The "23 Minutes to Refocus" Stat Has No Study Behind It

"It takes 23 minutes and 15 seconds to refocus after an interruption." You've seen this line. It's in productivity blog posts, corporate wellness slides, LinkedIn infographics, and at least a few articles on this very topic elsewhere on the internet - always the same oddly specific number, always attributed, loosely, to "research" or "a University of California Irvine study." The precision is what makes it convincing. Nobody invents a number that specific without a study behind it.

Except, as far as anyone has been able to trace, nobody did invent it with a study behind it - not this exact figure, anyway.

Why the Specificity Is What Sells It

There's a general pattern worth naming in why "23 minutes and 15 seconds" reads as more credible than a rounder number like "about 20 minutes" would. Precision signals measurement - nobody invents a number down to the second unless they're reporting something actually measured, or so the intuition goes, which is exactly the intuition this figure exploits without anyone necessarily intending to exploit it. Round numbers read as estimates. Oddly specific numbers read as data. That heuristic is usually a reasonable one for evaluating claims quickly, which is precisely what makes it a liability here: the heuristic pushed people toward trusting the figure more, not less, the more suspiciously precise it became.

Where the Trail Actually Leads

The number traces back to a 2006 interview UC Irvine researcher Gloria Mark gave to Gallup, discussing her genuinely real and genuinely rigorous research on workplace interruptions. Somewhere between that interview and the thousands of articles that now cite "23:15" as an established finding, the figure calcified into a fact with a phantom citation. People have gone looking for the primary paper that contains this specific number - checked the studies most commonly listed as the source - and it isn't in any of them. Not because Mark's research doesn't exist or isn't credible; her body of work on interruptions and attention is extensive and well-regarded. It's that this one hyper-precise figure appears to have been generated somewhere in the retelling, then repeated by enough people, enough times, that repetition did the work citation was supposed to do.

That's a fairly common failure mode for statistics that travel well: a real finding gets rounded, simplified, or misquoted once, and every article that copies the previous article instead of the original source inherits the error, with growing confidence, until the number looks more solid the further it gets from the place it actually came from.

How a Number Like This Survives Fact-Checking

What's striking about "23 minutes and 15 seconds" isn't that it went unchecked - plenty of people have gone looking for the source, precisely because the specificity invited scrutiny. It's that the checking itself rarely changes how the number gets used afterward. Someone writes an article noting the citation doesn't hold up; the number keeps appearing in new articles regardless, often published after the debunking existed and was findable. That pattern says less about anyone's carelessness and more about how differently a precise-sounding statistic and a caveat about its sourcing compete for space in a piece of writing. The number is short, quotable, and immediately usable. The correction requires several sentences of context to land properly, and several sentences of context rarely survive the edit down to a punchy line in someone else's article.

A Broader Pattern Worth Watching For

This isn't really a story about one number or one researcher - it's a fairly generalizable case study in how any sufficiently precise, sufficiently citable-sounding figure can outrun its own sourcing once it starts getting repeated by writers who are citing the previous article rather than the original data. The tell is usually the same each time: a number specific enough to feel authoritative ("23 minutes and 15 seconds" rather than "about 20 minutes"), attached loosely to an institution or researcher's name rather than a specific paper, appearing across a wide range of otherwise unrelated sources with suspiciously identical phrasing. None of that proves a given number is wrong on its own. It's just worth treating as a prompt to actually look for the primary source before repeating it - a habit that would have caught this particular figure decades before it accumulated thousands of citations built on nothing more solid than each other.

What's Real, Versus What's Just Well-Traveled

This isn't an argument that interruptions are harmless - Mark's actual published research is fairly unambiguous that switching tasks carries a real cost, and separate work of hers has tracked how average unbroken attention on a single screen has fallen sharply over the past two decades. The interruption-recovery cost is real. The specific number "23 minutes and 15 seconds," attached to every interruption regardless of length, task, or person, is not something that traces back to a measurement - it's closer to folklore that happens to be phrased like a statistic.

What a Trustworthy Version of This Claim Would Look Like

A more defensible version of the underlying point doesn't need a dramatic, precise number to be compelling - it just needs to be honest about its own uncertainty. Something closer to "interruptions carry a real, measurable recovery cost, and that cost appears to compound with frequency and task complexity" is less quotable than "23 minutes and 15 seconds," but it's the version that actually survives someone checking the sourcing. The irony is that the honest version is, if anything, more alarming once you sit with it - a cost that varies and compounds based on your actual work, rather than a flat, universal tax that's the same for everyone regardless of what they're doing.

The distinction matters for a practical reason, not just a pedantic one: a made-up universal constant invites you to shrug ("well, every notification costs 23 minutes, so what's one more") in a way that a more honest, messier truth doesn't. The honest version - that switching costs vary, compound, and are worse for some kinds of work than others - is less quotable. It's also the one worth actually acting on.

Acting on it usually means cutting the interruptions that are actually frequent and actually optional, rather than the mythical constant attached to all of them equally. For most people that's notifications and, more specifically, short-form feeds - which is the one category built to interrupt as often as physically possible. Batching notifications helps with the first. Something like Dam It, which stops the automatic reload inside Shorts, Reels and TikTok, targets the second.