How Much Screen Time Is Too Much? The Honest Answer Has No Number In It

Short answer: for adults, no evidence-based threshold exists. None. Not two hours, not four. The only screen time limits published by a health authority are for children under five, and they were never intended to apply to you. When researchers went looking for the harm in the largest datasets available, they found an association so small that it sat alongside eating potatoes.

That is not a claim that screens are harmless. It is a claim that the hours number is the wrong instrument, and this article is about what to use instead.

The study that reset the question

Orben and Przybylski took three large datasets, applied specification curve analysis, and reported every defensible result rather than the one that made the best headline.

Then they did the thing that made the paper famous: they benchmarked that effect against other variables in the same datasets, so the size could be understood rather than merely reported.

The association we find between digital technology use and adolescent well-being is negative but small, explaining at most 0.4% of the variation in well-being.

Orben and Przybylski, Nature Human Behaviour, 2019

The potatoes line gets quoted for the joke, but the glasses line is the more useful one. Nobody proposes screen limits on the basis of an association that size when it attaches to spectacles.

Three honest caveats, because this study gets over-claimed in the other direction too. It is about adolescents, not adults. It is correlational. And "small on average across a population" is compatible with "large for some individuals" — which is exactly why an average threshold was never going to be the right tool.

Where the numbers people quote come from

The WHO guidance is for under-fives. No sedentary screen time for children under one; no more than one hour for ages two to four, less being better. That is the actual published limit, it concerns early childhood development, and it is the ultimate source of most "experts recommend" claims applied to adults.

The two-hour rule was a paediatric recommendation for older children, which the American Academy of Pediatrics itself moved away from in favour of a family-media-plan approach, precisely because a single number did not survive contact with how devices are actually used.

"The average person spends 7 hours a day on screens" is a market-research figure, usually covering all screens including work. It describes what people do. It has never described what is healthy, and a description is not a threshold.

What to measure instead of hours

If the total is uninformative, something has to replace it. Three things carry more signal, and all three are visible in the same settings screen.

1. Pickups. How many separate times you unlocked the device. This measures fragmentation, which is the mechanism with actual experimental support behind it. Forty pickups spread through a working day damages sustained attention in a way that a two-hour evening film does not, and the hours total cannot tell those apart.

2. First and last use of the day. Screen use in the hour before sleep and the first ten minutes after waking has outsized effects on sleep timing and on how the day is framed. Two specific moments beat a daily aggregate.

3. Chosen versus delivered. Did you open the app, or did a notification open it for you? This is the distinction between using a device and being used by one, and it is the only one of the three that tells you something about agency.

There is also a fourth thing worth knowing that does not appear in any dashboard. Ward and colleagues found that the mere physical presence of your own smartphone reduced available cognitive capacity, and that participants did not notice it happening. Zero minutes of screen time, phone face-down on the desk, cost still paid. No hours-based metric can capture that, which is another reason the hours are not the variable.

So when is it actually too much

The threshold is functional, not numerical. Screen time is too much when it is displacing something, and the displacement is measurable where the hours are not.

That last distinction does most of the work. An hour of video-calling your family, reading, or navigating a new city is not the same hour as an hour of fragmented scrolling, and every metric that adds them together is discarding the only information that mattered.

If your answer to the first two questions is yes, you have found your limit, and it is personal rather than published. The practical version of acting on it is in how to reduce screen time, and if the specific worry is your eyes rather than your wellbeing, the dark mode evidence covers what actually causes screen strain.

Children and adults are genuinely different cases

Collapsing these is the most common error in this topic, in both directions. The evidence is not the same and neither is the reasoning.

For young children, the concern is displacement during a developmental window: screen time competing with the sleep, movement and face-to-face interaction that early development depends on. The WHO guidance follows from that logic, it is specific, and it is not controversial.

For adolescents, we have the large datasets above, and they show an association so small it sits between potatoes and spectacles. The honest reading is not "screens are fine for teenagers" but "total hours is a poor instrument for finding whatever effect exists."

For adults, there is no threshold research at all, because nobody has designed a study that could produce one. What exists concerns specific mechanisms — evening light and sleep timing, notification-driven fragmentation, posture and dry eye from sustained near work — and each has its own answer that is not a number of hours.

What none of the research can measure

Three limitations, worth knowing before you weigh any study in this area including the one above.

Self-reported screen time is unreliable. Most of the large datasets asked people to estimate their own use. Studies comparing estimates against logged device data find the estimates are poor, and poorly measured variables produce small associations almost by construction. Some of that 0.4% may be measurement error rather than a genuinely tiny effect.

Passive and active use get counted identically. A video call with a grandparent and an hour of algorithmic scrolling are the same minute on a dashboard. If those have opposite effects, averaging them across a population produces approximately zero, which is roughly what the studies find.

Correlational data cannot settle direction. If people who feel worse use their phones more, that produces the same statistical pattern as phones making people feel worse. The datasets above cannot distinguish those, and the authors do not claim to.

So the position is not "the research proves screens are harmless". It is that the studies large enough to find a population-level effect are also the studies least able to detect the specific effects that matter, and both of those facts should make you more sceptical of anyone quoting a threshold in either direction.

The one place a hard number is defensible

Children under five. The WHO guidance exists, it is specific, and it rests on developmental evidence about the early years that does not generalise to adults or to teenagers.

Everywhere else, a page confidently telling you that four hours is fine and five is harmful has produced that boundary from nothing. There is no study behind it, because the study that would establish it has never been done, and the largest attempt to find the effect at all found something the size of a potato.