The Social Media Moral Panic Is Like Every Other Moral Panic

the evidence isn't compelling and the ban proposals are poorly thought out, but it's totally different this time

Australia has become the first country to ban social media for minors under the age of 16. Norway is planning to follow suit by banning social media for under 15s and the United Kingdom’s Technology Secretary briefly considered a ban before quickly pivoting away from it.

What brought this all about? Well, South Australian Premier Peter Malinauskas’ wife read Jonathan Haidt’s psychology book The Anxious Generation, where Haidt argues that smartphones and social media are causing a rise in teen mental illness, and she told the Premier that he “better effing do something about this’”. Then Malinauskas’ announced a social media ban proposal that was taken up by NSW Premier Chris Minns and then by the Australian Prime Minister Anthony Albanese. Yes, that’s genuinely how this started. The funniest part is that we know this because Malinauskas publicly spoke about it as if he thought people would believe that legislating a ban because his wife read a book is sound policy-making.

The ban is simply the next step for a moral panic built on weak, cherry-picked evidence and media hysteria.

Cherry-Picked Evidence

Malinauskas opened his Social Media Summit, ostensibly an event to consult experts on the relative impact of social media on teens and different regulatory options, by declaring that “the results are in and the science is settled.” In reality, the evidence is weak and mixed at best.

Jonathan Haidt and Jean Twenge are the two big social psychologists ringing the alarm bells and arguing that social media is driving a teen mental health crisis (especially for teen girls). But they are decidedly in the minority. Most meta-analyses and systematic reviews on the relationship between social media use and teen well-being indicate that the effect is small, mixed or statistically insignificant. Many of the findings suggesting a negative impact indicate a weak effect size and those that do indicate a strong negative effect have not been replicated or have been criticised for being of poor quality.

From what I’ve read, these are the three major live questions in this literature:

1. Does the evidence shows causation between smartphone/social media use and mental illness in teens or mere correlation?

2. Does the evidence conclusively indicate a positive, negative or mixed effect?

3. Does the evidence demonstrate a weak or strong effect size?

Most researchers, who admittedly are not as famous and don’t have as big of a popular audience as Haidt, seem to answer “We don’t know yet, no, and if there is an effect then it’s weak”. That’s not to say that Haidt’s thesis must be wrong, just that Malinauskas is definitely wrong to claim that the science is settled.

To cherry pick one piece of writing myself, let’s look at Sara Abrahamsson's widely-lauded paper where she claims to demonstrate that “banning smartphones significantly decreases the health care take-up for psychological symptoms and diseases among girls”. The paper was received by some in the press to show that banning smartphones in schools is a great idea. But when you actually sit down and read the paper, you discover that the ‘smartphone ban’ in most cases assessed in the study amounted to requiring students to put their phone on silent. Hardly what most people would call a smartphone ban. Even then, the results don’t tell us very much. Girls had their grade point average go up by 0.08 standard deviations and the author found “no effect on the boys’ mental health, GPA, their average grades set by teachers, or on the probability of them attending an academic high school track”. A bigger effect was found on bullying incidents, where the author found that they “decline by 0.25–0.35 standard 19 deviation two to four years after a smartphone ban is implemented (p-values of 0.067 and 0.094)”. The most hyped-up effect found was on visits to medical practitioners for psychological issues:

“One year of exposure to a smartphone ban reduces the number of consultations by 0.98 visits (p-value 0.044), and three and four years of exposure lead to a reduction of 2–2.7 visits (p-value 0.011 and 0.008 respectively). […] Moreover, girls exposed full-time in middle school to the smartphone ban have 0.22 (p-value 0.076) fewer consultations for psychological symptoms and diseases at their GP.”

The most eye-catching way to put this is to say that specialist visits for middle-school girls went down by nearly 60% compared to the pre-treatment mean (3.4 visits) after 3+ years post-’ban’. But these p-values are underwhelming, and visits actually went up for boys! The results are not statistically insignificant but the effect sizes are small at best.

Maximised Harms, Minimised Benefits

What usually happens during a moral panic is that the pearl-clutchers exaggerate the real and potential harms of the vice that they’re up in arms about while minimising the real and potential benefits. That our best available evidence seems to indicate that, at most, social media and smartphone use among teens has a small negative impact on their mental health is not enough. One must operate under the assumption that smartphones and social media are killing children. If you don’t agree with smartphone and social media bans, you must support mass child suicide. And don’t you know, these things are responsible for everything under the sun: mental health problems, poor attention spans, bad grades, turning kids gay and trans, killing gay and trans kids, indoctrinating kids with left-wing propaganda, indoctrinating kids with right-wing propaganda, childhood obesity and so on.

The standard move is to claim that there are no benefits or to claim that the negatives are so bad that to even mention any benefits is morally wrong. Never mind that social media use might have a small positive effect on well-being, especially on feelings of social connectedness. Or that it might encourage creativity, improve writing skills or facilitate civic engagement. Maybe social media use predicts academic success and increases self-motivation for learning. All these things don’t exist, or the evidence for them is weak – totally unlike the evidence for the claim that smartphones and social media are destroying kids lives. This is absolutely different from every other moral panic. It wasn’t true for the internet in the 90s and 00s, or video games, or television, or comic books, or popular fiction books. This time it’s different. Apparently.

Half-Baked Solutions and Political Stunts

Australia’s social media ban was a carefully thought out piece of legislation. Not that it’s based on the best available evidence or that it’s the best solution to the problem. No, I mean that the Albanese government carefully thought through the political implications of the ban. It polls well and has gotten a lot of media attention, which serves as a useful distraction from the cost of living crisis and the Albanese government’s waning popularity. It also doesn’t come into effect until at least 12 months after receiving Royal Assent, so the government and the opposition have a year to figure out how to make it work or quietly shelve it.

The legislation is remarkably light on details. What counts as an ‘age-restricted social media platform’ is more or less up to the Communications Minister.¹ The government also skirts around the issue of how exactly the ban is going to be enforced, only stating that social media companies will need to take reasonable steps to ensure that people under 16 aren’t creating or using accounts.

Almost no consideration was given to the potential downsides, whether the ban violates the right to privacy or children’s rights to access information and express themselves. Australia’s Human Rights Commissioner and the National Children’s Commissioner note that:

“The Australian public was given just one day to make submissions to the parliamentary committee considering the proposed laws. The committee today held just three hours of public hearings and is required to report tomorrow. This is not meaningful consultation and does not allow for a thorough consideration of the views of experts, parents or children.”

Haidt himself is notably dismissive of the potential downsides, suggesting that even if he’s wrong banning smartphones and social media alike will just mean kids spending more time doing things like paying attention in school, playing outside and spending less time on those gosh darn screens.

I happen to believe that it is bad when governments implement poorly designed policies because a Premier’s wife read a book one time or because it’s politically advantageous to do so. Those who do believe that smartphones and social media are having a big negative impact on kids lives should be even more concerned. If this policy dies in 12 months or morphs into something disastrous or unpopular when the government comes up with an enforcement mechanism, then it will discourage political will for more sensible regulation in the future.

I’m Not Saying It Doesn’t Matter

I’m not a libertarian. I don’t think governments shouldn’t regulate private companies or services. Nor do I think that regulating or restricting access to things that have negative impacts is always a bad idea or that the treatment is always worse than the disease.

Likewise, the fact that something is a moral panic does not mean it isn’t an issue. I’ve noted that the evidence we currently have is weak and mixed, but that doesn’t mean it’ll remain like that. Maybe Haidt is right and his critics are missing something or imposing impossible standards on the studies he’s citing. I am not strictly opposed to governments restricting access to social media platforms or smartphones, nor do I think that we should have a totally laissez-faire approach to the internet. But what does bother me is politicians declaring that the science is settled when there is no expert consensus that social media and smartphones are driving a teen mental illness epidemic. It is worrying when politicians pass legislation that is simultaneously vague and heavy-handed as a stunt to distract the media and woo the public. And, ultimately, it is bad that the future of social media regulation may depend on the long-term success of this poorly thought out legislation.

1. The bill does make a laughably bad attempt to define ‘age-restricted social media platform’ that ends up including almost every online service, but it’s largely up to the minister’s discretion.