Some apparent violations of the representativeness heuristic in human judgement

According to the representativeness heuristic, the probability that an element is an exemplar of a given class is judged to be high to the extent that the element is representative of the class with respect to its salient features. In three experiments involving situations previously called upon in support of representativeness theory, questionnaire responses from 265 university students demonstrated systematic biases that deviated sharply from the obvious predictions of the theory. One such bias, the students' misinterpretation of proportion information as absolute-number information, is comparable to Piaget's concrete operations. The implications of representativeness theory are discussed in terms of the theory's relationship to concrete thinking, the importance of task characterisitcs, and the difficulty of a priori specification of the salient features with respect to which representativenss is assessed.