The Representativeness Heuristic

Definition

The representativeness heuristic is the substitution of "how well does this match my stereotype of the category?" for "how probable is this, actually?" It routinely produces the conjunction fallacy — rating a specific, detailed scenario (A and B) as more probable than a general one it's logically contained within (A alone), because "when you specify a possible event in greater detail you can only lower its probability," yet detail makes for a more coherent, representative-feeling story. Kahneman and Tversky designed the "Linda problem" specifically "to provide conclusive evidence of the role of heuristics in judgment and of their incompatibility with logic."

In the Book

Linda is described as "thirty-one years old, single, outspoken, and very bright... majored in philosophy... deeply concerned with issues of discrimination and social justice." Given a list of possible occupations, participants rated "bank teller and active in the feminist movement" as more probable than "bank teller" alone — even though every feminist bank teller is, by definition, a bank teller, so the conjunction cannot be more probable than the broader category. In an early between-subjects version the effect already showed; in a later within-subject version Kahneman found, to his considerable surprise, that all of roughly ten early respondents ranked "feminist bank teller" as more likely than "bank teller" even when both items sat side by side on the same questionnaire — he calls his memory of discovering this "a flashbulb memory." A companion example, the "Tom W" problem, shows the same mechanism: given a personality sketch that resembles the stereotype of an engineering student, people ignore the base rate of how few students are actually in that field and rank it highly likely anyway, based purely on resemblance to type.

Why It Matters

The representativeness heuristic explains a specific, testable failure that's easy to demonstrate and hard to unlearn: adding vivid, plausible-sounding detail to a story makes it feel more likely even though detail can only make an event logically less likely, never more. It shows up wherever a stereotype or narrative template exists to be matched against — stock pitches, fraud narratives, hiring judgments based on "fit," and forecasts that feel compelling because they're specific and coherent rather than because they're statistically well-supported. The corrective isn't more vivid reasoning but the opposite: checking whether the "obvious" story is a strict subset of a more general, boring, base-rate-respecting one.