The school phone bans worked. The test scores haven’t moved

The largest controlled look at U.S. classroom phone-ban policy yet — drawing on data from roughly 4,600 schools — found that the policies are working at the thing they were built to do. In schools with bell-to-bell restrictions, teachers reported phone use in class dropping from 61 percent to 13 percent in three years. Thirty-seven states and the District of Columbia now require some version of the ban. The Los Angeles Unified School District is preparing to extend the policy this fall to laptops and tablets. The compliance figures are something the ban movement could not have invented in its early days.

What the same data set will not give the ban movement is the achievement story. Test scores have not moved. Bullying rates have not moved. Self-reported attention has not moved. Attendance is flat. The ban camp will say the dose-response curve has not had time to show; the skeptic camp will treat the null result as proof that the whole thing was moral panic in shop-class clothes. Both readings miss what is actually here. The bans worked. The metrics meant to validate them never had anything to do with what the bans were really delivering.

Anyone with a child in a school district that has banned phones knows the texture of the change without needing a study. The cafeteria is louder. The hallway feels less like a row of small private cinemas. Kids who would have spent lunch in their hand are talking to each other, or at least staring at each other, which is the first half of talking. The complaint that nothing has gotten better depends on what you thought school was supposed to deliver. If the answer was higher test scores, the data is right and the ban is irrelevant. If the answer was anything else, the data is irrelevant and the ban is a small victory.

The American case for restricting phones in school was built, deliberately, on metrics that made the policy legible to elected officials. Jonathan Haidt’s argument in The Anxious Generation tied phone-driven anxiety to school performance, partly because school performance is the language education policy understands. State legislatures wrote phone-ban bills using the same vocabulary — they would not have moved if the only argument on offer was that kids were spending less time with each other. Adults wrote bills the way adults write bills, which is in numbers that show up on dashboards.

The trouble with that framing is that test scores were already moving in the wrong direction for reasons that have nothing to do with phones. Pandemic learning loss has not fully recovered. Teacher shortages persist. Curriculum fights have eaten time. Math instruction has been in slow erosion across much of the OECD. Asking the phone ban to lift achievement metrics out of all that is asking one variable to do the work of twelve.

But it is also asking the ban to deliver something it was not really aimed at. The teenager who hands her phone to the teacher at the start of the day is not entering a regime of intensified academic effort. She is entering a regime of restored attentional and social availability. The change shows up in the social texture of the building, not in the algebra grades, because algebra was never the point. The point was the building.

We have become awkward about saying that. The post-Coleman-Report tradition of American education research has gone to lengths to insist that schools are knowledge-transfer institutions whose quality is measurable. There are good political reasons for that tradition. School funding follows test scores. Accountability runs on test scores. The unsaid thing in American education is that, for most students most of the day, school is a structured social institution whose academic output is a byproduct of having a few hundred children spend their waking hours in a building together. The phone ban is the rare policy that takes the byproduct out of the equation and works the building itself.

A serious objection to this reading deserves to be made in its strongest form. The objection is that calling school a social institution is itself an evasion — that the real failure of the ban is to be sentimental about analog teenagers. The skeptic case, made by writers who have looked closely at phone use and learning, runs roughly like this. Phones are not a foreign object in modern adolescence; they are how a generation already learns to read, write, organize, and find each other. Banning the device that does most of that work and then asking whether anything is better sets up a false test. The honest answer is that something has been removed and nothing has been put in its place. The kid in 2026 still needs digital literacy, still needs to manage a notification stream, still needs to learn how to be reachable without being captured. The ban does not teach any of that. It delays the lesson. The empirical null in the study is not a confused metric; it is the absence of an intervention that does the harder thing.

The steel-man is right about one piece and wrong about the rest. It is right that the ban is not, by itself, a digital-literacy programme. Children leaving a phone-banned school do walk into an adult world that runs on phones, and the case for actually teaching them how to handle that — through curriculum, not abstinence — is real and ongoing. The bans do not fill that gap, and nobody serious has claimed they do.

Where the objection breaks is in the assumption that the removal achieved nothing. The removal achieved the only thing a school institutionally can: it cleared the channel. The act of clearing the channel is the act of insisting that a school day is a different category of time than the time on either side of it. The same insistence is what the four-day-week experiments are trying to make about non-work hours. The Boston College trial of the four-day week, the largest of its kind to date, did not find that productivity jumped because of the missing day. It found that workers reorganized their lives because the missing day told them what their week was for. Schools are doing the analogous thing with the phone hour by hour. They are not raising the achievement ceiling. They are reasserting the category.

This is the unpopular implication. If the ban worked, and the ban worked on the social and attentional ground rather than the academic one, then the policy debate has to change languages. The next decade of phone-policy work in schools cannot keep promising achievement deliverables it cannot prove. It has to defend the time itself — the right of a teenager to four or six hours a day in which she is not findable, not pinged, not visible to her network. The right not to be reachable. That is the actual product. That is what parents who support the bans are buying.

It is also what parents who support the bans hesitate to say in public, because the argument sounds soft to a budget committee. Soft arguments, as the saying goes, do not survive contact with line items. So the bans were sold as achievement boosters, and now the achievement story is going to be used to take them apart. The lesson of the data is not that phones in classrooms were fine. The lesson is that schools are now, irreducibly, the last building most teenagers spend serious unmediated time in. The phone is not a teaching aid that got out of hand. It is the medium through which the rest of the world keeps reaching in. The ban is the door closing.

The simplest reading of the new study is that the ban is a partial success that cannot show up in the wrong yardstick. The harder reading is that schools have stopped being the kind of institution whose product can be measured in achievement, if they ever were. What the bans deliver is not a higher number. It is an interval of life in which the device is not the third presence in the room. That used to be the default condition of being a child. It is now a policy. The policy is correct. The yardstick is wrong. The next reform anyone wants to argue for has to start by saying what the yardstick should be instead.