The very same device can be a brilliant maths tutor and a concentration-destroying machine — sometimes within the same ten minutes. So the anxious question "how much screen time?" turns out to be almost the wrong one. What the research keeps showing is that it isn't the screen that matters, but what's on it and how it's used. Here's how to tell the helpful from the harmful.
A thirteen-year-old sits down to do maths on a tablet. For ten minutes it's wonderful: an adaptive app feeds him questions at just the right level, marks them instantly, and nudges him when he slips. Then a notification slides in. He glances at it, replies to a friend, watches a clip, and surfaces twenty minutes later having done three more problems and forgotten where he was. The device didn't change. It hosted a first-rate tutor and a first-rate distraction back to back — and most of the time, the distraction wins.
Parents are told, in the same week, that "edtech is the future" and that "screens are rotting children's brains." Both can't be simply true. This article cuts through it with what the evidence actually shows — that the device is neutral, and everything depends on what runs on it and how. It draws heavily on the mechanisms covered elsewhere in this series, especially spaced and retrieval practice and working memory, because that's exactly where good and bad screen use part ways.
The defining feature of the modern maths device is that it is two things at once. On one side: adaptive practice apps, instant feedback, video explanations you can pause and rewind, intelligent tutoring systems, and now AI that can explain a step in three different ways. On the other side: an endless feed, group chats, games engineered to be hard to put down, and a steady drip of notifications, each one a tiny hook for the attention the maths needs.
Because both live on the same rectangle of glass, families end up arguing about the rectangle — "less screen time!" — when the real question is which of its two faces is showing. A child doing focused, well-designed practice and a child half-watching videos while replying to messages are having opposite experiences, even though a screen-time tracker logs them identically. Lumping them together is how the debate gets so confused.
Here is the idea that clears the fog.
"Screen time" treats minutes as the unit of harm or benefit, as if all minutes on a device were the same. They are not — not even close. A minute of focused retrieval practice and a minute of distracted scrolling differ as much as a minute of reading differs from a minute of being shouted at. The screen is a delivery mechanism, not an activity. Asking "how much screen time?" is like asking "how much paper time?" — paper can carry a maths textbook or a comic, and the paper isn't the point.
Think of a screen as a kitchen knife. In one hand it's an indispensable tool; in another it's a hazard. Nobody asks "how much knife time is healthy?" — they ask what you're doing with it. The same knife chops the vegetables and cuts the finger. A maths app delivering spaced, feedback-rich practice and a notification-stuffed game are the chop and the cut. Counting the minutes tells you nothing about which one your child just had. The useful question is always: what is this screen doing right now — building focus, or breaking it?
This reframing matters because it changes the job from policing a number to curating an experience. You're not trying to minimise minutes; you're trying to maximise the helpful kind and minimise the harmful kind. And, encouragingly, the research is fairly clear about which is which.
The evidence on technology in maths is more nuanced than either the boosters or the doom-mongers admit. A few findings anchor it.
In a large randomised study across seven US states, Pane and colleagues (2014) tested a well-designed "intelligent tutoring" maths system — software that adapts to each student and gives targeted feedback. The result is instructive: no effect in the first year, but a meaningful positive effect in the second, once teachers and students had learned to use it well. Good edtech, in other words, isn't a magic download; it's a tool that pays off when it's properly embedded — and it works by doing the mechanisms we already know help.
Reviewing decades of experiments, VanLehn (2011) found that sophisticated "intelligent" tutoring systems were nearly as effective as human tutors, and far better than simple answer-checking software. It's a genuinely hopeful finding for well-built apps. But two things are worth holding onto: the cheap "did you get it right? try again" apps did much less, and human tutoring still came out on top — the relationship and responsiveness of a person remain hard to fully automate (more in our piece on what one-to-one tutoring really does).
The OECD's PISA studies have found that students who report being distracted by digital devices in maths lessons — their own and others' — tend to score lower, and that the relationship between device use and achievement is far from "more is better." Meanwhile, researchers such as Kirschner and De Bruyckere (2017) have dismantled the comforting myth of the "digital native multitasker": there is no effective multitasking, only rapid, costly switching. Every notification answered mid-problem is a small tax on the working memory the maths depends on.
And what about AI? Honesty matters here, because the technology is new and the high-quality evidence is still thin. Generative AI tutors are genuinely promising as patient, infinitely available explainers — but they introduce a specific danger: cognitive offloading, where a child lets the AI do the thinking and ends up with fluent-looking answers and almost no learning. The mechanisms in the rest of this series tell us why that's fatal: learning is built by effortful retrieval and productive struggle, and an answer machine removes both. Used to explain and quiz, AI can help; used to bypass the struggle, it quietly guts the learning.
Strip away the novelty and the picture is simple: technology helps when it delivers the things that have always helped, and harms when it attacks the things learning depends on.
Good tech helps by industrialising the proven mechanisms. A well-built app can give instant feedback (which a busy teacher can't always), serve spaced and interleaved practice automatically, adapt difficulty to keep a child in the productive zone, and offer rewatchable worked examples. These aren't new ideas — they're the greatest hits of learning science (spacing and retrieval, worked examples, feedback) — delivered at scale and on demand.
Bad tech harms by fragmenting attention. Maths is unusually fragile to interruption because it leans so hard on working memory — the small mental workspace where a multi-step problem is held together. A notification doesn't just cost the seconds you look at it; it can collapse the whole half-built calculation, forcing a restart. Even a silent phone on the desk has been shown to pull on attention. Distraction isn't a character flaw in the child; it's the predictable result of engineered interruption meeting a fragile mental workspace.
And screens harm indirectly, by displacement. Every hour lost to a feed is an hour not spent on focused practice, outdoors, with family, or — crucially — asleep. Late-night scrolling that eats into sleep is a double hit, because sleep is when the day's maths gets consolidated (see our piece on sleep and learning). The harm often isn't in the screen itself but in what it quietly crowds out.
Countries are running very different experiments on screens in school, and the results are starting to inform policy. Tap through five.
The UK has moved toward restricting phones in schools amid concern about distraction, while still investing in maths software and online practice. It captures the modern tension neatly: clamp down on the distraction face of the device, lean into the tool face. The direction of travel — phones away for focus, good apps used deliberately — fits the evidence well.
The United States is where much of the intelligent-tutoring research was done (Pane, VanLehn) and where adaptive maths platforms are most widespread. It also shows the flip side: heavy device saturation and growing worry about attention. The American experience underlines the core lesson — the technology can genuinely help, but only when it's the focused-tool kind, not the always-on-distraction kind.
France introduced one of the strictest school smartphone restrictions, banning phones in schools for younger pupils. The rationale is straight from the distraction research: remove the engineered interruption so attention — the scarce resource maths depends on — can recover. It's a natural experiment worth watching, and an example of treating the device's two faces very differently.
Singapore pairs its strong maths curriculum with deliberate, structured use of educational technology rather than a free-for-all. The emphasis is on tech that serves clear learning goals — practice, feedback, visualisation — embedded into good teaching. It's a useful model of edtech as a servant of sound pedagogy rather than a substitute for it.
Across OECD countries, PISA data tell a consistent story: moderate, purposeful use of devices for learning can help, but heavy use and device distraction are associated with lower performance. The relationship is not "more technology, better results" — it's a curve, where the right tool used the right amount helps and saturation hurts. The international signal matches the kitchen-knife principle exactly.
The global pattern is converging on a sensible middle: protect attention from the distraction face of the device, and use the tool face deliberately and in moderation. Neither blanket ban nor unlimited access fits the evidence; curation does.
You don't have to win the whole screens war. You just have to tilt your child's device time toward the helpful face and away from the harmful one. These moves do most of the work.
Before letting a tool into your child's routine, ask one question: "Does this make my child do the thinking, or does it do the thinking for them?" A good practice app makes them retrieve, attempt, and get feedback on their own work — the effort stays with the child. A bad one (or a misused AI) serves up answers and animations while the child watches. Tools that keep the cognitive effort with the learner help; tools that remove it, however slick, do not. It's the same principle behind productive struggle.
For educators, the task is to harness the tool face of technology while defending the classroom from its distraction face.
Use tech to deliver feedback and practice at scale. Adaptive platforms shine at giving every student instant, individualised feedback and endless calibrated practice — things one teacher with thirty students cannot do alone. Deployed for those jobs, and given time to embed (as Pane's two-year result suggests), good software genuinely extends what a teacher can offer.
Protect attention deliberately. Phones away during problem-solving isn't anti-technology; it's pro-concentration. Given how fragile maths is to interruption, defending the focused stretches of a lesson is one of the most valuable things a teacher does — and modelling single-tasking matters as much as enforcing it.
Teach AI as a study partner, not a ghostwriter. Students will use AI regardless, so the high-value move is teaching them to use it well: to explain, to generate practice, to check their reasoning after they've attempted a problem — never to skip the attempt. Naming the offloading trap explicitly is the best protection against it.
Our practice portal is deliberately the "tool" face of the device: focused, single-purpose practice with instant feedback and worked solutions, no feeds, no rewards-for-rewards'-sake, no notifications. It's the kind of screen use the evidence actually supports.
Open the practice portal →"Less screen time automatically means better learning."
Minutes are the wrong unit. A minute of focused, feedback-rich practice and a minute of distracted scrolling are opposites. Judge the activity, not the screen-time total.
"Kids today are great at multitasking, so notifications are fine."
Effective multitasking is a myth — the brain switches rapidly at a cost. In maths, an interruption can collapse a half-built calculation. Even a silent phone nearby drains focus.
"An AI that gives the answers is a great maths tutor."
Learning is built by effortful retrieval and struggle. An answer machine removes both, producing fluent-looking work and little real learning. AI helps when it explains and quizzes, not when it solves.
It's tempting to want a single verdict — screens good, screens bad — but the honest answer is more useful than that: it depends entirely on what the screen is doing. The same tablet that can deliver world-class, individualised practice can also serve up the most attention-shredding experience ever engineered, and a child can swing between the two in minutes. Your job isn't to ban the device or to surrender to it. It's to keep tilting the balance toward the tool and away from the trap — one focused session, one phone-in-another-room, one "explain it, don't solve it" at a time. Do that, and technology becomes what it should be: a genuinely powerful ally for a young mathematician, rather than the thing standing between them and the work.
"Screen time" is the wrong measure. The same device can host a well-designed practice tool that genuinely helps and a stream of notifications that wrecks concentration. What matters is what's on the screen and how it's used — active, focused practice with feedback helps; passive video, multitasking and constant alerts harm. Judge the activity, not the minutes.
Some do, when built on sound principles — spaced retrieval, instant feedback, worked examples, adapting to the child's level. A large trial of one well-designed intelligent tutoring system found meaningful gains, though only in the second year of proper use. Good apps help by delivering the mechanisms we know work; flashy, game-heavy apps with little real practice mostly don't.
It can, but with a real risk. Used as a patient explainer — to clarify a step, generate practice, or check reasoning — AI can be useful. Used as an answer machine that does the thinking, it produces fluent-looking work and almost no learning. The evidence is still early, so the safe rule is: AI may explain and quiz, but the child must do the actual problem-solving.
Because there's no such thing as effective multitasking — the brain rapidly switches rather than truly splitting attention, and each switch carries a cost. Maths leans heavily on working memory, where you hold a problem together. A notification that pulls attention away can collapse a half-finished calculation. Even a phone sitting silently nearby has been shown to drain focus.
For focused maths work, keeping the phone in another room is one of the simplest high-value moves a family can make — not as punishment, but to protect the concentration maths needs. The device can come back for a deliberately chosen, well-designed practice tool. The aim isn't zero screens; it's screens used as a tool, on purpose, rather than as a constant background distraction.
The free assessment includes practical guidance on the tools and routines that actually help your child's maths — and the habits quietly getting in the way. Calm, specific, no obligation.
Book the free assessment