Causal Schemata in Judgments under Uncertainty.

摘要

In contrast to the normative theory of evidence, where the impact of data is determined solely by their informativeness, this paper develops the thesis that the impact of evidence on intuitive judgments of probabilities depends critically on whether it is perceived as causal, diagnostic or incidental. The first part of the paper shows that people assign greater impact to causal data than to diagnostic data of equal informativeness. When the same datum has both causal and diagnostic implications, the former dominate the latter. The ease with which people explain unexpected facts and the reluctance to revise old conceptions in the light of new facts are related to the dominance of causal over diagnostic reasoning. The second part of the paper analyzes the use and neglect of base-rate data in terms of the role of these data in causal schemata. It is shown that base-rate information which is given a causal interpretation affects judgments, while base-rate information which cannot be interpreted in this manner is given little or no weight.