Why Would Social Media Increase Schizophrenia?

There has been a 67% increase in youth schizophrenia since 2010.

Look to the bottom of this graph from Jonathan Haidt’s The Anxious Generation. The line for schizophrenia on the graph looks flat, but the percentage increase is large.

Look at the “X% increase since 2010” indications, rather than population-level prevalence.

How come there’s been a 67% increase in schizophrenia since 2010?

Many people point to social media and smartphones as explanations for the increased prevalence of mental illness among young people.

But why would social media lead to an increase in schizophrenia?

And if social media isn’t what led to the increase in schizophrenia, then could whatever else led to the increase in schizophrenia also be influencing the increase in other forms of mental illness?

On Twitter, Matt Grossman noted he thinks the increase in mental illness among young people likely has a multi-causal explanation. I think many people hold that view, myself included. The contention comes in when trying to figure out how big or small of a role different factors (i.e., social media vs. other factors) play.

In response, Matt Blackwell observed that since it’s not apparent that social media would lead to an increase in schizophrenia, perhaps the increases in schizophrenia could be used as a negative control.

In response to both Matts, Zvi Mowshowitz elaborated:

The 67% increase in schizophrenia seems like it shouldn’t obviously be smartphone related, and could be a proxy for measurement adjustments, as could the 57% for bipolar. Then you need to explain how all this isn’t causing much ADHD, which is only up 72%, but seems like something phones would make much worse and also something with a big diagnosis and stigma shift.

As Zvi notes, we could treat the increases in schizophrenia and bipolar prevalence as proxies for the increases in measurement. If we do that, roughly 50-60% of the observed increase in mental illness could be accounted for by increases in measurement.

But it might even account for more than 50-60%. As Matt Grossman pointed out, the factors increasing observed mental illness among youth probably build on themselves. Based on that potential feedback loop mechanism, the portion of the increase in psychopathology accounted for by decreases in stigma could be even higher.

Does a decreased-stigma→ increased-reporting feedback loop explain the increased prevalence of psychopathology?

Based on the above, the alternative explanation for the increases in the prevalence of measured psychopathology among youth would be:

1. Decrease in stigma leads to an increase in reporting¹

2. Increases in reporting lead to a further decrease in stigma

3. Repeat steps 1 and 2 over and over

This alternative explanation seems compelling to me.

Has Haidt addressed this?

Jonathan Haidt has engaged with a part of this argument but not the full mechanism². Along with collaborators Jean Twenge and Zack Rausch, he maintains an extremely useful Google Doc where they compile evidence about the increased prevalence of mood disorders in adolescents since 2010. In the “Cautions and Caveats” section of this document, they acknowledge that Gen Zs are “probably more willing than previous generations to admit to symptoms and seek out diagnoses” and that “this may contribute to rising trends in some datasets, particularly for depression and anxiety, without necessarily indicating a real change in underlying rates.”

However, Haidt and his colleagues do not engage with the feedback loop argument there.

What role does social media play in all this?

Social media would facilitate the mechanisms at play in this alternative explanation as well. It’s just not as nefarious of a story.

In this story, social media just increases the extent to which information is shared and helps spread the impact of the reduction in stigma. Then, more people (especially younger, more open people) share when they experience symptoms of mental illness. And that, in turn, helps further reduce stigma, and the cycle continues.

Could this mechanism explain half of the reported increase in mental illness among young people? Or maybe even more, to the extent that the feedback loop is at play? Or is there something else I’m missing that makes this implausible? I’m keen to hear different perspectives on this.

Also, soon after I wrote this, the New York Times released a piece about a related hypothesis: the prevalence inflation hypothesis. I’ve written more about that hypothesis in relation this hypothesis here.

1. By reporting, I mean increases in individuals self-identifying or being diagnosed as having a specific psychological disorder. Some subset of people who self-identify and/or receive a diagnosis are also open about this with the people in their lives, which leads to the subsequent reduction in stigma.

2. As far as I know. If you know of somewhere where Haidt or his co-authors have addressed this, please point me to it!