What's Really Behind the Scientific Debate On Screens & Kids' Mental Health
It's another culture war: Science vs. The Movement
An acerbic scientific debate has been reignited this past month with the publication of The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness by
. The funny thing about this debate, though, is that it doesn’t really exist - at least not in the way that you’ll read about on X or Substack. There, you’ll read passionate and outraged defenses of the book and its conclusions. You’ll read about the the critics who don’t believe screens are destroying our kids’ mental health.The truth is, those critics don’t actually exist. Instead, the vast, vast majority of scientists actually agree with the broad conclusions of the book. Yes, screens can contribute to problems with mental health. Yes, we should spend less time on screens. Duh.
If I’m right, then why the increasingly shouty defenses of the book by the author and his allies? Why does it feel like a fight between believers and doubters, or like a particularly nasty partisan political debate?
Science isn’t supposed to be a movement.
The answer, I believe, lies in my choice of the word allies. As I wrote in the NYT review of the book, while it cites research and leverages data, it is also fundamentally crusading. It’s the rallying cry of a Mission with a capital M, which is to “free the anxious generation.” There are billboards, posters, and even art installations across the country proclaiming exactly that. It’s a movement for digital absolutism. It’s this that scientists rankle at because science isn’t supposed to be a movement.
And for good reason. To see why, we can draw on Haidt’s own metaphor for moral decision making: the elephant and the rider. Emotions, intuitions, and subconscious motives - all the feels - are the elephant, whereas our conscious, rational decision making is the rider. And because of the overpowering strength of the elephant, it’s too easy for the rider to let the elephant lead, adjusting to the urgency of its goals, all the while convincing themselves that it is they who are in the drivers seat, setting the direction.
What critics of the book are saying without saying is that the core idea of the book is an elephant, dressed up like a rider.
Science is one of humanity’s greatest tools for balancing the scales in this eternal struggle. We apply its rules and norms to give more strength to the rider. And science has been wildly successful in this endeavor.
What critics of the book are saying without saying is that the core idea of the book is an elephant, dressed up like a rider.
That core idea can be presented as a simple equation:
smartphones + overprotective parenting = mental illness in children.
Scientists don’t disagree that these factors are important. They disagree with the oversimplification of the algorithm. Yes, it might feel right, because both digital experiences and parenting certainly influence mental health. But it completely lacks context and consideration of individual differences - all of which are absolutely necessary to understand the complexities of mental health. Otherwise, everyone who experiences a trauma, or who spends too much time on screens, or has neglectful or abusive parents would be mentally ill.
And it’s no coincidence that it’s social psychologists like Haidt, not clinical or child psychologists, who are raising this alarm. Do clinical psychologists or child healthcare professionals just care less? Of course not. But they think about mental health in a specific way. It’s called the biopsychosocial model:
This model states in no uncertain terms - it’s complicated. Social and environmental experiences interact with people’s individual biological and psychological differences, and mental health emerges at the intersection of this stew. And the differences among people in both risk and resilience are as important as the commonalities.
But this isn’t how The Movement thinks of digital media and mental health. In graph after graph (this can be found on Haidt’s Substack and in the book), it looks like a clearly rising tide:
This is what rankles researchers. While it’s important to track these trends, a thousand trend lines don’t get at biopsychosocial determinants of mental health. To boot, interpretation of these data are sometimes frustratingly off. For example, the graph on the left is based on a survey which asked undergrads to self-report on whether they have a mental illness or are receiving treatment. It’s not as the graph is labeled “Percent of undergrads WITH a mental illness.” That suggests an actual diagnosis. Self-report of mental illness is highly vulnerable to error, especially at a time when kids are self-diagnosing on TikTok. What might be rising in this particular graph is the tendency to self-identify with a mental health problem. We can’t know.
Now, the graph on the right is more compelling because the outcome of interest is documented emergency room visits. The clinical scientist, however, will still ask a million questions: What else other than smartphone adoption was happening during that period before rates rose? What was going on in the family context? What’s going on with the 99.5% of kids who are also increasingly exposed to social media and smartphones but are not self-harming (i.e., 589/10,000 is a little over 0.05%).
These differences in norms are at the heart of the culture war driving this debate - the culture of science versus the culture of a movement. They can be summarized like this:
Science states caveats while movements state ‘where there’s smoke there’s fire’. In science, if your findings are incomplete, flawed, or inconclusive in any way, you go with the more conservative interpretation1. In contrast, movements aren’t interested in being subtle or putting caveats around conclusions. That doesn’t serve the goal of changing hearts and minds. If they see smoke, they have to assume there’s a fire and activate people to do something about it.
Science avoids causal language while movements embrace the starkest language possible. If this book were a scientific study, the subtitle would reflect what meta-analyses always find: Small to moderate associations between social media use and mental health and patterns of equivocal results across studies. Not sexy. But to the scientific eye, it kind of is because it translates into: “Whoa! social media are part of the problem for some people. Let’s suggest caution and learn more, stat!” For a movement, however, strong statements are the only ones that matter because they raise the alarm. A movement has exactly the kind of subtitle we see in this book: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness. This is the kind of powerful elephant that can move the movement forward.
Science responds to all critiques while movements respond to the ones that make their point. In science, when you publish a study or apply for a grant, you get pummeled with critiques from peer reviewers. Your job, if you want to get published or funded, is to thoughtfully and accurately respond to all the critiques. Not just some of them, and not just the weakest ones. All of them. But when a movement responds to a debate, typically online and with the standards of science not in effect, you can pick and choose how and to what you respond. Something that would be stopped dead in the peer review system is a powerful bulldozer in the public square.
Indeed, the best science shows that it’s not that simple at all - and believe me they’ve looked. She, as a scientist, is pointing out that his statements don’t match the evidence. So, what is Haidt arguing against?
For example, in an interview with The New Yorker, Haidt claims that Dr. Candice Odgers, in her review of the book in the journal Nature, “…said I had no evidence.” He then goes on to argue why she’s wrong about that. But that’s not what she wrote. Here’s the section: “Moreover, findings from the Adolescent Brain Cognitive Development study, the largest long-term study of adolescent brain development in the United States, has found no evidence of drastic changes associated with digital-technology use7.” She’s accurately stating that there is no existing evidence that digital tech use “rewires” youths’ brains in ways that cause mental illness. Indeed, the best science shows that it’s not that simple at all - and believe me they’ve looked. She, as a scientist, is pointing out that his statements don’t match the evidence. So, what is Haidt arguing against?
It might be hard to understand why scientists care. Don’t we all want less digital in our lives? Who cares if while fighting for change, we overstate things a bit or get incendiary in our language. That’s what it takes to grab the headlines. The outrage machine gets eyeballs and drives action.
Maybe so. But it doesn’t help with what comes next. What comes next after you ban the phones? How do we contend with the fact that parents are perpetually phubbing their own kids? - which by the way is another viable, but ignored, explanation for why kid’s mental health has worsened after smartphone adoption increased. Will kids end up any better off at cultivating healthy digital lives than adults are today? And what about focusing resources on those kids who are most at risk for harm, rather than asking that all kids make digital chastity pledges?
As noted in a recent interview about the book, one researcher noted, “Rather than asking, is it a net negative or positive, which is an absurd discussion,” he said, “it’d be much nicer if we could ask: What are the impacts? To who? And which thing does it, and how can you change it?”
I think it’s clear which culture I belong to in this culture war. But I’m also all in for a movement to fight for less digital disruption and more meaningful IRL in all of our lives. Here’s the thing, though. I also really care about where the movement goes next. I want to invite more people in, not just the ‘true believers’, to have conversations about solutions and strategies. I want doubt and questions to be allowed, rather than defended against and systematically shot down.
What comes next? Maybe we should figure that out together.
That’s because statistics, which we use to test our predictions, don’t actually prove anything. Instead, they show us probabilities - how likely we are to be wrong, how confident we can be that there are enough people in our study to detect a real effect, and the degree to which we can predict the outcome given what we know. And when it comes to the social sciences with their messy subject matter - humans! - full of error and individual differences, these probabilities are never that satisfying. There are always people who don’t conform to the trendline. There is always error in our measurement.