Does It Really Take 23 Minutes to Refocus After a Distraction?

You've seen the number. A notification pulls you away from your work, and somewhere a graphic appears claiming it now takes 23 minutes and 15 seconds to get your focus back. The citation, when there is one, usually points to a 2008 paper by researcher Gloria Mark. I went and read that paper. The number isn't in it. Not close to it, not rounded, not anywhere.

That's not a conspiracy, it's a citation that got detached from its source somewhere around 2010 and has been photocopied ever since. The real story is more useful than the fake one, because the actual research says something different from what the soundbite implies.

The 2008 paper people cite doesn't have the number

The paper usually linked for "23 minutes to refocus" is "The Cost of Interrupted Work: More Speed and Stress," by Gloria Mark, Daniela Gudith, and Ulrich Klocke, presented at CHI 2008. I read the full text. It is a real, careful study, and it has nothing to do with the 23-minute claim.

It's a lab experiment with 48 university students in Germany, who role-played an HR manager answering email, with some groups interrupted by phone or instant message and others left alone. The finding, stated plainly in the abstract: people who were interrupted finished their work faster, with no drop in quality, by writing less and working at a quicker pace. The cost wasn't speed. It was how they felt: more stress, more frustration, more time pressure, more effort, all measured on standard workload scales and all statistically significant.

That's a genuinely interesting, counterintuitive result. It is not, in any sentence of the paper, a claim about 23 minutes. The word "minutes" appears in the paper only to describe task durations of about 20 minutes, the opposite direction of the popular claim.

What the 2005 study actually found

The closer match is an earlier paper by the same lead author, "No Task Left Behind? Examining the Nature of Fragmented Work," published at CHI 2005 by Mark, Victor González, and Justin Harris. This one is field research: 24 information workers at two tech companies, shadowed and timed as they actually worked.

That 25-minute-26-second figure is the real, peer-reviewed number closest to the popular claim, and it's still not 23:15. It also comes with a detail almost nobody repeats: the standard deviation was 54 minutes 48 seconds, nearly double the average itself. That means the data is heavily skewed. Most interrupted work was probably picked back up fairly quickly; a smaller number of long gaps, hours in some cases, dragged the average way up. A single average summarizing that kind of spread is misleading if you treat it as "what happens to me."

Where "23:15" really comes from

The exact figure comes from neither paper. It's from a June 2006 Gallup Business Journal interview with Mark, conducted by Jennifer Robison. Asked about her research, Mark said:

"The good news is that most interrupted work was resumed on the same day, 81.9 percent, and it was resumed, on average, in 23 minutes and 15 seconds."

This is a real statement from the researcher herself, just not one that appears in a peer-reviewed paper with a methods section you can check. It's plausibly a more current cut of her ongoing fieldwork at the time of the interview, close to but not identical to the 25:26 published the year before. In the same interview, she added the detail that matters most and gets dropped every time the number is recycled:

"When you're interrupted, you don't immediately go back to the task you were doing before you were interrupted. There are about two intervening tasks before you go back to your original task."

That sentence is the whole story. The 23 minutes was never a measurement of lost focus. It was the clock time before returning to a specific task, and most of that time was spent doing other work.

What this number does and doesn't mean

Put the two studies together and a more honest summary looks like this:

• It's not 23 minutes of impaired thinking. It's elapsed time until you're back on the original task, filled with roughly two other legitimate activities, not a blank fog.
• The average hides wide variation. A standard deviation of 55 minutes on a 25-minute mean means a lot of interruptions were resolved in a few minutes, and a few outliers skewed the number upward.
• It describes 24 office workers in the mid-2000s. A small field study at two companies is real evidence, not a universal law that applies identically to your job, your brain, or a TikTok notification twenty years later.
• Neither study is about phone notifications specifically. This is office interruption research: phone calls, coworkers, instant messages at a desk. Phone-checking research exists separately, and I've covered what it actually shows in why you keep checking your phone.

What Gloria Mark says now

Mark is still an active researcher, a Chancellor's Professor at UC Irvine, and in 2023 she published a book, Attention Span, summarizing two decades of this work for a general audience. Reporting on the book describes her landing closer to the peer-reviewed figure: about 25 minutes to bring attention back to a task after an interruption, not 23.

She's also been clear on a related point that fits her own data: people interrupt themselves more often than anyone else interrupts them. That matches other research on the checking habit, where roughly 9 in 10 phone pickups turn out to be self-initiated rather than triggered by a notification, which I dig into in the piece linked above.

Frequently asked questions