Representativeness
Definition
The representativeness heuristic is the tendency to judge the probability of an event or the likelihood of category membership by how similar the case is to a prototype — by how representative it is of its category. The hard question (“how probable is X?”) is replaced by an easier one (“how much does X resemble the typical member of category Y?”).
Tversky and Kahneman demonstrated the heuristic through the Tom W. problem: subjects told that Tom W. is “orderly, self-contained, has a need for order and clarity, and a passion for systems” rated him more likely to be a computer science student than a humanities student — ignoring that computer science departments are much smaller. The description fit the CS stereotype, so probability tracked representativeness rather than base rate.
Why it matters
Base rate neglect
The most consequential failure of the representativeness heuristic is base rate neglect: when a description fits a stereotype, the base rate of the category is ignored or drastically underweighted. The question becomes “does this look like X?” rather than “given how common X is, what is the probability?”
This matters in high-stakes domains:
- Medical diagnosis: a symptom that fits the dramatic disease is judged probable even when the disease is extremely rare and common conditions explain the symptom better.
- Personnel selection: a candidate who “looks the part” is judged qualified even when the base rates of success in the role are low.
- Investment: a company with a compelling growth narrative feels like a winner even when base rates for sustained growth are low.
The conjunction fallacy
Representativeness also produces the conjunction fallacy: judging a conjunction (A and B) as more probable than one of its components (B alone). The Linda problem is the canonical example: Linda is described as politically active and socially conscious; subjects rate “Linda is a bank teller and active in the feminist movement” as more probable than “Linda is a bank teller” — even though the conjunction cannot exceed the probability of either component.
The conjunction feels more representative of the description, so it is judged more probable. Probability is being conflated with typicality.
Causal vs. statistical base rates
A critical refinement: people do not ignore all base rates equally. Causal base rates (this cab company has 80% blue cabs — a fact about the world that causally affects outcomes) are used naturally. Statistical base rates (85% of cabs in this city are green — a frequency statistic) are systematically underweighted.
The implication: when base rate information is framed as a specific causal fact rather than a population statistic, it is given far more weight in judgment. This is why storytelling about base rates is more persuasive than presenting them as numbers.
Key takeaways
Key takeaways
- Representativeness heuristic: probability is judged by similarity to a prototype, not by statistical frequency — 'how likely?' is replaced by 'how typical?'
- Base rate neglect: when a case matches a stereotype, the frequency of the category is systematically ignored — prototype fit dominates base rate.
- Conjunction fallacy: a conjunction (A and B) is judged more probable than B alone when A adds representativeness — a direct violation of probability logic.
- Causal vs. statistical base rates: causal base rates (specific facts about the world) are used; statistical base rates (population frequencies) are underweighted.
- High-stakes domains: base rate neglect in medical diagnosis, personnel selection, and investment produces systematically wrong probability judgments.
- Correction: explicitly ask 'what is the base rate for this category?' before evaluating how well the case fits the category description.
Mental model
Read it as: When asked “how probable?”, System 1 answers “how similar to the stereotype?” High similarity triggers a probability judgment that ignores base rates — because the substituted question (typicality) has already been answered. Two signature errors follow: base rate neglect (the frequency of the category doesn’t register) and the conjunction fallacy (more-typical descriptions feel more probable, even when they are logically less probable than simpler alternatives).
Related lessons
Jump to…
Type to filter; press Enter to open