Stroop test online

Each part asks for the same kind of answer under a different load. Part 1 shows a colour word in plain off-white letters, so the task is word reading. Part 2 shows a bar of colour with no word, which makes it pure colour naming. Part 3 prints each word in an ink colour that never matches it. Comparing part 3 with part 2 isolates the cost of a word you cannot help reading.

An item where word and ink agree is called congruent, and one where they clash incongruent. The bars of part 2 are colour patches, a baseline with no word. The slowing in the incongruent part is the Stroop effect, a reaction time cost in milliseconds.

Classic test in three parts has parts of 12, 24 or 48 items; the default of 24 makes 72 answers in all. The colour set has three, four or six colours: by default red, green, blue and yellow, while the largest adds orange and purple.

Each part opens with its own button and a fixation cross shown for 600 ms. Answers come from buttons, written as colour words or shown as colour swatches, or from the number keys. A press counts as the finger or key goes down, timed from the screen frame that drew the item.

Every colour is the answer equally often within a part, never twice in a row. A wrong answer shows a brief red mark, and the item is not repeated.

The interference score is the median time of part 3 minus the median time of part 2. Put simply, it is how much longer naming the ink of a clashing word takes than naming a plain bar. Only right answers are timed. Wrong ones are counted and shown, but they stay out of the times. The median is the middle time of a part, which a single slow or lucky answer barely moves.

The result also gives the interference as a percentage of the bar time, so faster and slower players can be set on the same footing. A part only counts when at least 50% of its answers are right; below that, no interference figure appears, since a part with that many slips says little about speed. Accuracy and errors sit beside the times, since a fast score full of slips means something different from a careful one.

Take an illustrative run, not a norm, with a bar median of 690 ms and 930 ms for clashing words: the interference is 240 ms, or 35% slower.

A second line sets the bars against the words, 130 ms here: naming a colour is usually slower than reading its name. Results stay in this browser's local storage, with a personal best for each combination of settings and your last ten runs. With no norms, the fair comparison is with your own earlier runs.

Because a practised reader reads a familiar word without meaning to, and that word offers an answer before the ink has been named. Holding back that wrong answer takes time.

The University of Edinburgh's open day practical sums up four explanations. Relative speed of processing: words are processed faster than colours, so the word arrives first. Selective attention: naming a colour needs more attention than reading. Automaticity: reading happens whether wanted or not. Parallel processing: both pathways run at once and compete.

The last grew into parallel distributed processing models, where pathways of different strength compete and attention tips the balance. MacLeod's 1991 review weighed these accounts against some 400 studies.

In current terms the task calls on response inhibition, stopping a strong, ready answer in favour of a weaker, correct one. That is part of cognitive control, within executive function, the abilities that plan and steer behaviour.

John Ridley Stroop described the effect in 1935 in the Journal of Experimental Psychology. His cards held 100 printed items each, read aloud against a stopwatch.

In his first experiment, 70 students read words printed in black (41.0 seconds) and in clashing ink (43.3 seconds). The ink added only 2.3 seconds, or 5.6%, a difference Stroop called not reliable. Reading barely notices the ink.

The second experiment reversed the task. 100 students named the colours of squares (63.3 seconds) and then the inks of clashing words (110.3 seconds): 47.0 seconds more, reported by Stroop as 74.3%. Whole cards were timed, so these figures do not compare with on-screen milliseconds.

In a third, 32 students practised for eight days, 200 reactions a day, and the interference shrank but stayed. Stroop's inks were red, blue, green, brown and purple; yellow was hard to print at equal strength. One participant whose Ishihara test showed red-green trouble was excluded.

The roots go further back. The 1991 review credits Cattell (1886) with showing that naming colours is slower than reading their names. Jaensch is credited with first setting word against colour in 1929. Matching items and single timed answers came after Stroop.

The timed challenge is the quick Stroop test game here, lasting 30 seconds, 60 seconds or 2 minutes. Items arrive in shuffled blocks of eight, and 50% of each block are matching words in their own ink.

The score is the count of right answers, shown with wrong ones, accuracy and the median times for matching and clashing words. When time runs out, the item still on screen is not counted.

Matching words usually speed naming up a little, an effect called facilitation, which MacLeod found far smaller than the cost of a clash. Many studies also add neutral trials, such as a row of letters. In the classic test the bars play that role.

The reverse task, set under "What you name", turns things around: the player reads the word and ignores the ink, and part 3 is set against part 1. A reverse Stroop effect, ink slowing the reading, is possible but small. The 1991 review says it usually needs a change of task.

Pure speed is a separate skill from choosing between colours. The click speed test measures how fast a finger can press with no decision to make.

The Stroop effect is a common practical in A-level psychology, because it gives a clear result with simple kit. A write-up usually follows the same frame.

• Aim: to find out whether clashing ink slows colour naming.
• Hypothesis: naming the ink of incongruent words takes longer than naming colour patches.
• Independent variable: the condition, bars of colour against clashing words.
• Dependent variable: the median reaction time of right answers, with errors recorded.
• Design: every participant takes every condition, a repeated measures design.

After a test, the "Copy table for a spreadsheet" button copies one row per part: items, right and wrong answers, median and mean. Tabs separate the values, so they paste into a spreadsheet as cells.

The "Print three cards" button gives a printable sheet for the paper method. It holds a word card in black, a card of bars and a card of clashing words, 50 items each in 5 columns of 10. Each print is shuffled, with inks darker than on screen so yellow shows on paper.

Pressing a key or button gives a smaller Stroop effect than speaking. MacLeod reports one study with 177 ms of interference for spoken answers and 98 ms for manual ones. Crump and colleagues found 293 ms online with typed answers. These are single studies, not standards.

Screens, keyboards and browsers add a few milliseconds. In the timing study by Bridges and colleagues, most packages kept a precision under 10 ms in every browser. A steady delay largely cancels when one median is subtracted from another, though each single time is approximate.

Taps on a phone are timed the same way. If the page is hidden mid-run, the run pauses, and carrying on shows the item again with a fresh clock.

Colour vision deficiency, often called colour blindness, matters in a test built on inks. According to nidirect, it affects about one in 12 men and one in 200 women, mostly with red, yellow and green. Every colour set here includes red and green.

Remembering a colour is a different skill from naming it under pressure. The colour memory game asks a player to rebuild a colour seen for a few seconds, with no words to get in the way.

Clinics use published versions of the task, and this game is none of them. Scarpina and Tagini's review describes the Stroop Colour and Word Test with a word page, a colour page and a colour-word page, each timed for 45 seconds.

The Lancaster Glossary of Child Development notes that performance is rated by differences in errors and reaction times across conditions. In clinical use, results are read against norms for the person's age and background, which this game does not have.

North Bristol NHS Trust explains that a neuropsychologist runs and interprets an assessment of attention, concentration and processing speed. A browser game cannot stand in for that, and it does not detect or measure ADHD, dementia, concussion or stress.