Does multitasking actually cost you?

What does switching cost?

Time, reliably. Rubinstein, Meyer and Evans ran four experiments in which people alternated between tasks such as solving maths problems and classifying shapes, and measured what alternation cost against doing the same work in sequence.

Switching produced consistent time costs. Those costs grew as the rules got more complex, shrank when a cue told people which task was coming, and were larger when moving from a familiar task to an unfamiliar one than the other way round.

Their account has two stages: shifting the goal, then activating the rules for the new task. The second stage is where the complexity penalty lands, which explains why switching between two hard things costs more than switching between two easy ones.

So is multitasking bad for you?

That is a different question, and it is the one that fell apart. A widely repeated study in 2009 reported that people who habitually consumed several media at once performed worse on tests of filtering out distraction, and the finding travelled fast.

Wiradhany and Nieuwenstein ran two replications with an average statistical power of 0.81. Across 14 tests, only five reached significance in the predicted direction, and only two survived a more conservative Bayesian analysis.

They then pooled 39 assays across 14 studies and found an association of 0.17. Small, but present.

What happened when they checked for bias?

It vanished. The funnel plot showed small studies clustered toward large positive effects, and Egger's test confirmed the asymmetry at p equals 0.005.

After correcting for that pattern using trim-and-fill, the association became nonsignificant. The authors' conclusion questions whether an association between media multitasking and distractibility in laboratory tasks exists at all.

This is the third time this pattern appears on this site. Ego depletion and growth mindset both shrank the same way, for the same reason: the studies that found something got published, and the ones that found nothing did not.

What survives?

The switching cost, which was always the better-evidenced half. Alternating between tasks takes longer than doing them in sequence, and that holds regardless of what kind of person you are.

What does not survive is the claim that habitual multitasking has damaged you as a general capacity. The measurement used for that, a self-report index of how much media people combine, may be insensitive to how long anyone actually multitasks.

The practical implication changes accordingly. The cost sits in the act rather than in your wiring, which means it is paid fresh each time and avoided the same way.

What is worth doing?

Sequence instead of alternating. The cost is per switch, so two tasks done one after another cost less than the same two interleaved.

Cue the next task before you leave the current one. Costs fell when people knew what was coming, so a note about where you stopped is doing real work.

Group similar work together. Switching penalties rose with rule complexity, which means moving between two tasks that share rules costs less than moving between two that do not.

And stop treating your attention as damaged. The attention span evidence lands in the same place: the behaviour changed, and the capacity is intact.