Instagram's Mental Health Emergency

Facebook has recently paused the development of their “Instagram Kids” project after a whistleblower leaked internal documents showing that Facebook's own research finds a link between poor mental health and Instagram use. “We make body image issues worse for one in three teen girls,” said a slide in a presentation given to Facebook executives, “Teens blame Instagram for increases in the rate of anxiety and depression,” said another. Not surprisingly, the Wall Street Journal’s expose has led to a bipartisan backlash against Facebook, with law-makers, child advocates, psychologists, and users themselves describing the harmful effects of social media as the “public health issue of our time.”

For years psychologists have been warning the public that social media, and Instagram in particular, were contributing to large increases in depression and anxiety in teenagers, especially girls. Jean Twenge, psychology professor at San Diego State University whose research looks at generational trends found in very large data sets, described in her 2017 book, “iGen,” sharp decreases in behaviours that prepare young people for adulthood such as driving, dating, socialising at parties, and working.

American 12th grader experiences with driving, alcohol, dating and working from iGen by Jean Twenge.

These trends have occurred in parallel with sharp increases in anxiety and depression, and general feelings of loneliness for this cohort. Twenge’s hypothesis in 2017 was that these trendlines corresponded with widespread changes in behaviour driven by smartphones, as well as the ubiquitous adoption of social media by teens. No other plausible hypothesis for these trends has emerged, and since 2017, they have become more pronounced. Now Facebook’s own research appears to confirm Twenge's hypothesis.

Percentage of American teens (12-17 years) experiencing a major depressive episode in the past 12 months. Source: National Survey of Drug Use and Health

Appearing on this week's Quillette podcast to discuss Instagram and mental health, is New York University social psychologist, Jonathan Haidt, who co-wrote The Coddling of the American Mind with Greg Lukianoff. Haidt has hypothesised that Instagram exacerbates the normal anxieties felt by teenage girls about their appearance, their status within their peer groups, as well as fuelling the subtle bullying that is done via gossip and social exclusion. On the podcast, Haidt explains why Instagram is different from other forms of social media, contrasting synchronous real-time social behaviour, like playing video games, and asynchronous post-and-comment social interactions. Rather than getting lost in a fantasy world with friends, like many teenage boys do with interactive, multiplayer video games, Instagram encourages users to post impressions of their lives and then wait for social approval, intensifying the normal self-consciousness that pervades adolescence.

Podcast #168: Jonathan Haidt on Instagram’s Mental Health EmergencyQuillette podcast host Jonathan Kay speaks with NYU professor and Coddling of the American Mind Co-Author Jonathan Haidt about the emotionally destructive effect of social media on many young users.

But the effects of Instagram go beyond the individual user. Haidt has warned that Instagram has an impact on girls even if they don’t use it very often, describing the systemic effects of the app as placing every girl into a “prestige economy based on her appearance.” The implication being that social prestige at school or university now frequently derives from a girl's Instagram followers, as status on the app is transferring into status in real life.

Psychological stress arises from Instagram through the relentless social comparisons it induces, alongside its addictive design. Instagram, like Facebook and Twitter, doesn’t stop. Features such as “infinite scrolling” are designed to addict users to an endless newsfeed. The little red dots and circles hook us up to a drip feed of rewards—all we have to do is press or click, and our repetitive behaviour is reinforced.

The first insights into reward-seeking behaviour came from a series of classic experiments in a lab at McGill University in the 1950s. Two scientists, James Olds and Peter Milner, found that if they hooked up electrodes to certain sections of a rat’s brain, the rat would voluntarily press a lever giving itself tiny jolts to stimulate itself.

A rat performing electrical brain stimulation, Scientific American, 1956

The scientists were interested in which parts of the brain the rats wanted to stimulate, and how far the rats would go in order to receive this stimulation. Disturbingly, Olds and Milner discovered that when the lateral hypothalamus was stimulated, the rats pressed the lever several thousand times per hour, for days, preferring to press the lever than to eat or drink. The rats even crossed electrified grids—scorching themselves—in order to get their hit. Observing this, Olds assumed the rats were in a state of bliss, describing his findings in Scientific American with an article titled "Pleasure Centres of the Brain." But this hypothesis was not quite right. Scientists now know that it is not the reward itself, but the anticipation of a reward that powerfully drives goal-directed behaviour. The rats were not in a state of bliss, they were in a state of desire.

Dopamine, the feel-good hormone and neurotransmitter that is implicated in the reward circuitry of the human brain, rises in anticipation of a reward such as social approval, but not necessarily in response to a reward itself. Robert Sapolsky, professor of biology at Stanford, explains, “dopamine is not about pleasure, it’s about anticipation of pleasure, it’s about the pursuit of happiness, rather than happiness itself.” It also rises much higher in the anticipation of an uncertain or intermittent reward—a mechanism well known to the designers of poker machines.

Every time we open our phones, and see a red dot which indicates that we have a new “like” or a new message on a social app, we experience feel-good chemicals in the anticipation of a reward. The intermittent nature of these rewards means that our behaviour is reinforced more strongly than if the rewards were consistent or certain. One day we may receive 30 likes on a post, the next day two. Intermittent reinforcement helps explain why we compulsively check these apps, even though they may make us feel bad about ourselves. In some cases, we may go back even though we are being bullied. Like a bleary eyed slot machine player, or a rat on the hook in a science lab, our brains keep telling us to press that lever.

Prior to the discovery of Facebook’s own research linking Instagram to teenage depression and anxiety, evidence of its harms came from mostly correlational studies. One study, looking at 120,115 English adolescents found that light use of social media was associated with greater wellbeing than no use, but this dropped off when usage exceeded two hours, with heavy users (five hours or more) having lowest levels of wellbeing in the sample. Another study published in 2019 reported that:

[H]eavy users (vs. light) of digital media were 48 percent to 171 percent more likely to be unhappy, to be low in well-being, or to have suicide risk factors such as depression, suicidal ideation, or past suicide attempts. Heavy users (vs. light) were twice as likely to report having attempted suicide.

Evidence from experiments where participants are randomly assigned to different groups suggests that these are not spurious correlations. In one experiment, students were randomly assigned to a group where they limited their social media to 10 minutes a day, in the other they were instructed to use social media as they normally would. The group which limited themselves to 10 minutes a day had significant decreases in feelings of loneliness, depression and anxiety when compared to the group that didn’t change their behaviour. Another experiment in which participants deactivated their Facebook accounts entirely found that those who deactivated had increased subjective wellbeing, socialised more with friends and family, and when they eventually returned to Facebook, spent much less time on the platform than they had done previously.

However it is also important to note that some studies have found little harmful effect of social media use, and others have found that certain ways of using social media, such as passive use, is more predictive of suicidal ideation than use in general. Further analyses have found that when controlling for variables such as experience of online bullying, sleep, and seeing friends, the ill effects of social media use disappear entirely.

Yet the impact of social media does not just occur at the individual level. Like the weather, our social world is a complex system, where even tiny changes in one variable can have cascading, systemic effects. And some behaviours predicting poor mental health may have a contagiousness to them. If mood disorders like anxiety and depression are on the rise, it is plausible that social media creates a feedback loop where young people are increasingly exposed to others who are depressed and anxious—making them more likely to be depressed and anxious themselves. Sociological studies of network effects do indeed show that depression can be contagious among adolescent social groups. Social contagion is also known to drive self harm and suicide. And the statistics showing an increase in self harm in teenage girls are stark. In Australia, the years between 2007 and 2008 the Australian Institute of Health and Welfare (AIHW) records that 359 per 100,000 girls were hospitalised for self harm. In the years 2016–2017, after Instagram had become the most popular social media platform for teens, this had doubled to 686 hospitalisations per 100,000. Rates of hospitalisation for self-harm have also doubled in the in the US, and have tripled in the UK over the same time frame.

In Australia, these disturbing mental health trends have been compounded by prolonged lockdowns due to the COVID pandemic. In the first six months of 2021, the attempted suicide rate for Victorian teenagers reportedly sky rocketed by 184 percent. The Australian has reported that hospital admissions for acute mental health care increased by only two percent for the general population in 2021, but for children and young people, admissions have surged by 43 percent. But while lockdowns can be lifted, and sleep can be adjusted, many of the variables that impact mental health remain difficult to address both at the individual and societal levels. We know that genes affect our subjective wellbeing as well as the likelihood of developing depression and anxiety. We also know that poverty, family stability, childhood abuse, neglect, as well as trauma, all play a significant role in the development of mental illness.

One reason why it is important to know that social media use is related to anxiety and depression is that it is comparatively easy to fix. It is not easy for us to fix our genes—or our family's socio-economic-status—if we happen to be born into disadvantage. We can, however, deactivate our social media accounts. And if the available evidence is any guide, this quick and easy intervention may have the potential to bring about significant increases in happiness to a large number of people.

Adam Mosseri, head of Instagram, has likened social media to the automobile industry. Cars, of course, create tremendous value by getting us from A to B, and making our lives easier. But cars also create accidents and road fatalities.

Instagram boss likens social media to cars, says people will dieAdam Mosseri, head of Facebook-owned Instagram, made the unfortunate argument that social media is like cars and that casualties will be common.

Similarly, social media shortens the distance between friends around the world, and helps us maintain connections we would easily lose through distance and the busyness of life. There is also no question that social media has made it easier for people to monetise their hobbies, something that is truly beneficial for those seeking unconventional career paths. And it is true that when used judiciously, and not as a substitute for real world interaction, social media can strengthen relationships and reduce isolation.

The magazine in which this essay appears—Quillette—likely wouldn't exist without social media technologies, including Facebook. In the past, only corporations with vast capital could afford the distribution networks that would allow a publication to reach audiences around the globe. With the connectivity of the internet, and the ability of like minded people to find each other, these traditional barriers to entry have been broken down, allowing companies like ours to thrive.

When criticism of Facebook has come from corporate media organisations, it is often dismissed as being motivated by professional resentment. It is often assumed that journalists working for legacy (pre-internet) media companies are threatened by the new (post-internet) media landscape, which has democratised journalism, allowing a thousand blossoms to bloom. This might be true in some cases. But journalists are also parents and citizens—as are the entrepreneurs who have built their businesses on the back of social media. Aside from having an interest in living in a healthy and functioning society, most content creators who use these platforms for their daily jobs, also have a commercial interest in social media companies exercising quality control. A professional singer, for example, might be grateful that she gets to sing in a bar where patrons provide lucrative tips. But she will become discouraged if the bar plies minors with alcohol and continually serves drunks more booze. Once these toxic externalities become too large for the wider community to bear, the bar will likely be shut down, making the singer worse off financially than she was before. Social media is no different. Platforms like Instagram create eco-systems where content creators can thrive. But if that eco-system becomes unmanageably toxic, its entire economy is at risk.

The main ethical problem posed by Instagram is that young people are not mature enough to give informed consent in having their preferences harvested and fed back to them by a corporation that specialises in attention manipulation. It is one thing for adults to voluntarily consent to being dazzled by a newsfeed of glossy, unattainable images, but quite another for a 14 year old child.

Mosseri’s car analogy has been harshly criticised, but in many ways it is accurate, and should be considered carefully. We do recognise cars as providing great value, but we also recognise them as lethal killing machines in certain contexts. Over the years, societies have developed regulations for cars ensuring their safety and roadworthiness including requirements for brake lights, indicators, headlights, fog lights, and seatbelts. We also require drivers to be licensed, to have number plates, to obey drink driving laws and follow the road rules. And, crucially, we don’t let kids drive. In Australia, it usually takes a 16 year old about 12 months and 100 hours of driving practice to get their full license, and even then, their license is only provisional. In contrast, an 11 year old can sign up to Instagram in just a couple of minutes, without any oversight and without any attempt on behalf of Facebook at age verification. In light of this, Mosseri’s automobile analogy should be taken seriously. If Instagram wants to be seen as analogous to the car industry, we should meet them halfway and legislate accordingly.

Listen to Quillette's Jonathan Kay with NYU's Jonathan Haidt on Instagram’s Mental Health Emergency here:

Podcast #168: Jonathan Haidt on Instagram’s Mental Health EmergencyQuillette podcast host Jonathan Kay speaks with NYU professor and Coddling of the American Mind Co-Author Jonathan Haidt about the emotionally destructive effect of social media on many young users.

This is a companion discussion topic for the original entry at

Another adverse effect of social media is that it crystallizes socially affirming signals into widespread dogma. An example is Quillette’s position on vaccines where articles have divided people into vax versus anti-vaxx. The subject is emotional because it relates to personal safety. But it should not be over-simplified as it is fast-changing and complex. Most people including journalists are not scientifically literate and unable to get their heads around details of immunology, epidemiology, physiology and health policy. They get taken in by political talking points that are amplified through social media.

Social media has proven destructive to nuanced discourse relating to the pandemic. Quillette’s owner, unfortunately, has been unduly influenced by Instagram.

1 Like

interesting that Claire left twitter and remains active on instagram, twitter is a worse evil - in a signle day (today) this trashy analysis in a high impact journal was tweeted 16,000 times: Increases in COVID-19 are unrelated to levels of vaccination across 68 countries and 2947 counties in the United States | SpringerLink

1 Like

Of course it is terrible that social media leads to such human misery, anxiety and depression, but the far greater fear should be that it is also deranging us- nothing sells click bait more quickly than the latest outrage porn. The attention economy is driven most by the negative engagement of partisanship asking us to believe things about the people who support the other party which are often little more barefaced lies and rarely possess anything more than a shred of truth.

It shouldn’t surprise- at least in America- where the legacy media has been playing this self-same game for over twenty years. Matt Taibbi writes about it in his book Hate Inc. Of course, in the past this rage porn fix was little more than laudanum or the modern equivalent of the opium den, but now social media has accelerated the fix, giving anyone who wants it a fentanyl hit on demand, and is tearing the world apart in the process…

As usual, my essays are available on my Substack, and free to view and comment:


There’s a trend in the level of social measures imposed to restrict the spread of Covid in the most vaccinated countries.

For western countries incl former East European countries with the lowest vaccination rates not so much

Looking closer at what’s happened in Hungary and Czechia, which now have lower levels of social restrictions, both countries have had very high death rates, which might mean they’ve nearly reached natural herd immunity.

Agreed, but legislators can be bought. I would suggest going further with the auto industry comparison. Instagram/Facebook should be liable for damages, particularly given that:


This may surprise you, but I don’t actually get my information about vaccines from Instagram.


Yep. I wonder if class actions are now on the table? If you lost your daughter to suicide, and Instagram knew that this is or was likely to happen, surely they can be held liable in a court of law.


I work in a hospital in the U.S. Every clinician makes mistakes at some point (myself included). Layers of process prevent the vast majority of errors from getting thru to the patient, but it does happen. It hurts the patient, and it hurts the clinician. The feelings of guilt can be terrible. We didn’t go into healthcare to hurt people, quite the opposite.

An industry that knowingly causes very real damage to children? I have no sympathy for them.


I am honestly undecided. Although personally I get a great deal of schadenfreude from their current predicament. I detest the company.

In all honestly I am not left wondering why Facebook behaves the way it does. I am left wondering “Why do we allow children to give consent to EULAs in the first place”?


Not that I want to defend Facebook or Instagram, because they both suck in my opinion and I don’t have an account for either one, but from what I read, Facebook had obtained data based on their own internal studies that indicated 20-30% of teenaged girls using Instagram thought it made them feel worse about themselves, and people at Facebook were in fact concerned about what could be done to remedy the situation. I guess they get a few brownie points at least for undertaking the study in the first place and then at least brainstorming a bit about what to do with the results. I mean, they’re still basically a bunch of 21st century tobacco salesmen, I know, but at least they’re concerned about the cough it’s giving us.


This made me laugh.


I do applaud the relative honesty of both Instagram and Facebook about this issue. It’s very different from the tobacco or auto industries back in the day. We can’t say we weren’t warned.

(BTW listen to the latest podcast episode with Jonathan Haidt, covering these very issues.)

1 Like

Hmm. The first graph that shows a decline in twelve graders behavior concerning driving, dating, working, etc. is showing a serious decline about 15 years before the smartphone and about 10 years before wide spread use of mobile phones. Maybe Facebook et. al. is making a bad situation worse but imho they are exonerated by this graph.

Having my children being born in 1988 and 1990 I can see a much clearer reason why our kids, mostly the girls, are so fragile: feminist overly protective mothers.

The role of the father has been significantly diminished which has led to a focus on safetyfying (rubber tiles on tame fenced playgrounds) and equitizing (everybody wins) society, reducing the buildup of robustness against adversity, the part mostly fathers used to add.

Add to this the burden put on girls to compete on masculine traits while making the use of feminine traits suspect (how vicious can you be?), while any masculine traits of boys are tried to be changed from a very young age and I think it is clear the destruction of our youth started long before Facebook was a dream of Z.


I agree that social media is not the only culprit here & to be fair Haidt makes it clear there is a parenting safety component as well. But what has created that in my experience is increased access to news & its overwhelming negativity bias. I was born in 1970 & it was very common that the news was on during dinner time through out my child hood & beyond. A steady diet of news horrors through out a lifetime is bound to have repercussions on parenting styles particularly women given neuroticism is a more common personality trait for the purposes of raising children. Our monkey brains weren’t exactly designed to have our fears pumped 24/7 with unlikely atrocities that must inbue an excessive sense of the world being a very dangerous place which not only potentially induces safety culture but generational anxiety. And with access to the internet their daughters have even more exposure to the news as well influences from alarmist activism. There really needs to be some solid research on the cultural forces on parents to fill in the blanks of safety culture.
If anything feminism promotes the opposite of safety culture. Freeing women to do anything, wear anything & f%#* anyone they want is hardly safety culture but of course it only becomes so when when women set the terms….


Wow, Ella we seem to almost not disagree for the first time! Although I quibble about ‘our monkey brains’. I think evolution gave most males the natural stoicism/stupidity to ignore dangers and the testosterone to, often foolishly, attack their dragons. See YouTube … After all, who cares about men? Evolutionary seen they are mostly disposable. Women’s higher evolutionary value requires them to be more careful.

Howeve, imho it’s the male role that is destroyed by too many mothers that were hyped up by feminism to believe that they were smarter and needed no mainsplaining about how to raise kids; men stupid enough to disagree were dealt with by a sisterhood of tens of thousands of journalists worldwide, whose sole task it is to ensure no man survives unscathed stepping out of line. See Damore and Strumia and the almost complete silence of men during the #metoo monologue.

We really have a big problem that is costing our kids dearly.

Great article. An unfolding disaster, as chronicled by the WSJ and Matt Stoller’s anti-monopoly Substack ( The whistle-blower who just testified doesn’t get the larger policy aspects at all. Repealing Section 230 and limiting this business model are not just imperative, they’re no-brainers. Facebook must be broken up. It should never have been allowed to buy Instagram and WhatsApp.

I’m waiting for Quillette to further connect more dots as part of the Instagram story … the connection with the explosive transgender and “transitioning” fad among adolescent girls of the last several years, as documented by Abigail Shrier and in contrast to most transsexualism (male > female).


I had this in my first draft, as well as the George Floyd riots, but decided to cut these points to keep the article focused on depression & anxiety. But yes. The social contagion aspect of Instagram is very very interesting.


That reminded me of clips I’ve seen of monkeys being hunted and killed by chimps and of how baboons form and move in a defensive troop, males always alert for threats.

In fact our monkey brains are exactly designed to be mindful of threats 24/7, it’s the reason we are so quick to take any possible threat seriously, and so quick to react with such caution, we evolved in a world that was a very dangerous place.

1 Like