Kids These Days

What we are really arguing about

From Jonathan Haidt’s The Anxious Generation

Kids these days have it great. Compared with decades ago, young children benefit from more time with their parents and much more time with their fathers. Teenagers drink less, smoke less, and are less likely to die in car accidents. Life is much better for racial and sexual minorities—when I was in high school, someone rumored to be gay would be bullied savagely.

Children and adolescents are less likely to experience physical and sexual violence, and less likely to be murdered. They have access to cheap travel, better food, more information of all sorts, and limitless entertainment. And the internet offers an escape from solitude for those rejected in the real world, allowing weirdos, nerds, and the painfully shy to make social connections in ways that would have been impossible for their counterparts in earlier decades. There is no better time to be young.

For many readers, this is a ridiculous claim. There are good reasons to conclude that this is a terrible time to be young. Compared to their counterparts in previous decades, children and adolescents are more likely to spend their time alone; they devote hours each day to staring at their phones; they play less, sleep less, and read less for pleasure. They have listless and coddled lives. High school students today are more likely to be anxious and depressed (that is particularly true for girls), and are more likely to kill themselves. Our children are living in a time of crisis.

So who’s right? I’ve been reading about these topics for an academic chapter that I’m writing—which might, who knows, turn into a book proposal—and have some ideas that I’d love to get feedback on.

---

The question of whether there has been improvement or decline since, say, the 1980s, is an important one. Suppose things really have gotten worse. If so, then it’s an argument for radical change. The point is nicely made in No Country for Old Men, where Anton Chigurh says this to a man he is about to kill.

What sort of radical change? Well, perhaps we should restore the values, practices, and policies of the earlier period before things fell apart. We might want to dismantle capitalism, cut back on the welfare state, restore religious values, crush movements like feminism and anti-racism … or double down on them. (Everything depends on one’s preferred theory of why we are in this pickle.)

Alternatively, if things are going better, we should just stick with the plan, confident that, in the long run, the state of the world will continue to improve. (For a good defense of this view, see Steven Pinker’s Enlightenment Now.) Now this sort of optimism is currently unpopular, but it’s always been unpopular—particularly when it comes to children and adolescents. The Kids Are Never All Right.

Some might look back fondly on the lives of teens in the 1980s, before TikTok and smartphones, when kids hung out in shopping malls, listened to mixtapes, and waited in arcades to play Pac-Man. Maybe they are right to do so, but in the 80s, people were complaining that things had gone to hell. Here’s a popular list (the facts are totally fabricated, by the way) distributed at the time (from the Eibach and Libby chapter in this book).

My bet is that in the 1940s, adults really missed those great teens in the 1890s, who knew the value of work (only half of them went to school) and weren’t yet coddled by indoor toilets.

None of this means that the worriers are mistaken or that we are in the grip of a moral panic (though I might make this argument in another post). But it does suggest that the popularity of despair can’t be taken as good evidence that this despair is warranted.

---

But still, things really are rough for kids today. I opened this post with a graph from Jonathan Haidt’s Anxious Generation. Here’s another, looking at American children ages 10-14.

And another:

Haidt blames the screens, and particularly social media. He calls the period between 2010 and 2015 “The Great Rewiring”, where we moved from a “play-based childhood” to a “phone-based childhood”. Blaming the screens has caught on. Besides hatred of data centers, what holds Americans together is the view that you should keep social media away from the kids.

It’s hardly a crazy view. The timing of increases in anxiety, depression, and suicide does seem to implicate smartphones and social media. And the argument by Haidt and others for how they are bad for teenagers makes a lot of sense, at least to me.

But then you get to the research.

There are many, many studies now that look at the relationship between social media use and declines in mental health. Now, as everyone who does this research acknowledges, findings from correlational research are hard to interpret. Suppose you find a large association. This could mean that exposure to social media makes teenage girls depressed, just as Haidt and others claim. But it could also mean that depressed girls are more likely to go on social media. Or maybe certain personality traits in girls lead to both depression and social media use, but there is no causal relationship between the two outcomes.

We might not have to deal with this interpretative issue, though, because many investigators believe that there is either no relationship or a negligible one. One well-known study by Amy Orben and Andrew K. Przybylski found that the effect of digital technology on well-being is equivalent to that associated with eating potatoes and smaller than that associated with wearing eyeglasses or eating breakfast. Another more recent review, put together by the National Academies of Sciences, Engineering, and Medicine, examined all relevant research, including cross-sectional correlations, in which you look at social media exposure at a certain period and its later outcomes. They conclude:

Feelings of sadness, anxiety, depression, and stress are some of the most frequently studied outcomes related to social media use. Studies across adolescent samples in the United States and other countries do not find a consistent pattern... The larger sample sizes and increased statistical power associated with meta-analysis generally find small and inconsistent effects.

The issue isn’t settled, though. Replying to the potato paper, Jean Twenge, Jonathan Haidt, and colleagues argue that if you zoom in on girls (not adolescents in general) and on social media (not screens in general), you do find a small but real association between social media use and certain symptoms of depression. In response, Orben and Przybylski accuse Twenge et al of cherry-picking and using arbitrary cut-offs, and they emphasize the weakness of the effects, noting that the data suggest that “adolescents might have to use screens for upwards of 11 h 14 min each day before they would be able to perceive any negative effects.”

Well, screw correlations—what about actual experiments? There is an increasing body of studies that directly test the causal effects of social media exposure (for instance, by examining the effects of school-hour phone bans). These avoid the interpretative problems of correlational studies—though they have other issues.

And what do we find from these studies? One review of this literature by Christopher Ferguson finds that the effects of social media in these studies are “statistically no different from zero”. But Zach Rausch and Jonathan Haidt wrote a blistering critique of Ferguson’s meta-analysis that makes some convincing points. So maybe it’s more than zero? Still, by all accounts, the effects in question are not large. Commenting on the debate, Mike Males summarizes the issue as:

How tiny is the tiny effect social media has on users’ mental health? Is it next-to-nothing insignificant, or truly-nothing insignificant?

---

Many psychologists take these findings (or rather, the lack of findings) to mean that social media has no negative effects. Many of them are upset with Haidt for manufacturing a crisis with what they see as little or no empirical support.

I am not one of these critics. I think it’s entirely plausible that screens have a real negative effect on teenagers, but studies can’t find them because (a) the effects need not be dose-dependent (maybe two hours on your phone isn’t worse than one hour), and (b) given that everyone else is on social media, giving it up can have a negative effect. To expand on the second point, it’s possible that teenagers would be better off if social media didn’t exist, but, given that all of their peers have adopted it, there is a serious social cost to opting out.

If so, this means that the psychological research could all come up empty, but Haidt could still be right: correlational studies comparing light users to heavy users are comparing harmed kids to harmed kids, and deactivation experiments are uninformative—quitting doesn’t restore the pre-smartphone social world; it just turns the kid into a hermit.

For what it’s worth, there is evidence that college students themselves think of certain internet distractions this way—as collective action problems. A 2023 study of “collective traps” found that users would need to be paid to deactivate TikTok and Instagram if others kept using their accounts. No surprise—they use these platforms, so they must value them. The interesting finding is that they would pay to have them deactivated, so long as they were deactivated for everyone. This sort of finding grounds Haidt’s own recommendations in The Anxious Generation—the solution has to be laws, bans, and coordinated restrictions, not individual teenagers opting out.

I like this story—it’s the general presence of social media that harms adolescents, not the extent of individual use—but I have to be honest: there is some evidence that I’m wrong. Look at this graph (modified from one presented by Peter Gray in his forthcoming book).

The European kids got social media at the same time as the Americans, but their suicide rate doesn’t rise; it drops. Also, the rise in U.S. teens' suicide rate stops around 2017–2018, and rates have been falling since—first among girls, and in the past few years among boys. It's not like they got off the screens.

Haidt and Rausch have responses to the European data—they point out, reasonably enough, that comparing absolute suicide rates is too crude, and, more controversially, that other, more nuanced analyses favor their view. But I’ve never heard a convincing explanation (from anyone on either side of this issue) for why things started to improve about eight years ago.

I think the safest summary, then, is that, to put it mildly, one cannot confidently say that psychological evidence supports the claim that screens have messed up the kids.

---

But isn’t it obviously the screens? Here are some facts about smartphone use.

Many children and adolescents are never without these devices. They steal away time that could be spent with friends and family, in playgrounds, gyms, and in nature. Engagement with them is almost bizarrely passive—often just vacant staring. And work by Stanislas Dehaene and colleagues finds that long-term use restructures children’s brains, taking over a patch of the left ventral occipitotemporal cortex. Furthermore, it’s not just children who are affected; many adults (including me) are addicted to such devices, at the cost of lost physical activity and social engagement and, of course, sleep. When considering whether to ban these devices for children, do we really have to wait for a consensus from psychologists?

Well, maybe. All the facts in the above paragraph are true, but they aren’t about smartphones—they’re about books. (Thanks to Alison Gopnik for this rhetorical trick.) Books suck us into made-up worlds; they take us away from real-world social and physical interaction; the opportunity costs are massive. But books are good—when I listed the “decline of reading for pleasure” as a negative about kids today, I bet few of you balked. One summer, many years ago, one of my teenage sons became obsessed with Russian novels and spent many sunny weekend afternoons slumped on the sofa reading Turgenev. His brother devoted himself to Call of Duty, spending these hours shouting at fellow platoon members and perfecting his sniper skills. Can you guess which one I was mad at?

The point of this example isn’t that books are the same as first-person shooters. I personally think books are better. But if your big complaint about smartphones is the loss of socialization and time outside, why do books get a free pass?

Let’s do this again. Forget about video games and social media. Just focus on TikTok and YouTube videos today. How much time do you think teenagers spend just passively staring at them? The best estimates are about two and a half hours a day, a huge chunk of time that’s not spent socializing or goofing around—or studying, for that matter. Do you really need a meta-analysis to conclude that this is causing them great harm?

Again, the numbers are correct, but it’s the same sort of trick. It’s not “TikTok and YouTube videos today”. The data are from studies of teenagers' television watching in the 1990s.

---

I wonder if our feelings about kids today are influenced less by data about trends in anxiety and depression and more by our conception of what a good life is.

Some of these values are shared. I’d guess that for most parents of teenagers, books are better than video games; playing sports is better than watching television; and watching television is better than pornography and pot.

What’s interesting is how much we differ. For some adults, the shift to what Jean Twenge calls a “slow life strategy” of delayed dating, drinking, working, driving, and sex is mostly good news, a beneficial product of our prosperous and safe environments. It’s nice that teenagers still like to hang out with Mom and Dad! For others, it is a cause for concern. Why aren’t teenagers more independent, more adventurous, getting into more trouble? I’m old enough to remember when conservative commentators fretted that teenagers were having too much sex; now they worry that they’re not having enough.

I’ve met people who are sad about the decline of hunting and fishing; others find this a mark of progress. I’ve seen people online who are livid that so many adolescents identify as nonbinary; others are upset that so many people are upset. Some worry about teenage boys getting into crypto and sports betting; some find it kind of cool. I’ve written about my concerns about teenagers having AI companions instead of actual human friends; some techno-optimists don’t see the problem.

My friend Frank Keil has a terrific new book out (buy it here), in which, among many other things, he mourns the loss of tinkering. Children and teenagers used to take things apart and put them back together again—toasters, bicycles, and especially cars. Keil argues that the decline of tinkering has had profound consequences for children and adolescents’ understanding of the world. But reading his rich descriptions, you get the sense that he sees a childhood of curiosity and exploration—think of a 6-year-old taking apart an old alarm clock that would otherwise go in the trash or a group of teenage boys getting together to buy an old car and rebuild the engine—as having real intrinsic value.

I was more of a bookish kid, and this bias certainly influenced my own parenting—I have the nicest memories of taking my boys to the New Haven Barnes & Noble at midnight for the release of the newest Harry Potter book. And there is room here for the value of fun and shared experience. My teenage sons and I would sit together and watch every episode of Breaking Bad the minute it aired, and while I wouldn’t recommend watching television as an essential part of parenting, those were great times.

There are real clashes here. I just read Peter Gray’s forthcoming book, Restoring Childhood, and have sympathy for his appreciation of autonomy and freedom, and his anger at the academic stresses that we put children and teenagers through, of childhoods lost to homework and tutoring and test preparation. (Gray argues that these stressors are the primary cause of the uptick in suicide, depression, and anxiety, not the phones.) But a while ago, I read Amy Chua’s Battle Hymn of the Tiger Mother, which defends pretty much the opposite view. While I lean towards Gray, I do see the appeal of time spent mastering different abilities, and I appreciate Chua’s paean to discipline and ambition.

---

The children have obtained what their parents and grandparents longed for — greater freedom, greater material welfare, a juster society; but the old ills are forgotten, and the children face new problems, brought about by the very solutions of the old ones, and these, even if they can in turn be solved, generate new situations, and with them new requirements — and so on, for ever — and unpredictably.
—Isaiah Berlin

When I write my chapter, I’m going to end by making two points.

First, we make our best decisions with the right data, but the sort provided by psychologists and pollsters is often thin gruel. Of course, we need to continue paying attention to Gallup polls and articles in journals such as Developmental Psychology and the American Economic Review. But when thinking about the different worlds that children and adolescents could live in—and trying to think clearly about which world we want to provide for them—aggregate measures fall short. We should engage more with the detailed descriptions provided by anthropologists, sociologists, and journalists. And also with novels like David Copperfield and My Brilliant Friend, movies like Boyhood and Moonlight, and television shows like The Wire and Adolescence. These serve not as technically accurate descriptions, but as ways of sharpening our picture of childhood’s possibilities—what it looks like when it goes right, and what it looks like when it is wasted or warped.

Second, the claim here is that arguments about smartphones, bullying, and homework are arguments about values, often moral values. And so the debate over whether phones should be banned isn’t like deciding which drug best treats an infection; it’s more like a debate about the age of sexual consent, medically assisted suicide, or abortion.

Now, to say that these debates are infused with values doesn’t have to be a conversation-ender; it doesn’t mean “anything goes.” Value decisions are important, and one can make them well or make them poorly. We can work to avoid being swayed by factors that shouldn’t carry much weight, such as unreflective nostalgia. And we can continue to look skeptically and thoughtfully at psychological research on outcomes such as anxiety and depression; such research (assuming one can get causality right) plainly matters. If your moral view is indifferent to how many children and adolescents are committing suicide, well, you should go get yourself another moral view.

But in the end, these issues will not be resolved by data. And decisions have to be made—sometimes by the children and adolescents themselves, but often by parents and governments. Often, these will be coercive. And so, as Berlin points out in the quote above, we should expect such decisions to be unstable, constantly under revision, and never fully satisfactory.

---