Why Multitasking Doesn't Work: The Evidence in Plain English

Multitasking, as most people picture it, is not something the brain does. Hand it two tasks that both need thought, such as answering an email during a video call, and it does not run them side by side. It takes turns, and every turn costs a little time. That is why multitasking doesn’t work. In four experiments published in 2001, US psychologists Joshua Rubinstein, David Meyer and Jeffrey Evans found that switching between tasks cost time, and cost more when the rules were complex or the new task was unfamiliar.¹

There is a narrow exception: one task on autopilot and one that needs thought, like walking while you talk. For everything else the fix is to switch less often and on purpose: group similar jobs, give unfamiliar work one unbroken stretch, and keep calls away from writing. The wider habits of concentration are gathered in a practical guide on how to focus better.

Prefer to watch? The video covers the same ground as this article.

VideoWhy Multitasking Doesn't Work (and What to Do Instead)WiserHours

Your brain switches between tasks instead of sharing them

When you try to do two thinking tasks at once, your brain switches back and forth between them, and each switch has a price. Psychologists measure that price with task switchingtask switching: Moving attention from one task to another and later back again. What people call multitasking is usually rapid switching of this kind, and each switch and return tends to cost some time or accuracy.Full entry in the glossary experiments. The American Psychological Association (APA) sums up the core result: responses are slower on a trial that follows a change of task than on one that repeats it, even when people know exactly when the change is coming.²

Rubinstein, Meyer and Evans split every switch into two steps. First comes goal shifting: deciding to do this now instead of that. Then comes rule activation: turning off the rules for the old task and turning on the rules for the new one. In their experiments, people moved between different rules for sorting geometric shapes, or between different operations for solving arithmetic problems.¹

Getting ready helps, but only partly. Stephen Monsell, based at the University of Exeter in the UK, reviewed the lab evidence in 2003: responses are substantially slower and usually more error-prone right after a switch, and a chance to prepare shrinks this cost without removing it. Some of the old task’s setup carries over into the new one.³

Each cost is small on its own. The APA’s summary puts it at sometimes just a few tenths of a second per switch, an amount that can add up to a lot when you switch back and forth repeatedly.²

Picture building a slide deck while answering chat messages as they arrive. Each time you go from slide to message and back, you drop one set of rules (the layout, the argument of the slide) and pick up another (who is asking, what they need). No single jump feels expensive. Across a morning you pay for dozens of them, and the task you know least well pays the most.

So group work that uses the same rules, and give unfamiliar work one long stretch instead of dipping into it between other things. Two related costs have their own articles: the mental fog that follows a task switch, and what an interruption costs the task you left.

Why multitasking doesn’t work when both tasks need thought

Two thinking tasks collide because one step in the mind handles a single thing at a time. Harold Pashler, a psychologist at UC San Diego in the US, reviewed experiments on doing two simple tasks at once and in 1994 described a stubborn bottleneck that covers choosing what action to take and probably retrieving things from memory in general.⁴ Both tasks can be under way, but when both need a decision at the same moment, one of them waits.

Driving simulators give the clearest everyday test. In a 2001 study at the University of Utah in the US, David Strayer and William Johnston found that listening to the radio or to a book on tape did not disrupt simulated driving. Phone conversations did: drivers missed about twice as many simulated traffic signals and reacted more slowly to the ones they caught, whether the phone was handheld or hands-free. The authors suggest that a conversation pulls attention into a context of its own, away from the road; in one version of the task, the cost grew as the driving got harder.⁵

A broadcast asks nothing of you. A conversation keeps asking you to understand, decide and reply, which is exactly the kind of step that has to queue.

The same collision happens at a desk. On a call while writing an email, you can follow the words, but the moment someone asks you a direct question, your typing stops, or your answer comes out half-formed. One task has been put on hold without your deciding it.

So a headset, a second monitor or a faster laptop leaves the bottleneck where it is. If both tasks need you to choose words or make decisions, do them one after the other.

When doing two things at once works

Doing two things at once works when one of them runs on autopilot. Walking while talking is the classic case. Walking is one of the everyday skills described as automaticautomaticity: How far a behavior runs by itself, triggered by a familiar context with little thought or effort. Habit researchers use it to measure how strong a habit has become, usually by asking people to rate it.Full entry in the glossary: practiced enough that it no longer occupies the mind, which leaves room to attend to something else.⁶ Automatic is not the same as free. In their 2011 meta-analysismeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary of walking studies, Emad Al-Yahya and colleagues at Oxford Brookes University in the UK reported that a mental task added to walking disturbed it, most clearly its speed, and that in healthy people the slowing appeared strongly linked to age.⁷

The APA’s summary offers a household version: throwing in a load of laundry while talking to a friend will probably work out fine.²

With practice, even two simple decisions can sometimes run together. In three experiments reported in 2001 by Eric Schumacher, David Meyer and colleagues in the US, some participants reached virtually perfect time sharing between two basic reaction tasks after relatively modest practice.⁸ Real work is rarely that simple or that well practiced. And a rare few people seem to pay almost no cost: in a 2010 US simulator study at the University of Utah, about 1 in 40 drivers showed no drop at all when they also did a demanding memory task.⁹

Autopilot also has limits, as the harder simulated drives above suggest. On an icy path or at a busy junction, let the conversation wait.

Small moves that cut the switching

The practical answer is to switch less often and more deliberately. The moves below apply the lab findings above; they have not been tested as habits in offices.

The logic is simple. Grouping similar tasks cuts the number of rule changes, since each switch between different rule sets carries its own cost, and protecting unfamiliar work matters most, because switching into the less familiar task is the slower direction.¹ Keeping calls apart from thinking work follows from the simulator studies: the conversation, not the device, is what competes.⁵

To find your own weak spot, keep a tally for one morning: a mark on paper each time you change task. The count shows which switches you chose and which simply happened to you.

Where the “40 percent” multitasking figure comes from

The popular claim that multitasking costs 40 percent of your productivity is an estimate, and the page that reports it cites no study that measured it. It appears in a line of the APA’s 2006 summary of switching research, which reports that David Meyer has said brief mental blocks from shifting between tasks can cost as much as 40 percent of someone’s productive time.² The 2001 experiments measured time lost at each switch in a lab, not the share of a working day.¹

Myth

Multitasking wastes 40% of everyone's productive time.

Fact

That figure is a researcher's estimate, quoted without a supporting study. What experiments measure is a cost per switch; your total depends on how often you switch.

A fixed percentage invites a shrug, as if the loss were simply how work goes. The per-switch view puts the cost where you have some control: the number of switches. Two people with identical to-do lists can pay very different amounts, depending on whether they jump between tasks every few minutes or finish one piece before starting the next.

Are heavy media multitaskers worse at focusing?

Whether heavy media multitaskers are worse at focusing is still contested. The first famous study replicated poorly, later pooled analyses find a link, and none of the evidence shows which comes first. The idea comes from a 2009 Stanford University study by Eyal Ophir, Clifford Nass and Anthony Wagner. US students who reported using many media at once were more distracted by irrelevant items in lab tasks, and slower to switch tasks, than students who rarely did. Each comparison involved only a few dozen students, and the authors themselves left cause and effect as an open question.¹⁰

Later researchers tried to repeat the result, which is the normal test of a surprising finding.

The study

Mixed evidence

Retested in Indonesia and the Netherlands, the Stanford effect mostly faded (2017)

Wisnu Wiradhany and Mark Nieuwenstein of the University of Groningen in the Netherlands repeated the Stanford tests with students at Universitas Gadjah Mada in Indonesia and at Groningen. Of 14 tests, only five showed heavier media multitaskers as more distractible, and only two of those held up in a stricter statistical analysis. Pooling 39 effect sizes from 12 published studies and their own two experiments, they found a weak link that became non-significant once they corrected for small studies’ tendency to report larger effects.¹¹

Wiradhany and Nieuwenstein conclude that any link between media multitasking and distractibility in lab tasks is likely very small, too small for studies of a few dozen people to detect reliably.¹¹ Melina Uncapher and Anthony Wagner, one of the original Stanford authors, took a cautious middle line in a 2018 review: on balance, heavier media multitaskers perform worse in some areas of thinking, often on tasks that need sustained attention, but many studies find no difference, and the direction of cause remains unknown.¹²

Two larger pooled analyses published since did find a link. In 2023, Fanchang Kong and colleagues at Central China Normal University in China combined 43 studies of teenagers and young adults and found that heavier media multitaskers had weaker inhibitory control and working memory, though no clear difference in cognitive flexibility.¹³ A 2025 analysis of 33 studies by Huohong Chen and colleagues at Hunan Normal University, also in China, linked more media multitasking with poorer attention.¹⁴ Both pool correlations, so neither can say which comes first.

So if you are a habitual tab-juggler, none of this shows that multitasking has rewired your brain: where a link turns up, the arrow may point the other way. In a 2013 study of undergraduates at the University of Utah in the US, the students who multitasked most had lower scores on a test of executive control, and about 7 in 10 rated their own multitasking ability above average. The authors suggest that people often multitask because they are less able to block out distractions, not because they are good at it.¹⁵

Treat the feeling that you are juggling well with suspicion.

The evidence behind each claim, graded

The switching cost itself has the most consistent evidence here; the claims built on top of it are weaker.

The bottom line

Multitasking fails on thinking work because the brain takes turns, and each turn costs a little time, more for hard or unfamiliar tasks. Pair a thinking task only with something on autopilot, and give everything else its own stretch of time.