Robust Tolerance for Ambiguity

We argue first that "ambiguous prospects" should be defined in a way that is compatible with dictionary definitions and other technical uses of ambiguity. We then define an ambiguous prospect as a disjunction of risky prospects. Then to harmonize some findings in the literature we test the hypothesis that people are ambiguity indifferent up to moderate amounts of ambiguity and are ambiguity averse only for large amounts of ambiguity. To test the hypothesis we found probability equivalents to risk for prospects of varying degrees of ambiguity. Up to an ambiguity range of .7 the equivalents were near that of a normatively equivalent risky prospect while the equivalents of highly ambiguous prospects differed. We interpret the data in terms of competing desires for minimum effort and maximum accuracy.