A colour word printed in ink that may disagree with it. Tap the colour of the letters, not the word. Thirty quick trials, and we measure what the conflict alone costs you.
It measures stopping an automatic answer — your ability to hold back a response that arrives on its own and put the correct one in its place.
For anyone who can read, reading is automatic and cannot be switched off at will. When you see BLUE written in green letters, the word reaches you before the colour does, whether you want it to or not. You then have to stop the first answer before giving the second — and that stop takes time. The time is what we measure.
The effect was first described in 1935, and it remains one of the most reliably replicated results in experimental psychology.
A note on terms: the scientific name for this process is inhibition. We mention it for anyone who wants to read further, and use the plainer phrase — stopping an automatic answer — everywhere else on this page.
It is the headline number in your result:
Interference = your time when word and ink conflict − your time when they match
Why the difference rather than the raw time? Because the difference separates stopping from speed. A slow hand is slow on both kinds of trial, so subtraction cancels it out, and what remains is what the conflict alone cost. Nobody is punished for a slow hand, and nobody is flattered by a fast one.
Lower is better: a small interference means the stop was cheap for you on the day.
Buttons labelled BLUE or GREEN would add a second act of reading on top of the first, so the test would measure a compound interference of unknown size and unknown source. Swatches keep the response colour against colour with no language in it — so what is measured stays the stop itself.
The buttons also keep fixed positions across all trials, so your hand learns them and searching adds no noise. We are measuring stopping, not searching — searching has its own test.
Thirty trials: half with word and ink matching, half conflicting. The split is controlled rather than random, because pure randomness can deal nine conflicts in twelve trials and distort the very difference we are measuring. No more than two trials of the same kind run back to back either, since a run builds expectation and expectation shrinks interference.
Incorrect trials are excluded from the timing and counted only in accuracy. That follows standard practice: the time of a wrong answer does not describe a stop, because no stop happened.
This task depends on colour vision by its nature. In our reaction time test we solved that by showing the target colour as a swatch next to its name — here that is impossible, because showing the colour is giving the answer.
So we chose a palette that never puts red against green, the most common form of colour vision deficiency. That softens the problem; it does not erase it.
If you have a colour vision deficiency, your score here describes your eyes, not your focus. Take from this site what describes you, not what wrongs you — and note that none of the other tests here uses colour at all.
Interference is more sensitive to your immediate state than most measures here: sleep, caffeine, noise around you, even mood. Tens of milliseconds of difference between two sittings is ordinary. The trend across weeks says more than the gap between two attempts an hour apart.
Your interference on this specific task. You will learn not to look at the word at all, and the conflict will shrink. Whether that carries over to resisting distraction at work or in study is far less certain — the evidence for transfer is weak, and we would rather say so than sell you otherwise.
It is free, and it runs directly in your mobile or desktop browser with nothing to install and no sign-up. Signing in is optional, and all it does is save your results so you can follow them from any device.