How to stop checking your phone while working, and what the distance is really worth

Maxi Heitmayer sat 22 people in a soundproof room for two five-hour working days and changed one thing between them. On the first day the phone stayed on the desk within arm's reach, inside a rectangle of red tape. On the second it sat on a table 1.5 metres away, far enough that reaching it meant standing up. Phone use roughly halved.

The honest answer to how to stop checking your phone while working starts with where that time went. Leisure minutes on the phone fell from 23 to 16 and non-work pickups from 17.5 a session to 6.5, while leisure use of the computer rose from 27 minutes to 1 hour 6. Total work time did not change, total leisure time did not change, and the number of work and leisure activities people moved through did not change either.

Distance works, narrowly. Moving your phone out of reach cuts how often you pick it up by more than half, and it does not change how much of the day you spend not working. Every page currently on the first page of Google for this question recommends the distance anyway, and not one of them mentions the study that watched what happened next.

Put the phone out of reach and the habit finds your laptop

Heitmayer's participants brought their own devices, kept their own notification settings, and were told they could get up and use the phone whenever they liked as long as they put it back in the red zone afterwards. This was not a detox. It was the single intervention every guide recommends, run for ten hours a person with high-resolution cameras recording the whole thing.

What moved was the device, not the behavior. The interviews afterwards showed where the time had gone: people who normally open WhatsApp or Messenger on the phone opened them on the computer instead, and Instagram and TikTok got scrolled in a browser tab. Heitmayer's own summary is blunter than anything on that first page of results. "The study shows that putting the smartphone away may not be sufficient to reduce disruption and procrastination, or increase focus," he said. "The problem is not rooted within the device itself, but in the habits and routines that we have developed with our devices."

It is one study of 22 people in a soundproof room, and Daantje Derks, who was not involved, told New Scientist that larger studies tracking people in their normal working environments are needed before anyone treats this as settled. That limit is real. Take it as the best available evidence rather than the last word, which is still more than the advice it contradicts has ever carried.

Seven ways to stop checking your phone while working, sorted by evidence

Here is the standard advice, sorted by what has been measured about it rather than by how sensible it sounds.

The rows are not all measuring the same thing, and the table says which is which. Ward 2017 is a lab study of cognitive tests, Heitmayer 2025 is ten recorded hours of real knowledge work, and the last two are field trials that ran for a fortnight.

Notifications are not the thing interrupting you

The most useful number in this literature comes from Heitmayer's earlier work. With Saadi Lahlou, he gave 37 people wearable cameras and collected 200 hours of first-person footage covering 1,130 separate smartphone interactions. Then they coded what had started each one.

The user started 89 percent of them. Eleven percent were triggered by a notification, and of that eleven, 59 percent were silent visual alerts the person happened to glance at. At work the notification share fell further, to 7 percent, against 17 percent when people were not working. One interaction in six was a lock screen check that lasted 5.2 seconds on average and turned up nothing.

That changes what the advice is for. If your phone is already on silent, which most people's is, turning notifications off is a fix aimed at 7 percent of the problem. The rest is you, reaching, about every five minutes, mostly without deciding to.

None of this is visible from the inside, which is the part that makes it hard to fix by trying harder. Sally Andrews and her co-authors logged what 23 people did over two weeks and compared it with what those people guessed: 84.68 phone uses a day against an estimate of 37.20, with the guesses showing no significant correlation with the counts.

Why the other-room study is weaker than it looks

Almost every "put it in another room" article traces back, knowingly or not, to a single 2017 paper. Adrian Ward, Kristen Duke, Ayelet Gneezy and Maarten Bos randomly assigned 520 people to leave their phone on the desk, in a pocket or bag, or in a separate room, then ran working memory and fluid intelligence tests. The other-room group scored highest and the desk group lowest, p = .002 between those two. A second experiment with 275 people found the same ranking on working memory.

Read the details and the advice gets narrower. The pocket condition was never significantly different from either end in either experiment, so the popular version of this finding, that a bag beats a desk, is not in the paper. Switching the phone off changed nothing, which rules out incoming notifications as the mechanism. There was no effect at all on sustained attention. The effects on capacity were small, partial eta squared between .014 and .026, and they showed up only in people scoring high on smartphone dependence: at low dependence, phone location did nothing.

Nobody noticed any of it. Asked whether the position of their phone had affected their performance, 75.9 percent of participants said "not at all," and 85.6 percent said it had neither helped nor hurt them.

Then in 2022 Ana Ruiz Pardo and John Paul Minda ran a preregistered direct replication of the second experiment with a bigger sample, 453 recruited and 383 in the final analyses. It did not replicate. No effect of phone location on working memory, none on the go/no-go task, and no moderation by dependence either. Their conclusion runs to one clause: "the mere presence of one's smartphone may not be enough to affect cognitive performance."

A small effect that failed one well-powered replication is not nothing. It is also not a foundation to build your working day on.

Where the numbers in these guides came from

Three claims turn up on the pages ranking for this question right now, and each one is worth tracing before you repeat it.

TaskGate's guide, published on 13 May 2026, calls the mere presence effect "this 'attention residue,' described by Sophie Leroy (2009)." Leroy's 2009 paper in Organizational Behavior and Human Decision Processes is about switching between work tasks. Her participants solved a word puzzle and then appraised a job candidate, and her finding was that finishing the first task under time pressure is what lets people disengage from it. No phones appear in it.

IntuiWell's 4-Window Reset guide, published on 5 February 2026, says "people who batched their phone checks at fixed times reported 23% less stress." It names no study. The closest real experiment is the Fitz trial in the table above, which reports effect sizes rather than percentages and whose no-notification arm left people measurably more anxious. The same page credits the 85-checks-a-day figure to "research from the University of California." It is Andrews and colleagues, at Nottingham Trent and Lincoln, with 23 British participants.

The 23-minutes-to-refocus line on that page has travelled furthest of the three. The paper people mean is Mark, Gonzalez and Harris in 2005, which reports 25 minutes 26 seconds before an interrupted piece of work was resumed, with 2.26 other pieces of work in between. Their 2008 follow-up found something nobody quotes. Across 48 participants answering email under three conditions, the interrupted groups finished the task faster than the uninterrupted one once the interruption time itself was subtracted out, 20.31 and 20.60 minutes against 22.77, while reporting more stress, frustration, time pressure and effort. The cost of the interruption was real. It just wasn't time.

Schedule the check, and name what you are coming back to

The one intervention in the table with a decent effect and a plausible mechanism is the boring one. In a two-week field trial with 237 people, Nicholas Fitz, Kostadin Kushlev and colleagues intercepted notifications and released them in three batches a day. Inattention dropped, concentration rose, unlocks fell, and people reported feeling more in control of the device. Batching hourly did nothing measurable. Switching notifications off completely made things worse.

Three a day is a starting point rather than a prescription, and what makes it worth copying is the predictability rather than the number. You stop wondering whether anything came in, because you know when you will find out. If you want somewhere to hang those windows, we have written about what the time blocking apps do when a block overruns, which is the part of scheduling that decides whether a schedule survives past Tuesday.

The other half of it is smaller and almost nobody bothers. Mark's 2005 observations found that when those office workers did return to interrupted work, they had passed through an average of 2.26 other activities on the way, which is what makes the return expensive rather than the interruption itself. So write the next physical action down before you stand up, not the project name. "Draft the opening paragraph" survives a trip to the kitchen. "Client proposal" does not. Keeping the whole day down to three commitments makes the same problem smaller, and we argued that case separately.

If you want the heavy version, it exists and it has numbers attached. A 2025 randomised trial blocked all mobile internet on participants' phones for two weeks while leaving calls and texts alone, and in the intervention group daily screen time fell from 314 minutes to 161 while objectively measured sustained attention improved by roughly what ten years of age-related decline costs. Then look at the compliance. Of 467 people who signed up, 266 installed the app and 119 kept the block running for at least 10 of the 14 days. A quarter. That is the realistic ceiling on anything this demanding, and a decent argument for the version you will actually run.

Which leaves three lines worth trying on Monday:

• Pick two or three points in the day when you deal with everything, put them in the calendar, and treat anything arriving outside them as waiting rather than missed.
• Write the next physical action before you get up, in the words you would use to start it, so the return costs you nothing.
• Move the phone if you want, then point the same rule at the browser, because the tab is where the time went last time. If a focus timer is doing that job, the thing worth comparing is what each app does when you abandon a session.