J. Ridley Stroop · 1935

The Stroop colour-word test

When a word names one colour but is printed in another, can people ignore the word and simply say the colour of the ink?

  • George Peabody College for Teachers, Nashville, Tennessee (Experiments 1–2); University of Arizona (Experiment 3)
  • Replicated

The question

Long before 1935, psychologists knew something odd: people read a colour word aloud faster than they can name a patch of colour. Cattell, working in Wilhelm Wundt’s laboratory, had reported this in 1886. He thought that reading had become automatic through practice, while naming a colour still needed a deliberate choice.

J. Ridley Stroop, a doctoral student at George Peabody College in Nashville, pushed the idea one step further. What if the two were put in conflict? If a word and its ink colour disagree, which one wins, and how much does the loser suffer? He was interested less in naming speed for its own sake than in interference: what happens when two well-learned reactions compete.

He was not quite the first to combine the two. According to a 1966 review by Jensen and Rohwer, summarised by Colin MacLeod in 1991, the German psychologist Jaensch had already set words and colours against each other in 1929. MacLeod suggests Jaensch deserves at least partial credit. But it was Stroop’s paper, published in English, that set the method everyone later used.

How it was done

Stroop printed the names of five colours (red, blue, green, brown and purple) on cards. On the key cards, every word appeared in an ink colour that did not match it: “red” might be printed in blue, green, brown or purple, but never in red. Each card held 100 items, and each person went through two cards per condition, saying the answers aloud as fast as possible while Stroop timed each card.

He ran three experiments:

There was no congruent condition, with “red” printed in red. Stroop never used one, and he timed whole cards rather than single words. Both are now standard, which is one reason later versions of the task can look quite different from the original.

What came out

Reading was barely affected. In Experiment 1, reading words in clashing inks took on average 2.3 seconds longer than reading them in black, an increase of about 5.6%. Stroop considered this difference not reliable.

Naming the ink was hit hard. In Experiment 2, people took on average 63.3 seconds to name 100 colour squares, but 110.3 seconds to name the ink of 100 clashing words. That is 47 seconds more, an increase of about 74%. Differences between people also grew, roughly in proportion to the slower times.

Practice helped, but did not cure it. In Experiment 3, eight days of practice cut the time on the clashing cards by about a third (from 49.6 to 32.8 seconds per 50-item half-sheet). Plain colour naming improved much less. And something new appeared: after all that colour practice, reading the words in clashing ink suddenly became slower, by about 79%. This “reverse” interference faded quickly, mostly gone a day later.

Stroop’s interpretation was that the difference came from training. People have read words countless times, so the link from a printed word to saying it aloud is very strong; the link from a colour to its name is weaker. When the two compete, the stronger habit intrudes. He also found that women named colours faster than men, and suggested that this too came from upbringing, since girls were more often encouraged to name and talk about colours. Later work, reviewed by MacLeod, found no reliable sex difference in the interference itself.

What really happened (replications and critiques)

One of the most robust findings in psychology. The task then sat largely unused for decades before researchers revived it in the 1960s. By 1991, MacLeod could review some 400 studies (and refer to more than 700 related articles), and he listed 18 findings solid enough that any theory had to account for them. His review also showed that interference is larger than the speed-up when word and colour match; the effect rises in early school years as reading becomes fluent, falls through adulthood, and rises again in old age.

In 2016 the Many Labs 3 project ran a simplified Stroop task as one of ten known effects, in 20 university participant pools in the US and Canada plus an online sample. Across 3,279 people, responses to clashing items were clearly slower, with an effect size (Cohen’s d) of 0.88, which is large by usual standards. It hardly changed with the time of semester, which is what the project was testing.

Why it happens is still argued. Three families of explanation have dominated:

A robust effect is not automatically a good personal test. In 2018 Hedge, Powell and Sumner pointed out a paradox. Effects like Stroop interference are reproducible precisely because nearly everyone shows them in a similar way. That same uniformity makes it hard to tell individuals apart, so the interference score can be a weak measure of differences between people, even though the effect itself is rock-solid. This matters whenever the task is used to rank or compare individuals.

Ethics

There was no deception: participants were told exactly what to do, read words or name colours, and did it. The task is frustrating but harmless. The paper says nothing about consent, except that the students in the first and third experiments volunteered, and there was no debriefing because there was nothing hidden to explain. Today the study would count as minimal risk, and the main change would be a recorded informed consent.

Stroop himself left research soon afterwards. According to a digital archive at Abilene Christian University, he spent around forty years at David Lipscomb College in Nashville as a teacher and administrator, and became known as a preacher and Bible scholar. After 1935 his output in psychology was small.

What it means today

The Stroop task is now a basic tool for studying attention and self-control, and versions of it are used in neuropsychological assessment, for example after brain injury and in research on ageing and ADHD. Its many offshoots replace colour words with other material, such as emotional words or pictures.

The everyday message is simple: a well-practised habit runs even when you have decided not to use it, and holding it back costs time and effort. Some practical ideas follow from this, though they go beyond what Stroop tested:

A Stroop score on its own does not say how attentive or capable a person is. Professionals interpret it alongside other information, and only in context.

Would you have passed?

Coming soon. You’ll see colour words one at a time and name the ink, not the word, first on plain colour blocks and then on clashing words. At the end you’ll see how much slower you were on the clashing words, next to Stroop’s 74%. One honest caveat: Stroop timed whole cards read aloud, while a screen task times single clicks or key presses, so the numbers are not directly comparable. Almost everyone shows the effect; that is the point.

Questions people ask

What is the Stroop effect?

It is the slowdown, and the extra mistakes, that happen when you have to name the ink colour of a word that spells a different colour, such as the word RED printed in green. Reading the word is so practised that it gets in the way of naming the colour.

Has the Stroop effect been replicated?

Yes, many times. A 1991 review covered some 400 studies, and the Many Labs 3 project (2016) found a large effect across more than 3,000 participants at many sites in the US and Canada.

Does practice make the Stroop effect go away?

Not completely. In Stroop's own practice study, eight days of training cut colour-naming time on the clashing cards by about a third, but interference remained. The training also made word reading suffer from colour interference for a short time.

What is the Stroop test used for?

Researchers use it to study attention and self-control. Versions of it appear in neuropsychological assessment, for example after brain injury or in research on ageing and ADHD. A single score is not a diagnosis; it is one piece of information that a professional interprets alongside others.

Sources

  1. Stroop, J. R. (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology, 18(6), 643–662. doi:10.1037/h0054651
  2. MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative review. Psychological Bulletin, 109(2), 163–203. doi:10.1037/0033-2909.109.2.163
  3. Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On the control of automatic processes: A parallel distributed processing account of the Stroop effect. Psychological Review, 97(3), 332–361. stanford.edu/~jlmcc/papers/CohenDunbarMcC90.pdf
  4. Dunbar, K., & MacLeod, C. M. (1984). A horse race of a different color: Stroop interference patterns with transformed words. Journal of Experimental Psychology: Human Perception and Performance, 10(5), 622–639. uwaterloo.ca/memory-attention-cognition-lab/sites/default/files/uploads/files/jepdunbar84.pdf
  5. Ebersole, C. R., Atherton, O. E., Belanger, A. L., Skulborstad, H. M., Allen, J. M., Banks, J. B., et al. (2016). Many Labs 3: Evaluating participant pool quality across the academic semester via replication. Journal of Experimental Social Psychology, 67, 68–82. doi:10.1016/j.jesp.2015.10.012
  6. Hedge, C., Powell, G., & Sumner, P. (2018). The reliability paradox: Why robust cognitive tasks do not produce reliable individual differences. Behavior Research Methods, 50(3), 1166–1186. doi:10.3758/s13428-017-0935-1
  7. Weir, K. (2025, March 24). The Stroop effect's long (and colorful) influence. American Psychological Association. www.apa.org/research-practice/conduct-research/stroop-effect
  8. Abilene Christian University, Milliken Special Collections. About John Ridley Stroop. John Ridley Stroop Digital Archive. blogs.acu.edu/johnridleystroop/about-john-ridley-stroop/