Margen
You have four cards in front of you. Each has a letter on one side and a number on the other, though you only see one face of each: E, K, 4 and 7. Someone tells you a rule —"if a card has a vowel on one side, it has an even number on the other"— and asks you a single question: which cards must you turn over, and only those, to find out whether the rule is false.
Almost everyone answers the same: the E and the 4. The E is obvious; if it has an odd number behind it, the rule falls apart. The 4 is the trap. Turn it over and whatever comes up, the rule still stands: if you find a vowel you confirm it, and if you find a consonant nothing happens either, because the rule says nothing about what consonants carry on their backs. Turning the 4 cannot change anything. It's work done for pleasure.
The card almost no one touches is the 7. If there is a vowel behind the 7, the rule is disproved on the spot: a vowel with an odd number. The correct answer is E and 7. In the original experiments about one in ten people got it right.
The setup is by Peter Wason, a psychologist at University College London, who published it in 1966 and repeated it for years with students, professors and professional logicians. He called it the selection task, and what it measured was not intelligence: it was where people go to look when they want to check something. They go looking for what fits. They pick the cards that, if they turn out well, prove them right, and not the ones that, if they turn out badly, take that away.
That is confirmation bias, and this is its laboratory version: it's not that you reject the evidence that contradicts you, it's that you don't even go looking for it. The card with the 7 is on the table, within reach, and you don't touch it.
Now the part that changes things. In 1982, Richard Griggs and James Cox, at the University of Florida, set up the same test in different clothes. Four cards again, but now each one is a person sitting in a bar: on one face is what they are drinking and on the other their age. The rule is "if someone is drinking beer, they are over nineteen". The visible cards are: drinking beer, drinking soda, twenty-two years old, sixteen years old.
This one you solve in three seconds. You have to look at the beer drinker and the sixteen-year-old. The soda drinker's age doesn't matter, and the twenty-two-year-old's drink doesn't matter. Most participants got it right —around three in four— and it is exactly the same logical problem as the one with the letters and numbers. The only thing that has changed is that now you know what you are looking for: whoever is breaking the rule.
Why this happens is still debated. One explanation says the brain has specific training for catching whoever breaks a deal, because living in a group demands it. Another, more sober, says nothing so exotic is needed: when the problem has familiar content and a rule about permissions and obligations, you understand what the question is about, and with letters and numbers you don't. The two explanations agree on the practical point: the abstract problem goes unsolved because you don't know what the proof that you are wrong would look like.
And there is the usable part. When someone defends an idea —a colleague, a headline, you yourself— they almost always bring the E and the 4: the cases that turned out well. A method that worked with two clients. A remedy that did wonders for your brother-in-law. None of that is a lie, and none of it is worth much, because they are cards that cannot disprove anything.
So turn it into a question you can write down. Before you close the conversation or buy what you are being sold, take a piece of paper and finish this sentence: "I would know this is false if I saw ." If you can't fill in the blank, the idea isn't an idea yet, it's a preference. And if you can fill it in, go and look for that first, before any success story: that's your 7 card.