Social media and teen suicide: If you’re going to discuss the “other side,” can you at least get the facts right?
A New Yorker article beautifully tells the stories of grieving families, but misrepresents the research on mental health
Is social media responsible for the increase in teen suicide and depression?
That’s the question tackled by an article in The New Yorker this week. Most of the article tells the horribly tragic stories of teens who have taken their own lives and the devastated parents they left behind. The article handles these stories with heart-wrenching detail and compassion, as well as demonstrating the role that social media may have played in the teens’ deaths.
Unfortunately, the last part of the article felt the need, presumably for “balance,” to cover the “other side” of the issue. The “both sides” model is a by-product of covering politics, where journalistic rules require that both sides should have a say. However, it doesn’t work very well for questions with empirical evidence.
What frustrates me even more is that coverage of this so-called “other side” routinely gets the facts wrong – and not just a little wrong, but completely, utterly wrong, often on important points. Here are some examples from the New Yorker article, to set the record straight:
1. The article states, “Nonetheless, research has failed to demonstrate any definite causal link between rising social-media use and rising depression and suicide.” The use of the verb “failed” here strongly implies that several studies have tried to demonstrate a causal link, but have not found it – they have “failed to demonstrate” it.
That is not true, for a very simple reason: It is not possible to determine a definite causal link between rising social media use and rising depression and suicide because there is no way to do a random-assignment experiment with over-time trends. To do so, you would have to randomly assign people to be born at different times and then measure rates of depression and suicide. It’s impossible. The studies haven’t “failed” – they can’t be conducted.
The closest anyone has come is what is sometimes called a natural experiment, where people are exposed to technologies at different times and you measure the outcomes. Those studies support the link: The introduction of Facebook on college campuses coincided with more depression, and the deployment of high-speed internet access across Spain led to more mental illness hospitalizations among girls.
2. “Working with Amy Orben, a Cambridge psychologist, [Andrew] Przybylski has also noted that lower life satisfaction correlates slightly more strongly with wearing glasses than with digital-technology use.”
This refers to the UK Millennium Cohort Study of teens. It’s easy enough to find the correlation between wearing glasses and low life satisfaction in this dataset. Here it is for girls:
r = .05
Since the statement refers to “digital technology use,” let’s look at the average of the correlation between low life satisfaction and internet, gaming, and social media use for girls:
r = .20
Thus, the statement that “lower life satisfaction correlates slightly more strongly with wearing glasses than with digital-technology use” is demonstrably false. In fact, the correlation with digital technology use is four times larger than the correlation with wearing glasses. This isn’t new – these numbers are from Table 3 in this paper published 4 years ago. And when you zoom in on social media for girls, the number is even higher, r = .24. The correlation is similar for social media use and symptoms of depression. This translates to a tripling of the risk of depression from no use to heavy use of social media among girls in this dataset, whether you use fancy statistics or simpler ones.
Not only that, but the correlation between low life satisfaction and digital media use among girls is larger than the correlation between low life satisfaction and obesity (r = .09), marijuana use (.18), hard drug use (.06), and exercise (.11). If we’re going to write off social media, we might as well write off a lot of other health and lifestyle factors related to teens’ well-being.
3. “In North America, rates of depression and anxiety in young people have been rising for at least eighty years. ‘Why weren’t people in the 1980s or ’90s asking why adolescent depression was at an all-time high?’ the Johns Hopkins psychologist Dylan Selterman has written.”
The first sentence implies that rates of depression and anxiety have risen steadily over the last 80 years, but they have not. Yes, depression and anxiety rose among young people from the 1950s to the early 1990s. But teen depression was unchanged or down from the early 1990s to about 2011, before it shot upward (see Figure 1).
Figure 1: Depression among U.S. teens (8th, 10th, and 12th graders), 1991-2022. Source: Monitoring the Future surveys of 8th, 10th, and 12th graders
Suicide rates show an even more pronounced pattern, decreasing from the early 1990s to the late 2000s before rising again.
And this: ‘Why weren’t people in the 1980s or ’90s asking why adolescent depression was at an all-time high?’ is absurd: People were asking that question in those decades. Here are two examples from high-profile journals documenting the rise in teen and young adult depression. Book chapters were written on the subject. The rise in teen suicides during the period was extensively covered in newspapers. The book Prozac Nation, a memoir of teen depression, was a bestseller in 1994.
When I saw that quote I made an educated guess that Dylan Selterman had never done research on trends in mental health. A look at his online CV shows that is indeed the case. So why quote him? To offer an “other side” perspective that is laughably wrong?
4. The article quotes Laurence Steinberg, an expert on the psychology of adolescence, as saying, “Given the widespread eagerness to condemn social media, it’s important to remember that it may benefit more adolescents than it hurts. . . . If other factors that have contributed to the rise in adolescent depression are being overlooked in the rush to point the finger at Facebook, we may be contributing to the very problem we hope to solve.”
If social media is benefiting more adolescents than it hurts, why has depression gone up? Why are heavy users of social media twice as likely to be depressed as light users? Where are the adolescents who have benefited so much?
And what are the “other factors that have contributed to the rise in adolescent depression”? He doesn’t say. He previously pointed to increases in academic pressure and homework time, but academic pressure and homework time haven't increased among U.S. teens.
5. “Orben … is at pains to emphasize that tech is not the only thing that has happened in the past fifteen years. ‘The world is on fire,’ she said, adding that politicians have focused on social media because it makes for a simple, popular target. ‘It is a lot easier to blame companies than to blame very complex phenomena.’”
What complex phenomena? That’s not clear. And is the world really more on fire now than it was during the Great Recession or during the Iraq and Afghanistan wars, when teen depression was lower than it is now? Maybe she’s referring to climate change, but that doesn’t fit the data either (see next point).
6. To be fair, the article does mention a few other possible causes elsewhere. Some of these might certainly be true for individual cases, but they can’t explain the rise in teen depression. For example:
• Climate change. If climate change was the cause, why did teen loneliness also increase? How exactly would worries about climate change cause loneliness? Why wouldn’t teen depression start rising in the 1990s when concern about the environment peaked among teens? Why wouldn’t the response to a crisis be action, not depression, the way it’s been for other crises young people have faced?
• Intolerance for their gender, race, or sexuality. These factors have improved in the past few decades. If these are the cause, why would the increase in teen depression begin when the country re-elected our first Black president (2012) and continue after same-sex marriage was legalized (2015)?
• Bleak financial prospects and diminished social mobility. Inflation-adjusted median incomes among young adults are at all-time highs, and, according to the St. Louis Fed, Millennials and older Gen Z’ers have earned 25% more wealth than Boomers or Gen X’ers at the same age. So social mobility has increased, not decreased. The economy improved steadily 2011-2019 while teen depression doubled. Even if this is about perception instead of reality, if financial causes were paramount, why would the largest increases in depression and self-harm appear among 10- to 14-year-old girls?
• Acute loneliness, even among those who appeared not to be lonely. Exactly: They appeared not to be lonely because they had 500 followers on Instagram and no one to actually talk to.
• A sense of impotence and purposelessness. That’s how many people feel after spending too much time on social media. Ditto for the quote from British political theorist David Runciman: “They feel powerless about climate change, war, misery,” he said. “That is a toxic combination: permissionless access to information, and relative powerlessness over the topics to which that information pertains.” “Permissionless access to information” is a pretty succinct definition of social media. This is a mechanism for social media leading to depression, not an alternative explanation.
7. “Gina Neff, the Cambridge technology researcher, grew up in the hill country of eastern Kentucky. ‘Kids who are gay in Appalachia find the Internet and it is a lifesaver,’ she said.”
If it’s a lifesaver, why have suicidal thoughts among LGBT young adults increased, not decreased, as social media became more popular and acceptance of LGBT identities grew? Why would more LGBTQ+ young adults than straight say that social media has a negative impact on their mental health? And why are gender diverse heavy social media users more likely to be depressed than light users?
Here’s my request of journalists: If you’re going to quote the academics who argue that social media has nothing to do with the rise in teen depression and suicide, can you at least check the facts?
And why can’t any of these critics provide a plausible answer to this question: If it’s not smartphones and social media, what is it?
It’s been seven years since I first publicly argued that smartphones and social media may have caused the increase in teen depression, and that question has yet to be answered with anything that fits the data.
I appreciate anyone mentioning the “other side.” However, Jean Twenge’s refutation here is not convincing.
The New Yorker article Twenge cites is typically superficial. The teenagers are mere props, too unimportant to merit serious, hard-hitting investigation into what really underlay their suicides. This is about justifying grownups. The shallowness of articles like these is infuriating for anyone like me who worked for years inside families and communities and saw the family stresses what we call “at risk” teenagers endure.
Of course, the New Yorker writer could have chosen the vastly more common adult tragedies affecting teens – or, conversely, the vastly greater numbers of teens from among the 60-80% who told Pew Research they use social media to connect with people who “help them get through tough times” as examples of teens saved by social media. Instead, he adopts the view that only his grownup peers are entitled to have their grief recognized and justified. This is a society that doesn’t care about kids. This article, aping others, declares: we grownups are the victims here of kids' social media and terrible behaviors.
The same pop-media authors and substack commentators who summon understandable sympathy for parents who lose a child to suicide seem unable to summon any sympathy – in fact, only callous indifference – toward the vastly greater numbers of children and teens who lose tens of thousands of parents and nearby adults to suicide and overdoses every year and live daily with millions of parents’ serious troubles and abuses.
Over the last three years, 1,781 teens ages 10-14 died from suicide and self-inflicted overdoses. All are tragedies worth understanding; their parents’ grief is afforded intense attention and lamenting. However, during that same period, 38,492 Americans ages 40-44 died from suicides and self-inflicted overdoses, most of these parents, parents’ partners, family members, and other grownups who powerfully affect teenagers. Yet – no mention, no sympathy, no attention. Look at the thousands of book, media commentaries, substack pages – absolutely no mention of the teenagers who suffer grownup tragedies and abuses.
Another example: six times more children and teenaged youths are shot to death at home than at school, overwhelmingly by shooters 21 and older. There are no anguished commentaries on these young gun violence victims; only fleeting mention in buried news articles, if that. We only care about kids if their parents do.
Even when the Surgeon General finally issued an advisory on parents’ widespread mental health and addiction issues that directly challenges Twenge's claim that parents today are healthy, the few commentaries it got ignored the effects on children and teens. Twenge once again dodges the most convincing explanation for teenagers’ increased depression: exploding drug abuse and widespread violent and emotional abuses inflicted on teens, especially girls and LGBTQ youth, by their parents and household adults.
Among ages 25-64 from 2010 to 2021, deaths from overdoses surged from 27,399 to 93,107, hospital ER overdose cases rose from 440,000 to 1,040,000, and annual ER cases involving drugs and alcohol rose from 2.8 million to over 5 million. Contrary to Twenge’s dismissals, these mammoth numbers, surging in the US and all across Anglo countries during exactly the period teens became more depressed are not “small.”
Think correlation strongly suggests causation, as Twenge states? Soaring parent-age drug/alcohol crises and teens’ rising depression and suicide rates correlate at r=0.80 plus, orders of magnitude higher than the weak, near-nothing social-media correlations (0.20 and lower) Twenge cites.
We now have further estimates of the grownup-caused difficulties a majority of America’s youth, especially girls and LGBTQ teens, face in their homes. The Centers for Disease Control’s massive, just-released 2023 survey of 19,000 teens finally reveals how appallingly widespread parents’ and adults’ addiction, depression, suicidality, jailing, absence, domestic violence, and emotional and violent abuses are associated with teenagers’ poor mental health. This survey is devastating – which is why I suspect it is receiving no attention. I will detail on my substack how the new CDC survey demolishes conventional discussion of teens, depression, social media, and home life within a few days.
Thanks for taking the time and energy to call out inappropriate ways of using data in the general media. As professionals, we need to hold them accountable to cite research accurately.