The study
Mark, G., Gudith, D., & Klocke, U. (2008). The Cost of Interrupted Work: More Speed and Stress. CHI '08 Proceedings.
The counterintuitive result
People completed interrupted work faster, not slower. That is the opposite of the usual claim.
The compensation was measurable and unpleasant: higher stress, higher frustration, greater time pressure and more effort. Participants worked faster to make up the lost time, and paid for it in strain.
So the honest framing is not "interruptions make you slow". It is "interruptions make the same work cost more".
About the 23-minute figure
The widely quoted "23 minutes and 15 seconds to return to a task" comes from Gloria Mark's broader research programme at UC Irvine, not from a single line in this paper.
It gets repeated everywhere with no attribution at all, which is how a real finding turns into folklore. If you cite it, cite Gloria Mark and link her work.
Why this matters for prioritising
If switching is expensive, then the number of times you re-open and re-scan your task list is itself a cost. Every re-scan is a small context switch and a fresh round of deciding.
This is the argument for working from a fixed ranked order rather than re-choosing from the full list each time. You pay the deciding cost once.
Related guides
The Research Behind Lumin
Lumin's design rests on three well-evidenced findings: plan-making removes the intrusive pull of unfinished…
Why Writing Tasks Down Works
Unfinished tasks intrude on attention until they are resolved. Masicampo and Baumeister found that writing a…
Is Decision Fatigue Real?
The evidence is genuinely contested. A 23-lab preregistered replication involving over 2,000 participants…
Frequently asked questions
Does this mean I should never be interrupted?
No — and that isn't achievable for most jobs. The useful takeaway is that the cost is real but paid in stress and effort, so reducing self-inflicted switches is worth more than trying to eliminate external ones you don't control.
Is checking my to-do list a context switch?
A small one. Re-reading the full list means re-evaluating options you already evaluated. That's part of why a ranked order beats a grouped list.