In 1983, two psychologists named Amos Tversky and Daniel Kahneman introduced the world to a woman named Linda. Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.
Here’s the question: which of these two statements is more probable?
- Linda is a bank teller.
- Linda is a bank teller and is active in the feminist movement.
If you’re like 85% of the people who’ve ever taken this test, you picked option 2. It just feels right, doesn’t it? The description screams “feminist activist.” Option 2 matches the image in your head. Option 1 seems incomplete, almost wrong — “just a bank teller? But look at her story!”
Here’s the problem: option 2 is mathematically impossible to be more probable than option 1. If Linda is a bank teller AND active in the feminist movement, then she is necessarily a bank teller (period). The conjunction of two events can never be more likely than either event alone. It’s like saying “it’s more likely that it will rain AND be a Tuesday than just that it will rain.” It makes no logical sense.
🧠 How the Heuristic Works
This is the representativeness heuristic in action. When asked to judge probability, our brains don’t actually compute probabilities — we substitute a much easier question: “How well does this outcome represent the stereotype?” Linda’s description perfectly matches the stereotype of a feminist activist, so option 2 feels more “representative” — and therefore more likely — even though it’s logically impossible.
Tversky and Kahneman studied this extensively. In another version, they asked about a man named Bill — “intelligent but unimaginative, compulsive, and generally lifeless” — and asked whether he was more likely to be an accountant or an accountant who plays jazz. Same result: people chose the more detailed option because it fit the stereotype, even though adding details can only make something less likely.
The effect is so powerful that even Stephen Jay Gould, the famous evolutionary biologist, admitted: “I am particularly fond of this example [the Linda problem] because I know that the conjoint statement is least probable, yet a little homunculus in my head continues to jump up and down, shouting at me — ‘but she can’t just be a bank teller; read the description.’”
🤔 The Counterintuitive Twist
The most unsettling thing about the representativeness heuristic is that it doesn’t just affect people who are “bad at math.” It affects everyone — including experts. In one study, policy experts were asked to predict the probability that “the Soviet Union would invade Poland and the United States would break off diplomatic relations.” They gave it a 4% probability. Another group of experts was asked to rate just “the United States will break off relations with the Soviet Union.” They gave it only 1%. Adding the condition about Poland somehow made the scenario seem more plausible, even though it should have made it strictly less probable.
The effect persists even when real money is on the line, and even when the problem is presented with intuitive physics scenarios instead of stories. Your brain doesn’t want to do probability — it wants to tell a good story.
🔗 Why It Matters
For anyone building products, writing content, or making decisions, the representativeness heuristic cuts both ways. On the consumer side, it explains why a detailed product description can feel more convincing than a vague one, even when the details add no real information. On the business side, it’s why brand imagery works — if your product “looks like” a premium product, people assume it is one, regardless of the actual specs.
For someone working with divination and personality systems — like Ba Zi (八字) or tarot — this heuristic is especially relevant. The Barnum effect (vague statements that feel personally accurate) and the representativeness heuristic work together: when a reading contains specific, representative details that match the user’s self-image, those details make the entire interpretation feel more “probable” and “accurate,” even though the statistical connection between the details and the conclusion may be nonexistent.
The key insight? A good story will always beat good statistics in the court of human judgment. Knowing this doesn’t make you immune — it just makes you aware of the homunculus in your head that keeps jumping up and down.
🎲 Bonus
Here’s a fun party trick: the representativeness heuristic explains why people fear plane crashes more than car accidents, even though car accidents are vastly more deadly. Plane crashes are “representative” of catastrophe — they make the news, they have dramatic images, they “feel like” a disaster. Car accidents feel like everyday life. Your brain judges the probability of an event by how easily it can picture it, not by the actual statistics. This is also why Jack the Ripper is more famous than the far more prolific — but less “representative” of a serial killer — Dr. Harold Shipman (who killed over 200 patients and looked like your friendly neighborhood GP).