Ambiguity, Probability, Preference, and Decision Analysis.

The possible influence of ambiguity on decisions about probabilities has received considerable attention, with most modeling efforts focusing on the probabilities. In this article, it is argued that effects of ambiguity on decision making often operate via the decision maker's preferences rather than probabilities. Therefore, a natural way to model ambiguity is in terms of preferences, via the consequence space and utilities. Some difficulties in developing ambiguity-related models to encompass the full range of realistic decision-making problems and in implementing such models are discussed briefly. Finally, the question of whether ambiguity should matter prescriptively is addressed. Copyright 1991 by Kluwer Academic Publishers