Is Social Media Actually Bad for Your Wellbeing?

In 2019, two researchers at Oxford set out to answer a question that had already produced hundreds of contradictory headlines: does using social media damage teenagers' wellbeing? Instead of running one more study with one more set of choices, Amy Orben and Andrew Przybylski did something unusual. Using large datasets covering tens of thousands of adolescents in the UK and the US, they ran the same basic question through thousands of different combinations of reasonable analytical decisions - different ways of measuring technology use, different ways of measuring wellbeing, different sets of statistical controls.

This approach has a name: specification curve analysis. It exists because a single published study normally reports only the one analysis its authors happened to choose, which hides how much the conclusion depends on unremarkable decisions made along the way. If a result only shows up under one specific combination of choices out of thousands of equally defensible ones, that result is fragile. If it shows up almost everywhere, it is probably real, whatever its size.

A real effect, and a small one

The result, published in Nature Human Behaviour, was not a clean yes or no. Across the overwhelming majority of specifications, there was a statistically real association between digital technology use and adolescent wellbeing - but a tiny one. Technology use accounted for a sliver of the differences in how teenagers reported feeling - at most 0.4% of the variation in wellbeing, by the authors' own estimate. An effect that small is easy to detect as "statistically significant" in a dataset with tens of thousands of people without being practically meaningful, which is exactly the gap that this kind of large-scale analysis is designed to expose.

Orben followed up with a colleague, Tobias Dienlin, using longitudinal data published in the Proceedings of the National Academy of Sciences - tracking the same teenagers across years rather than a single survey moment. The pattern held up over time: whatever effect existed did not swing wildly depending on when researchers looked, but it also never grew into anything large. That is a more boring conclusion than most headlines about phones and mental health, and it is probably closer to accurate.

Scrolling versus talking

Zooming out from teenagers specifically, one of the more useful ideas in this research is not "social media: yes or no" but how you use it. In 2015, Philippe Verduyn and colleagues combined an experience-sampling study - texting people at random points during their day to ask what they were doing and how they felt - with a controlled experiment. People who passively browsed a social network, scrolling through other people's updates without interacting, reported feeling worse afterward. People using the same platform actively - messaging friends, commenting, posting - did not show that drop.

The proposed mechanism is not exotic: passive scrolling invites upward comparison (other people's curated highlights against your ordinary Tuesday) without the offsetting warmth of an actual conversation. Later work from the same research group frames this as an active-passive distinction rather than a single verdict on an entire category of apps, which is a messier message than either "social media is destroying a generation" or "it's all fine" - and probably the more honest one.

A very loud disagreement

None of this has made the public debate quieter. In his 2024 book The Anxious Generation, social psychologist Jonathan Haidt argues that smartphones and social media are a central cause of a well-documented rise in teenage anxiety and depression, particularly among girls, roughly coinciding with the spread of smartphones in the early 2010s. In a 2024 commentary in Nature titled "The great rewiring," developmental psychologist Candice Odgers pushed back, arguing that the evidence Haidt leans on is largely correlational, that specification-curve and longitudinal work like Orben's keeps finding effects too small to explain something described as an epidemic, and that other candidates - sleep loss, economic pressure on families, changes in how mental health difficulties are measured and reported - deserve just as much attention.

Neither side denies that adolescent mental health changed over the past fifteen years, and neither denies that phones changed too. The disagreement is about how much of the first trend the second one actually explains, and that is precisely the kind of question that two rising lines on a graph cannot settle by themselves.

What this means for you

If you use social media and want something more useful than a debate about teenagers in general, the research suggests asking not "how many minutes" but "what am I doing during those minutes, and how do I feel afterward." Scrolling passively before falling asleep is a different behaviour from messaging a group of people you actually know, even though a screen-time counter cannot tell the two apart. If you want to look at your own pattern rather than take a stranger's word for it, our phone habit test and social media type test are built around exactly this distinction, and our FOMO test looks at what tends to pull people back to the app in the first place.

The honest summary is unglamorous. For most people, social media is neither a spotless tool nor a mental health emergency on its own. It is one input among many, its measured effect is smaller than most headlines suggest, and how you use it seems to matter more than whether you use it at all. If you notice it consistently and clearly worsening your mood, sleep, or ability to concentrate, that pattern is worth taking seriously - and if it persists, it is worth raising with a professional rather than settling it from an article.