The Stroop Effect: Why Your Brain Reads a Word Before It Sees the Colour

Write the word RED in blue ink, show it to anybody who can read, and ask them to name the colour of the ink. They will get it right. They will also be noticeably slower than if you had shown them a blue square, and they will feel the hesitation happen. Something in the head says "red" first, and the correct answer has to push past it.

That is the Stroop effect, and it is one of the most reliable findings in psychology. It works on you when you know exactly how it works. It works on the person who designed the test. There is no trick to seeing through, which is what makes it interesting: it is a demonstration that a large part of what your mind does is not under your control at all.

The effect is named after John Ridley Stroop, who published the experiment in 1935 as part of his doctoral work. He did not discover the interference — James McKeen Cattell had noticed as far back as 1886 that naming colours takes longer than reading the words for them — but Stroop measured it properly. In the key comparison, naming the ink colours of a list of conflicting colour words took his participants about 74 percent longer than naming the colours of plain patches. Over a hundred items, that is most of a minute lost to nothing but the words being wrong.

The part people skip is the other half of the experiment, and it is the half that explains everything. Stroop also asked participants to read the words and ignore the ink. That barely slowed them down at all. The interference only runs one way: colour cannot stop you reading, but reading can stop you naming a colour.

The usual explanation is automaticity. For a fluent reader, converting letters into a word is not a task you perform, it is something that happens to you. You cannot look at a word in a language you read and choose not to read it. Naming a colour, by contrast, is a deliberate act — you have to look at it, find the word, and produce it. So the two processes are not evenly matched. They both start, the automatic one arrives first with the wrong answer, and your executive control has to suppress it before the right answer can get out. The delay you feel is the cost of that suppression.

Which is why the Stroop task ended up doing a second job. If the slowdown measures how well you inhibit an unwanted response, then it measures inhibitory control — one of the core executive functions — and versions of it sit in standard neuropsychological test batteries for exactly that reason. Interference tends to be larger in young children who have only recently become fluent readers, smallest in healthy young adults, and larger again in older adults. It is also larger under fatigue, and in various clinical conditions. Worth saying plainly, though: this is a pattern across groups, not a diagnosis for an individual. A slow score on a phone game on a Tuesday night means you are tired.

There is a detail here that matters for anyone raising a child. Pre-readers show no Stroop effect at all. Show a five-year-old who cannot yet read the word RED in blue ink, and they will tell you it is blue without a flicker of hesitation, because to them it is a blue shape. The interference arrives as a side effect of learning to read, and it grows as reading becomes fluent. The test is, in a backwards sort of way, a measure of literacy: being slowed down by it is evidence that reading has become automatic.

For Malaysian readers there is a second version worth knowing about. What happens when the word is in one language and you are answering in another? The answer is that the interference is still there. Show a Malay-English bilingual the word MERAH in blue ink and ask for the colour in English, and it still slows them down. This is one of the standard pieces of evidence that a bilingual does not switch one language off — both are active, all the time, and the word reaches its meaning regardless of which language it arrived in.

The interference is usually a bit smaller across languages than within one, which is what you would expect if the pull comes partly from the word form and partly from the meaning. And the size of it tends to track how strong each language is: the language you read most fluently interferes most, because that is the one you cannot help reading. If you want to know which of your languages is really dominant, that asymmetry is a more honest answer than your own estimate.

This is a good place to repeat something from an earlier post on bilingualism. Doing this every day does not turn you into a person with superhuman focus — the evidence for a broad bilingual advantage in executive function is much more mixed than the headlines of a decade ago suggested. What the cross-language Stroop shows is narrower and better established: both languages stay switched on, whether you want them to or not.

Now the honest bit about the version on CerdikCerdik. The classic test needs you to say the colour out loud, and a phone cannot listen to you, so Color Match uses the standard adaptation: it shows you a colour word in some ink colour and asks a single question — do they match? Press YES or NO. Half the rounds match and half do not, drawn from six colours, and the clock gives you 30 seconds.

Turning it into a two-button judgement does change the task a little. You are no longer suppressing a spoken word, you are comparing two pieces of information that arrive at different speeds and disagree with each other. But the interference is the same interference, and the feeling is unmistakable — the mismatched rounds cost you time, and the errors cluster there. What the format buys you is a score you can compare against yourself, which speech never gave anyone outside a lab.

Two things to try. First, notice where your mistakes land. Almost everyone is near-perfect on matching rounds and loses points on mismatched ones, and if your accuracy is high but your total is low, the cost is showing up as hesitation rather than errors. Second, and this is the good one: play a round in English, then switch the site to Bahasa Malaysia and play again. Same game, same colours, different words. Most people are clearly worse in one of them, and it is usually the language they read most easily — the opposite of what you would guess.

It takes 30 seconds. Try it at cerdikcerdik.com/games/color-match, in both languages, and see which one your brain refuses to stop reading. If you want to separate raw speed from the cost of interference, cerdikcerdik.com/games/reaction-time measures the first with no words involved at all — the gap between the two is roughly what the reading is costing you. And for more on what living in two languages does and does not do to your head, we went through the evidence at cerdikcerdik.com/blog/bilingual-benefits-malaysia.