Read this word aloud, and then name the colour of its ink: BLUE. If you hesitated on the second part, you just felt one of the most replicated results in psychology.
In 1886 the American psychologist James McKeen Cattell noticed something odd: people read words faster than they name colours or objects. Reading is so practised that it runs by itself. Naming a colour takes a little more effort.
In 1935 John Ridley Stroop turned that gap into an experiment, published in the Journal of Experimental Psychology. He had people name the ink colour of colour words printed in a clashing colour, such as the word "green" printed in red. They were clearly slower than when naming plain colour patches. Going the other way, reading the words while ignoring the ink, the clash made no real difference.
The reason is not that colours are hard. The word wins the race to your brain, and the wrong answer it supplies has to be suppressed before you can say the right one. Stroop's own conclusion was that the difference came down to training: you have read words far more than you have named colours.
Each round has 12 items. Tap the ink colour, not the word. Do all three rounds and compare your speed.
Psychologists have offered a few explanations. The speed-of-processing idea says the word is simply read faster than the colour is named, so it arrives first and the colour answer arrives late. The automaticity idea goes further: reading has been practised so heavily that it starts on its own whenever letters are in view, and it uses up attention that naming the ink also needs. Most researchers favour some form of the automaticity account, and Stroop's own conclusion about training points the same way.
A neat consequence is that the effect depends on you being able to read the script. Someone who has never learned to read a given writing system has no automatic reading to suppress in it, so the words there are just shapes. Your round 1 blocks are the same idea: with no word attached, there is nothing competing with the colour.
Compare round 2 with round 3. In both you see a colour word, but in round 2 the word helps you and in round 3 it works against you. The gap between those two bars is the interference. Pay attention to how it feels, too. In round 3 you may notice your mouth, or your finger, starting toward the wrong answer a split second before you correct it. That correction is the conflict being resolved.
Try going faster on purpose. Many people find that pushing for speed makes the clash worse, because the automatic response is quickest of all, and slowing down by a hair gives the control process time to step in.
Brain imaging studies of Stroop tasks keep pointing to the anterior cingulate cortex, a region that helps detect conflict between competing responses, working alongside prefrontal areas that hold the goal "name the ink". The slowdown is the cost of that conflict being resolved.
That is why the test is still used. Versions of it are used in clinics and labs to measure selective attention and how well people suppress an automatic response.
Any habit that runs on autopilot can interfere with a goal that does not match it. Signs that point left while an arrow points right, a familiar road you take by mistake on the way somewhere new, a word you can't help reading on a billboard: these are Stroop effects in the wild. If you were slower in round 3 than round 1, your brain was not failing. It was doing exactly what years of reading trained it to do.