Base-rate neglect is the tendency to ignore prior probability — how common something actually is — when a vivid, specific, or representative piece of information is available. Kahneman and Tversky demonstrated it repeatedly: told that a described individual is tidy and detail-oriented, people guess “librarian” over “salesperson,” ignoring that salespeople vastly outnumber librarians. In their taxicab problem, participants judging a witness’s report systematically discounted the base rate of cab colors in favor of the witness’s stated accuracy. The mind substitutes representativeness — how well the case fits a pattern — for probability — how likely it truly is.
As a manipulation lever, the tactic is simple: supply a compelling specific and withhold the prevalence. A wellness pitch profiles a dramatic recovery without noting how rarely it occurs; a fear campaign spotlights a horrifying rare event as though it were common; a scam paints a vivid winner while the odds of joining them go unstated. The most consequential everyday form is the misreading of test results — treating a positive screening result as a near-certain diagnosis without accounting for how rare the condition is and how often the test produces false positives, which for uncommon conditions can make a “positive” more likely wrong than right. In every case the vivid particular does the persuading while the missing base rate does the deceiving.
Recognition rests on hearing the silence where a number should be. If a claim leans on a gripping case, a matching profile, or a scary specific but never states how common the outcome is, the base rate has been suppressed — and the effect is hard to resist precisely because specifics feel more convincing than statistics. The defense is to ask the omitted question directly: how common is this actually? Anchor on the prior probability first, combine it with any test’s error rate, distrust the case that fits a stereotype too neatly, and restate the odds in plain natural frequencies. The vivid story rarely survives being placed next to the real denominator.