Nobody Asked Where
You will be asked about a colour. The side of the screen it appeared on is random, it is never mentioned, and it will answer anyway. This page measures how long that lasts.
A coloured square appears, off to one side. One colour means the left key, the other means the right key. That is the whole task.
The side it appears on is random, it carries no information, and nothing in the instruction refers to it. You will still be slower when the side disagrees with the hand. Simon and Rudell found this in 1967 and it has never stopped working on anybody, including on people who have just been told about it.
a difference, so your hardware cancels out of it
One number cannot tell you what kind of thing it is. So this page cuts your own reaction times into five speeds, your fastest fifth through to your slowest, and asks how big the effect is inside each one.
If an irrelevant location were an extra stage costing thirty milliseconds, that picture would be flat. It is not. It slopes down, and in plenty of people the slowest bin goes negative. Which means it is not a stage, it is a race against a clock: the side of the screen reaches your hands early and by itself, and then it fades.
Three blocks are fitted and then closed. The last block fades the colours until they are hard to read, which slows you down and does not touch the location at all, and before it opens two rival readings of the closed data each name a number:
They disagree about the direction, which is what makes it a test. The intuition that more time means more interference is the one on the block here.
There is a boring explanation for all of this: a bright thing off to the side yanks your attention, and yanked attention is slow. If that were the whole story it would not matter which way the square was offset.
So one block moves it up and down instead, with the keys still left and right. Same onset, same distance, no left or right in it for a hand to agree with. That block is scored against the one leak there is, which is that up and right feel like the same end of a scale.
The colour decides the key, and this pair is drawn fresh for every session, so nobody's preference for one colour can ride along in the effect. The side decides nothing. Six blocks, about seven minutes, the first one unscored, escape aborts at any time. On a phone the two buttons do the same job and the headline still works, because it is a difference and your touchscreen's lag is in both halves of it. A run below 82 percent correct is refused rather than scored.
Everything runs in your browser. Nothing is recorded. Nothing leaves the page.