It is a hot topic in the media, and I have written about it several times before: the relationship between social media and young people’s mental health. Mental health among young people has been deteriorating, and for some time now we have been looking for explanations (deliberately plural). Anyone looking for a simple answer to this issue will usually get disappointed. Yes, I’m looking at you, mister Haidt.

This summer, I wrote about a Swedish study that followed more than 3,000 young people for five years. The study found that the amount of time young people spent on social media did not predict their later mental health once the researchers accounted for differences already present at the start. My conclusion at the time: perhaps we are focusing too much on how much young people use social media and too little on what they do there.

A new study by Alva Tang and colleagues adds another piece to the puzzle. They do not show that social media causes depression after all. Instead, they approached the issue from a different angle. Perhaps we should not only look at how much young people use social media, but also at changes within the same young person. And perhaps there is a possible intermediate step between social media use and mental health problems? The researchers examined cognitive control as a possible candidate for such a step.

They use cognitive control as an umbrella term for mental processes that help us regulate our thoughts and behaviour. E.g.  inhibiting an automatic response, keeping your attention focused on something, or keeping relevant information active in working memory. These abilities continue to develop during adolescence, and previous research has already linked them to mental health. When reading this, I immediately thought of executive functions. This is hardly surprising given the considerable overlap between both concepts.

For their analyses, the authors used data from the large US ABCD study. Their analyses included 9,747 children who were followed for 4 years, starting around ages 9–10, with 3 measurement points. The young people themselves reported their social media use and overall screen time (yes, that is a limitation!). Parents reported depressive and anxiety symptoms (another limitation). Cognitive control, on the other hand, was not measured using a questionnaire but with two performance-based tasks (better).

Every study has limitations, but fortunately studies often have strengths too. One of the strengths here, besides the enormous amount of data, is that the researchers tried to make a clear distinction between two things: differences between young people on the one hand and changes within the same young person on the other.

Why is this relevant? Young people who use social media heavily may differ in many ways from peers who do not. That is a different question from what happens when a particular young person suddenly starts using social media more than they usually do.

So, what did the researchers find? When young people used social media more than usual at the first measurement, this predicted slightly poorer cognitive control two years later. And it was precisely this lower cognitive control that subsequently predicted slightly more depressive symptoms another two years later. But, as in much previous research, the associations were small!. In both cases, β was about -.06. But the indirect association was statistically significant. They also found a similar pattern for overall screen time.

If we try to capture this in a simple diagram, we get:

more social media/screen time than usual
-> slightly poorer cognitive control
-> slightly more depressive symptoms later

So the idea of a possible intermediate step does receive some empirical support here. A potential (negative) effect may not run directly from screen use to depression, but could perhaps operate through other developmental processes.

But we need to be careful. This study does not demonstrate that social media impairs cognitive control and thereby can lead to depression. The effects are, as said, small, and even with this kind of longitudinal research, you cannot rule out all alternative explanations. The authors themselves mention sleep and fatigue, for example, as possible factors related to both screen use and cognitive control. This is something I personally agree with.

There is another finding that nicely illustrates why we should remain careful. The researchers also checked what happens if you compare young people with each other, rather than looking at changes within the same person. In that case, young people who used more social media on average actually had better cognitive control. You read that right! Also, their average social media use also didn’t significantly correlate with depressive or anxiety symptoms.

This may seem like a paradox, but it does not have to be one. It mainly shows that “Who uses a lot of social media?” is a different question from “What happens when someone starts using more social media than usual?”

For girls, the researchers also found another pattern in overall screen time. More screen time predicted slightly more depressive symptoms two years later. Those depressive symptoms, in turn, predicted both more screen time and more anxiety two years after that.

Earlier research suggested that it may matter when social media are used during adolescence. Other research reminded us that social media are only one part of a much bigger picture. This study adds another question: through which mechanisms might any effects emerge? Cognitive control could be one piece of that puzzle, but much more research is needed.

The more we learn about this topic, the harder it becomes, in any case, to reduce the story to: “social media make young people depressed.”