Blog · Data · 12 min

Phone addiction statistics: which numbers survive a check

Most phone addiction statistics you have read trace back to a company press release, a sample of ninety-odd people, or a source that nobody has ever been able to find. This page does the checking. Some famous numbers fall apart. A few hold up well enough to build a decision on, and those are the ones worth knowing.

There is no shortage of lists like this one. Type the query and you get forty pages of the same figures, copied sideways from each other until nobody remembers who measured what. So this is not another list. Every number below comes with the study it came from, the number of people in it, the year it was taken and, where it matters, the reason it is weaker than it sounds. Three of the most repeated statistics in the field do not survive that treatment. We start with the most famous one.

The most quoted phone number in the world came off 94 Android handsets

You have seen it: we touch our phones 2,617 times a day. It is in TED talks, in books, in the opening slide of every digital wellbeing deck ever made. It is real, in the sense that somebody genuinely counted. It is just not what people think it is.

The figure comes from Mobile Touches, a 2016 study by dscout, a US research platform. They built software that logged every tap, swipe and pinch on the phones of 94 Android users for five days, around the clock. About two thirds of those people completed all five days; the rest managed two to four. That is the entire evidentiary base for a number that has been repeated for a decade as if it were a law of nature.

None of that makes dscout dishonest. Their deck is upfront about the sample and openly says the true figure is probably higher, because taps on a locked screen could not be captured. It is a good piece of company research. It is not a national statistic, it is not peer reviewed, it says nothing about iPhone users, and it describes a phone landscape from before TikTok existed. Quoting it in 2026 as "we touch our phones 2,617 times a day" is quoting a five-day snapshot of 94 Android owners taken ten years ago.

The most useful thing in that study is a number almost nobody repeats. Before seeing their data, participants were asked to guess. One said, "I will probably touch my phone 500 times today." The real figure was five times that. dscout also found that nearly half of all phone sessions happened without the user ever unlocking the device, and that the average person opened 76 separate sessions a day. The gap between what people believe about their own use and what their phone records is the single most reliable finding in this entire field, and we will come back to it.

Four questions that break most phone addiction statistics

Before you quote anything from here on, including our own numbers, run it through four questions. Most figures fail at least one.

Who measured it, and what do they sell? A phone insurer, a screen time app and a university lab all have reasons to publish a number, and only one of them has a peer reviewer standing behind it. Company research can be perfectly good, as dscout's is, as long as it is labelled as company research rather than dressed up as science.

Was it measured or was it asked? This one matters more than any other. Passive tracking reads the phone directly. A survey asks a person to estimate. Those two methods produce different numbers from the same population, and the difference is not small.

How many people, and for how long? A five-day study of 94 people and a nationally representative survey of 1,391 people are both legitimate and answer different questions. Neither can be turned into "the average person" without losing something.

How old is it, and who was in it? A figure drawn from university students in 2014 tells you very little about a forty-year-old in 2026. If the age of the sample is never mentioned, that is usually because mentioning it would weaken the headline.

Two more famous numbers fail these questions badly, and both are still in circulation this year.

The 23-minute rule is not 23 minutes

The claim goes like this: after an interruption it takes 23 minutes and 15 seconds to get back to what you were doing. It is attributed to Gloria Mark, a researcher at the University of California, Irvine. The attribution is right. The number is not hers.

The paper is No Task Left Behind?, by Gloria Mark, Victor González and Justin Harris, presented at CHI in 2005. Researchers sat next to 24 information workers with a stopwatch and a notepad and timed every switch they made. What the paper reports is an average of 25 minutes and 26 seconds before an interrupted piece of work was picked up again, with a standard deviation close to 55 minutes. It also found that 57% of tasks got interrupted at all, and that 77% of interrupted work was resumed the same day.

So the research is real, and the effect is real. But the figure everybody repeats is not in the paper, the sample was 24 people in two offices, and the fieldwork happened before smartphones existed. The honest version of this statistic is: in one careful observational study of office workers, getting back to an interrupted task took around 25 minutes on average, and it varied enormously. That is still worth knowing. It just is not a rule, and it certainly is not accurate to the second.

The mechanism underneath it is better established than the number attached to it. Returning to a task is not instant, and the cost is paid in the recovery, not in the glance at the screen. We wrote about what that does to a working day in how to rebuild an attention span the phone took apart.

The goldfish statistic, and where it actually came from

The other one you have definitely met: human attention spans have fallen from 12 seconds to 8 seconds, which is shorter than a goldfish. It has run in Time, the Telegraph, the Guardian, USA Today and the New York Times.

In 2017 the BBC programme More or Less went looking for the source. Every trail led to a 2015 report by Microsoft Canada's consumer insights team. Microsoft did run real research for that report, surveying 2,000 Canadians and measuring brain activity in 112 people. But the attention span figure was not theirs. It appears in their report with a citation to a website called Statistic Brain. Statistic Brain listed its own sources as the US National Library of Medicine and the Associated Press. When the BBC contacted both, neither could find any record of research that produced those numbers. Statistic Brain never replied.

The goldfish half is wrong too. Fish researchers point out that goldfish learn and remember perfectly well, and have been shown doing so in scientific papers going back more than a century.

This is what a zombie statistic looks like: a number with no measurable origin, laundered into credibility by one big brand name, then repeated until challenging it feels contrarian. If a figure about attention cannot name the people it was measured on, treat it as decoration.

The phone addiction statistics that do hold up

Now the other side of the ledger. These four are the ones we would still quote after reading the method sections, and each comes with the caveat its own authors put on it.

237 notifications a day, measured off the phones themselves

Common Sense Media's Constant Companion, published in 2023, did what almost no other study in this field does: it read the data off the devices. App tracking software ran on the phones of 203 young people aged 11 to 17 for a week. On a typical day those participants received a median of 237 notifications. They actually saw or engaged with about a quarter of them, a median of 46. They used their phones for a median of almost four and a half hours a day, and picked them up a median of 51 times, with individuals ranging from 2 to 498 pickups.

Two details from that data are worth more than the headline. About 23% of those notifications arrived during school hours. And almost 60% of the participants used their phone at least once between midnight and 5am on a school night. The authors are also straight about the limits: 203 young people is not a national census, and app tracking of this kind could only be done on Android handsets, which skewed the sample towards lower income households. They say so themselves, in the report.

If you have a teenager at home, the number that matters there is not the four and a half hours. It is the 237 interruptions. We went into what to do about it in what to do with a teenager glued to their phone.

A phone on the desk costs you something, even face down

The strongest experimental finding in the field is Brain Drain, by Adrian Ward, Kristen Duke, Ayelet Gneezy and Maarten Bos, published in the Journal of the Association for Consumer Research in 2017. Two experiments, 520 people in the first and 275 in the second. Participants sat cognitive tests with their phone either on the desk, in a pocket or bag, or in another room entirely. Performance on working memory and fluid intelligence tests dropped as the phone got closer, and it dropped even for people who reported not thinking about their phone at all.

This is proper experimental work with a real sample, and it is the one result here that points directly at what to do. Silencing the phone did not fix it. Turning it face down did not fix it. Distance did. Hold on to that, because it is the whole argument at the end of this page.

Sleep: the number the press keeps inflating by a factor of ten

A 2023 meta-analysis in the Journal of Clinical Sleep Medicine by Chu and colleagues pooled 17 studies covering 36,485 participants. The heavy-use group had an odds ratio of 2.28 for poor sleep quality, with a confidence interval of 1.81 to 2.89. Odds of sleeping badly rose by about 4.2% for every extra hour of daily use.

You will see that 2.28 written up as "a 228% higher risk". That is wrong. An odds ratio of 2.28 means the odds were roughly 2.28 times as high, which is not the same thing as a 228% increase in risk, and the two only converge when the outcome is rare. Poor sleep is not rare.

There is a second correction to make, and this one comes from the authors. Almost every study in the pool was cross-sectional, so the direction of the arrow is unknown: heavy phone use at night and bad sleep travel together, but the data cannot tell you which one is dragging the other. Heterogeneity between studies was high. They found evidence of publication bias. They rated their own overall evidence quality as very low. Stated properly it is still a finding worth having. It is just not proof that your phone is causing your insomnia.

What to do about it does not depend on settling that argument. If the phone is the last thing you touch at night, the fix is the same either way, and we laid it out in how to stop using your phone at night.

What people say about their own use, which is its own kind of data

Pew Research Center surveyed 1,391 US teens and their parents between September and October 2024, published as Teens, Social Media and Mental Health. 45% of teens said they spend too much time on social media, up from 36% two years earlier. 44% said they have cut back on their smartphone use, and the same share said it about social media. Among girls it was around half; among boys, 40%.

This is self-reported and Pew says so. It cannot tell you how many hours anyone actually spent. What it can tell you, on a properly weighted national sample, is that the discomfort is widespread and growing, and that a lot of people are already trying to do something about it on their own. Those are two different facts and both are useful.

Why the averages never agree

Read four articles on this and you will get four different daily averages. Nobody is lying. They are measuring different things.

The first source of variation is method. Ask people and you get one number; read it off the device and you get a bigger one. dscout's participants guessed 500 touches and made 2,617. That gap is not a rounding error, it is the central fact of the subject, and it is why a survey average and a tracking average should never be quoted in the same sentence as if they were rivals.

The second is the sample. A study of 11 to 17 year olds and a study of the general adult population will disagree by hours, and both will be right about their own group. The third is definition. Some studies count social apps only, some count all screen time, some count the tablet too. Change the definition and the average moves before anybody has changed their behaviour.

That is why nothing on this page has been blended into one tidy figure. Where two good sources disagreed, we kept both and said why.

Why knowing your number changes nothing on its own

dscout recorded one more thing, and it is the least comfortable finding in this whole article. When participants were shown their real touch counts, most of them needed fewer than ten seconds to go from shock to resignation. Almost nobody changed anything.

That is not a character flaw, and it is not unique to their 94 volunteers. Feeds are built on variable reward: you do not know what the next swipe brings, and the not-knowing is the hook, exactly as it is on a slot machine. Layer a negativity bias on top and you get the pull described in why you doomscroll and how to stop. The practical consequence is that you almost never decide to start scrolling. You notice you already are.

Which is why advice built on awareness has such a low ceiling. It asks for willpower at the precise moment you have least of it. Statistics are awareness. They are the easy half.

What the evidence says actually reduces it

Strip this field down to what has been demonstrated rather than asserted, and one thing survives: friction. The harder an app is to open, the less it gets opened.

Brain Drain is the cleanest demonstration. Across 795 participants in two experiments, what improved cognitive performance was not silencing the phone or flipping it over. It was putting it in another room. Physical distance beat mental effort in a controlled test, which is about as close as this subject gets to a usable instruction.

We have our own data on this, and it deserves the same labelling we have applied to everyone else's. We interviewed our first 100 Monk customers: 95 of those 100 reported cutting their screen time by around 30%. That is a customer survey, self-reported, run by the company selling the product, on people who had already chosen to buy a barrier. Read it as what it is. It tells you something about people who put a physical obstacle between themselves and an app; it does not tell you what would happen to a random stranger.

The Monk is that obstacle. A physical disc about the size of an AirTag, no battery, that blocks the apps you choose. Three steps: you get the disc, you pick the apps, and from then on getting back into one of them means scanning the disc. There is no soft unlock: no ignore-for-today button, no code to type. What the app does hold is three emergency unlocks, and they are finite — a spare key, not a back door. Calls and texts are never blocked and the rest of the phone carries on as normal. Leave the disc in the kitchen and laziness finally starts working for you instead of against you.

With the same honesty as the rest of this page: on iOS there is no block that cannot be worked around. Safari is still there, and so are workarounds. Monk does not claim to be unbreakable. It kills the automatic gesture, which is where the hours in that table actually go. The first three days are the hard part, and everyone finds that.

Frequently asked questions