Choice Reaction Time Test

Give someone more options and they get slower, in a very orderly way. Hick measured this in 1952. Hyman found the same pattern a year later, working on his own.

The pattern is not what you might guess. Going from one option to four does not make you four times slower. Each time you double the options you add roughly the same small amount of time. So four choices cost most people 130 to 160 ms more than a single choice.

That gap is the number worth having. Take the simple reaction time test and subtract its median from this one. What is left is what deciding costs you. It is the part of your score that has nothing to do with your eyes, your nerves, your mouse or your monitor. Two people can match on the simple test and differ hugely here.

The four squares are drawn in the shape of the four keys. W sits above the middle of the other three on screen, exactly as it does under your hand. Matching those two shapes matters more than it sounds.

Scramble the mapping so the left key lights the right square and everyone loses 100 ms or more. The test stops measuring decisions and starts measuring how well you memorised the layout.

W, A, S and D take one hand and no reaching. Anyone who has played a game on a keyboard already knows where they are, and you never have to look down. If the cluster is new to you, expect the first few trials to be slow and treat your first run as practice.

The squares are not clickable, and that is deliberate. A mouse answer times how far the cursor had to travel as much as how long the choice took. Two of those numbers are not comparable, so this test only accepts one of them.

You can cheat this test in a way you cannot cheat the simple one. Start your press before you have properly worked out which square lit and your median drops by 40 ms. You pay for it by getting one trial in seven wrong.

So we show both numbers side by side and never blend them into one score. Only correct trials count towards the median, because the timing of a wrong answer measures nothing.

If your accuracy is under about 90 per cent, your median cannot be compared with anyone else's. Slow down and run it again.

This task is one half of the Deary-Liewald pair. They are two free tests, built so reaction time could be measured the same way across very large studies.

Ageing researchers use it because choice reaction time slows with age faster than simple reaction time does. The gap between the two tracks that change better than either number alone.

For how the simple version shifts across a lifetime, see average reaction time by age. For why one number is a weak summary of any of this, see what counts as a good reaction time.