CHAPTER 7

Anxiety and Depression

Depressed people often stick pins into their own life rafts. The conscious mind can intervene. One is not helpless.

ANDREW SOLOMON, The Noonday Demon: An Atlas of Depression1

The second of our six explanatory threads is the rise in rates of depression and anxiety among American adolescents in the 2010s. These mood disorders have many close relationships with the three Great Untruths.

Here is a first-person account of depression. It is not from an adolescent, but it illustrates Andrew Solomon’s statement above, about how the conscious mind can intervene:

I had spent the day scouring websites for ways to kill myself. At almost every turn, I found stories about how a method could fail, leaving you still alive but permanently injured. This even applied to shooting yourself. I could not risk that, so I went to the hardware store across the street, looking for strong plastic bags and metal wire. The idea was to crush up all the sleep meds, tranquilizers, and anti-anxiety meds I had, take them all at once, and then wrap my head so that even if the pills did not kill me, suffocation would. But it had to be strong enough that I could not claw my way out of the bag if I had a change of heart.

I needed to go through with it now, as quickly as possible. Because . . . why? Because it was the right thing to do, and if I waited, I might not go through with it, and I needed to go through with it while I had the will. If I felt better later, it would somehow be a lie. I had a powerful sense that I was in touch with some dark, larger truth: that I needed to die.

I don’t know if it was briefly sensing how strange this thought was that gave me that tiny flash of sanity that caused me to call 911. First, I started to explain what I had planned in a detached way, but soon I was crying. The voice on the other side of the line told me to get myself to a hospital right away. I listened.

I spent the next three days of December 2007 at a psychiatric facility in North Philadelphia. I was already scheduled to move from Philadelphia, where I felt utterly isolated, back to New York City, where I had friends and family. I found a doctor who was the first person in years to reduce—rather than increase—my meds. And I started cognitive behavioral therapy as soon as I moved to New York.

At first, it seemed to make little difference. The doctor showed me time and time again how I used every bit of brain power to support a view of myself—a schema—that said I was a hopeless, broken person. I did my CBT exercises twice a day, and I gradually came to recognize my angry, flailing, defensive mind trying to protect that nasty vision of myself.

There was no “eureka” moment. My rational mind could understand that my thoughts were distorted, but nothing changed until it simply became a habit to hear the cruelest, craziest, and most destructive voices in my head without believing I had to act on them. When I stopped letting those voices win, they got quieter. Thanks to CBT, my mind is now in the habit of hearing my worst thoughts as if they are speaking in silly cartoon voices. While I still get depressed, the frequency and severity of those bouts are nowhere near as powerful as they used to be.

The author of this account is Greg. He believes that CBT saved his life. In a matter of just a few months, he began to learn how to catch his own distortions. And once he learned to spot them in himself, he started to hear them coming from other people, too. Once you are accustomed to looking for them, it’s not very hard to identify catastrophizing, dichotomous thinking, labeling, and all the rest.

Almost as soon as he started practicing CBT, in 2008, Greg noticed, in his work as the president of FIRE, that administrators on campus were sometimes modeling cognitive distortions for students. Administrators often acted in ways that gave the impression that students were in constant danger and in need of protection from a variety of risks and discomforts (as we’ll discuss in chapter 10). But back then, Millennial students mostly rolled their eyes at administrative overreaction. It was only when the first members of iGen started entering college, around 2013, that Greg began to notice this more fearful attitude about speech coming from the students themselves. In the new discussions about safe spaces, trigger warnings, microaggressions, and speech as violence, students often employed arguments and justifications that seemed to come right out of the CBT training manual. That’s why Greg invited Jon to lunch in 2014, and that’s why we wrote our Atlantic article in 2015.

In that essay, we briefly discussed changes in childhood in the United States, such as the decline in unsupervised time and the recent rise of social media, but we focused our attention on what was happening after students arrived at college. At the time, we had just begun to hear the first alarms being raised by college mental health professionals, who said they were being overwhelmed by rising demand.2 We suggested that perhaps some of the very things colleges were doing to protect students from words and ideas ended up increasing the demand for mental health services by inadvertently increasing the use of cognitive distortions.

By 2017, however, it was clear we had misunderstood what was going on. Colleges were not the primary cause of the wave of mental illness among their students; rather, the students seeking help were part of a much larger national wave of adolescent anxiety and depression unlike anything seen in modern times. Colleges were struggling to cope with rapidly rising numbers of students who were suffering from mental illness—primarily mood disorders.3 The new culture of safetyism can be understood in part as an effort by some students, faculty, and administrators to remake the campus in response to this new trend. If more students say they feel threatened by certain kinds of speech, then more protections should be offered. Our basic message in this book is that this way of thinking may be wrong; college students are antifragile, not fragile. Some well-intended protections may backfire and make things worse in the long run for the very students we are trying to help.

In this chapter, we explore recent findings on the declining mental health of American adolescents. There is some evidence that similar trends may be happening in Canada4 and the United Kingdom,5 although the evidence in those countries is not as clear and consistent as it is in the United States.6 In all three countries, girls seem to be more affected than boys. How is mental health changing, on campus and off, and why did the new culture of safetyism emerge only after 2013?

iGen

In the 2017 book iGen (which we discussed briefly in chapter 1), Jean Twenge, a social psychologist at San Diego State University, gives us the most detailed picture yet of the behavior, values, and mental state of today’s teenagers and college students. Twenge is an expert on how generations differ psychologically and why. She calls the generation after the Millennials iGen (like iPhone), which is short for “internet generation,” because they are the first generation to grow up with the internet in their pockets. (Some people use the term Generation Z.) Sure, the oldest Millennials, born in 1982, searched for music and MapQuest directions using Netscape and AltaVista on their Compaq home computers in the late 1990s, but search engines don’t change social relationships. Social media does.

Marking the line between generations is always difficult, but based on their psychological profiles, Twenge suggests that 1994 is the last birth year for Millennials, and 1995 is the first birth year for iGen. One possible reason for the discontinuity in self-reported traits and attitudes between Millennials and iGen is that in 2006, when iGen’s oldest were turning eleven, Facebook changed its membership requirement. No longer did you have to prove enrollment in a college; now any thirteen-year-old—or any younger child willing to claim to be thirteen—could join.

But Facebook and other social media platforms didn’t really draw many middle school students until after the iPhone was introduced (in 2007) and was widely adopted over the next few years. It’s best, then, to think about the entire period from 2007 to roughly 2012 as a brief span in which the social life of the average American teen changed substantially. Social media platforms proliferated, and adolescents began using Twitter (founded in 2006), Tumblr (2007), Instagram (2010), Snapchat (2011), and a variety of others. Over time, these companies became ever more skilled at grabbing and holding “eyeballs,” as they say in the industry. Social media grew more and more addictive. In a chilling 2017 interview, Sean Parker, the first president of Facebook, explained those early years like this:

The thought process that went into building these applications, Facebook being the first of them . . . was all about: “How do we consume as much of your time and conscious attention as possible?” . . . And that means that we need to sort of give you a little dopamine hit every once in a while, because someone liked or commented on a photo or a post or whatever. And that’s going to get you to contribute more content, and that’s going to get you . . . more likes and comments. . . . It’s a social-validation feedback loop . . . exactly the kind of thing that a hacker like myself would come up with, because you’re exploiting a vulnerability in human psychology.7

Earlier in the interview, he said, “God only knows what it’s doing to our children’s brains.”

In short, iGen is the first generation that spent (and is now spending) its formative teen years immersed in the giant social and commercial experiment of social media. What could go wrong?

Twenge’s book is based on her deep dives into four surveys that stretch back several decades. One survey focuses on college students, two of them focus on teenagers more generally, and one samples the entire U.S. adult population. Her book contains dozens of graphs she created from these four datasets, showing changes in teen behavior and attitudes since the 1980s or 1990s. The lines mostly amble along horizontally until some point between 2005 and 2012, at which point they arc upward or plunge downward. Some of the trends are quite positive: members of iGen drink less and smoke less; they are safer drivers and are waiting longer to have sex. But other trends are less positive, and some are quite distressing. The subtitle of the book summarizes her findings: Why Today’s Super-Connected Kids Are Growing Up Less Rebellious, More Tolerant, Less Happy—and Completely Unprepared for Adulthood—and What That Means for the Rest of Us.

Twenge’s analyses suggest that there are two major generational changes that may be driving the rise of safetyism on campus since 2013. The first is that kids now grow up much more slowly. Activities that are commonly thought to mark the transition from childhood to adulthood are happening later—for example, having a job, driving a car, drinking alcohol, going out on a date, and having sex. Members of iGen wait longer to do these things— and then do less of them—than did members of previous generations. Instead of engaging in these activities (which usually involve interacting with other people face-to-face), teens today are spending much more time alone, interacting with screens.8 Of special importance, the combination of helicopter parenting, fears for children’s safety, and the allure of screens means that members of iGen spend much less time than previous generations did going out with friends while unsupervised by an adult.

The bottom line is that when members of iGen arrived on campus, beginning in the fall of 2013, they had accumulated less unsupervised time and fewer offline life experiences than had any previous generation. As Twenge puts it, “18-year-olds now act like 15-year-olds used to, and 13-year-olds like 10-year-olds. Teens are physically safer than ever, yet they are more mentally vulnerable.”9 Most of these trends are showing up across social classes, races, and ethnicities.10 Members of iGen, therefore, may not (on average) be as ready for college as were eighteen-year-olds of previous generations. This might explain why college students are suddenly asking for more protection and adult intervention in their affairs and interpersonal conflicts.

The second major generational change is a rapid rise in rates of anxiety and depression.11 We created three graphs below using the same data that Twenge reports in iGen. The graphs are straightforward and tell a shocking story.

Adolescent Depression Rates

FIGURE 7.1. Percent of adolescents aged 12–17 who had at least one major depressive episode in the past year. Rates have been rising since 2011, especially for girls. (Source: Data from National Survey on Drug Use and Health.)

Studies of mental illness have long shown that girls have higher rates of depression and anxiety than boys do.12 The differences are small or nonexistent before puberty, but they increase at the start of puberty. The gap between adolescent girls and boys was fairly steady in the early 2000s, but beginning around 2011, it widened as the rate for girls grew rapidly. By 2016, as you can see in Figure 7.1, roughly one out of every five girls reported symptoms that met the criteria for having experienced a major depressive episode in the previous year.13 The rate for boys went up, too, but more slowly (from 4.5% in 2011 to 6.4% in 2016).

Have things really changed so much for teenagers just in the last seven years? Maybe Figure 7.1 merely reflects changes in diagnostic criteria? Perhaps the bar has been lowered for giving out diagnoses of depression, and maybe that’s a good thing, if more people now get help?

Perhaps, but lowering the bar for diagnosis and encouraging more people to use the language of therapy and mental illness are likely to have some negative effects, too. Applying labels to people can create what is called a looping effect: it can change the behavior of the person being labeled and become a self-fulfilling prophecy.14 This is part of why labeling is such a powerful cognitive distortion. If depression becomes part of your identity, then over time you’ll develop corresponding schemas about yourself and your prospects (I’m no good and my future is hopeless). These schemas will make it harder for you to marshal the energy and focus to take on challenges that, if you were to master them, would weaken the grip of depression. We are not denying the reality of depression. We would never tell depressed people to just “toughen up” and get over it—Greg knows firsthand how unhelpful that would be. Rather, we are saying that lowering the bar (or encouraging “concept creep”) in applying mental health labels may increase the number of people who suffer.

Adolescent Suicide Rates (per 100,000)

FIGURE 7.2. Suicide rate per 100,000 population, ages 15–19, by sex. (Source: CDC, Fatal Injury Reports, 1999–2016.15)

There is, tragically, strong evidence that the rising prevalence of teen depression illustrated in Figure 7.1 is not just a result of changes in diagnostic criteria: the teen suicide rate has been increasing in tandem with the increase in depression. Figure 7.2 shows the annual rate of suicide for each 100,000 teens (ages fifteen to nineteen) in the U.S. population. Suicide and attempted suicide rates vary by sex; girls make more attempts, but boys die more often by their own hand, because they tend to use irreversible methods (such as guns or tall buildings) more often than girls do. The boys’ suicide rate has moved around in recent decades, surging in the 1980s during the gigantic wave of crime and violence that receded suddenly in the 1990s. The rate of boys’ suicide reached its highest point in 1991. While the rise since 2007 does not bring it back up to its highest level, it is still disturbingly high. The rate for girls, on the other hand, had been fairly constant all the way back to 1981, when the dataset begins, and although their rate of suicide is still substantially lower than that of boys, the steady rise since 2010 brings their rate up to the highest levels recorded for girls since 1981. Compared to the early 2000s, nearly twice as many teenage girls now end their own lives. In Canada, too, the suicide rate for teen girls is rising, though not as sharply, while the rate for teen boys has fallen.16 (In the United Kingdom, there is no apparent trend for either gender in recent years.17)

Confirming this increase in mental illness with a different dataset, a recent study looked at “nonfatal self-inflicted injuries.”18 These are cases in which adolescents were admitted to emergency rooms because they had physically harmed themselves by doing such things as cutting themselves with a razor blade, banging their heads into walls, or drinking poison. The researchers examined data from sixty-six U.S. hospitals going back to 2001 and were able to estimate self-harm rates for the entire country. They found that the rate for boys held steady at roughly 200 per hundred thousand boys in the age range of fifteen to nineteen. The rate for girls in that age range was much higher, but had also been relatively steady from 2001 to 2009, at around 420 per hundred thousand girls. Beginning in 2010, however, the girls’ rate began to rise steadily, reaching 630 per hundred thousand in 2015. The rate for younger girls (ages ten to fourteen) rose even more quickly, nearly tripling from roughly 110 per hundred thousand in 2009 to 318 per hundred thousand in 2015. (The corresponding rate for boys in that age range was around 40 throughout the period studied.) The years since 2010 have been very hard on girls.

Antisocial Media?

What is driving this surge in mental illness and suicide? Twenge believes that the rapid spread of smartphones and social media into the lives of teenagers, beginning around 2007, is the main cause of the mental health crisis that began around 2011. In her book, she presents graphs showing that digital media use and mental health problems are correlated: they rose together in recent years. That makes digital media a more likely candidate than, say, the global financial crisis and its associated recession, which began in 2008. By 2011, the economy and the job market were steadily improving in the United States, so economic factors are unlikely to be the cause of deteriorating adolescent mental health in the following years.19

Simple correlations are suggestive, but they can’t tell us what caused what. Lots of things were changing during that time period, so there are many opportunities for what are called spurious correlations. For example, the annual per capita consumption of cheese in the United States correlates almost perfectly with the number of people who die each year from becoming entangled in their bedsheets, but that’s not because eating cheese causes people to sleep differently.20 That correlation is “spurious” because it’s just a coincidence that both numbers rose steadily over the same period of time.

To avoid getting fooled by spurious correlations, we need to consider additional variables that would be expected to change if a particular causal explanation were true. Twenge does this by examining all the daily activities reported by individual students, in the two datasets that include such measures. Twenge finds that there are just two activities that are significantly correlated with depression and other suicide-related outcomes (such as considering suicide, making a plan, or making an actual attempt): electronic device use (such as a smartphone, tablet, or computer) and watching TV. On the other hand, there are five activities that have inverse relationships with depression (meaning that kids who spend more hours per week on these activities show lower rates of depression): sports and other forms of exercise, attending religious services, reading books and other print media, in-person social interactions, and doing homework.

Notice anything about the difference between the two lists? Screen versus nonscreen. When kids use screens for two hours of their leisure time per day or less, there is no elevated risk of depression.21 But above two hours per day, the risks grow larger with each additional hour of screen time. Conversely, kids who spend more time off screens, especially if they are engaged in nonscreen social activities, are at lower risk for depression and suicidal thinking.22 (Twenge addresses the possibility that the relationship runs the other way—that depression is what causes kids to spend more time with their screens—and she shows that this is unlikely to be the case.23)

Part of what’s going on may be that devices take us away from people. Human beings are an “ultrasocial” species. Chimpanzees and dogs have very active social lives, but as an ultrasocial species, human beings go beyond those “social” species.24 Like bees, humans are able to work together in large groups, with a clear division of labor. Humans love teams, team sports, synchronized movements, and anything else that gives us the feeling of “one for all, and all for one.” (Ultrasociality is related to the psychology of tribalism that we talked about in chapter 3. The trick is to satisfy people’s needs to belong and interact without activating the more defensive and potentially violent aspects of tribalism.) Of course, social media makes it easier than ever to create large groups, but those “virtual” groups are not the same as in-person connections; they do not satisfy the need for belonging in the same way. As Twenge and her coauthors put it:

It is worth remembering that humans’ neural architecture evolved under conditions of close, mostly continuous face-to-face contact with others (including non-visual and non-auditory contact; i.e., touch, olfaction), and that a decrease in or removal of a system’s key inputs may risk destabilization of the system.25

This idea is supported by Twenge’s finding that time spent using electronic devices was not generally harmful for highly sociable kids—the ones who spent more time than the average kid in face-to-face social interactions.26 In other words, the potentially negative impact of screens and social media might depend on the amount of time teens spend with other people. But electronic devices are harmful not just because they take kids away from face-to-face interactions; there are more insidious effects, which are felt more strongly by girls.

Why Is It Mostly Girls Who Suffer?

The previous graphs show that mental health has deteriorated much further among iGen girls than among iGen boys. Furthermore, to the extent that social media seems to bear some of the blame, that may be true only for girls. For boys, Twenge found that total screen time is correlated with bad mental health outcomes, but time specifically using social media is not.27 Why might social media be more harmful for girls than for boys?

There are at least two possible reasons. The first is that social media presents “curated” versions of lives, and girls may be more adversely affected than boys by the gap between appearance and reality. Many have observed that for girls, more than for boys, social life revolves around inclusion and exclusion.28 Social media vastly increases the frequency with which teenagers see people they know having fun and doing things together—including things to which they themselves were not invited. While this can increase FOMO (fear of missing out), which affects both boys and girls, scrolling through hundreds of such photos, girls may be more pained than boys by what Georgetown University linguistics professor Deborah Tannen calls “FOBLO”—fear of being left out.29 When a girl sees images of her friends doing something she was invited to do but couldn’t attend (missed out), it produces a different psychological effect than when she is intentionally not invited (left out). And as Twenge reports, “Girls use social media more often, giving them additional opportunities to feel excluded and lonely when they see their friends or classmates getting together without them.” The number of teens of all ages who feel left out, whether boys or girls, is at an all-time high, according to Twenge, but the increase has been larger for girls. From 2010 to 2015, the percentage of teen boys who said they often felt left out increased from 21 to 27. For girls, the percentage jumped from 27 to 40.30

Another consequence of social media curation is that girls are bombarded with images of girls and women whose beauty is artificially enhanced, making girls ever more insecure about their own appearance. It’s not just fashion models whose images are altered nowadays; platforms such as Snapchat and Instagram provide “filters” that girls use to enhance the selfies they pose for and edit, so even their friends now seem to be more beautiful. These filters make noses smaller, lips bigger, and skin smoother.31 This has led to a new phenomenon: some young women now want plastic surgery to make themselves look like they do in their enhanced selfies.32

The second reason that social media may be harder on girls is that girls and boys are aggressive in different ways. Research by psychologist Nicki Crick shows that boys are more physically aggressive—more likely to shove and hit one another, and they show a greater interest in stories and movies about physical aggression. Girls, in contrast, are more “relationally” aggressive; they try to hurt their rivals’ relationships, reputations, and social status—for example, by using social media to make sure other girls know who is intentionally being left out.33 When you add it all up, there’s no overall sex difference in total aggression, but there’s a large and consistent sex difference in the preferred ways of harming others. (At least, that was Crick’s finding in the 1990s, before the birth of social media.) Plus, if boys’ aggression is generally delivered in person, then the targets of boys’ aggression can escape from it when they go home. On social media, girls can never escape.

Given the difference in preferred forms of aggression, what would happen if a malevolent demon put a loaded handgun into the pocket of every adolescent in the United States? Which sex would suffer more? Boys, most likely, because they would find gunplay more appealing and would use guns more often to settle conflicts. On the other hand, what would happen if, instead of guns, that same malevolent demon put a smartphone, loaded up with social media apps, into the pocket of every adolescent? Other than the demon part, that is more or less what happened between 2007 and 2012, and it’s now clear that girls have suffered far more. Social media offers many benefits to many teens: it can help to strengthen relationships as well as damage them, and in some ways it is surely giving them valuable practice in the art of social relationships. But it is also the greatest enabler of relational aggression since the invention of language, and the evidence available today suggests that girls’ mental health has suffered as a result.

iGen Goes to College

The first members of iGen started arriving on college campuses in September 2013; by May 2017, when the eldest members began graduating, the student body at U.S. colleges was almost entirely iGen (at least in selective four-year residential colleges). These are precisely the years in which the new culture of safetyism seemed to emerge from out of nowhere.

These are also the years in which college mental health clinics found themselves suddenly overwhelmed by new demand, according to many newspaper and magazine articles profiling the lengthening waiting lists for psychological counseling at universities across the United States.34 At the time, these profiles of crises at individual universities seemed somewhat anecdotal. When we were writing our Atlantic article, there was no nationally representative survey documenting the trend. But now, three years later, there are several.

A 2016 report by the Center for Collegiate Mental Health, using data from 139 colleges, found that by the 2015–2016 school year, half of all students surveyed reported having attended counseling for mental health concerns.35 The report notes that the only mental health concerns that were increasing in recent years were anxiety and depression. Confirming these upward trends with a different dataset,36 Figure 7.3 shows the percentage of college students who describe themselves as having a mental disorder. That number increased from 2.7 to 6.1 for male college students between 2012 and 2016 (that’s an increase of 126%). For female college students, it rose even more: from 5.8 to 14.5 (an increase of 150%). Regardless of whether all these students would meet rigorous diagnostic criteria, it is clear that iGen college students think about themselves very differently than did Millennials. The change is greatest for women: One out of every seven women at U.S. universities now thinks of herself as having a psychological disorder, up from just one in eighteen women in the last years of the Millennials.

Percentage of College Students Who Say That They Have a Psychological Disorder

FIGURE 7.3. Percentage of college students responding “yes” to the question “Do you have [a] psychological disorder (depression, etc.).” (Source: Higher Education Research Institute.)

These years also saw a rise in self-reports of anxiety as the reason for seeking help. One large survey of university counseling centers found that only 37% of students who came through their doors in 2009 and prior years had complained about problems with anxiety—roughly on a par with the two other leading concerns, depression and relationships.37 But beginning in 2010, the percentage of students with anxiety complaints began to increase. It reached 46% in 2013 and continued climbing to 51% in 2016. It is now by far the leading problem for which college students seek treatment. These years also saw substantial increases in rates of self-injury and suicide among college students,38 so while part of the increase may be due to students being more willing to self-diagnose, once again, we know that the underlying rates of mental illness were increasing. Something was changing in the lives and minds of adolescents before they reached college, and when growing numbers of depressed and anxious students began arriving on campus, beginning around 2013, it was bound to have some effect on university culture and norms.

You can see why it was hard for us to make a strong case that universities were causing students to become anxious and depressed by teaching them disordered ways of thinking. Anxiety and depression rates were already rising for all teenagers before they arrived at college, and for those who never attended college as well. Clearly universities were not causing a national mental health crisis; they were responding to one, and this may explain why the practices and beliefs of safetyism spread so quickly after 2013. But safetyism does not help students who suffer from anxiety and depression. In fact, as we argue throughout this book, safetyism is likely to make things even worse for students who already struggle with mood disorders. Safetyism also inflicts collateral damage on the university’s culture of free inquiry, because it teaches students to see words as violence and to interpret ideas and speakers as safe versus dangerous, rather than merely as true versus false. That way of thinking about words is likely to promote the intensification of a call-out culture, which, of course, gives students one more reason to be anxious.

Depression and anxiety tend to go together.39 Both conditions create strong negative emotions, which feed emotional reasoning. Anxiety changes the brain in pervasive ways such that threats seem to jump out at the person, even in ambiguous or harmless circumstances.40 Compared to their nonanxious peers, anxious students are therefore more likely to perceive danger in innocent questions (leading them to embrace the concept of microaggressions) or in a passage of a novel (leading them to ask for a trigger warning) or in a lecture given by a guest speaker (leading them to want the lecturer disinvited or for someone to create a safe space as an alternative to the lecture). Depression distorts cognition, too, and gives people much more negative views than are warranted about themselves, other people, the world, and the future.41 Problems loom larger and seem more pervasive. One’s resources for dealing with those problems seem smaller, and one’s perceived locus of control becomes more external,42 all of which discourages efforts to act vigorously to solve problems. Repeated failures to escape from what is perceived to be a bad situation can create a mental state that psychologist Martin Seligman called “learned helplessness,” in which a person believes that escape is impossible and therefore stops trying, even in new situations where effort would be rewarded.43 Furthermore, when people are depressed, or when their anxiety sets their threat-response system on high alert, they can succumb to a “hostile attribution bias,” which means that they are more likely to see hostility in benign or even benevolent people, communications, and situations.44 Misunderstandings are more likely, and more likely to escalate into large-scale conflicts.

Screen Time: A Caution About Caution

The rise in adolescent mental illness is very large and is found in multiple datasets, but the percentage of that rise that can be attributed to smartphones and screen time is small, and the evidence is more indirect. Twenge uses the data available, and those datasets report crude measures of what kids are doing—mostly the approximate number of hours per week spent on various activities, including using devices. Twenge finds relationships that are statistically significant yet still generally small in magnitude. That doesn’t mean that the effects of smartphones are small; it just means that the amount of variance in mental illness that we can explain right now, using existing data, is small. If we had better measures of what kids are doing and what is happening to their mental health, we’d be able to explain a lot more of the variance. These problems are very new, and a lot more research is needed before we’ll know why rates of mood disorders began rising so quickly in the 2010s.

One conclusion that future research is almost certain to reach is that the effects of smartphones and social media are complicated, involving mixtures of benefits and harms depending on which kinds of kids are doing which kinds of online activities instead of doing which kinds of offline activities. One factor that is already emerging as a central variable for study is the quality of a teenager’s relationships and how technology is impacting it. In a recent review of research on the effects of social media, social psychologists Jenna Clark, Sara Algoe, and Melanie Green offer this principle: “Social network sites benefit their users when they are used to make meaningful social connections and harm their users through pitfalls such as isolation and social comparison when they are not.”45

So 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. In the meantime, as we’ll say in chapter 12, there is enough evidence to support placing time limits on device use (perhaps two hours a day for adolescents, less for younger kids) while limiting or prohibiting the use of platforms that amplify social comparison rather than social connection. There is also a strong case to be made for rethinking device use in the context of one’s overall parenting philosophy, especially given everything we know about children’s overarching need to play. We take up those topics in the next two chapters.

In Sum