Judgments under uncertainty: Representativeness or potential surprise?

Tversky and Kahneman's (1983) account of conjunctive probability judgment in terms of the representativeness heuristic is questioned. Instead, potential surprise (Shackle, 1969) is proposed as an important mechanism underlying subjective probability judgment. Study I reveals that, consistent with Shackle's theory, probabilities assigned to conjunctions are predominantly determined by the magnitude of the smaller component event probability. Also consistent with Shackle, Study 2 shows that for disjunctions, this role is performed by the larger component. Study 3 again contrasts the relative roles played by the component events in determining the value assigned to the conjunction. The results of the study are consistent with two reasoning processes: one analytically based, in which due account is taken of both component events, the other heuristic in nature and consistent with Shackle's theory of potential surprise. The implications of these results for a range of different types of judgment, e.g. with regard to person perception, stereotyping, categorization, and typicality, are evaluated.