The Cost of Switching. Part One: What It Actually Costs

Abstract

The claim that multitasking makes you slower is one of the few pieces of cognitive science to have escaped the laboratory and become common knowledge. It is also usually stated too crudely to be useful. What exactly costs you — the moment of switching, or something else?

This first part examines a 2018 experiment by Hirsch, Nolden, Declerck and Koch in Advances in Cognitive Psychology, which measured the two standard paradigms of multitasking research in the same people and asked which mental operations they share.

The answer contains a detail that changes how the finding should be applied. The largest part of the cost is not paid at the moment of switching. It is paid for merely being in a situation where a switch might be required.

1.Two Ways of Doing Two Things

The literature separates multitasking into two paradigms, and keeping them apart is the precondition for reading any of it sensibly.

In task-switching, you alternate rapidly between tasks — classify by colour on one trial, by shape on the next. In dual-tasking, two tasks overlap and are performed more or less simultaneously. Both produce measurable performance costs relative to doing one thing alone, and both have accumulated substantial literatures.

The problem the authors identify is that those literatures grew up apart. Costs in task-switching and costs in dual-tasking have largely been explained independently of one another, which leaves an obvious question unanswered: are the same cognitive control components doing the work in both, or different ones?

Their approach was to build highly comparable versions of both paradigms, run the same participants through both, and correlate the costs. If two costs are driven by a shared mechanism, people who pay a lot of one should pay a lot of the other.

2.Two Kinds of Cost, and the Logic That Separates Them

This is the part worth slowing down for, because the distinction is elegant and almost never survives into popular accounts.

Switch costs are calculated as the difference between switch trials and repetition trials within mixed-task blocks — that is, within a block where both tasks occur, comparing the trials where the task changed against those where it did not.

Mixing costs are calculated differently: repetition trials in mixed blocks are compared against trials in single-task blocks. Both sides of that comparison are repetitions. Nothing switched. The only difference is the context — in one case the other task was also in play, in the other it was not.

As the authors put it, since none of the trial types contrasted in mixing costs have a switching requirement, the shifting component is held constant. What remains is the cost of holding two task sets available at once. On their reading, mixing costs mainly reflect working memory processes of maintaining and updating task sets, while switch costs primarily reflect the shifting component.

The practical translation is uncomfortable. You do not only pay when you switch. You pay throughout, simply for keeping the second thing live. A morning spent on one task with email merely available is not the same as a morning spent on one task.

3.What the Correlations Showed

The study's contribution is the pattern of within-subject correlations across the two paradigms. Their summary is precise: mixing costs and dual-task costs are markers of working memory updating, while switch costs — including switch costs in the second task of a dual-task pair — index the shifting component, though partly the updating component as well.

Two things follow. First, dual-tasking is not a separate phenomenon from switching; the presence of switch costs within dual-task trials indicates that shifting is involved there too. Earlier work drew the same conclusion, reasoning that time is needed to disengage from the first task and engage with the second.

Second, and more usefully, the two families of cost load different mental resources. What makes mixed contexts expensive is memory — holding several task sets available and updating which one is current. What makes the switch itself expensive is the shift of attentional set. These are separable, and they respond to different remedies.

4.Preparation Buys Some of It Back

One further finding deserves attention because it is the only actionable one in the paper. Switch costs shrink when there is more time between trials — longer intervals reduce them relative to short ones.

The authors are careful about why, and the care matters. A reduction with longer intervals can be attributed to two different things: active preparation for the upcoming task, or simply the decay of activation belonging to the previous one. Separating them requires manipulating the cue-stimulus and response-cue intervals independently, which the paradigm allows but which most reporting of this literature ignores.

So the honest version is that a pause before switching helps, and part of that help is you getting ready while part of it is the previous task fading on its own. Both are reasons to leave a gap. Neither justifies a claim that a technique trains the cost away.

5.What This Study Cannot Support

Thirty-two participants took part, twenty-nine of them women, aged seventeen to thirty-eight with a mean age of 22.7. That is a small, young, overwhelmingly female sample from a single laboratory.

It is also a correlational design. Finding that two costs covary across people is consistent with a shared mechanism but does not demonstrate one; a third factor — general processing speed, motivation on the day, sleep the night before — could produce the same pattern.

The authors describe their work as a first step towards understanding the relationship between the cognitive control components involved in task set control, which is the right register. It is a well-designed contribution to a specific theoretical question, not a demonstration about office work.

We would add a caution about the tasks themselves. Classifying stimuli by colour or shape on cue is not writing a report while a chat window blinks. The paradigms are built for experimental control, and every step from them to a workday is an extrapolation the study does not make.

6.What Survives Translation

Three things, in our reading, and they are worth more than the usual advice about focus.

The cost of a mixed context is separate from, and often larger than, the cost of the switch itself. Availability is expensive even when unused. That argues for removing the second task from the situation rather than resolving to ignore it — a different intervention from trying harder.

The two costs load different systems: memory for holding sets available, attention for shifting between them. A person who feels scattered in a mixed context and a person who feels slow at the moment of transition are describing different problems.

And preparation time genuinely reduces the switch cost, for at least two distinguishable reasons. That is the closest thing to a lever this literature offers, and Part Three asks whether practice can do more. First, though, Part Two looks at what the brain is doing during all this — where a much larger body of evidence produces a considerably sharper answer.

Editorial Comment

MindHeaven® makes no claim that any product reduces switch costs, mixing costs, or dual-task costs. This study did not test a supplement, and we are not aware of evidence that any compound in our formulations affects these measures. We publish it because the distinction between the two kinds of cost is genuinely useful, and because it is almost always lost in the summary.