Know Yourself
Most tests on the internet hand you a number and hope you share it. These hand you something you can use, and they are honest about what they cannot measure.
How much your working memory holds
Digits appear one at a time. When they stop, type them back in order. The rows get longer until you miss two at the same length. This is a standard clinical task, unchanged since the 1930s.
What this actually measures — and what it does not
This is working memory: the small board your mind writes on while it is using something. It is not intelligence, not knowledge, and not how good your memory is for your own life.
Most adults land between five and nine. The famous phrase is George Miller's “seven, plus or minus two”, from a 1956 paper — but Miller himself called it a joke about a number that kept following him around. Later work by Nelson Cowan put pure capacity closer to four items when you are prevented from rehearsing them under your breath. Almost everything above four is technique, not capacity.
Which is why the number moves so much: whether you were rehearsing, whether the room was quiet, whether you grouped the digits into pairs. Grouping is the whole trick. It is also why a long score here does not transfer to remembering names at a party.
Three ways your mind takes a shortcut
Nothing here is scored. Answer, then find out what happened to the people this was tested on, and why. The point is not to learn how you did. It is to watch the mistake happen from the inside.
Is the Danube longer or shorter than km? Now write down your best estimate of its actual length.
The number you were shown was chosen by a coin flip.
You wrote km. The Danube runs about 2,850 km through or along ten countries.
This is anchoring. In the original experiments the two groups' estimates pulled apart sharply, and they did so even when people were told the number was random, even when they were warned about the effect, and even among experts estimating in their own field. Knowing about it does not switch it off. The only reliable defence is to work out your own number before you look at anyone else's.
A disease affects 1 person in 100. A test finds it in 90 out of 100 people who have it, and wrongly flags 9 out of 100 people who do not. Your test comes back positive. What is the chance you have the disease?
About 9%. Count 1,000 people.
Ten of them have the disease, and the test catches nine. The other 990 are healthy, and the test wrongly flags about 89 of them.
So about 98 people get a positive result — and only nine of them are ill. Nine out of ninety-eight is roughly one in eleven.
The trap is that the 90% is so loud that the 1% disappears. This is base rate neglect, and it is not a beginner's error: when researchers put versions of this to doctors, large numbers of them gave the wrong answer too. It is the single most useful piece of arithmetic to carry into any conversation about screening.
Tom is 34, quiet, reads a great deal and was good at mathematics at school. Which is more likely?
The first one — always, and no description could change that.
Every person who plays an instrument and works in IT is already inside the group of people who play an instrument. The second group cannot be larger than the first. It is a subset. No amount of detail about Tom can reverse that.
Yet most people pick the second, because it makes a better story. Amos Tversky and Daniel Kahneman named this the conjunction fallacy in 1983; in their best-known version around 85% of people chose the conjunction.
This is worth noticing in yourself, because it is exactly how a detailed, vivid, well-told scenario comes to feel more probable than a vague one — when detail can only ever make something less likely.