The Illusion of "Doing Two Things at Once"
You keep a chat window open during a lecture. You check your phone while driving. You code with YouTube playing. We call this "multitasking" and assume the brain is processing two things at once. Cognitive science paints a different picture.
The human brain almost never does true parallel processing. What we call multitasking is actually rapid task switching. The moment you glance at chat, the lecture pauses; the moment you return, the chat pauses. The switching is so fast it feels simultaneous, but both tasks aren't running at the same time. Decades of research, including work by David Meyer and David Kieras, consistently point this way.
Switching Cost
Every task switch carries an invisible tax called switching cost. When you move from one task to another, your reaction speed dips briefly and accuracy drops. The cost is typically measured in tens to hundreds of milliseconds. One switch looks small, but if you flip to chat several times a minute, the daily loss adds up fast.
Why does this cost exist? To perform a task, its rules, goals, and current state have to be loaded into working memory. Switching tasks means unloading the previous state and reloading the new one. This "reset" costs time and accuracy. Even switching between two identical tasks incurs a cost; different content costs more.
Working Memory Runs the Switch
This is where working memory comes in. Task switching is a job working memory handles directly — holding the current goal, suppressing interference from the previous task, and quickly loading new rules.
Several studies find that smaller working memory capacity means larger switching costs. On the same switching task, people with roomier working memory adapt faster, while those for whom it's a strain take longer to recover from each switch. So "being good at multitasking" isn't the brain doing two things at once — it's recovering from switching costs quickly.
Why It's Dangerous in Real Life
A switching cost of a few hundred milliseconds can be fatal in everyday settings.
- Driving + phone: at 100 km/h, a car travels about 28 m in one second. A 300 ms switching cost alone means roughly 8 m traveled with no reaction. Recognizing a light change or a braking car is delayed by that much.
- Lecture + chat: the 2 seconds you spend on chat misses one sentence of the lecture. If that sentence set up what came next, even returning leaves the understanding broken.
- Coding/work + notifications: deep-focus tasks are especially costly to switch. Some research suggests it can take over 10 minutes to return to the original state.
The common thread: one side is delayed or dropped entirely. Because the brain alternates rather than parallelizes, gaps open between switches. In driving those gaps become accidents; in lectures they become holes in understanding.
Single-Tasking Is Faster and More Accurate
The flip side: focusing on one task at a time is the fastest, most accurate mode, because you pay no switching cost. The 암산 game is a good example. Mental arithmetic is a single task — see a number, compute, answer — driven end to end. Interrupt it with anything and you feel speed and accuracy drop immediately.
Try it and the difference is clear. Solve 5 arithmetic problems while checking chat once per problem, then solve 5 straight and check chat after. Most people find the latter faster and cleaner. It's not that the brain can't do two things — it's that avoiding the switching cost is simply better.
Are There "Good Multitaskers"?
Sometimes someone claims, "My efficiency doesn't drop when I multitask." But studies tend to find that people who rate themselves as good multitaskers actually show larger switching costs. Subjective feeling and objective performance don't line up. You feel like you're doing both; measurement shows one side lagging.
There are exceptions. When one task is so automated it needs almost no conscious control — walking while holding a light conversation, say — combining is possible, because one side barely uses resources. But two tasks that both demand cognitive attention are a different story. If both need attention, switching cost is unavoidable.
In Short
In cognitive science, human "simultaneous" processing is mostly rapid task switching. Each switch causes a brief dip in speed and accuracy, and working memory bears that load. Driving and phone use, lectures and chat, are dangerous because of these gaps. Single-tasking is the fastest, most accurate mode, and a single-task game like 암산 lets you feel it directly. Claims that "brain training improves multitasking" rest on weak evidence — what's measured is the size of switching cost, not the brain starting to parallelize.
Note: results from this site are not a medical diagnosis or a substitute for an attention assessment. They're a reference point for observing your own state.