Reaction Time Tests
Five ways of measuring the gap between a signal and your finger.
Reaction time is the cleanest number in cognitive testing. A signal appears, you move, and the stopwatch does the rest. There is very little left over to argue about. That is why the measure has survived from 1868 to now almost unchanged.
The five tests here are not one test with different paint. Each adds exactly one stage to the chain and puts a price on it.
Simple reaction time prices spotting and moving, and nothing else. Auditory reaction time swaps the eye for the ear, and the ear is quicker. Choice reaction time adds a decision between four options, which costs you well over a hundred milliseconds.
Go/no-go keeps that decision but sometimes asks you to do nothing, which is harder than it sounds. The target flick test adds aiming, so the clock keeps running while your hand travels.
Take them in that order. The gaps between your own scores tell you more than any single score does. The gap between your simple and choice times is what deciding costs you. The gap between your go trials and your false alarms is what stopping costs you.
The 5 tests in this domain
Numbers and reading in this domain
What every test on this page is actually timing
Your score is not the speed of your nervous system. It is the total of a chain of events.
- Light leaves the screen and reaches your eye.
- The signal travels to the part of your brain that handles vision.
- A decision gets made, or skipped.
- Your brain sends the command to move.
- Your muscle contracts and the switch in the mouse closes.
- The browser finally notices.
Why a third of your score is not you
Roughly a third of a typical 265 ms score is spent in those last two steps. Neither has anything to do with your brain.
That is not a reason to distrust the number. It is a reason to compare it only against itself, and against people measured the same way.
A browser test overstates your true reaction time by roughly 30 to 80 milliseconds. How much depends on your screen, your mouse and your operating system. The methodology page says exactly how we handle that. We do not quietly subtract a flattering amount the way some sites do.
Why the tests get slower as they get more complicated
More options make you slower, but not in proportion. Four choices cost more than two, and nothing like twice as much. Each doubling adds roughly the same small amount.
This is the most solid finding in the area, and you can reproduce it here in about six minutes. Take the simple and four-choice tests back to back.
The interesting number is the gap, not the total. Two people with identical simple times can differ by 150 ms on the four-choice version. That gap tracks how fast you decide, rather than how fast your nerves or your mouse are. Donders was doing exactly this subtraction in 1868, and it is still why these tasks are worth taking seriously.
How to get a number worth keeping
Reaction time is noisy. A single trial tells you almost nothing. That is why every test here runs several and reports the median rather than the average. One distracted 700 ms trial moves a five-trial average by 80 ms. It barely moves the median.
Tiredness, caffeine, time of day, sleep debt and which hand you use all move the number more than the differences you are trying to spot.
So take the same test at the same time of day on the same hardware, three days running, and compare medians. And if you want to know whether you are slower than last year, you needed last year's number, taken the same way. That is the whole argument for keeping a history.
What the score is not
Across large groups, faster reaction times go with slightly higher scores on thinking tests. The link is far too weak to say anything about one person.
A fast score does not make you clever. A slow one does not make you unwell. If your reaction time has changed in a way that worries you, take it to a doctor rather than to a website with a stopwatch.