Since the early 2010s, the curves of anxiety, depression, and self-harm among young people have sharply risen, especially among girls. Coincidence? The data says otherwise. Here are the specific mechanisms, study by study, and the concrete levers to take action.
Teenage anxiety did not emerge yesterday. Since the 1950s, rates of anxiety and depressive disorders among young people have fluctuated, with ups and downs, without anything spectacular. Then, in the early 2010s, the curve abruptly changes shape: it rises sharply, like a hockey stick.
This shift is reflected in several indicators at once, which is rarely the case with simple diagnostic fashion trends. Rates of depression among teenage girls soar, self-harm increases, and psychiatric hospitalizations rise as well. Around 2013, in the United States, the United Kingdom, and other English-speaking countries, psychiatric services begin to receive a disproportionate number of young girls.
What happened in the meantime? One massive, global, and very dated thing: the transition from basic mobile phones to smartphones, with the internet in your pocket, all the time. This shift has been accompanied by four underlying damages, on which everything that follows rests: social deprivation (less real-time interaction with others), lack of sleep, attention fragmentation, and addiction.
To understand what follows, a point of comparison prior to social networks is needed. There exists a long-term study conducted in Framingham, Massachusetts, initially focused on cardiovascular health. Sociologist Nicholas Christakis and political scientist James Fowler identified data that allowed them to track the flow of emotions within a group, year after year.
Unexpected result: happiness snowballs. When one person becomes happier, the likelihood that their friends will also become happier increases, and the effect sometimes extends even to friends of friends, or even further. Emotions do not remain confined to the mind of the person experiencing them; they travel along social ties.
By analyzing the same data with psychiatrist James Rosenquist, they tested negative emotions. Two striking observations emerged. First, negativity proves stronger than positivity: depression is significantly more contagious than good mental health. Secondly, it only spread among women: when a woman began to feel depressed, the probability that her close friends, both male and female, would follow suit increased by 142%. When a man became depressed, there was no measurable effect on those around him.
The explanation put forward by the researchers is simple: women express their emotions more and communicate their feelings more effectively to those close to them. When men get together, it is more often to do things together than to talk about how they feel. A noteworthy detail: this study ended in 2001, just before the advent of social networks.
Now place the two pieces of information side by side. Depression is more contagious than happiness, and it primarily spreads through people who share their emotions the most. What was supposed to happen starting in 2010, when entire communities of teenage girls became much more interconnected than the adults of Framingham?
A sudden explosion of anxiety and depressive disorders was logically to be expected. And that’s exactly what happened around 2012, when a large number of young girls signed up for Instagram, Snapchat, Tumblr, and other sharing platforms. Instagram surpassed 50 million users in 2012, becoming the default venue for adolescent social life.
Another dated marker: the percentage of American high school students reporting "accepting themselves as they are" began to decline from 2012. In other words, it’s not just mental health that is deteriorating; it’s the way these young people view themselves. By 2015, Snapchat introduced its filters that, with a click, plump lips, slim noses, and reshape eyes.
Psychologically, girls and boys are very similar. However, some differences appear in most cultures, and one of them sheds light on the entire subject: the distinction between agency and the need for communion.
Agency is based on the desire to stand out, assert oneself, and make an impact: effectiveness, competence, self-assertion. Communion is based on the desire to belong to a larger social group by caring for others: kindness, cooperation, empathy. Both motivations drive us all and intertwine, but research has long shown that boys and men lean more towards agency, while girls and women lean towards communion.
In adolescence, these two needs become central: we build our identity both by integrating into groups and by showing those groups what we are worth individually. Platforms promise easy connection, so they seem to meet the need for communion... and disappoint it.
A direct consequence on our topic: girls are more willing to share their emotions and difficulties. When social life moves online and circles expand to hundreds of people, those who suffer from anxiety or depression can influence a very large number of other users, leading them to feel and express the same symptoms.
The first engine of propagation is called emotional contagion. We are permeable to the emotions of others, without even deciding it: the mood of a friend changes ours, and that of a friend's friend can reach us by ricochet. This is a banal mechanism, observable offline for as long as we can remember.
What has changed is not the mechanism, but the channel. In the time of Framingham, contagion occurred through direct conversations, in small numbers. Today, a teenager is connected to hundreds of accounts, continuously seeing the moods, crises, bad news, and distress of others, and she shares her own in return. The network multiplies the speed and reach.
Add to this the already mentioned principle, very solid in psychology: negativity shows itself to be stronger than positivity. Anxiety and depression spread better than calm and balance. A news feed is therefore not a neutral environment where good and bad moods would compensate for each other; it is an environment that structurally favors the diffusion of the negative.
The second engine is more counterintuitive. Humans learn by imitation, but not from just anyone: we first identify who has the highest prestige in the group, and that is whom we copy. This is what researchers call the prestige bias, a very old and effective rule of social learning... and very easy to subvert.
On social networks, however, to amass followers and "likes," one must be more outrageous than the average. The person who displays the most impressive symptoms gains visibility faster, thus gaining prestige, and becomes a model to imitate. This phenomenon has a name: "audience capture." The individual is gradually conditioned by their audience to become an increasingly extreme version of what that audience wants to see. And when most members of a circle have adopted the same behavior, a third mechanism kicks in, the conformity bias.
Algorithms accelerate everything. Researchers from the Center for Countering Digital Hate opened dozens of fake TikTok accounts in the names of 13-year-old girls: within just a few weeks, the platform inundated them with tens of thousands of weight loss videos, many of which showed emaciated young women promoting extreme diets, from the "corpse bride diet" to the water diet.
One documented case summarizes the mechanics. An American teenager opened her first account at age 11, lying about her age, and initially focused on fitness. Within six months, the content pushed by the algorithm shifted from fitness to photos of models, then to diet recommendations, and finally to posts glorifying anorexia. Two years later, she was hospitalized for anorexia and depression.
A word to explain here, as it is at the heart of the subject: "sociogenic" means "generated by social forces," as opposed to a biological cause. In 1997, researcher Leslie Boss reviewed the historical and medical literature dedicated to these sociogenic epidemics and distinguished two recurring forms. An "anxious variant," characterized by abdominal pain, headaches, dizziness, fainting, nausea, and hyperventilation. And a "motor variant," which manifests as hysterical dances, convulsions, nervous laughter, and pseudo-epileptic seizures, like those "dancing manias" in medieval European villages where people danced until exhaustion.
In each modern episode, medical authorities found no toxins or pollution to explain the symptoms. Instead, two regularities emerged: it is girls who are most affected, and these epidemics often break out after a distressing event or a threat to the community. Boss had even predicted, at the dawn of the Internet, that new mass communication tools would increase the amplification capacity of these epidemics.
She was visionary. A specific example: German psychiatrists, led by Kirsten Müller-Vahl, saw an influx of young people claiming to suffer from Tourette syndrome, a neurological disorder characterized by motor and vocal tics that typically appears between ages 5 and 10 and affects 80% of boys. However, none of these new patients actually had it: no childhood history, different tics, and above all, strangely identical symptoms from one patient to another. They were mimicking the tics of a German influencer who genuinely had the condition, whose videos were going viral on YouTube. Researchers proposed calling this "mass social media-induced illness."
The phenomenon crossed gender boundaries as soon as it became popular online: English-speaking teenage girls began shaking their heads and shouting the word "beans," mimicking the tics of a British influencer. The first advice from doctors in these cases: stop using social media. The lockdowns of 2020 amplified the situation, adding a global threat and extra hours spent on TikTok. By August 2023, videos tagged #mentalhealth had accumulated over 100 billion views, and #trauma over 25 billion.
Time matters, and it matters a lot. Among American high school students, the proportion of those reporting spending more than 40 hours a week on social media—equivalent to a full-time job in addition to classes—has been measured since 2013. By 2015, in the United States, one in seven girls had reached this figure. Before 2010, when very few teenagers owned a smartphone, such a volume was practically impossible.
A large British survey, which followed nearly 19,000 children born around 2000 until the end of adolescence, links this time to depression measured on a thirteen-point scale. For boys, knowing their screen time doesn’t reveal much: the curve only begins to rise beyond two hours a day. For girls, the relationship is clear and consistent: as screen time increases, the risk of depression rises. Those who report five hours or more per day are three times more likely to suffer from depression than those who do not use it at all.
In practice, what do these hours look like? Among active girls aged 11 to 15, average daily usage is around 2 hours and 39 minutes on TikTok, 2 hours and 23 minutes on YouTube, 2 hours on Snapchat, and 1 hour and 32 minutes on Instagram. These times accumulate, encroaching on sleep, homework, and moments spent with family and friends. A useful benchmark for parents: it’s not just the content that poses a problem; it’s the volume.
In the early 2010s, boys and girls both spent more time online, but not in the same places. Boys turned to YouTube videos, text-based platforms like Reddit, and especially multiplayer online games. Girls were drawn to visual platforms: Instagram first, then Snapchat, Pinterest, and Tumblr.
This difference is significant, as not all platforms are equal. A British survey from 2017 asked teenagers to assess the impact of major platforms on their anxiety, loneliness, self-image, and sleep. Instagram came last among the five platforms evaluated, followed by Snapchat, while YouTube was the only one to receive an overall positive rating.
The reason is structural. These visual platforms follow the same business model: maximizing user time to extract as much data as possible and increase their value to advertisers. An internal study commissioned by Facebook, which remained unpublished and was later revealed by whistleblower Frances Haugen, stated bluntly: the teenagers surveyed spontaneously attributed the rise in their anxiety and depression rates to Instagram, and this reaction was consistent across all groups. Social comparison was "worse" there than on competing apps, with Snapchat's filters focusing on the face, while Instagram heavily emphasizes body image and lifestyle.
In total: girls spend more time on social networks, and the networks they prefer are the most harmful to mental health. Even with identical psychological profiles, one would therefore expect a greater increase in anxiety among them.
The classic objection is legitimate: correlation proves nothing. Perhaps depression drives girls to social networks, not the other way around. Maybe a third variable, such as genetics, a very permissive upbringing, or loneliness, produces both at once. To settle this, scientists use randomized controlled trials (RCTs): they randomly assign who receives a "treatment" and who belongs to the control group, instructed not to change their habits.
In this field, the treatment often involves asking young people to reduce or stop using social networks for a few days or weeks. One study did exactly this with students: after three weeks, the experimental group showed significantly lower levels of loneliness and depression than the control group. Another study randomly exposed teenage girls to Instagram selfies, either original or edited to be more attractive: exposure to manipulated photos directly led to a depreciation of their own bodies. Across dozens of gathered experiments, the conclusion converges: the use of social media is not just a simple correlate; it is a cause of anxiety and depression.
"Quasi-experiments" further reinforce the picture by taking advantage of gradual technological rollouts. Facebook was initially opened to only a small number of universities: institutions that gained access saw the mental health of their students decline, with an increase in depressive symptoms, greater reliance on psychiatric care, and repercussions on academic performance, the effect being more pronounced among women. The mechanism identified by the authors: the platform fosters unfavorable social comparisons.
The same logic applies with the arrival of high-speed Internet, region by region. Of five studies conducted worldwide, all five show a negative impact on mental health. In Spain, a 2022 study examined hospitalizations of adolescents for mental and behavioral disorders during the rollout of fiber optics: a significant net effect among girls, absent among boys, with an increase in addictive Internet use and a significant decrease in sleep, homework, and social life with family and friends. Finally, a word on the scale: when we stop mixing all digital activities and all adolescents, and isolate social networks among girls, the correlation is no longer anecdotal. We move from the level of "eating potatoes" to something closer to binge drinking or cannabis consumption.
There remains a major limitation to all these experiments: they measure the effect on individuals taken in isolation, as one would test sugar. However, social networks are not sugar. They do not only affect those who consume them; they reconfigure the entire group.
When smartphones entered middle and high schools in the early 2010s, school culture changed. Students began to talk less during class breaks, recess, and lunch, absorbed by micro-dramas unfolding all day on their screens. They looked at each other less, laughed together less, and lost the habit of conversing. In other words, social life deteriorated even for students who did not engage with these platforms. This abrupt transformation of group dynamics can be called the Great Rewiring of Childhood.
This explains why researchers who deem the effects "too weak to explain such increases" are mistaken in their scale: they are looking at the individual, not the group. Moreover, the mechanics reverse: a teenager who disconnects alone while all their friends remain online becomes more isolated, and despite this handicap, several studies show that their mental health still improves. Imagine then the effect if twenty entire schools put their phones in a locker every morning.
The last piece of the puzzle, and not the least: the quantity of connections has exploded, while the quality has collapsed. By dividing their hours and attention among hundreds of contacts, teenagers invest less in each relationship, and digital substitutes reduce the desire to socialize without meeting emotional needs. Clinical psychologist Lisa Damour summarizes it well for friendships among girls: quality trumps quantity. The happiest are not those with the most friends, but those who have forged strong and supportive friendships, even if there is only one.
Understanding the causal link is only meaningful if it changes something. First lever: timing. The age at which a child receives their first smartphone impacts their mental health in adulthood, and the earlier the introduction (elementary school, middle school), the worse the outcome. Delaying access, reserving smartphones for high school entry, and banning screens in bedrooms while keeping a shared computer in a common area are simple measures. However, be aware that these can be circumvented: children may lie about their age during registration, download the app and then delete it, or hide it behind a calculator icon. Control alone is not enough; it is essential to explain and talk.
Second lever: the duration of a break if you attempt to wean off. A few days are almost useless; the brain needs at least three weeks to reset, and it is from this timeframe that studies observe a decrease in loneliness and depressive symptoms. In cases of tics that suddenly appear after immersion in video content, the first medical recommendation is clear: cut off social media.
Third lever: act in groups rather than alone. An adolescent deprived of platforms while their entire class remains on them is socially penalized. A collective rule, at the level of a class, a school (phones in lockers in the morning), or a group of parents, has a much greater effect than an isolated ban. And always prioritize what restores real life: outings, activities, long conversations with one or two close friends.
Final point, often misunderstood. Repeating to a teenage girl that "everything is edited, this is not real life" has a limited effect. French researchers exposed young women to images of very thin or average-sized women for barely twenty milliseconds, a time too brief for conscious awareness: those who saw the slender silhouettes became more critical of their own bodies. Social comparison operates without our consent, in a part of the brain that reasoning does not govern. That is why reducing exposure matters more than multiplying warnings.
Author Audrey on 01 October 2024
Health and Beauty : Establishment
Author Nelly Daubord on 05 August 2024
Health and Beauty : Care and Beauty