causes of the epidemic. OK, well, what are those real causes? We, along with Jean Twenge, have addressed over a dozen alternative explanations for the changes in poor youth mental health since 2010 and why they cannot be the primary driver of the youth mental health crisis. A number of these alternatives—from American school shootings to the opioid epidemic—certainly impact some young people’s mental health, but we are trying to figure out what changed between 2010 and 2015 to cause the upsurge in mental health problems, especially among adolescent girls in many nations around the world at roughly the same time.

4. The correlational studies are clear: heavy users of social media are much more likely to suffer from depression and anxiety

There is little dispute that social media use is correlated with depression and anxiety, especially for girls and heavy users (see Section 1 of our Review document). In our review, we list dozens of correlational and longitudinal studies (note that there are a number of longitudinal studies that control for reverse causation). These large datasets provide correlation coefficients that indicate how variables are related and highlight when and for whom those relationships are stronger. The studies consistently show that heavy social media users are at a much higher risk of mental illness or poor mental health compared to others. For instance, a widely cited 2018 study of 14-year-olds found that girls who spend five or more hours on social media per day are three times as likely to be depressed as those who use it minimally or not at all. For boys, the risk ratio is closer to two-to-one. Additionally, a meta-analysis of 26 studies found that the risk of depression increased by 13% for each additional hour spent on social media, with the increase being even higher for girls (also see this study by Jean and Jon). Over recent years of debate, there has been a convergence of opinions on the size of the correlation coefficient, as Jon explains in Section 4 of this post.

5. The experiments mostly show that reducing social media usage improves mental health — as long as you stay off for at least a week

Researchers who claim that social media is unlikely to be harmful to teens on a large scale routinely point out that “correlation does not imply causation.” They want to see experiments using random assignment, with a control condition, before they will grant any claim about causation. It is worth pointing out that there has never been and never will be an experiment in which hundreds of groups of children are randomly assigned to have a phone-based childhood while hundreds of other groups of children are randomly assigned to have a play-based childhood. That’s the experiment that would come closest to capturing what happened to children in the great rewiring. 

Instead, we have access to a few dozen experiments that tried something much more modest: They randomly assigned individuals (mostly college students or adults) to substantially reduce their social media usage, while others were randomly assigned to the control group, where no changes in social media behavior were mandated. What happened in these studies? In those that used a very short period of abstinence, aspects of well-being such as momentary satisfaction generally went down. This is not surprising; when people who are heavy users of something are deprived of their dopamine-producing substance or activity, they don’t feel relief or pleasure; they undergo withdrawal symptoms, which include anxiety, insomnia, and a general feeling of unease (dysphoria). But when we limit the analysis to only those studies that required (and obtained) at least a week of reduced usage and measured symptoms of depression or anxiety, nearly all of them find mental health benefits. Experiments such as these strongly imply causation. The Surgeon General is on very strong ground.

(We note that a meta-analysis of these kinds of studies was published in May that purported to show that, taken together, these experiments produced an effect that was statistically no different from zero. However, we and others have found several problems with that meta-analysis, especially the inclusion of short-term deprivation studies, the failure to include a few studies, three major errors in calculating effect sizes, and a number of other methodological choices that bury true effects. We will publish a report on this meta-analysis soon. See this recent analysis of it for now.) 

6. Parents, educators, and Gen Z say that the phone-based childhood damages adolescent mental health.

Most people who spend time with young people—particularly parents and educators—have witnessed firsthand the negative impacts of social media. They see the harm and understand that it goes beyond anxiety and depression; it interferes with sleep, friendship, family life, and attention. The harms can range from public shaming and anonymous cyberbullying to sextortion schemes and ongoing sexual harassment. Most teens know someone (if not themselves) whose despair is related to something happening on social media. Many young people report that they would prefer to live without social media as long as everyone else did too. 

Young people are aware of the harms of the phone-based childhood. Many blame social media as the primary reason for the decline of mental health for their generation.

There is even evidence from the companies themselves. In an internal study from Facebook, the company’s researchers reported that:

“Teens blame Instagram for increases in the rate of anxiety and depression,” said one slide. “This reaction was unprompted and consistent across all groups.”

Numerous youth-led organizations have spontaneously emerged in response to the harms that young people have experienced due to the design and nature of social media platforms (these organizations, collectively, work with thousands of young people). We have looked for young people who think the phone-based childhood has been good for their generation and have found hardly any. 

Those skeptical of the negative impact of social media on youth mental health seem to believe any appearance of harm is just correlation, not causation. But few members of Gen Z seem to agree with them.

As one Gen Z author (and After Babel staff writer), Freya India expressed,

Gen Z were the guinea pigs in this uncontrolled global social experiment. We were the first to have our vulnerabilities and insecurities fed into a machine that magnified and refracted them back at us, all the time, before we had any sense of who we were. We didn’t just grow up with algorithms. They raised us. They rearranged our faces. Shaped our identities. Convinced us we were sick.

It Is Time to Act

Perhaps the most powerful passage in Dr. Murthy’s essay was this one:

One of the worst things for a parent is to know your children are in danger yet be unable to do anything about it. That is how parents tell me they feel when it comes to social media — helpless and alone in the face of toxic content and hidden harms. 

Millions of parents all across America (and the world) feel helpless; many have lost hope that anything can be done, ever. But Dr. Murthy’s call for warning labels is a huge step forward, for he is saying that we should treat social media as we treat other defective consumer products:

It doesn’t have to be this way. Faced with high levels of car-accident-related deaths in the mid- to late 20th century, lawmakers successfully demanded seatbelts, airbags, crash testing and a host of other measures that ultimately made cars safer. This January the F.A.A. grounded about 170 planes when a door plug came off one Boeing 737 Max 9 while the plane was in the air. And the following month, a massive recall of dairy products was conducted because of a listeria contamination that claimed two lives.

Social media is harming children and teens at a level many orders of magnitude higher than airplanes and dairy products, yet at the federal level, America has done nothing whatsoever to recall, regulate, or even warn parents about the consumer product that dominates our children’s lives and that thousands of parents believe was implicated in  their children’s deaths. To take just one class of harm: The FBI recently investigated 13,000 reports of sextortion that took place on social media platforms. They identified 20 boys whose suicides were directly related to the trauma and shame of this crime, which means the true number of sextortion suicides is many times higher. The social media companies, which are among the richest and most powerful in the world, must be made to bear some degree of responsibility for their actions, for their inactions, and for the harm to children that they facilitate. 

Thank you, Dr. Murthy, for recognizing the struggles that American parents and teens are living through. Thank you for standing up for common sense and rallying parents, state legislators, and perhaps even Congress to take collective action. We can change this if we act together. 

1. Some may argue that delaying kids' entry into the virtual world will leave them less prepared. However, there is no evidence to support that delaying access would be harmful to teens. In fact, they would have more time to develop the social, emotional, and regulatory skills needed to manage online life better. Others might argue that kids will rebel and find ways around these bans anyway—just as they do with cigarettes and alcohol. The goal is not perfection; it is to reduce access so that children do not feel compelled to spend time on social media platforms simply because everyone else is doing so.

2. The Gallup World Poll includes data collected annually since 2005 from 168 countries. But it has very small year-to-year sample sizes, especially when broken down by age. The Global Burden of Disease is not actual data; it is an effort to estimate the levels of various disorders using other available variables. In one Substack post, we show that the GBD systematically and substantially underestimates official national statistics and fails to detect the large increases in poor mental health and suicide that occurred in the 2010s. The WHO Global Health Estimates estimate suicide rates in countries around the world. However, the WHO Global Health Estimates have underestimated actual suicide rates in many nations since 2010, including the U.S., UK, Australia, Canada, and France.