"The Anxious Generation" and the false comfort of easy diagnoses

Content warning here: this post talks a lot about mental health and has a reference to self-harm. If you are struggling with this, please know you aren’t alone. If you’re thinking about suicide, are worried about a friend or loved one, or would like emotional support, the 988 Lifeline network is available 24/7 across the United States.

The latest issue of Casey Newton’s vital Platformer newsletter covered the debate over Jonathan Haidt's "The Anxious Generation" and is worth your time. Haidt’s book, which documents the genuine statistical rise in mental health problems among kids and teens, argues that smartphones and social platforms bear a considerable share of the blame. The work and ensuing media tour on morning shows and Podcast Alley has reignited a polarizing debate over mobile phone technology in the lives of our kids.

We’ve been here before. Haidt has been banging this drum in the public discourse for a few years along with Jean Twenge, who has been making similar arguments in her own public scholarship. There’s an eye-popping data point underlying Haidt’s thesis: a prodigious spike in mental health issues among teens that coincides with the mass adoption of smartphones and the launch of Instagram in the early 2010s. From there he makes the argument by tying together unconnected research that correlation is causation in this case, that the evidence is overwhelming that smart devices and constant social connection are stalling the development of kids. His offered solutions are straightforward: ban smartphones in schools, age-gate social media until kids are 16 years old, and have kids spend more time outside.

The main criticism I keep hearing of Haidt's work tends to follow two related lines of thought.

First, he relies heavily on the volume of loosely (at best) connected scholarship to draw causal lines without the benefit of a study that finds causality itself. There’s a reason they warn us in intro to psychology classes not to fall for the idea that correlation equals causation (the classic example I heard as a student was that “popsicles cause drowning deaths” because we can sync up charts showing a spike in pool drowning deaths and consumption of popsicles …. spoiler alert: it’s because both tend to happen with greater frequency in the summertime).

Haidt does acknowledge here that he’s doing the correlation-equals-causation thing but argues that the sheer volume of studies decreases the likelihood of false positives in linking smartphone access to mental health struggles. And look, that is a thing in social science, which rarely offers us a single causal solution. Sometimes the volume of evidence, carefully linked together, can argue for causation without having definitive proof. But those last six words are crucial: the evidence might point that way, but that’s a far cry from definitive causality. Haidt’s thesis has a type of stridence that ignores the reality of other potential factors in the crisis.

The second main criticism tends to be that Haidt’s work is the latest in the long history of classic moral panics around new forms of communication technology, a replication of a historical phenomenon we see time and again as new ways of communicating emerge.

One example: A pamphlet written by Johannes Trithemius in 1492 ("de laude scriptorium" or "In Praise Of Scribes") that argued that the Gutenberg Bibles were inferior to hand-copied scriptures. His particular argument was that monks hand-copying Bibles required contemplation of the scriptures, which made it an act of thinking reverence compared to the automated stamping of a printing press. Notice the casual inference made here, that the robotic automation of a production process somehow is dehumanizing and subtly evil by removing the human mind from the act of transcription. Clay Shirky hilariously points out in Here Comes Everybody that Trithemius had to use a printing press to get his message out because hand-copied versions of his pamphlet couldn’t compete with the speed with which Gutenberg Bibles spread all over Europe.

The Abbot's position would have been mere reactionary cant ("We must preserve the old order at any cost") but for one detail. If, in the year 1492, you'd written a treatise you wanted widely disseminated, what would you do? You'd have it printed, of course, which was exactly what the Abbot did. De Laude Scriptorum was not itself copied by scribes; it was set in movable type, in order to get a lot of copies out cheaply and quickly­ something for which scribes were utterly inadequate. The con­tent of the Abbot's book praised the scribes, while its printed form damned them; the medium undermined the message.

I look at Trithemius’ argument as a correlation-equals-causation case. He wrote his pamphlet in 1492, about 50 years after Gutenberg unleashed his printing press on Europe. This was 25 years before Martin Luther kicked off the Reformation, but the winds of change were already blowing. Bibles in the hands of common folks became a thing after centuries when owning a Bible was a luxury for elites due to the high cost of hand-copying. You can draw a direct line between access to information and the Reformation; when people can read for themselves, they question things. Trithemius’ world, where the Catholic church was in control by controlling access to information, was falling apart. And he did what moral panic purveyors due: the problem is this new technology that lets people connect with new ideas and consider the world for themselves.

A more contemporary example is the ad campaign in the 1990s from conservative groups, fronted by celebrities such as Steve Allen, that argued television is corrupting our kids (leading them down a moral sewer, no less!). A snapshot of one ad that ran in newspapers big and small during the '90s.

Wild stuff. It's the printie version of your crazy family member that forwards all-caps chain letters. But it’s important to understand how seductive these arguments are because they invite so little thought on our part as we look for solutions. The ‘80s and ‘90s were a time of intense social change, and it was easy to look around and seek easy causality. This provides an opening for charlatans to exploit fear by making evidence-light arguments about why. In this case, notice the blanket labeling of television, which flattens all differences between types or genres of programming, and goes straight to demonizing the whole medium.

That TV example is how Haidt's book reminds me of the assertions made in Robert Putnam's famous book Bowling Alone in the 1990s. Putnam wrote about the decline of civic engagement (joinership) in American life, and he said television was the primary driver of this trend. It was an easy thesis in some ways. For example, all you have to do is imagine millions of Americans sitting transfixed every night in their living rooms and understand that this means they aren’t out connecting in bowling leagues, PTA meetings, and other civic happenings. Cause meets effect with efficiency.

But a closer look at Putnam’s book causes his arguments to fall apart and look a lot like the arguments against TV and printing presses. He made the same diagnostic mistakes others made in past moral panics by treating new communication technology as a monolith and being uncurious about the complexities of a particular medium. That is, sure, some TV might be bad for us. That's a programming problem and thus a human decision-making issue, not a technology problem. All of these moral panics end up being reductive to the technology, though. The printing press is making us dumb, radio is rotting our brains, television is turning our kids into miscreants, smartphones are addicting our children, and social platforms are making our kids anxious.

Technology arguments are seductive because they offer easy solutions. Ban them, or limit access. You can just cut a whole product category out and be done with it, not having to deal at all with nuance, and left feeling good that you solved the problem. But if it’s a content problem, we’re talking about human decision-making and thus talking about regulation or more complex solutions that aren’t as easy to implement, and that’s a tougher sell because it involves a lot of thinking, hard work, assessment, and iteration. More importantly, the content problem is less in our own hands and is a matter of choices by humans we can’t control. More broadly, I’d posit we’re uncomfortable asking the obvious question about human choices: what if the problem is us? Not them, the evil content makers, or them the other parents who are doing it wrong, but rather us as parents or leaders tasked with rearing kids?

Later scholarship kicked the tires on Putnam’s work and bore out what it looks like to examine something like television with an eye on complexity, including work in the '90s and '00s contextualizing the question of TV and civic life. An example is this piece from Dhavan Shah that examined social capital (the lack of social connections and trust Putnam was concerned about) and types of TV use. Some types of content, it turns out, were pro-social! Watching social dramas such as the then-popular “ER” show, for example, correlated with more trust and participation. People can learn from TV about different experiences and that helps them be better co-workers or neighbors.

The point here is that the freakout over mobile phones and social platforms looks an awful lot to my eyes like the cycle of demonization of new communication technologies. It casually draws causal lines based on recency bias by connecting short-term social trends to a new technology. It finds two new(ish) trends (in this case spikes in mental health issues and the adoption of smartphones among teens) and links them together for an easy cause/effect relationship, and it gives us something to blame other than ourselves and our antisocial uses of those tools.

I worry about this not only because of the reductiveness of the claims but also because of the easy ways shortcut diagnoses and solutions take root and become coda due to repeated exposure. Putnam was wrong about a lot of things, but his ideas still have resonance nonetheless. He still sells a lot of books. I was at a lecture this week that had the speaker quoting Bowling Alone uncritically. The thing about easy diagnoses and simple solutions is they are easy to accept in the public mind, particularly when you're afraid and unsure of how to solve a complex problem.

And that's the thing about Haidt’s argument that is so problematic. He offers easy solutions that lack any sort of empirical support, all to fix a diagnosis that itself lacks support. We can make a lot of bad policy that way, and in implementing poor solutions we end up not fixing the thing we care about all while having the veneer of having done something about it. Evidence-free policy fixes are mostly for show and create a false sense of comfort, a lazy and ineffective way to solve problems.

One thing I see a lot in social platform research is that the context is just as complex as the TV picture Putnam missed. We think about use and effects from our own experience and ignore the parts that are empowering.

For example, the scholarship on LGBTQ teens. These are people who have a difficult time finding support in their own homes and communities at times. I’m mindful of the data about LGBTQ kids showing they are 40% more likely to attempt suicide than their non-LGBTQ peers. This has long been considered not something inherent to those kids but rather a sign of how they are treated in society and by the people close to them. It is a social problem, and one benefit of the internet for this group is they’ve been able to forge online connections that help them weather the difficulties they face as teens. One person’s tool of disconnection is another person’s lifeline, and I am moved by the stories my LGBTQ friends consistently tell me about the community and support they have found over decades on social platforms such as Tumblr and Instagram. Even if something is going on with teens and social platforms to support Haidt’s thesis, LGBTQ kids are a significant counterpoint that tells us it’s not a technology problem and it’s way more complex than a device issue. Haidt argues that the tool is disconnecting, but these kids are telling us there is bona fide community for them in online spaces.

More broadly, some problems come with default assuming online social connection is harmful. We know that it's not. Just like the content problem on TV, it's not the technology it's what humans do or don't do with it. Social platforms can be learning and empathy engines for teens, and social connections can serve good purposes. For example, when I was in grad school we did some team research on a local disabled living community and found that a significant portion of the residents saw the internet as a lifeline to social connection because they had mobility issues that made social connection in person difficult. The context of an individual’s situation matters so much when looking at the effects of online connections.

From a sample size of one, I can say this has been my own online experience. The internet has helped me grow as a person. I’ve become connected to folks who aren’t like me and have different lived experiences. This has 100% made me a better person. But there are downsides. In the social network age, I’ve had to think a lot about network construction and what different platforms are good for or not. In the past couple of years, I’ve made a real effort in particular to use things like Facebook, Instagram, LinkedIn, or microblog sites like X or Bluesky in distinct ways that let me think of both message and audience more specifically, so that the content I’m both sharing and receiving in those spaces matches the particular needs I bring to those platforms. When done well, this can work. It’s not the platforms; it’s about the use case and our approach to the technology. But this requires thought and planning.

If I had more grounding in media psychology, one thing I'd like to know is whether the mental health struggles Haidt documents are one of the side effects that come from learning more on the internet in general. That is, is this just one of the things societies deal with when we have more access to information and learning? In this case, I wonder about the negative mental health effects that come with learning about the wider world and a broader array of experiences, and then realizing the people you know in the everyday are harming folks you are connecting with online. It’s the problem The Good Place tries to get at; sometimes knowing more helps you do better, but it opens up new frontiers of moral struggle that are difficult to square with simple answers. The things a lot of these kids are learning about online include better access to LGBTQ life experiences, racial justice problems, and climate change. Problems that have deep roots in the failures of adult leadership just as they’re coming of age with the energy to solve problems that have heavy impacts on their generation, but also with the realization of the social dynamics that limit their ability to make change. I am given in my internet culture classes to quote Ecclesiastes 1:18, which says, "for in much wisdom is much grief." The struggle with knowing things is part of life but this isn’t a call for heads in the sand.

The solution here, then, can’t simply be “know less.” We wouldn’t make that argument about books or the news (even as Trithemius made this very argument about Gutenberg Bibles in 500+ years ago!), so why smartphones? Maybe the mental health crisis, in other words, is a realization we know more and can’t always fix it. That isn’t a downside of the internet or technology; it’s a call to work together to fix problems in more productive ways. It's easier for adults who either don't care or feel completely unempowered to solve problems bigger than individual choice to demonize phones and tell kids to stay off TikTok, because it feels like doing something. But this is a collective empowerment and collective action problem, not something that can be solved by telling kids no TikTok until 16 (as if they'd listen anyhow).

The solution I'd offer as a scholar is much more difficult to implement than what Haidt offers up but is rooted more in what we can learn and know. We need to mentally distinguish between pro-social uses of phones and social platforms from ones that harm us. We have to look at types of content and think about what we are making and distributing. Then if we want to talk about regulation or education efforts, we can craft specific solutions that target specific harms. But it has to be a society-wide project that doesn’t sluff off responsibility onto schools or social platforms.

There are all kinds of regulatory levers we can pull that don't involve banning devices or platforms. We can do more with public pressure around platform decision-making about content. We can do a lot more with education as technology continues to work its way into the digital lives of kids and teens. Why are we talking about banning phones while simultaneously letting Meta off the hook, for example, about algorithmic feed construction on Instagram? You cannot keep kids off these platforms. Bans on devices or unenforceable platform access denials are a fig leaf.

So anyhow, my point is this picture is way more complex than the one Haidt paints, and easy diagnoses/solutions are both seductive and ineffective. If we take seriously the spikes in mental health problems, let's enlarge it past media use and not fall for the thin and simplistic explanation. I do think Haidt is raising important questions about mental health in kids, and I think that’s an issue most everyone cares about to some extent. There certainly are aspects of digital life that require more thought. But simple-sounding solutions here are just a recipe for not addressing the problem, all while offering us the false comfort of doing something. It would, to my eyes, be more harmful to be baited into thinking a simple solution has solved the problem because it tempts us to not regard the mental health problems in front of us as something we all have to grapple with.

Jeremy Littau is an associate professor of journalism and communication at Lehigh University. Find him on Bluesky, which is now open to the public.