Depression, anxiety and suicidality have all sharply increased in adolescents over the past decade1. So, too, has the amount of time that young people spend online (see ‘Troubling trends’). Partly because of fears that there’s a link between these trends, governments around the world are under pressure to do more to regulate technology companies.
In the United Kingdom, the Online Safety Bill, currently being debated by Parliament, seeks to protect children from harmful content online. Last year, the European Union approved the Digital Services Act — which, among other things, has introduced tougher mandates requiring companies to remove illegal content from their websites. And in 2021, the US surgeon-general called for social-media companies to prioritize adolescent health and well-being at “all stages of product development”1.
A difficulty facing policymakers, however, is that most of the scientific evidence on the impact of social media and other online activities on adolescent mental health is inconsistent2. Some studies might report similar effects, such as small negative correlations between time spent on social media and measures of well-being, but researchers differ in how important they think such findings are3.
There might be many reasons why psychologists, psychiatrists, computational scientists and others have failed to obtain a clearer picture of what is going on2. Many have called for more detailed, objective assessments of what activities users engage in during their time online — an issue being addressed in part by smartphone apps that track the amount of time people spend on certain platforms4. Others say that what makes any one person vulnerable to the negative impacts of social media needs to be better understood5.
All these arguments make sense. But in our view, there is another gap: researchers have not systematically interrogated, using large-scale data sets or experiments, how the relationship between social-media use and mental health changes with developmental stage.
Developmental stage matters
To explore this, in 2020 and 2021, we analysed longitudinal data from two UK data sets6. The data had been collected every year between 2011 and 2018: 17,409 participants had been asked about their social-media use and life satisfaction, either in interviews or in online questionnaires, once per year for up to seven years. At the time of the first surveys, the participants’ ages had ranged from 10 to 21.
To establish how social-media use and level of life satisfaction relate to each other over time, we looked for a connection between participants’ estimates of their time spent on social media at ages 10, 11, 12 and so on until age 20, and the level of life satisfaction that they reported a year later. In other words, we used age as a proxy for developmental stage.
The effect sizes were small, but we found that social-media use (the participants’ estimates of how much time they spent each weekday, on average, interacting with friends through a social website or app) did predict levels of life satisfaction a year later — but this was true only for participants at certain developmental stages.
Time for the Human Screenome Project
In participants who self-reported as female, increased social-media use (meaning an increase in a participant’s use compared with their own average across the data-collection period) at 11, 12 or 13 years old predicted decreased life satisfaction (compared with their average) a year later. The same pattern occurred in participants who self-reported as male when they were 14 or 15 years old. These age ranges align with when young people go through puberty; on average, girls enter puberty earlier than boys7.
For both female and male participants, increased social-media use at 19 years old — which is soon after most young people first leave home and gain independence — again predicted lower levels of life satisfaction a year later.
There are several reasons to be cautious when interpreting these results. First, the study needs to be replicated, and more sophisticated measures used to track both social-media use and what developmental stage a person is at. Also, this study and others indicate that social media’s effects on mental health vary substantially between individuals8. Finally, the impact of social media seems to be bidirectional and complex: for male and female participants of all ages, reporting a life-satisfaction level in any one year that was lower than their personal average predicted a small increase in social-media use a year later.
Still, the idea that people’s sensitivity to online social environments might be linked to certain developmental changes fits with what we know about adolescence from neurocognitive studies and other research.
A time of change
Early adolescence is characterized by wide-reaching hormonal changes, as well as physiological changes throughout the body. At the same time, all sorts of neural, cognitive and social shifts are happening. These changes could make social-media environments, such as those provided by Snapchat or TikTok, particularly alluring, but also especially impactful on mental health9–11.
Several developmental-psychology studies have shown, for instance, that adolescents — particularly those in early to mid-adolescence — place increased importance on being able to interact with their peers, and on what their peers think of them12. Other studies suggest that although young children tend to view themselves positively, as they become adolescents, their ideas about themselves come to more closely align with what they perceive others to think of them13,14. Still more work has shown that being rejected or not being included has a greater impact on mood for those in early to mid-adolescence than for people older than 2515.
Social media provides new ways for adolescents to quantify social approval — for instance, through the number of ‘likes’ that they receive after posting something online, or by how many seconds, minutes or hours they have to wait before receiving feedback. For some people, being able to constantly track feedback from peers could heighten anxieties about self-worth or amplify the impact of judgements from peers. Some researchers have proposed that digital innovations, such as games or social-media platforms including TikTok, might even impact adolescents’ development of a sense of self, how they perceive others’ opinions of them or what habits they develop around social-media use14,16,17.
Adolescents’ social environments tend to undergo particularly dramatic changes at certain stages, too.
In the United Kingdom and elsewhere, most young people move from primary schools to secondary schools, where social networks are larger and more unstable, when they are 1118. Being able to access peers through communication pathways on social media — many of which are eminently available, highly public, permanent, asynchronous and lacking social cues, such as facial expressions and body language — could be especially impactful at this time9. The same is true when young people leave school for work or college. In both early and later adolescence, young people become more independent, and social pressures might increase.
Plugging the gap
We urge more psychologists, psychiatrists and other behavioural-science researchers to examine the effects of social-media use at specific developmental stages. In many studies, the effects of using social media are averaged over a broad age range. This means that potential fluctuations in the impacts of social media as adolescents age might be missed. Some investigators have collected data from smaller age ranges, or from participants who are all the same age, but they often generalize their results to the entire adolescent age range.
Smartphones are bad for some teens, not all
To try to pinpoint how exactly developmental stage affects how young people interact with and experience online spaces, researchers could use methodologies from fields such as developmental neuroscience or developmental psychology. In some psychology studies, for instance, researchers give participants feedback from fictional participants, and assess changes to people’s sense of self-worth over time19. Other studies measure whether participants’ opinions or actions (such as how much they donate to charity) are influenced by their peers’ opinions or actions, or how much participants’ self-judgements are affected by how they perceive others to judge them. In principle, such experimental paradigms could be used in combination with digital data sources to investigate how the use of social media and development interact to affect the fragility of young people’s sense of self and their self-worth.
Likewise, measures of pubertal hormones collected through saliva samples or estimates of puberty stage in self-reported questionnaires in major data-collection efforts, such as the Adolescent Brain Cognitive Development (ABCD) study, could be used to establish whether people’s age or pubertal stage is most important when it comes to determining sensitivity to social media. (The ABCD study is the largest long-term study of brain development and child health in the United States.)
So far, investigating the role of online environments in the rise in mental-health problems in young people has been challenging and frustrating — not least because social-media companies are generally reluctant to share their data with researchers. But mental-health difficulties experienced in adolescence can affect someone through every life stage: around 48% of people with a mental-health disorder, such as depression, first experience symptoms before they are 18 years old20. What’s more, new digital technologies will keep emerging.
It is therefore crucial that psychologists, neuroscientists and other researchers keep refining their approaches to better understand what kind of online experiences prime some young people for depression, anxiety, self-harm and a suite of other mental-health problems.
A.O. is an unpaid member of governmental and non-governmental organizations (including the UK Department for Culture, Media and Sport, and The British Academy). She has also provided unpaid and paid talks and consultancy work to organizations that will not gain or lose financially from this publication. S.-J.B. currently receives funding from Wellcome, the MRC, the Jacobs Foundation, the Wellspring Foundation and the University of Cambridge. In the past five years, S.-J.B. engaged in a paid consultancy with Cognita International Schools Group and has provided paid expert witness work for UK charities, legal organisations in the UK and USA and the UK government. S.-J.B. is the author of two books related to the brain, education and learning, for which she received an advance and royalties. S.-J.B. gives talks in schools, in the state and private sector, as well as at education conferences and for education organizations, and other public, private and third sector organizations (some talks are remunerated). S.-J.B. is a member of the Rethinking Assessment group, the Steering Committee of the Cambridge Centre of Science Policy, the Technical Advisory Group for the UK Government Department of Education ‘Education and Outcomes Panel-C Study’ and the Singapore Government National Research Foundation Scientific Advisory Board. She was a member of the Times Education Commission in 2021-22.