The Descriptive Validity of Normative Decision Theory in Auditing Contexts
探讨审计师在不确定性下的决策是否遵循预期效用理论，指出实际决策常偏离该理论，并以阿莱悖论为例说明这种偏离。
Research on the optimal decisions of auditors under conditions of uncertainty has relied on normative models such as expected utility theory to develop optimal decision rules for auditors or to assess the quality of auditors' actual decision in experimental settings (e.g., Kinney [1975a; 1975b] and Joyce and Biddle [1981a; 1981b]). It is well known, however, that individuals generally do not follow assumed optimal decision models either in real life or in experimental studies. For example, see Einhorn and Hogarth [1981] for a general review and Biddle and Joyce [1981] for a discussion of deviations by auditors. One of the more interesting deviations has been termed the Allais paradox. Assuming individuals maximize expected utility (according to the Savage [1954] axioms), they should choose options Al and A2 or options B1 and B2, given the following decision problem: I. Do you prefer situation Al or B1? Situation A,: Certainty of receiving 1 million dollars Situation B1: 10 chances in 100 of gaining 5 million dollars 89 chances in 100 of gaining 1 million dollars 1 chance in 100 of gaining nothing II. Do you prefer situation A2 or B2? Situation A2: 11 chances in 100 of gaining 1 million dollars 89 chances in 100 of gaining nothing