Can You Guess Your Own Screen Time? What the Research Found

Ask most people how much time they spent on their phone yesterday, and they'll give you a number with total confidence. A team of six researchers wanted to know how close that number usually is to what a phone's own log actually shows, so they gathered every study they could find that measured both.

What the researchers did

Douglas Parry, Brittany Davidson, Craig Sewall, Jacob Fisher, Hannah Mieczkowski and Daniel Quintana published their pre-registered systematic review and meta-analysis in Nature Human Behaviour in 2021. They searched six academic databases, hand-searched journals, and put out a public call for unpublished data, turning up 12,132 records in total. After screening, 47 studies made the final sample, each one comparing a person's self-reported estimate of their media use, how long or how often, with a logged measure of the same thing, taken from a phone's own tracking, a carrier's records or an app's usage data. Together, those studies yielded 106 separate comparisons.

What they found

• The match is moderate, not strong. Across 66 comparisons from 44 studies, covering more than 52,000 people, self-reported use correlated with logged use at r = 0.38, a middling relationship, well short of what the researchers would consider close enough to treat self-reports as a stand-in for the real log.
• Guesses about problem use fared worse. Comparing logged use with self-report scales for problematic or excessive use, 40 comparisons from 19 studies and over 5,500 people, brought the correlation down to r = 0.25, a weaker link again.
• Accuracy was the exception, not the rule. Of 49 comparisons between self-reported and logged averages, only 3, about 6%, landed within 5% of the logged number.
• No consistent direction. People were about as likely to overestimate their use as to underestimate it, 23 comparisons each out of 49, so the researchers couldn't conclude that people systematically round their use up or down.
• The kind of media made little difference. Phone use and social media use showed similarly loose correlations, and whether people answered a single estimate or a longer scale didn't reliably change the result either.

What the study can't tell you

This is a meta-analysis, so it inherits the limits of the studies it draws on. Most were convenience samples, and a minority were rated lower quality, mainly for small or poorly justified sample sizes. The studies also varied enormously in how closely their numbers lined up, more than the researchers could explain even after checking whether the medium, the self-report format or the study design mattered.

It cuts both ways, too: logged data isn't flawless either. The authors point out that tracking tools can misread background activity as active use, miss specific apps, or vary by device, so "more objective" doesn't mean perfect. And eligible studies had to be published in 2007 or later, the year the first iPhone came out, so some of the underlying data predates today's mix of apps and short-form video.

What it means for your phone

The practical lesson is simple: don't trust your memory over your phone's own number. If you've never actually opened your Screen Time report, the gap between what you'd guess and what it shows could run either way, and this research found the gap was wider still for people guessing at their own problem use. Before setting yourself a limit, check the real figure for a few days first.

Unrotten works from the measured number, not your guess. You choose the apps that eat your time, your rot apps, and Less scrolling gives them a daily limit: once the time Screen Time actually counted is used up, Apple's Screen Time shield locks them until the day resets. For more on reading your own numbers, see our guides to average screen time and to how much screen time is too much.