Does Your Brain Fight Itself?

weirdly accurate.

Does Your Brain Fight Itself?

Does your brain fight itself?

Yes, and you can time it. In the Stroop test you name the colour a word is lit in, and when the word spells a different colour you get slower, because a brain that can read reads before it looks. In the open data of a 2018 study of 103 university students, every one of them was slower on clashing words, and the middle student lost about 67 milliseconds.

This test is the classic Stroop colour–word task from 1935, set in a neon alley and turned into a climb. Each sign spells RED, YELLOW, BLUE or GREEN and glows in one of those four colours; you tap the colour of the light. After two warm-up signs with no clock come ten levels. At each level you get one matching and one clashing sign in a random order, and the light stays on for less time: 1.5 seconds at level 1, down to 0.5 seconds at level 10. Both right before the light goes out passes the level; the first miss gives you a fresh pair at the same level, the second ends the climb. Your result is the shortest light you kept up with. To compare, we replayed each of the 103 students in Hedge, Powell & Sumner’s open data on the same ten levels, 4,000 times, from their own answers.

It is not an IQ test, an attention test or a check for ADHD or any other condition. The students answered on a keyboard in a lab with no clock; you race a light by touch on your own screen, so your place among them is a reference rather than a precise rank. Two misses end a run, so one slip can cost a level, and the same person can do quite differently on another day. If you are red–green colour-blind the result won’t be accurate.

What the test measures

One thing: how short a light you can keep up with when the word in front of you often spells another colour.

The task

  • The Stroop colour–word task (Stroop, 1935): four colour words, RED, YELLOW, BLUE and GREEN, each lit in one of the same four colours
  • Two warm-up signs with no clock, then ten levels; at each, one matching and one clashing sign, and the light stays on for less time: 1.5, 1.25, 1.05, 0.9, 0.8, 0.72, 0.65, 0.59, 0.54 and 0.5 seconds
  • No colour or word repeats back to back, and a dark gap of 400–700 ms before each sign so the moment can’t be anticipated
  • A wrong colour or no tap before the light goes out is a miss; the first miss means a fresh pair at the same level, the second ends the climb, so a run is at most 24 signs

Your result

  • The shortest light you kept up with, in seconds, and how many of the ten levels you passed
  • How far that is among the 103 students in Hedge, Powell & Sumner (2018), replayed on the same climb from their own answers
  • The fight itself in milliseconds — your median time on clashing signs minus matching ones — when you answered at least five of each kind, set beside the 103 students’ scores
  • One of four verdicts, from “barely fought” to “a brawl”, cut where the replayed students fall into roughly four quarters
  • A shareable card and a link that challenges a friend to go faster

Before you start

  • About a minute, free, no sign-up
  • Works on a phone by touch; on a computer the keys D F J K or 1–4 answer left to right
  • Sound is optional: it only marks right and wrong
  • If you’re red–green colour-blind, the result won’t be accurate

Stroop test FAQ

Is this test free? Do I need to sign up to see my result?

Yes, it’s free, and there’s nothing to sign up for: no account, no email, no payment. Your result appears on this page the moment you finish, and you can share it or play again straight away.

What is the Stroop test?

The Stroop test asks you to name the colour a word is shown in while ignoring what the word says. John Ridley Stroop described it in 1935: people are slower and make more mistakes when the word spells a different colour. Colin MacLeod’s 1991 review counted over 700 studies that had repeated the effect, which is why it remains one of the best-known tasks in psychology.

What is a normal Stroop effect?

In the open data of Hedge, Powell & Sumner (2018), the middle of 103 university students was about 67 milliseconds slower on clashing words than on matching ones. Half of them fell between about 51 and 103 ms, and all 103 showed some slowdown. On this page’s climb, half of those students, replayed from their own answers, passed four of the ten levels, a light of 0.8 seconds, or more.

How does this Stroop test calculate my score?

This Stroop test scores how far you climb: ten levels, each with one matching and one clashing sign, the light shorter at each, two misses in all and the climb ends. Your result is the light’s time at the last level you passed. The percentile compares that with the 103 students in Hedge, Powell & Sumner (2018), each replayed 4,000 times on the same levels from their own trial-by-trial answers.

Why did I get a different Stroop score the second time?

Stroop results move a lot from one run to the next, even for the same person. That was the central finding of Hedge, Powell & Sumner’s 2018 paper “The reliability paradox”: the Stroop effect shows up in almost everyone, yet a person’s size of effect is hard to pin down on any single day. Here two misses end a climb, so one slip early can cost you a level or two; several runs tell you more than one.

Does the Stroop test measure IQ or ADHD?

No. This Stroop test measures one thing, how much a written word slows you down when you try to name a colour, and it is not an intelligence test, an attention test or a check for ADHD or any other condition. Clinical versions of Stroop-type tasks are standardised and given by professionals under controlled conditions; a one-minute run on your own screen is not one of them.

Can I take the Stroop test on my phone?

Yes. The Stroop test here is built for touch and takes about a minute; on a computer you can also answer with D F J K or 1–4. Timing starts from the frame each sign lights, but screens and touch panels differ, so compare runs on the same device. If you are red–green colour-blind, the test won’t be accurate for you.

Where the numbers come from

The comparison uses the first session of 103 students (Study 1: 47, Study 2: 56) from Hedge, Powell & Sumner’s public raw data, after the authors’ own exclusions and anyone under 60% correct. Each student was replayed 4,000 times on this page’s ten levels from their own answers, and their milliseconds were scored with this page’s function over all 480 trials. Only those summaries are kept; the trial data is not copied. They answered on keyboards with no clock, so the percentile is a reference, not a calibrated rank. The site does not publish its own players’ data.

How the score is worked out

The climb

Ten levels, each a pair of signs, one matching and one clashing. The light stays on 1.5 s at level 1 and 0.5 s at level 10. A pair passes when both are tapped right before the light goes out. Two misses in all end the run.

Your result

The light’s time at the last level you passed. Most of the time the clashing sign is the one that runs out of time, so the climb shows how fast you can go before reading wins.

Comparison

Each of the 103 students in Hedge, Powell & Sumner’s open data was replayed 4,000 times on this climb: at each level one of their own matching and one of their own clashing answers is drawn at random, and the pair passes if both were right and no slower than that level’s light. Your percentile is the share of those replays that passed fewer levels than you, ties counted half. Half of the replays passed four levels (0.8 s) or more.

The fight in milliseconds

Median time on correct clashing signs minus median on correct matching ones, between 200 and 2,500 ms. With fewer than five of either kind, no number is given.