Stroop Effect: The Stroop Test Explained, With the Original Data

What the Stroop task looks like

Picture the word RED printed in blue ink, then GREEN printed in red ink, then BLUE printed in green. The task is to say the ink colour — blue, red, green — as fast as possible, ignoring what the words say. Most people feel the pull of the printed word on every item. That pull, and the extra time it costs, is the Stroop effect. When the word and the ink match (RED printed in red) the item is called congruent; when they clash it is incongruent.

Stroop’s 1935 experiments, in numbers

Stroop ran three experiments at George Peabody College in Nashville, published in the Journal of Experimental Psychology. The first two are the ones still quoted; the third tested practice.

The asymmetry is the key finding. Colours hardly disturb reading, but words strongly disturb colour naming. In Stroop’s words, practice “decreased the interference of conflicting word stimuli to naming colors but did not eliminate it”. In his practice experiment, eight days of training cut the time on the conflicting cards by 33.9% of the first day’s mean.

Why the Stroop effect happens

For a fluent reader, a printed word triggers its meaning whether you want it to or not. Naming a colour is a slower, less practised route to the same kind of spoken answer. When the two disagree, they compete, and the brain has to suppress the faster, wrong response. Colin MacLeod’s 1991 review gathered half a century of research on this competition and the many variants of the task built since.

This is why the Stroop task is the standard example of selective attention and of the inhibition side of executive function: the information you must ignore is right in the middle of the thing you are looking at.

The Stroop test in clinical use

Psychologists use several published versions. The best known, the Stroop Color and Word Test after Charles Golden (1978), has three timed pages:

1. Word reading — colour names printed in black;
2. Colour naming — patches of colour to name;
3. Colour-word — colour names printed in a clashing ink, where the person names the ink.

In Golden’s version the score is the number of items named correctly in 45 seconds on each page, and an interference score compares the colour-word result with what the first two pages predict. Other versions, such as the Victoria Stroop Test, time how long each card takes and count errors instead. A 2017 review by Scarpina and Tagini found that none of the scoring methods in use fully captures the Stroop effect, and recommended looking at speed and accuracy on all three conditions together.

Results only mean something against norms. In a 2006 Maastricht University study of 1,856 healthy Dutch-speaking adults aged 24 to 81, the timed scores were profoundly affected by age, sex and education, while error counts were much less affected. Age matters in a specific way: a 1998 meta-analysis of 20 studies found that older adults are slower on every part of the test, but that their extra slowness on the conflict items is explained by general slowing rather than a Stroop-specific decline. That is one reason why processing speed has to be separated from interference before a score is read.

Is the Stroop test related to IQ?

Less than many people assume. Three findings set the limits:

• Inhibition is not where intelligence lives. In a 2006 study of young adults by Naomi Friedman, Akira Miyake and colleagues, inhibiting strong responses — measured with tasks including the Stroop — and shifting between tasks had only small, non-significant links with intelligence once their overlap was taken into account. Updating working memory was the executive function strongly tied to IQ.
• Stroop does not travel with other inhibition tasks. A 2018 University of Zurich study gave young and older adults 11 established inhibition tasks, including the colour Stroop. Each task was reliable on its own, but the tasks correlated only weakly with each other — the authors questioned whether “inhibition” is one measurable ability at all.
• A robust effect is not a precise individual score. Cardiff researchers tested seven classic tasks, including the Stroop, on two occasions. Test-retest reliability ranged from 0 to .82 across tasks and was surprisingly low for most of them — because the effect is so similar from person to person that small differences between people are hard to measure. They called it the “reliability paradox”.

So a fast Stroop time says you resist this particular kind of interference well. It does not say your IQ is high, and a slow time does not say it is low. Reasoning tests such as Raven’s Progressive Matrices measure the general ability that the g factor describes far more directly.

The Stroop test and ADHD

The Stroop task is often studied in ADHD, and the evidence depends on how interference is scored. A 2007 meta-analysis from Utrecht University found an average group difference of 0.24 standard deviations with the usual difference score, and argued that Golden’s correction for reading speed is inadequate. With a ratio score, it found more interference in ADHD groups consistently across 19 studies. These are averages across groups. Many individuals with ADHD score in the normal range, so the Stroop test cannot identify ADHD in a person.

The emotional Stroop task

In an emotional Stroop task the words are not colour names but words tied to a concern — for example threat or illness words — and the person still names the ink colour. A 1996 review by Williams, Mathews and MacLeod showed that patients with a range of clinical conditions are often slower to name the colour of words related to their own concern. Researchers use it to study attention to emotional material; it is a research tool, not a diagnostic test.

Can you take a Stroop test online?

Yes — many free versions exist, and they show the effect clearly. Two limits apply. Online versions usually ask you to click a button rather than say the colour aloud, which changes the task, and they have no norms, so the number you get cannot tell you how you compare with other people of your age and education. Treat the result as a demonstration. Practising one will make you faster at it, as Stroop himself showed, but there is no good evidence that this transfers to intelligence; our guide to dual n-back training covers the same transfer question for working memory.

The IQ Revealed test is an online reasoning test of 40 original items. It is not a Stroop test, uses no Stroop materials, does not measure interference control and does not diagnose anything. Our guides to how IQ tests work and the closely related Wisconsin Card Sorting Test explain what such tests can and cannot tell you.

Trust and scope notes

This page explains a psychological effect and the tests built on it in general terms. Clinical Stroop tests are part of assessments that only a qualified professional can administer and interpret for an individual; nothing on this page, or our own test, can stand in for that or for a diagnosis. Colour blindness, reading difficulties and some medicines can also change Stroop results. If you are worried about your own or someone else’s attention or thinking, a doctor is the right first step. Our IQ test is for adults aged 18 and over.

IQ Revealed is not affiliated with, endorsed by or connected to the publishers of any Stroop test version, the publishers of Raven’s Progressive Matrices or the Wisconsin Card Sorting Test, or the researchers and institutions named on this page. Test names are used only to identify the tests being described.