- A Book Review of The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness, by Jonathan Haidt.
When an academic writes a book for a popular audience, one of their main goals is to have an impact on the world. Jonathan Haidt’s new book The Anxious Generation is clearly already having an impact. The book has been sitting near the top of the New York Times bestseller list since it was released. But it’s already having an impact where academics care the most: public policy.
Consider the following set of recommendations that Arkansas Governor Sarah Huckabee Sanders recently gave in her state-of-the-state address:
- Experts suggest goals like no smartphones before high school; no social media before 16; phone-free schools; and more outdoor play and childhood independence.
I listened to that speech at the same time I started reading Haidt’s book. Here is how Haidt summarizes the suggestions in the conclusion of his book (page 290):
- 1. No smartphones before high school
- 2. No social media before 16
- 3. Phone-free schools
- 4. Far more unsupervised play and childhood independence
It seems quite clear that, either directly or indirectly, the expert that Sanders is getting her recommendations from Haidt. It’s almost word-for-word. She prefaced her list with this context:
- Study after study shows that too much social media exposure leaves our kids anxious and depressed. Suicide rates for young teens have tripled since 2007. Depression among teenagers is up 150%. 30% of teenage girls now seriously consider suicide.
This paragraph could easily have been a blurb for Haidt’s book. Clearly, Haidt is having an impact, and a very immediate impact on national conversations, that most academics could only dream of from a lifetime of doing research on a topic and trying to inject it into the public debate.
Given Haidt’s clear impact, it is useful to examine his book in detail, and the research that he (and others, like Governor Sanders) cites to support the conclusion that the mental health issues teenagers are experiencing in America (and elsewhere) are the direct result of social media use on smartphones.
And states are not waiting for more research to take action. Even before Sanders’ speech this year, Arkansas has already been at the forefront of restricting social media use for teenagers, with the 2023 “Social Media Safety Act” which requires age verification to use social media, and requires parental consent for anyone under age 18. While I have not seen Haidt comment on the Arkansas law, he praised a similar Utah law and said that every state should do the same.
From Free Play to Instagram Addiction
Haidt’s story of generational decline is neatly summed up by his four recommendations from above, but in reverse order. It starts with number 4 on his list, the decline of “free play.” For most of human history, kids had a lot of unsupervised play with kids their own age. Haidt tells us that this is an important part of childhood development, both in terms of learning how to interact with others (and resolve disputes) and to correctly wire young brains. Starting in the 1980s, parents began to give kids in the United States and other countries less time to engage in free play, in response to real and perceived threats to children in the world, driven by media stories about child abductions and murders.
That’s part one of the story. But the real problem doesn’t arise until decades later in Haidt’s telling. The decline in free play left a void in child development, but also a void in how children spent their time. One way that time was occupied was with more intensive parenting, as parents started spending a lot more time actively involved with their kids (rather than their kids playing with other kids). Structured “play” time was also increasingly introduced, such as organized team sports, music and dance lessons, and all variety of clubs for kids to engage their interests.
There were, of course, screens to occupy the time of kids. Lots of screens: TVs, VCRs and DVDs, video games, computers, computers with the internet. Then in the mid-to-late 2000s, a new and dangerous form of screen entered the scene: the smartphone. Around the same time, social media began to grow in influence, among both young and old. 2010 is a crucial year: the newest iPhone adds a front-facing camera and Instagram is introduced. After 30 years of mulling around with the decline of free play, kids finally found something to really occupy their time: spend 5 hours a day perfecting their online persona.
The lack of actual social connections with other young people, and the replacement of it with pseudo-, often toxic, social connections online has led to a crisis of mental health. Haidt documents in chapter 1 the “tidal wave” of suffering in the United States, especially among young people, and particularly among girls. This suffering includes not just self-reported or diagnosed mental health issues (e.g., 30 percent of teen girls with major depression, more than doubling since 2010), but obviously objectively bad outcomes, such as suicide, rising 167 percent in a decade for girls ages 10-14 in the United States.
A crisis, no doubt. The cause? Haidt is convinced and spends the rest of the book trying to convince the reader, that the double-whammy of declining free play in the 1980s and the rise of smartphones and social media in the 2000s is the culprit.
Screens Are Annoying
Any parent will tell you that screens are a double-edged sword. On the one hand, screens provide additional learning and entertainment opportunities, as well as temporary distractions for kids when adults need a break or need the kids quiet. This was true from TVs all the way up through the iPad (also introduced in 2010).
One of the biggest downsides of screens is that they are annoying. Most of the entertainment is pretty mindless, whether it was the Honeymooners in the 1950s or Cocomelon today. What’s worse is kids become dependent on the screens. They whine for them when they don’t have them, and they seem unable to sit still in normal social settings. Kids never misbehaved in the past, of course (I’m only half kidding).
Screens are not just disruptive for young people: they often disrupt adult social interactions too. The annoyance of screens, though, is not what Haidt’s book is about. Haidt is concerned very narrowly on smartphones with access to social media; this is clear from his first three recommendations. And he’s primarily concerned with mental health, not annoyance per se. Because Haidt is a good social scientist, he demands the best evidence, for he knows that two things can often follow each in time trends but either be unrelated, or both be caused by some third factor.
Is It Causal?
“Haidt argues very strongly that the relationship between social-media-equipped smartphones and teen mental health is causal, not a mere correlation. What is the evidence?”
Haidt argues very strongly that the relationship between social-media-equipped smartphones and teen mental health is causal, not a mere correlation. What is the evidence? For a book that asserts this so heavily, you might be surprised to learn that there is just one paragraph focusing on randomized-control trials that address this question. It can be found on page 148. In that paragraph he summarizes just two studies that attempt to measure the effect of social media on mental health. In a footnote, he tells us further that there are 14 RCTs showing harm, and another 6 that found no harm (but he regards these 6 as low quality studies), and then points us to an online Google Document that he put together with his collaborators.
There are multiple documents that he has put together on his website that relate to the research behind this book, but the one on Social Media and Mental Health runs 356 pages, longer than the text of the book itself! Haidt is to be applauded for putting this all online transparently, but as a social-science nerd, I would have liked to see this take up more than a single paragraph of the book. A few chapters perhaps? But that probably wouldn’t have landed the book on the New York Times bestseller list.
Anyway, back to the 356-page Google Document. The discussion of RCTs begins on page 168 (of the current version as I write—this is a living document) and runs for 20 pages. There are now 23 studies showing causal negative effects, and another 8 studies showing no effect, 11 more studies than when the book went to print just a few weeks ago. I won’t dive into all 31 of these studies, but given that Haidt is laser-focused for both the trends and policy recommendations on teenage girls, how many of the RCT studies would you guess are about teenage girls? The answer: just one. The other papers study college undergraduates, adults, or young adults.
It’s not that there is nothing we can learn about teenage mental health by studying people older than them. But what’s so striking about this fact—just one study of teenage girls!—is that Haidt is so confident in his overall hypothesis despite the evidence being so razor thin. And what of this one single study? Also interesting: it wasn’t even teens in the United States, but rather in the Netherlands. Again, there is nothing wrong about studies outside of the United States, but as I will argue below, the worst of the mental health problems seem confined to teenage girls in the United States, yet we have no studies of teenage girls in the United States.
The paper in question is well done. It randomly assigns girls to two groups, and one group shows manipulated Instagram photos that make the subjects more attractive. The girls that received the treatment reported lower body satisfaction, about 0.4 points on a 9-point scale. This result is statistically significant, but… is it enough to worry us? Is it good proof that social media is causing a mental health crisis, when you have a small change on the scale of body image from one single study of 144 teenage girls in the Netherlands, with no follow-up for long-term effects? This seems, to me, to be a very weak reed to build an entire apparatus of restricting phone use for teenagers around the world.
Is It Really Happening Everywhere?
While Haidt spends much of the book discussing evidence from the United States, he suggests that this is a global phenomenon. In Chapter 1, he spends three pages expanding his charts on teen mental health to other English-speaking countries (Canada, the United Kingdom, Australia). He has another two and half pages on “the rest of the world,” but the evidence here is pretty thin: a chart on psychological stress in Nordic nations, and a chart on alienation in school by broad regions (Asia, Europe, and English-speaking Latin America).
As with the summary of studies, there are other Google Docs to consult from Haidt and collaborators. For example, while the Nordic nations get just one paragraph and one chart in the book, the Nordic adolescent mood disorders online document runs over 100 pages. And much of that document is less certain and ambiguous than the text of the book. While there is plenty of evidence of rising mental health diagnoses and self-reports, evidence on self-harm and suicide doesn’t show increases. In some cases, it shows decreases. In Denmark, self-harm was reduced by almost 50 percent from 2007 to 2016 among teenage girls and boys—there had been a rise in the decade before 2007, but it came back down after that. Teenage suicides in Sweden exhibited a similar pattern, with a rise from 2000 to about 2008, then coming back down.
To go beyond the Nordic Nations, another Google Doc (Haidt is very thorough and transparent) on Adolescent mood disorders is useful. But the studies summarized in Section 3.6, looking outside of the Anglosphere (United States, Canada, United Kingdom, Australia, New Zealand) are not very convincing. The first study they summarize looks at suicides rates for ages 15-24 from 2006-2017 in several high-income countries. The only clear increases are in the Anglosphere, and even if we limit the analysis to girls, only Spain is added to the unfortunate group of rising youth suicide. France and Italy are declining, Germany and Poland are flat. There is no evidence that social media and smartphones have proliferated less in those countries than the Anglosphere. They also look at OECD data, but it is no more promising for their hypothesis: “Teenage suicides rates have, on average, declined slightly over the past two decades or so.”
There does seem to be something particularly bad happening in the United States and other large English-speaking countries, but our non-English-speaking peer nations aren’t seeing the same trends (though some are seeing the rise in reported mental health issues).
What Is to Be Done?
We can think of the question “What is to be done?” in two ways. Haidt is convinced that the evidence is overwhelming on the connection between social media, smart phones, and teen mental health. If you are also convinced, the thing to be done is find policy solutions or suggest changes in social behavior.
But the second way to think of “What is to be done?” is to think about what further research needs to be done to better understand the relationship between social media and teen mental health. Perhaps the two questions can be merged: targeted policy interventions could also produce good research results, to be used for future potential interventions.
On page 263-64, Haidt suggests just that merging of the questions, when he proposes that state educational authorities set up random-assignment of schools into one of four groups, such as phone-free, free-play, both, and a control group (status quo policy). What’s really important here is to note that no such studies exist or Haidt would have cited them. He is really making recommendations without much good evidence yet. But as Haidt notes earlier in the book (page 249), 77 percent of schools in the United States already say that they ban phones—they just aren’t enforcing the bans.
So clearly schools have this power (as they do have the power to limit all sorts of student behaviors and activities), they just aren’t using it. Haidt says: use that power but use it in a way that we can learn from it. If the governors of, say, Utah, Florida, and Arkansas (three states that have passed some restrictions on youth social media use) took this opportunity to conduct randomized experiments on schools, other states could learn from their experiments. It may seem cruel to treat school children as test subjects, but that’s actually we are already doing, we’re just doing it poorly and in a way that is hard to study the actual effects.
Unfortunately, it does not seem that any well-done studies have tried this randomized approach yet (you may have heard about a new paper supposedly on phone bans in Norway, but Haidt acknowledged on social media that the paper “does not really tell us much”).
What’s a Parent to Do?
Most of Haidt’s recommendations aren’t clearly directed at anyone, but rather are directed at everyone. The “no social media before 16” recommendation could be a call for laws. But the book can also be read as advice to parents. Indeed, Haidt’s final chapter before the conclusion of the book is pitched as advice to parents.
Haidt correctly identifies the nature of the issue: this is a collective action problem. Deciding on their own, most parents would probably follow Haidt’s recommendations about delaying the use of smartphones and social media. But one family alone making this decision is in a difficult spot, given that most other families are allowing their kids to use smartphones and social media. How to back out of this social dilemma?
For more on these topics, see
Haidt has a lot of advice for parents! Not all of it is backed by randomized controlled trials, but most of what we do as parents doesn’t have strong evidence (but thank you to Emily Oster for trying to write several books which do bring together the evidence that exists). He breaks down the evidence by the age of your children, and certainly some it is involves less screen time (or more productive uses of screens), but much of the advice is just good parenting advice. For example, for kids ages 6-13, Haidt recommends things such as encouraging sleepovers, walking to school, free play after school (instead of “enrichment activities), camping, sleepaway camps, and forming child-friendly neighborhoods. These suggestions have nothing to do with screens (though some are to avoid the temptation of screens), and are all good advice regardless of whether social media is causing a mental health crisis.
Parents are always making choices under uncertainty. That’s the nature of parenting. But hopefully public policy requires better evidence for imposing rules on the entire population, especially when public policy is precisely the area that has the power to conduct controlled experiments, so that we may possibly find out what is going on with youth mental health in the English-speaking world.
*Jeremy Horpedahl is Associate Professor of Economics at the University of Central Arkansas. He blogs at Economist Writing Every Day.
For more articles by Jeremy Horpedahl, see the Archive.