Base rate fallacy

The base rate fallacy is the error of underweighting a population’s general prevalence when interpreting specific evidence about an individual case.

Bayes’ theorem: A rule for updating probabilities by combining prior probabilities with the likelihood of observed evidence. It gives the mathematical rule that prevents evidence from eclipsing the prior prevalence.
Medical test: A procedure that detects, measures, or helps diagnose a medical condition. A positive result can seem decisive while the condition’s low prevalence keeps false positives consequential.
Conjunction fallacy: The judgment that two events occurring together are more likely than either event occurring alone. The Linda problem combines this probability error with neglect of statistical prevalence.
Bayesian diagnosis: Medical reasoning that updates disease probabilities using symptoms, tests, and prior prevalence. It makes explicit how prevalence and test accuracy jointly shape diagnostic conclusions.
Base rate: The frequency of a condition or event in a defined population before considering additional evidence. This is the population frequency that the fallacy neglects.
Mammography: Medical imaging of breast tissue, commonly used to screen for breast cancer. Screening illustrates how disease prevalence affects the meaning of a positive result.
Representativeness heuristic: A judgment shortcut that estimates likelihood by similarity to a prototype rather than by statistical frequency. Representativeness often drives the vivid, stereotype-based judgments that suppress base rates.
Forensic statistics: The use of statistical methods to evaluate evidence in legal investigations and proceedings. Courts must distinguish evidence strength from the probability that a person committed an offense.
Likelihood ratio: The ratio of the probability of evidence under one hypothesis to its probability under another. It measures how strongly specific evidence should shift odds from the base rate.
Linda problem: A judgment task in which people often rate a conjunction of events as more likely than one event alone. Its vivid description prompts judgments that neglect statistical frequencies and violate probability rules.

Linked from 8 pages