Does Screen Time Actually Make You Miserable?

The most careful analysis of this found an association that is real, negative and tiny: digital technology use explained at most 0.4 per cent of the variation in wellbeing, across 355,358 people, all of them adolescents.

The original version of this article said the equivalent work on working adults was thinner. That is no longer true. A 2024 systematic review of screen time and mental health in adults screened 1,695 articles and included 32, and found that most reported associations between excessive screen exposure and mental health. The two bodies of work do not straightforwardly contradict each other, and understanding why is more useful than either.

What the adolescent study did differently

The problem with screen time research has never been a shortage of studies. It is that any large dataset can be analysed in thousands of defensible ways, and researchers pick one.

Orben and Przybylski’s 2019 paper in Nature Human Behaviour attacked that directly. Rather than choosing an analysis, they used specification curve analysis to run every plausible one and look at the distribution of results, across three datasets totalling 355,358 participants.

Their framing of the problem: these datasets have so many variables and observations that they are often explored with an analytical flexibility that marks small effects as statistically significant, leading to potential false positives and conflicting results.

Their finding: the association between digital technology use and adolescent wellbeing is negative but small, explaining at most 0.4 per cent of the variation. Their conclusion: these effects are too small to warrant policy change.

According to the University of Oxford’s announcement of the study, eating potatoes had nearly as negative an association with adolescent wellbeing as screen use, and wearing glasses had a more negative one. That is a scale rather than a joke.

The adult evidence that now exists

A 2024 systematic review in the Journal of Technology in Behavioral Science by Santos and colleagues set out to expand understanding of the associations between screen time and mental health in adults. Registered with PROSPERO, it searched three databases in January 2023, identified 1,695 articles, read 90 in full and included 32, of which 19 predated the pandemic and 13 were conducted during it.

Its findings: most studies found associations between excessive screen exposure and impact on aspects of mental health in adults. Before the pandemic, high screen times were associated with depression, anxiety, stress, burnout and lower wellbeing.

And one finding that matters specifically for anyone whose job is a screen: watching TV was worse for mental health than computer or mobile device use.

The review reports no specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Why these do not contradict each other

Two reasons, and they are the point of this article.

The first is method. The 2019 analysis ran every plausible specification and reported the distribution. The 2024 review is a narrative synthesis of 32 studies that each chose one analysis. That is precisely the analytical flexibility the 2019 authors identified as generating both alarming and reassuring results from the same data. A finding that most studies report associations tells you what most studies chose to report; it does not tell you how large those associations are or how many other defensible analyses of the same data would have found nothing.

The second is measurement. The 2019 paper’s contribution was a size, expressed as variance explained. The 2024 review reports the presence and direction of associations. You cannot read “most studies found associations” as contradicting “explains at most 0.4 per cent of variation”, because they answer different questions.

What the adult review does establish is the activity distinction. Watching TV was worse than computer or mobile use, which suggests what you do with a screen may matter more than how long you look at one. That is closer to the finding in the digital detoxes literature, where the benefit ran through what filled the reclaimed hours rather than through absence itself.

What to ask of any screen time claim

The 2019 research is more useful as a calibration tool than as a verdict.

When you next see a headline linking screen use to some outcome, ask three things. How much of the variation does it explain. How many other analyses of the same data would have given a different answer. And was the analysis chosen before or after the researchers saw the results.

On the first, the most rigorous answer available is a fraction of one per cent, in adolescents. On the second, the same authors found you could produce a wide range of positive or negative associations, or no effect at all, from the same data depending on choices. On the third, a companion paper by the same group notes that only a few studies on technology effects have used a preregistered confirmatory framework, and names measurement issues, lack of transparency, little confirmatory work and overinterpretation of miniscule effect sizes as problems in the field.

When this stops being about screens

A small average association is entirely compatible with some individuals being badly affected.

If your mood, sleep or ability to function has changed in a way that worries you, that is worth taking to a doctor regardless of what an average effect size says. Population findings are a poor guide to individual situations, and a very small average is exactly the kind of result that can coexist with real problems in particular people. If the concern is work-related exhaustion specifically, after-hours work messages may be closer to the question you are actually asking.

Common questions

So screen time is harmless?

Not what the research says. The 2019 specification curve analysis found the association between digital technology use and adolescent wellbeing was negative but small, explaining at most 0.4 per cent of the variation across 355,358 participants. Negative but small is different from absent, and the authors’ conclusion that the effects are too small to warrant policy change is a claim about policy rather than individuals. Separately, a 2024 systematic review of 32 studies in adults found most reported associations between excessive screen exposure and mental health, with high screen times associated before the pandemic with depression, anxiety, stress, burnout and lower wellbeing.

Does the 0.4 per cent figure apply to adults?

No, and that is the most common misuse of it. Every figure in the 2019 analysis comes from adolescent datasets with adolescent wellbeing measures. Adult evidence does now exist: a 2024 systematic review searched 1,695 articles, included 32 studies of adults, and found most reported associations with mental health. But it is a narrative synthesis rather than a specification curve analysis, so it reports the presence and direction of associations rather than how much variation they explain. The two are not directly comparable, and neither gives you a per cent figure for adults.

Do the adolescent and adult findings contradict each other?

Not directly, because they answer different questions. The 2019 analysis measured effect size, running every plausible specification and reporting the distribution, and found at most 0.4 per cent of variation explained. The 2024 adult review reports how many studies found associations and in which direction. A finding that most studies report associations tells you what those studies chose to report; the 2019 authors’ whole argument was that analytical flexibility in large datasets marks small effects as significant and produces conflicting results. So a narrative count of significant findings is the thing the specification curve method was designed to interrogate.

Does it matter what you are doing on the screen?

The adult review suggests yes, and this is its most practical finding. It reports that watching TV was worse for mental health than computer or mobile device use. That points at activity rather than duration as the variable, which is consistent with digital detox research finding that benefits ran through what filled the reclaimed time rather than through absence itself. Note that these are associations from mostly observational studies, so the direction of causation is not established, and people who watch more TV may differ from people who use computers in ways the studies did not capture.

The number is a ruler, not a verdict

The reason the 2019 study matters is not that it settled whether screens are bad. It is that it showed how much of the disagreement was manufactured by analytical choice rather than by data.

Run every reasonable analysis instead of one, and the alarming results and the reassuring results both turn out to be members of a distribution whose centre is a very small negative association. That is a finding about how a field works as much as about screens.

What this revision adds is that adult research now exists and mostly reports associations, which is what you would expect from a literature of individually chosen analyses, and which the 2019 method was built to look past. The useful thing the adult review contributes is not a verdict but a variable: television appeared worse than computer or mobile use, which points at what you do rather than how long you do it. That is the same shape as the finding on media multitasking, where the association is small and the interesting question is which activity.