Does Screen Time Lower Cognitive Test Scores? Eight Years of Data Say No

The expectation going in was damage. Researchers at the Universities of Jyväskylä and Eastern Finland had eight years of measurements on a group of Finnish children, taken from primary school into the middle of adolescence, and a well-founded reason to think that the ones who had spent the most hours in front of screens would come out of it with weaker attention and weaker working memory. When the cognitive testing was done, the association ran the other way.

What the study actually did

The work comes from PANIC — the Physical Activity and Nutrition in Children study — a long-running Finnish cohort, and was published in Pediatric Exercise Science in 2026 by Petri Jalanko and colleagues under the title "Associations of Physical Activity and Sedentary Time From Childhood to Adolescence With Cognition in Adolescence". The analysed sample was 260 adolescents, 124 girls and 136 boys, first measured at ages six to nine and followed for eight years to a mean age of 15.8.

Two things were tracked across that period. Movement was captured both objectively, using a combined heart-rate and motion sensor worn by the children, and subjectively, through questionnaires in which they reported guided exercise, unsupervised exercise and screen time. Cognition was measured at the end with the CogState battery, a computerised set of tasks covering learning, attention and working memory — the kind of instrument that yields separate scores for how fast someone processes simple information and how much they can hold and manipulate at once.

That design matters for how much weight the finding can carry. It is observational: nobody assigned children to more or less screen time, so anything found is an association between how children already lived and how they later performed. It is also modest in size. Two hundred and sixty adolescents is a respectable longitudinal sample and a small one by the standards of a definitive claim.

• Longitudinal, not experimental — exposure was measured, not assigned, so reverse causation and third factors are both live explanations.
• Eight years of follow-up, which is long enough for childhood habits to have had a real chance to leave a mark.
• Two measurement methods for activity, which turned out to disagree with each other.
• One cohort, one country, one set of cognitive tasks — replication elsewhere is what would make this more than interesting.

The result that ran the wrong way

Higher amounts of screen time from childhood onwards were connected to better cognitive processing in adolescence. Not to worse attention, not to weaker working memory, and not to nothing at all — to a small advantage on the processing measures. The authors are careful about what that licenses. Their stated position is that intervention studies are needed to establish causal relations, and the lead author's summary of the practical implication was that screen time should not be regarded as solely harmful, and that the useful target is a balance between physical activity and screen activities that promote active thinking.

The physical activity findings, which were the study's actual subject, came out more tangled. Girls with more light-intensity physical activity since childhood tended to have better working memory; boys who had done more guided physical activity tended to have better working memory as well. Less self-reported unsupervised physical activity was connected to better cognitive processing — a result nobody would have predicted. And the device-measured activity, the more objective of the two methods, showed no associations at all.

Why "screen time" is a poor variable to test anything with

The finding is interesting mainly because of what it reveals about the measure. "Screen time" is a container, not a construct. It sums an hour of a well-designed maths game, an hour of a coding tutorial, an hour of a video call with a grandparent and an hour of an algorithmic feed, and then treats the total as a single exposure whose effect can be estimated. Almost nothing else in developmental research is measured that way. Nobody studies the effect of "paper time" on reading.

A container variable behaves in a predictable manner: it produces small, unstable and direction-flipping associations, because the things inside it are pulling against each other and their mix varies from child to child and from year to year. That is a fair description of the screen-time literature as a whole. It also explains why a single study finding a positive association is not a reversal of the field so much as a demonstration that the field is measuring the wrong thing.

The same caution applies in the other direction, and this is where an assessment desk has something to add. Cognitive test scores are affected far more reliably by the conditions of the day than by anything in a child's media diet. Sleep is the clearest case — we have set out what a short night actually does to test performance, and the effects on sustained attention and processing speed are large, well replicated and produced experimentally rather than merely observed. If screens matter to a score, the strongest route by far is through bedtime.

What a parent or a teacher can take from this

• Do not read a positive association as a recommendation. Nothing here says more screens produce better thinking; it says that in this cohort the children with more screen time did not test worse.
• Do read it as a reason to stop treating a raw hour count as the meaningful quantity. What is on the screen, and what it displaces, are the questions with content in them.
• Watch displacement rather than duration. Screen use that pushes out sleep is the case with the best evidence behind it, and it is measurable at home without any instrument.
• Be sceptical of confident causal language in either direction, including from people warning you. The honest state of this literature is small associations from observational designs with disagreeing measurement methods.
• And keep cognitive tests in proportion. A CogState processing score is a reliable measure of one narrow thing, not a verdict on a child.

The same restraint belongs on any single number attached to a child. A CogState processing score, like a score from a reasoning test, is an estimate with a range around it rather than a fixed property of the person who produced it, and the way to read one is to convert it into a position in a stated reference group — which is what a percentile conversion is for. How these tests are built explains why the range matters as much as the point: every serious instrument publishes a standard error of measurement, and it exists because one sitting is a sample of a person rather than a verdict on one.

The long view on childhood measurements

It is worth remembering how much and how little a cognitive score taken in adolescence tells you. Scores measured in childhood are among the most durable measurements psychology takes — they track meaningfully with health and outcomes decades later, at the level of group averages with wide individual spread. That durability is exactly why the environmental questions attract so much attention, and exactly why the answers need to be better than a container variable can supply.

What a study like this earns is a smaller and more honest claim than either camp wants. Eight years of Finnish data are not a licence for unlimited screens and are not evidence that the worry was invented. They are evidence that the quantity everyone is arguing about is too coarse to settle the argument, and that the next useful study will be the one that stops counting hours and starts distinguishing activities. The same caution belongs on the other end of the lifespan, where the age curves for different abilities turn out to point in different directions too.

Common questions