Special Collection

Terms of Service

2.6 billion people don’t use the internet. What’s stopping them?

How the internet changed news, according to The Onion

My anxious generation: The unforeseen toll of a digital childhood

The next big tech trend will start out looking like a toy

Replit CEO Amjad Masad on bringing the next 1 billion software creators online

Meet Thresh, the world’s first professional gamer

You’re thinking of the metaverse all wrong, says Matthew Ball

Hugo Mercier says we’ve been misinformed about misinformation

Perplexity, Google, and the battle for AI search supremacy

Constitutional warning shot for social media “deplatforming” laws

In 2017, Jean Twenge, a psychologist at San Diego State University, penned an article for The Atlantic titled “Have Smartphones Destroyed a Generation?” Her answer to that scary question was an unequivocal, “Yes.”

Analyzing large datasets, such as the Monitoring the Future survey, Twenge looked at the feelings, values, and behaviors of adolescents. She discovered that today’s teens were less likely to hang out with friends or engage in rites of passage such as getting a driver’s license and going out on dates. The only area where they seemed to be outcompeting previous generations was in feelings of loneliness, anxiety, depression, and suicidal thoughts. And the steepest declines began — according to Twenge’s graphs — right around 2007, the year Apple released the iPhone.

“[T]he twin rise of the smartphone and social media has caused an earthquake of a magnitude we’ve not seen in a very long time, if ever,” Twenge wrote. “There is compelling evidence that the devices we’ve placed in young people’s hands are having profound effects on their lives — and making them seriously unhappy.”

While Twenge certainly wasn’t the first to sound the alarm, her article and subsequent book, iGen (published that same year) brought the simmering disquiet to a cultural boil. Other researchers soon added their voices and research to a commentator’s call to arms. Among them was Jonathan Haidt, a social psychologist who came to prominence for his popular science books The Happiness Hypothesis (2006) and The Righteous Mind (2012), and Vivek Murthy, the US Surgeon General.

As Murthy told CNN, “[Adolescence is] a time where it’s really important for us to be thoughtful about what’s going into how they think about their own self-worth and their relationships, and the skewed and often distorted environment of social media often does a disservice to many of those children.”

However, many experts argue that Twenge sounded the alarm prematurely, and the research since 2017 hasn’t made the situation any clearer. They contend the data doesn’t support such forceful conclusions and the current dustup verges on becoming another moral panic over a new technology. Worse, politicians and scared parents acting on those fears may overstep in ways that don’t help kids and ultimately do more harm than good.

“Instead of panic and rash measures that take away young people’s agency and opportunities, we must focus on addressing the many and multifaceted mental health challenges young people face that are not limited to social media,” Eiko Fried, a psychologist at Leiden University, writes.

Who’s correct? Are teens taking an unacceptable risk when they click “I agree” on the terms of service, or is social media merely holding up a mirror to the challenges that adolescents have been facing for decades? And why can’t the experts agree?

The answers to those questions are complicated, and to explore them, we’ll need to ask a different question: Is social media more akin to heroin or potato chips?

Studies in heroin and potato chips

That word choice may seem odd and it is, but the analogy has its roots in two important studies that have come to embody the current debate.

The first is a 2019 study by Andrew Przybylski and Amy Orben, psychologists at the Oxford Internet Institute and the University of Cambridge, respectively. They studied three large datasets — among them the Monitoring the Future survey — to look at the relationship between various factors and mental health. They found a negative correlation for digital technology. That is, teens who reported spending more time using digital technology were less likely to report positive well-being.

However, the association was incredibly weak, so weak that the authors claimed it held “little practical value.” Their analysis showed that eating potatoes had a comparable negative association. Wearing glasses had a slightly larger negative association, and the largest negative associations were found in those well-documented risk factors for teens’ mental health, namely bullying and drug use.

Then in 2022 Twenge and Haidt performed their own study. Their statistical analysis looked specifically at the association between social media and mental health in girls (versus all digital technology use, across all teens). They found a much stronger negative association between social media use and mental health, which the authors said was more in the ballpark of binge drinking and hard drug use than eating potatoes.

“There is a clear, consistent, and sizable link between heavy social media use and mental illness for girls, but that relationship gets buried or minimized in studies and literature reviews that look at all digital activities for all teens,” Haidt writes.

Here’s the kicker: Despite these starkly different conclusions, both studies looked at the same three datasets.

A bevy of beauty filters and depression

We’ll get to why an analysis of the same datasets can lead to such different conclusions in a moment. For now though, let’s dive deeper into Twenge and Haidt’s argument as they aren’t hanging their premise on a single result.

Alongside Zach Rausch, a research scientist at the NYU-Stern School of Business, they have amassed a series of studies, systematic reviews, and meta-analyses they maintain to show a direct causal link between social media and the mental health struggles of GenZ. This research has been compiled into a public Google Document, and Haidt made it the subject of his latest best-selling book, The Anxious Generation: How the Great Rewiring of Childhood Is Causing an Epidemic of Mental Illness (2024).

According to them, the research shows that GenZ’s well-being began to decline after the introduction of smartphones and social media. This technological combo brought this change through several mediators. Many of these mediators were already known to affect teens’ mental health, such as bullying and sleep disruption, but making them an easily accessed, ever-present part of teens’ lives exacerbated their effects in subtle, yet hazardous ways.

We’ll consider one example: social comparison. Teenagers have always compared themselves to other teens, and this isn’t inherently a bad thing either. The adolescent years are a time for forming personal beliefs and identities. Exposure to others’ actions, opinions, lifestyles, and accomplishments can be inspirational and instructional. It can also serve as the foundation for new and lasting friendships.

The problem is that social media offers an endless stream of such comparisons, and online influencers manicure their feeds to present exciting lives of pixelated perfection. Unlike the magazine covers of the 1990s — which many argue triggered a wave of body dysmorphia in girls — such feeds aren’t presented as high-end fashion shoots either. They are posed as slices of Ferrari-driving, hotel suite-living, everyday life.

Beauty filters further allow teens — young women in particular — to enhance their appearances to Jessica Rabbit levels of impossible beauty. Preliminary research suggests that it’s not only viewing a parade of filtered facades that reduces mood and body satisfaction. The act of using these image editors does, too. Unfortunately, many young women feel compelled to use these filters — whether they want to or not — to feel relevant and seen in social media’s attentional arms race.

“Social media, as it is commonly used by teens today, increases the quantity of social connections and thereby reduces their quality and their protective nature,” Haidt writes in The Anxious Generation [emphasis his].

Combine this with other mediators, such as FOMO and fractured parental relationships, and it’s easy to see where Twenge and Haidt are coming from. Online culture seems prime to make teens “compare and despair” when their lives and accomplishments don’t measure up. The fallout may be depressive symptoms and lower self-esteem.

Twenge, Haidt, and Rausch also cite “natural experiments” that show mental health problems increasing among girls after the arrival of high-speed internet in local areas, such as rural Spain or even Amazon tribes, and surveys showing that the majority of teens believe social media to be the primary driver of young people’s mental health challenges today.

“This is the great irony of social media: the more you immerse yourself in it, the more lonely and depressed you become. This is true both at the individual and at the collective level,” Haidt writes.

A classic case of moral panic

In his 2024 book Unlocked: The Real Science of Screen Time (and How to Spend It Better), Pete Etchells, a psychologist at Bath Spa University, has a chapter sporting that now all-too-familiar question: “Are Smartphones Destroying a Generation?” His answer was an unequivocal, “Not so fast.”

As Etchells pointed out in an interview with Freethink, this kind of panic is nothing new. Only a few short decades ago, parents fretted that video games turned boys into murderous delinquents. In the 1920s, radio was blamed for zapping children of their vitality. In the 19th century, newspapers were condemned for spurring antisocial behavior. Social commentators even once feared that chess and books could cause insanity in young people.

This pessimistic history does not mean social media gets a free pass. It should be judged on what the research reveals its merits and demerits to be. But when Etchells and others have reviewed Twenge and Haidt’s arguments, they have found them wanting for three important reasons.

The first reason goes back to the Przybylski and Orben study. While studies in this area have shown a negative association between social media and mental health, most have either found the association to be weak or non-existent. In other words, small potatoes.

In a 2020 review of the research literature, Candice Odgers and Michaeline Jensen, psychologists at the UC Irvine and UNC Greensboro, respectively, found that the research was ultimately mixed but that the most rigorous, pre-registered studies report “small associations between the amount of daily digital technology usage and adolescents’ well-being that do not offer a way of distinguishing cause from effect.”

That note about “cause” brings us to the second important reason: Much of the research in this area comes from correlational studies, and as the internet is fond of reminding us, correlation does not imply causation.

That’s because correlational studies are non-experimental. They investigate the relationship between two variables, but the researchers don’t control either variable. This is for various reasons, such as it being impractical or unethical to do so. For instance, if you suspect that social media causes depression, it would be wicked to subject a group of teens to seven hours of doom-scrolling every day for months just to find out if your hypothesis is correct.

This lack of control means such studies can’t determine cause and effect. Even if a correlational study shows that social media use and depression are linked, it could be that social media erodes mental health or that poor mental health causes people to spend more time on social media or that some third factor is influencing both. There may also be lots of differences between the people who spend time on social media and those who don’t that are difficult to account for.

Now, if the size of the association is large, the findings are consistent across many studies, and there’s a plausible causal mechanism, then a case can be made that one thing causes the other. Such is the case between smoking and lung cancer. But as we’ve already seen, that’s not what the research currently shows.

That’s a quick-and-dirty explainer of a complex topic, so let’s look at an example. In Unlocked, Etchells points out that teen suicide-related hospital visits in the US began rising around 2012. That corresponds to just after the 2010 explosion of social media apps on smartphones, which may lead one to assume social media caused the uptick. However, it also corresponds to changes in medical screening practices around depression and self-harm. A similar uptick in 2017 followed a change in the way suicides were recorded, too. At the same time, countries such as France and Germany witnessed falling suicide rates among young people.

It could be that screening changes revealed widespread teen mental health issues that were always there, that both had a role to play in the rise of suicide-related visits, or that another factor is to blame. We simply don’t know.

“[R]esearch which attempts to understand the links between screen time and mental health is so fraught with methodological problems and statistical limitations that we simply don’t have anything near the sort of convincing evidence base necessary to come to firm conclusions about cause and effect,” Etchells writes.

No easy answers (but plenty of different ones)

Those methodological problems are the next reason why some researchers aren’t convinced by Twenge and Haidt’s argument. They also explain why the Przybylski/Orben study and the Twenge/Haidt one could reach such wildly different-sounding conclusions despite drawing from the same datasets.

Depending on what variables you’re looking for, how you define them, and what measurements the datasets provide that fit those definitions, researchers can conjure many different statistical analyses — even if they are interested in what sounds like the same question. If we care about “kids, screens, and mental health,” should we analyze smartphone use, social media use, or all “screen time”? Do we look at just boys, just girls, or all teens? Should we look at depression, anxiety, or self-reported mood? The question allows countless answers.

The larger the datasets, the more pronounced this problem becomes, because researchers will have more potential measurements to consider. In their 2019 study, Przybylski and Orben determined that there were 2.5 trillion analysis options for their three datasets.

“It’s a problem that is regularly found in the sciences, and particularly in new areas of enquiry, where researchers don’t yet have a strong theoretical or methodological base for investigating whatever it is they’re interested in,” Etchells writes. And without those, it becomes difficult to determine which is the best path to the answers we seek.

That said, he does add: “If you are looking at the results from a statistical analysis, and those results tell you that screens are worse for you than heroin, then either we’ve got the debate around screen time very wrong, or I would gently suggest that you’ve borked your analysis.”

Even when you ask adolescents directly about social media, the data and messaging can get messy. While many teens do say that social media harms mental health in general, they are also likely to say that it’s a problem for others and not themselves. When speaking of their own experiences, they are just as likely to report positive times with their friends and that the platforms offer opportunities to be creative and have fun.

“What I feel is missing from the debate is just an acknowledgement that we are scaring people, that we may be scaring them unnecessarily, and that we can have these deep, serious conversations without resorting to that,” Etchells said in our interview. “The thing that I find disheartening about the broad conversation that we’re having is this sense of despair and worry.”

Coming to terms with social media’s terms

Of course, Haidt, Twenge, and Raush will have their counterarguments. They may point to the fact that studies rarely show a positive association between social media and mental health. Even when weak, the association is primarily negative. They could also note that much of the research looks for what’s called a “dose-response relationship” between social media and mental health. It considers the size of the dose (time on social media) and then sees what the response is (mental health outcomes).

But Haidt notes that social media’s impact spreads through social networks in ways that mean even people who don’t participate are affected. Being the only one who opts out of social media might increase your social isolation, even if everyone would be better off without it. “At present, there is almost no way for parents to raise their kids the way they want, because the tech companies have put us in a trap where if we do what we think is right, we’re isolating our kid,” Haidt argued during an Intelligence Squared lecture.

Etchells, Odgers, and Przybylski will likely have their counterarguments to those counterarguments — and you see where this is going.

That leaves us, well, pretty much where we started. It will take more time and (hopefully constructive) conversations for the research surrounding social media to sharpen. In the end, we may need to develop new social norms surrounding social media, the panic may subside as the technology blends into the background of modern life, or both. And any conclusion won’t necessarily be the result of definitive answers.

There is one point, however, on which researchers on both sides of the debate agree, and it’s this: There are no easy answers when it comes to mental health, and no one solution that will solve all our problems.

In our interview, Etchells noted that mental health is the result of many things interacting in a myriad of ways. We won’t find one root cause to explain such a complex thing. Even back in her original 2017 article, Twenge acknowledged that no single influence can define a generation. “Parenting styles continue to change, as do school curricula and culture, and these things matter,” she wrote.

Because of this, Etchells believes we need to change our approach. Instead of focusing on whether social media is harmful or not, we should look at it as “an ecosystem of factors.” We should be asking: What factors of social media trigger adolescent vulnerabilities, and how do we support those teens who are most vulnerable to them? What factors help teens be creative, enliven friendships, and find help, and how do we promote those?

“Trying to figure out this complex relationship allows you to do more than if you are wholly focused on [social media] in terms of harm,” Etchells said.

Because when it comes to social media, the terms of service aren’t set once we click “I agree.” They evolve and change as we interact with and learn about these new social tools.