Representativeness heuristic
We judge how likely something is by how much it resembles our mental prototype, not by the actual odds.
What it means
The representativeness heuristic is the tendency to estimate the probability that an object or event belongs to a category by how closely it matches a typical or prototypical member, rather than by the relevant statistical evidence. Mechanically, the mind substitutes an easy similarity judgment for a hard probability judgment, so resemblance crowds out base rates, sample size, and prior probabilities. This shortcut generates a family of well-documented errors, including base-rate neglect, insensitivity to sample size, the conjunction fallacy, and the misperception of randomness. It is adaptive in a world where category membership genuinely predicts features, but it misfires badly when a description fits a vivid stereotype that is statistically rare. Critics note that performance improves when problems are posed in natural frequencies rather than single-event probabilities, suggesting the bias is partly a matter of representation rather than a fixed defect. It matters in practice because clinicians, recruiters, investors, and forecasters routinely confuse a convincing profile with a probable one.
Examples
Told someone is 'shy and tidy,' people guess 'librarian' over 'salesperson' — ignoring that salespeople vastly outnumber librarians, so the base rate swamps the resemblance.
A patient with textbook symptoms of a rare disease gets diagnosed with it, when those same symptoms are far more often produced by a common illness that simply looks less dramatic.
Six reds in a row at roulette does not look like chance, so players pile onto black — even though the wheel has no memory and the odds have not moved at all.
First described in Kahneman & Tversky (1972).