Born in Bradford participants spend hours each day on social media – just like teenagers everywhere – and the heaviest users tend to be less happy. Scientists have focused on the debate about correlation vs. causation, and whether the data suggest social media is bad for our mental health. While that debate is important, it is distracting us from a bigger question: no-one thinks all this screen time is a good idea, so what should be done about it?
Born In Bradford is a major research study following children and families across our city. As the children enter their teenage years, they’ve started using apps like TikTok and Snapchat, and we are fascinated to understand their experiences. Our latest data, published in BMC Public Health, shows a link between social media and mental health.
The chart below is from Born In Bradford surveys with 12-15 year-olds. It shows that teenagers who report more time using social media apps tend to have more symptoms of depression and anxiety.
Distribution of RCADS25 (depression and anxiety) scores for 12-15-years-olds in Bradford, stratified by self-reported daily social media use
Does this mean social media is harmful for teenagers?
Most teenagers are happy and well adjusted, and it is important to remember that the average teenager probably does not experience a great deal of harm from social media apps. Many enjoy or even benefit from certain aspects of social media. In our study, more young people reported positive experiences than negative experiences. Nonetheless the overall impact on society might be incredibly important and change the way we make friends and understand the world.
Our study is one of many that have tried to measure these impacts. They mostly show the same thing: people who are on social media a lot tend to have worse mental health.
By itself, this does not mean that social media is harmful. If you are interested in the effects of social media on young people, the fact screen time and mental health are correlated is clearly relevant. We’d probably be more relaxed if we found that heavy use was associated with being happy! But we have to be aware that the association might have arisen for other reasons. It could be due to reverse causality in which young people use social media due to their poor mental health. Or it could be residual confounding, in which the association could be explained by another variable. For example, teenagers with strong family support may use social media less and also be happier. In this case, reducing screen time might not improve mental health because the real driver is family connection rather than the apps themselves.
In our study teenagers reported a lot of screen time. On average they said they use social media apps for three and a half hours per day, with a quarter saying they use these apps for five hours per day or more. Even 12-year-olds reported a lot of social media use, and social media is not supposed to be available for under-13s in the UK. What other things feature this prominently in young people’s lives? Based on our participants’ reported weekday and weekend social media use, we made a rough estimate that an average teenager might spend 1,200 hours per year on social media; the same amount of time a student with full attendance spends at school.
To continue with this comparison, would we try to estimate an average effect of school on young people’s mental health? This might be difficult, but we would all accept that school is an important determinant of young people’s mental health, and in some cases harm probably occurs. We would probably also agree that schools are a good place for interventions to improve teenagers’ mental health. Do we feel social media should have a similar level of prominence in teenagers’ lives as school? Those are questions for society, not just scientists.
What’s next for research in this field?
Some people argue that we need more sophisticated research methods, particularly longitudinal studies that follow young people over time, rather than simpler cross-sectional studies like ours that measure social media and mental health simultaneously. We do need a variety of research designs to understand the problem from different angles, but more sophisticated methods won’t provide all the answers, for two reasons.
First, more sophisticated studies that follow participants over time can mitigate reverse causality, but focus on effects that happen over a specific duration, rather than providing an answer that is necessarily more authoritative. Imagine a teenager who was up all night on social media just before completing a survey. Their response might show high social media use and poor mental health, reflecting an almost immediate harm. A longitudinal method comparing the social media use reported in the survey to the subsequent change in mental health would miss this harm, even if it persists forever. The smaller effects typically found in longitudinal studies might just reflect the fact they focus on a specific lagged effect rather than shorter term mechanisms or harms that have already accumulated.
More convincing evidence could come from natural experiments, where people have different exposure to social media for reasons beyond their control. The most famous is the gradual roll-out of Facebook across US colleges in the mid-2000s, which was associated with a decline in mental health scores in student surveys. These natural experiments are hard to find, usually involving a roll-out of a policy across geographical areas, which is not how government policies in this field are likely to work.
Second, the average effect on the population is not really what we need. Even if we could prove that social media does not, on average, harm the mental health of teenagers, we would likely want things to change. We have a consensus among parents and teenagers that we all spend too much time on our phones, and a lot of that time is on social media apps. In our work developing The IRL Trial, teenagers in Bradford described social media as “addictive”. One said “I’ll get home after school and instantly, I’ll want to go on my phone, and then the next thing I know I’ve been doomscrolling for two hours when I could have been doing something else”. Another said “people are just on their phones a lot more which I think is not great ‘cause you miss out on quite a lot of social interaction. And like if you want to meet new people in real life it’s quite hard because people do seem to just be on social media in public quite a lot”.
Despite this emerging consensus, it’s unclear what policies or interventions might help.
This is where researchers can be more helpful: contributing evidence that could support policy to make our online lives safer and healthier. We have been amazed by the excitement about our plans for the The IRL Trial, which is an attempt to do this. Later this year we are going to try reducing the amount of time young Bradfordians aged 12-15 spend on social media. School year groups will be randomised to two conditions: either they continue as normal, or they allow a research app to restrict their social media use to a maximum of one hour between 7am and 9pm. We will measure the impact of this restriction on symptoms of anxiety. It only addresses one narrow question about our specific intervention, but we think people are excited because it’s trying to show what happens when you change something, rather than just describing a problem.
