“Screen time” is one of those phrases that sounds more precise than it really is. It gives us a number, and a number feels easier to judge than the messy reality of a child switching between homework, games, videos, group chats and whatever platform adults have only just heard about.

A new Finnish result complicates the familiar story. Across an eight-year follow-up, children who accumulated more reported screen use tended to perform better on some cognitive-processing tests as teenagers. It is a genuinely interesting finding. It is not evidence that adding another hour of YouTube will make a child smarter.

This is one study, not settled consensus. It was observational, the screen measure came from questionnaires, and nobody randomly assigned children to spend more or less time with a device.

What the Finnish team actually followed

The paper was led by Petri Jalanko and published in Pediatric Exercise Science. It drew on the long-running Physical Activity and Nutrition in Children study, known as PANIC, in Kuopio in eastern Finland.

The final analysis included 260 young people, 136 of them boys, who were 15 to 17 years old at the eight-year examination. Their screen use was recorded at the beginning of the study, two years later and after eight years. Parents completed the questionnaire at the first two examinations; the adolescents completed it at the last.

The researchers then used those three reports to estimate each participant’s cumulative exposure over the full period. That is a more serious design than asking a 16-year-old to reconstruct an entire childhood from memory. Still, it is not eight years of continuous monitoring. Three snapshots leave plenty outside the frame, and the final group of 260 may also differ from children who did not remain in the analysis.

One total covered very different digital lives

In this study, screen time was the sum of time spent watching television and videos, using a computer, playing video games, using a mobile phone and playing mobile games. That total does not tell us whether a child was building something, solving a puzzle, talking to a friend, watching a lesson or letting short videos wash past for an hour.

The calendar matters too. Baseline measurements began in 2007–09, before the current short-video and algorithmic-feed environment had taken shape. The media mix changed considerably while these children grew up.

So “more screen time” is accurate as a description of the variable, but not very descriptive of the experience. A stopwatch can count a strategy game and a passive video equally while missing nearly everything that makes them psychologically different.

What “better cognitive processing” meant here

At the final examination, the teenagers completed a computerised CogState battery. It included tasks covering psychomotor speed, attention, working memory, visual learning and an overall cognitive composite.

The clearest screen-time findings were shorter reaction times in one-back and two-back working-memory tasks, along with a higher overall cognition score. In the one-back task, a participant had to decide whether the card on screen matched the previous card. The two-back version required keeping one more step in mind.

Those are legitimate cognitive measures, but they are not school grades and they do not amount to a universal intelligence score. A teenager responding more quickly in a controlled card task has shown something specific about processing and working memory. That does not automatically tell us about reading, mathematical understanding, creativity, judgement or how well the same person learns in a classroom.

The associations were modest. The authors tested many relationships and used a Benjamini-Hochberg correction with a false-discovery rate of 0.2. In plain language, that is a fairly permissive threshold, designed to limit the expected share of false discoveries among the flagged findings to 20 per cent rather than eliminating that risk. It is another reason to treat the result as a lead worth following, not a verdict.

Accounting for earlier ability helps, but does not prove change

The researchers adjusted their models for age, sex, parental education and the children’s earlier result on Raven’s Coloured Progressive Matrices, a test of non-verbal pattern reasoning. Further adjustments for body fat, pubertal status and whether participants were in the PANIC intervention group did not materially alter the main associations.

That is meaningful. It makes the simple explanation that brighter children merely used screens more somewhat less convincing. But the baseline Raven test and the adolescent CogState battery were not the same measure. The study did not repeatedly administer an identical cognition test and show that one child’s score rose as that child’s screen use rose.

Nor can statistical adjustment erase every unmeasured difference. The home learning environment, sleep, existing interests, the kinds of games played and the content viewed could influence both digital habits and later test performance. The analysis found an association. It did not identify a cause.

Why credible studies can appear to disagree

A separate analysis of PISA 2022 data from more than 10,000 Finnish 15-year-olds linked frequent entertainment screen use with weaker mathematics, reading and science performance. That sounds like the opposite result.

Meanwhile, a 2022 US longitudinal study using ABCD data found small later gains on an intelligence measure associated with gaming and video viewing after the researchers accounted for genetic and socioeconomic differences. Neither study gives screens a single fixed effect.

The apparent clash becomes less mysterious once we stop treating all outcomes and all uses as interchangeable. Faster working-memory responses are not the same as academic achievement. A demanding game is not the same activity as late-night scrolling. Screen use that supports a friendship or a project may sit in a life differently from use that displaces sleep, exercise or schoolwork.

A 2024 meta-analysis of 100 early-childhood studies likewise found that context mattered: age-inappropriate content, background television and caregiver use during routines showed different associations from co-use. Those children were younger than the Finnish cohort, so the review is not a direct answer here. It does, however, show why duration alone is a blunt exposure measure.

The useful lesson is about measurement, not permission

I would not turn this paper into a new instruction to increase children’s screen time. I would use it to be more careful with a debate that often treats the timer as both diagnosis and explanation.

I reached a related caution while writing about Norway’s reversed Flynn effect. A cognitive score can be real and informative while remaining much narrower than the cultural story built around it. Measurement is not trivial, but neither is it the whole person.

The honest reading here is fairly simple. In this Finnish cohort, greater reported screen use tracked with faster responses on two working-memory tasks and a higher cognitive composite in adolescence, even after several important adjustments. The design cannot tell us that screens produced those differences, which digital activities mattered, or whether another group of children would show the same pattern.

That uncertainty does not make the finding useless. It makes the next question better. Instead of only asking how long a young person looked at a screen, researchers need to ask what happened on it, what it replaced and what the child brought to the activity in the first place.