Social Media is a Major Cause of the Mental Illness Epidemic in Teen Girls. Here’s the Evidence. Journalists should stop saying that the evidence is just correlational by Jonathan Haidt
https://jonathanhaidt.substack.com/p/social-media-mental-illness-epidemic?utm_medium=email
A big story last week was the partial release of the CDC’s bi-annual Youth Risk Behavior Survey, which showed that most teen girls (57%) now say that they experience persistent sadness or hopelessness (up from 36% in 2011), and 30% of teen girls now say that they have seriously considered suicide (up from 19% in 2011). Boys are doing badly too, but their rates of depression and anxiety are not as high, and their increases since 2011 are smaller. As I showed in my Feb. 16 Substack post, the big surprise in the CDC data is that COVID didn’t have much effect on the overall trends, which just kept marching on as they have since around 2012. Teens were already socially distanced by 2019, which might explain why COVID restrictions added little to their rates of mental illness, on average. (Of course, many individuals suffered greatly).
Most of the news coverage last week noted that the trends pre-dated covid, and many of them mentioned social media as a potential cause. A few of them then did the standard thing that journalists have been doing for years, saying essentially “gosh, we just don’t know if it’s social media, because the evidence is all correlational and the correlations are really small.” For example, Derek Thompson, one of my favorite data-oriented journalists, wrote a widely read essay in The Atlantic on the multiplicity of possible causes. In a section titled Why is it so hard to prove that social media and smartphones are destroying teen mental health? he noted that “the academic literature on social media’s harms is complicated” and he then quoted one of the main academics studying the issue—Jeff Hancock, of Stanford University: “There’s been absolutely hundreds of [social-media and mental-health] studies, almost all showing pretty small effects.”
In this post, I will show that Thompson’s skepticism was justified in 2019 but is not justified in 2023. A lot of new work has been published since 2019, and there has been a recent and surprising convergence among the leading opponents in the debate (including Hancock and me). There is now a great deal of evidence that social media is a substantial cause, not just a tiny correlate, of depression and anxiety, and therefore of behaviors related to depression and anxiety, including self-harm and suicide.
First, I must offer two stage-setting comments:
- Social media is not the only cause; my larger story is about the rewiring of childhood that began in the 1990s and accelerated in the early 2010s.
I’m a social psychologist who is always wary of one-factor explanations for complex social phenomena. In The Coddling of the American Mind, Greg Lukianoff and I showed that there were six interwoven threads that produced the explosion of unwisdom that hit American universities in 2015, one of which was the rise of anxiety and depression in Gen Z (those born in and after 1996); a second was the vast overprotection of children that began in the 1990s.
In the book I’m now writing (Kids In Space) I show that these two threads are both essential for understanding why teen mental health collapsed in the 2010s. In brief, it’s the transition from a play-based childhood involving a lot of risky unsupervised play, which is essential for overcoming fear and fragility, to a phone-based childhood which blocks normal human development by taking time away from sleep, play, and in-person socializing, as well as causing addiction and drowning kids in social comparisons they can’t win. So this is not a one-factor story, and in future posts I’ll show my research on play. But today’s post is about what I believe to be the largest single factor and the only one that can explain why the epidemic started so suddenly, around 2012, in multiple countries.
- The empirical debate has focused on the size of the dose-response effect for individuals, yet much and perhaps most of the action is in the emergent network effects.
Once you appreciate the extent to which childhood has been transformed by smartphones and social media, you can see why it’s a mistake to focus so narrowly on individual-level effects. Nearly all of the research––the “hundreds of studies” that Hancock referred to––have treated social media as if it were like sugar consumption. The basic question has been: how sick do individuals get as a function of how much sugar they consume? What does the curve look like when you graph illness on the Y axis as a function of daily dosage on the X axis? This is a common and proper approach in medical research, where effects are primarily studied at the individual level and our objective is to know the size of the “dose-response relationship.” (Although even in medicine, there are important network effects.)
But social media is very different because it transforms social life for everyone, even for those who don’t use social media, whereas sugar consumption just harms the consumer. To see why this difference matters, imagine that in 2011, just before the epidemic began, a 12-year-old girl was given an iPhone 4 (the first with a front-facing camera) and began to spend 5 hours a day taking and editing selfies, posting them on Instagram (which had launched the year before), and scrolling through hundreds of posts from others. This was at a time when none of her friends in 7th grade had a smartphone or any social media accounts. Suppose that Instagram does cause anxiety disorders in a dose-response way, but the size of the correlation with anxiety is smaller than the correlation of social isolation with anxiety. The girl spending 5 hours a day on Instagram finds her mental health declining, but her friends’ mental health is unchanged. We find a clear dose-response effect. If she were to quit Instagram, would her mental health improve? Yes.
But now fast forward to 2015, when most girls are on Instagram and all teens are spending far less time with their friends in person (as I showed in my Feb 16 post). Most social activity is now asynchronous—channeled through posts, comments, and emojis on Instagram, Snapchat, and a few other platforms. Childhood has been rewired—it has become phone-based—and rates of anxiety and depression are soaring (as I showed in my Feb 8 post). Suppose that in 2015, a 12-year-old girl decided to quit all social media platforms. Would her mental health improve? Not necessarily.
If all of her friends continued to spend 5 hours a day on the various platforms then she’d find it difficult to stay in touch with them. She’d be out of the loop and socially isolated. If the isolation effect is larger than the dose-response effect, then her mental health might even get worse. When we look across thousands of girls, we might find no strong or clear correlation between time on social media and level of mental disorder. We might even find that the non-users are more depressed and anxious than the moderate users (which some studies do find, known as the Goldilocks effect).
What we see in this second case is that social media creates a cohort effect: something that happened to a whole cohort of young people, including those who don’t use social media. It also creates a trap—a collective action problem—for girls and for parents. Each girl might be worse off quitting Instagram even though all girls would be better off if everyone quit.
An implication of this analysis is that the correlations we are about to look at probably underestimate the true effect of social media as a cause of the teen mental illness epidemic. But OK, let’s take a look.
1. The State of the Art in 2019
In The Coddling of the American Mind, Greg Lukianoff and I tried to explain what happened to Gen Z. We focused on overprotection (“coddling”), but in our chapter on anxiety, we included six pages discussing the possible role of social media, drawing heavily on Jean Twenge’s work in her book iGen. The evidence back in 2017, when we were writing, was mixed, so we were appropriately careful, ending the section with this:
We don’t want to create a moral panic and frighten parents into banning all devices until their kids turn twenty-one. These are complicated issues, and much more research is needed.
Our book came out in September 2018. Four months later, two researchers at Oxford University—Amy Orben and Andrew Przybylski—published a study that was widely hailed as the most authoritative study on the matter. It was titled The association between adolescent well-being and digital technology use. The study used an advanced statistical technique called “Specification Curve Analysis” on three very large data sets in which teens in the US and UK reported their “digital media use” and answered questions related to mental health. Orben and Przybylski reported that the average regression coefficient (using screen time use to predict positive mental health) was negative but tiny, indicating a level of harmfulness so close to zero that it was roughly the same size as they found (in the same datasets) for the association of mental health with “eating potatoes” or “wearing eyeglasses.” The relationships were equivalent to correlation coefficients less than r = .05 (where r = 1.0 indicates a perfect correlation and r = 0 indicates absolutely no relationship). The authors concluded that “these effects are too small to warrant policy change.”
It is impossible to overstate the influence of Orben & Przybylski (2019) on journalists and researchers. The comparison to potatoes was vivid and memorable. Here’s one writeup of the study:
Whenever you hear a journalist or researcher say that social media has been found to have little or no relationship with mental illness, you’re likely to find a link to that study. When I first read the study, I began to have doubts myself. After all, it was the largest and most impressive study ever done on the question, and it was published by researchers who had been studying social media far longer than I had. Might Greg and I have gotten it wrong? Might we have been contributing to yet one more unjustified moral panic over technology?
2. The Social Media and Mental Health Collaborative Review Doc
Many other studies came out in 2019, yielding conclusions on both sides of the question. It was a confusing time. So I decided to compile in one Google doc all the relevant studies I could find. I invited Jean Twenge to join me on the project since she was far more knowledgeable about the various datasets. We posted the Google doc online in February 2019 and invited comments from critics and the broader research community. Each section ends with a request to tell us what we have missed. One of the first comments we got was that some researchers doubted that the mental illness epidemic was real. That led us to create a second Google doc titled: Adolescent mood disorders since 2010: A collaborative review. (I described it in my Feb. 8 Substack post.)
We immediately found that there was a simple and obvious structure for the social media literature review: nearly all of the published studies fell into one of three categories: correlational, longitudinal, or experimental. We therefore structured the document around the three questions addressed by studies of those types. You can see the three questions in the Table of Contents. I’ve reproduced the first part of it in Figure 1. Please check out the doc itself, and especially our list of “cautions and caveats.”
Comments are closed.
