This autumn, six-year-olds in Norway are learning to read with books and pencils. Tablets have gone from the first years of primary school, phones are banned at every age, and the government has announced stricter screen rules for years 1 to 4, alongside more play and fewer subjects. The education minister, Kari Nessa Nordtun, has said that the digitisation of Norway's schools has failed.
The story has travelled fast. On LinkedIn and in staffrooms it is told as a verdict: a country gave every child an iPad, reading collapsed, and now it is putting things right. It is a story that takes correlation for causation.
Norway may well be right to take tablets away from children who are learning to read. The more useful question for a school leader is why it is right. The answer tells you what to do in your own school, and it is rarely "ban the screens".
A word that hides too much
Screen time has become one of the most powerful phrases in education, and a whole campaign has grown up around it. The trouble is that it is a term that counts minutes and ignores what actually happens in those minutes. An hour on Snapchat or TikTok is screen time. So is an hour on Times Tables Rock Stars or Sparx Maths. One is built to hold attention for its own sake. The others are built to give a child practice and to show the teacher what went wrong. If you think about it, referring to them in the same way makes no sense whatsoever.
The timing makes this matter. In September, PISA 2025 recorded the lowest average reading score the OECD has ever measured: 461, down 28 points since 2015. In January, the neuroscientist Jared Cooney Horvath told a US Senate committee that Generation Z is the first modern generation to score lower on standardised tests than the one before, and dated the turn to the spread of digital learning tools around 2010. His book, The Digital Delusion, is now being passed around schools. From what I have seen, it is also being used to lend weight to determined, single-issue lobbying of government. The same heads who nodded along to Jonathan Haidt's argument in The Anxious Generation are now nodding along to Horvath's.
Some of this is simply right. Haidt's case for phone-free schools has support in PISA itself, and I agree that school phone bans are a good thing. In the UK, where 80% of 15-year-olds attend schools that ban phones, pupils in those schools are about 40% less likely to report being distracted by devices in lessons. Across OECD countries, pupils who report digital distraction in science lessons score 11 points lower in science, after accounting for their background. A school that has not yet banned phones should weigh up their few benefits against their considerable downsides.
Paper wins
The strongest objection to screens in classrooms is the one I hear most often, and it is also true. For understanding text, paper beats the screen. A meta-analysis by Pablo Delgado and colleagues, with the memorable title Don't throw away your printed books, found that readers understood information texts better on paper, and that the gap was clearest when they read against the clock. Readers on screens were also more overconfident about how much they had understood. It is reminiscent of a trial in a Turkish secondary school, where pupils who practised maths with an unrestricted AI chatbot felt they had learnt more, yet scored 17% lower in their exams than pupils who never had it.
PISA 2025 adds a worrying detail. The sharpest falls in reading came on tasks built around longer texts, and the share of hasty readers, who answer quickly and wrongly, rose from about 7% in 2018 to 11% in 2025. That looks like a loss of sustained attention.
So the sceptic has a fair question. If paper is better for reading, and reading is what is falling, why put a tablet in a primary classroom at all?
My answer has always been that a tablet should join pen and paper, never replace them. Some activities are best done on paper and others suit a device. A teacher with good pedagogical knowledge decides which is which, checks the outcome and adjusts. I still believe that. But it puts the whole weight on teachers, many of whom have never been trained to make that choice, and it assumes that someone is checking the outcome. In a lot of schools, sadly, nobody is in charge of where pedagogy and devices meet.
The wrong unit
This is why screen time is the wrong unit for school policy. Minutes on a device tell you very little. Four other things tell you quite a bit: what the tool is for, how long it is used, at what age, and how the school will know it worked. Teachers answer those questions lesson by lesson. A digital strategy sets the limits they work within and trains them to work well inside those limits, which sometimes means paper.
From this perspective, Norway's decision about its six-year-olds is arguably what a good digital strategy would have recommended in the first place.
Correlation cuts both ways
The case against devices rests mostly on timing. Scores fell while devices spread, so devices must be to blame. That is a correlation, and it deserves the same scepticism we would give a supplier who pointed to rising results after a device rollout.
Norway itself complicates this argument with some awkward facts. Using army test data on three decades of conscripts, the economists Bernt Bratsberg and Ole Rogeberg showed that Norwegian IQ scores, which had risen for generations, began to fall with boys born around 1975. Those boys sat the tests at about 18, in the early 1990s. They went to school without devices, met email in their late teens and got their hands on their first mobile phone only in their twenties. They are about 50 now. The decline also appears between brothers in the same family, which means its cause is to be found somewhere in the world they grew up in.
To be fair, this is IQ rather than PISA, and it does not prove that devices are harmless today. It does, however, show that a decline in thinking skills was under way long before the devices arrived.
The OECD's own explanation for falling reading points the same way. This year's report describes four forces working together: pupils giving less effort and attention; classrooms that are harder to manage, with more bullying and more time spent keeping order; education systems struggling to meet more varied needs; and wider changes in society, including less reading for pleasure among parents as well as children, more time on social media and lower attendance since the pandemic. Devices in classrooms barely feature. I made a similar case earlier this year, when I looked at how attention is shaped outside school.
None of this will surprise anyone who has worked in schools. Policy, curriculum design, how we train and pay teachers, and how much a society values school will shape attainment far more than whether a child spends an hour a day on a device. That is my judgement rather than a measured result, but the OECD's analysis points in the same direction.
Purpose does the work
Where technology has helped, it has had a job to do. A large review of experimental studies by Maya Escueta and colleagues found that giving pupils devices did little for learning on its own, while software that adapts practice to each pupil, especially in maths, produced some of the strongest results in the field. In Delhi, an after-school programme called Mindspark, which combined adaptive software with small-group support, raised scores sharply in maths and Hindi in a randomised trial. The Education Endowment Foundation makes the same point in its guidance on digital technology: work out how a technology will improve teaching and learning before you adopt it, because the device itself is unlikely to raise attainment.
What Norway still tells us
Norway's own figures are less comfortable.
PISA asks 15-year-olds how long they spend on devices for learning at school. In Norway the answer is 2.6 hours a day, against an OECD average of 1.7. Norway's reading score fell by 24 points between 2022 and 2025, one of the largest drops of any country. Japan, at the OECD average of 1.7 hours, fell by about half as much. The UK, at 1.3 hours, held steady.
That is a correlation too. These are 15-year-olds rather than six-year-olds, and Norway left more pupils out of its sample than PISA's rules allow. But it shows something the phrase "screen time" hides. Norway's decline began before the devices arrived. What it seems to have lacked was anyone deciding what those hours were for, how many of them made sense, and how anyone would know whether they were working.
That is the job of a digital strategy. A good one starts with teaching, determines what each tool is for, and is willing to say no: to tablets for children learning to read, to phones in lessons, and to glittery tools that look great but whose purpose nobody can explain.
Four questions
Before anyone in your school uses the phrase "screen time", ask four questions of the tool in front of you:
- What is it for?
- For how long?
- At what age?
- How will we know it worked?
If they can't be answered satisfactorily, then go ahead and stick with paper. That is OK.
"Screen time" is simply an unhelpful concept in education, and I would put those four questions in its place. The success of a school day can then be measured by how much pupils learnt, rather than by how many minutes they spent on a device. Surely that is the right test.
Comments
Which tool in your school would struggle to answer the four questions, and which would pass easily?