Base Rate Fallacy

Definition

Imagine that you meet Tom one evening at a party. He is somewhat shy and reserved, is very analytical, and enjoys reading science fiction novels. What is the likelihood that Tom works as a computer scientist? The answer depends on both the knowledge you have about Tom and the number of computer scientists that exist in the population. Tom fits the stereotype of a computer scientist, but there are relatively few computer scientists in the general population compared to all other occupations. The knowledge you have about Tom is often called individuating or case-based information, whereas knowledge about the number of computer scientists in the general population is often called distributional or base rate information. When presented with both pieces of information—be it when judging the risk of contracting a disease, when judging the likelihood of a defendant's guilt, or when predicting the likelihood of future events—people often base their judgments too heavily on case-based or individuating information and underutilize or completely ignore distributional or base-rate evidence. Underutilizing or ignoring base-rate evidence in intuitive judgments and decision making is known as the base rate fallacy.

Background

The classic scientific demonstration of the base rate fallacy comes from an experiment, performed by psychologists Amos Tversky and Daniel Kahneman, in which participants received a description of 5 individuals apparently selected at random from a pool of descriptions that contained 70 lawyers and 30 engineers, or vice versa. Participants were asked to predict whether each of the 5 individuals was a lawyer or an engineer. The compelling result was that participants' predictions completely ignored the composition of the pool (i.e., the base rates, meaning whether the pool was made up of 30% lawyers or 70% lawyers) from which the descriptions were drawn. Instead, participants seemed to base their predictions of each person's occupation on the extent to which the description resembled, or was similar to, the prototypical lawyer or engineer. Relying on this representativeness heuristic led participants to completely disregard the base rates that should also have been incorporated into their predictions.

Results like these have been replicated in a wide variety of contexts since this initial demonstration. Underutilizing population base rates has been used, for instance, to explain why people are overly concerned about extremely rare events (such as dying in a terrorist attack or contracting a rare disease), why people pay for insurance they do not need, and why doctors misdiagnose their patients. However, broad conclusions about the general existence and robustness of the base rate fallacy in daily life have become quite controversial for two reasons. First, experimental results often show that people do indeed utilize base rates at least some of the time. Empirical research simply does not support the claim that people completely ignore base rate evidence when making judgments and decisions. Second, statisticians have pointed out the difficulty in determining exactly how much people [Page 103]should incorporate base rates into their judgments in daily life. It is therefore difficult, in some contexts, to argue that people should incorporate base rates into their judgments and decisions that they naturally ignore or apparently underutilize.

...