The phone-ban study measured grades. The case for it was never about grades

The first thing the teachers noticed was the noise. Not the bad kind: the hallways got louder, the lunchrooms filled with the particular racket of teenagers talking over one another, and the silence that had settled over a generation of corridors, each student bent into a private screen, broke up into something messier and more alive. This is the detail that survives in every account of a school that has cleared the phones out of the day, from English comprehensives to Dutch gymnasiums to public schools across Brazil. And it is the detail that does not appear anywhere in the number everyone is now arguing about.

That number is test scores, and the largest study yet conducted says they barely moved. A working paper released this spring through the National Bureau of Economic Research, written by economists at Stanford, the University of Pennsylvania, Duke and the University of Michigan and drawing on thousands of American schools, found that the effect of phone bans on academic achievement was, in the words of co-author E. Jason Baron, basically close to zero. No measurable lift in math or reading. No movement on bullying, on attendance, on whether students said they could pay attention. The skeptics treated this as a verdict, and the verdict was: the whole project was theatre. They are wrong, but not for the reason the other camp thinks. The study is not a measurement of whether phone-free schools work. It is a measurement of what we decided to count. And we decided to count the one thing the policy was never really about.

This matters to anyone who has watched a thirteen-year-old’s face go slack over a feed, or tried to teach a room that is physically present and mentally elsewhere, or simply remembers that an unsupervised afternoon used to be a place where a self got built. The question underneath the phone-ban fight is not whether children can be made to score higher. It is whether a few hours of a young person’s day can still belong to that young person, rather than to a platform engineered to harvest every idle second. Test scores cannot see that question. They were never going to.

Consider how strange it is that grades became the scoreboard at all. Nobody confiscates a teenager’s phone at the school gate hoping to raise the class average in algebra. Parents who back these policies talk about sleep, about anxiety, about a child who used to read and now scrolls, about a dinner table that has gone quiet. Teachers talk about eye contact. The grade was adopted as a proxy because it is the thing schools already measure obsessively, and a number you already have is always more tempting than a good one you would have to go and build. Judging a phone ban by its effect on test scores is like judging a public park by its parking revenue: you will get a precise figure, and it will tell you almost nothing about why the thing exists.

It is worth remembering why the test-score frame ever seemed plausible. A decade ago, economists Louis-Philippe Beland and Richard Murphy studied ninety-one schools across four English cities and found that banning phones lifted the scores of sixteen-year-olds by more than six percent, and by over fourteen percent for the lowest-achieving students, the ones most easily pulled off task. That finding became the load-bearing statistic of the entire movement, the empirical promise that getting rid of phones would close gaps and raise outcomes. The new study does not so much refute it as date it. The phone of the Beland and Murphy era was a distraction you carried; the phone of the current era is an attention economy you live inside, tuned by recommendation engines that did not exist when that research was done. A ban that recovers a few minutes of focus in a 2015 classroom is doing something smaller, against a far larger adversary, in 2026. The promise that it would show up in grades was always the weakest part of the case, and it has now quietly collapsed.

So look instead at the number the headlines skipped. The same body of research that found no academic effect found something else: student wellbeing got worse in the first year of a ban and then turned positive by the third. Read quickly, that is a wash. Read honestly, it is the most revealing measurement in the entire study, because it has a shape, and the shape is withdrawal. Things that are merely useless do not hurt to remove. You can take away a child’s broken calculator and no one grieves for a year. A tool whose removal stings for twelve months and only stops stinging once a new equilibrium sets in is, by definition, a tool that had a grip, one that was doing something to the nervous system it was attached to. The dip is not a cost that happens to precede a benefit. The dip is the evidence.

The strongest case against all of this deserves to be stated at full strength, because it is serious and it is widely held. The psychologist Candice Odgers, reviewing Jonathan Haidt’s The Anxious Generation in the pages of Nature, argued that the science does not actually support the claim that smartphones are rewiring young brains or driving an epidemic of mental illness, and that the correlation between rising phone use and rising distress may run partly in reverse, with struggling kids reaching for the phone rather than the phone making them struggle. On this view, the smartphone has become a moral lightning rod, a single villain that absorbs anxieties really caused by academic pressure, overscheduling, the disappearance of unsupervised play and plain economic precarity. And the school ban, the critics add, is the cheapest possible gesture: it governs the six hours a child is in the building and does nothing about the seven hours of scrolling that resume the moment they walk out. A policy that changes no measurable outcome while letting everyone feel they have acted is not a solution. It is an alibi.

This is a real argument, and it lands squarely on the people who promised that bans would raise scores. It does not land on the ban. Odgers is right that the smartphone cannot carry the whole weight of a generation’s troubles, and right that a school cannot fix a childhood. But a school was never claiming to. It claims jurisdiction over six hours, which happens to be exactly the scale at which it can actually do something rather than merely deplore. The alibi charge assumes that the only outcomes that count are the ones that fit in a regression. Yet the thing these policies most reliably produce is precisely the thing that does not fit: the loud hallway, the recovered lunch table, the unstructured social practice of being bored in a room with other people, which is how humans have always learned to be around one another and which the feed had quietly dissolved. That is not a rounding error in the data. It is the data the instruments were too crude to hold.

The international record makes the point in four languages. The Netherlands cleared phones from secondary classrooms at the start of 2024 and found, within a year, that three-quarters of schools reported sharper focus and a majority a better overall climate, improvements that lived in the texture of the school day, not the exam hall. France began its pause numérique by having younger students surrender their phones at the door, and the government now intends to extend the rule to the lycées for the 2026-27 year, a wager on attention rather than achievement. Brazil, which restricted phones nationwide in early 2025, offers the most honest ledger of all: more than eighty percent of students reported paying closer attention, while forty-four percent admitted to being more bored at break and nearly half of teachers noticed more anxiety. Those last two figures are usually cited as marks against the policy. They are better understood as its price tag: the discomfort of being handed back the empty time that a machine used to fill, and being expected to do something human with it.

None of this means the ban is a cure, and the honest version of the argument refuses that word. Clearing phones from a school is a small intervention with a narrow remit: it will not lift a struggling reader, mend a depressed teenager, or undo whatever is happening in the seven unsupervised hours at home. What it can do is fence off a single protected clearing in a child’s day, a stretch of waking life that no recommendation engine is mining, no notification is interrupting, and no metric of engagement is being optimised against. Whether that clearing is worth defending is a question about values, not test scores, and it is the question the spreadsheet was always going to dodge.

Which is why the lesson of the biggest study is not the one its loudest readers drew. The finding to keep is not that bans fail; it is that we have been defending them on the wrong ground, and the wrong ground gave way exactly as it should have. Stop promising parents better grades. Promise them what the policy actually delivers: a few hours in which their child’s attention is not for sale. A tool you can take away without anyone noticing is a tool nobody needed. The phones hurt to remove for a year — and that, not the flat line on the score chart, is the most honest measurement in the whole study.