In order to express an opinion on the fairness of the presentation of the financial statements, an auditor collects evidence from a variety of sources. Much of this evidence is in the form of sample results that must be evaluated as to the level of sampling risk present (AICPA [1981, sec. 26]). Auditors often select and evaluate samples judgmentally rather than on the basis of a statistical method.' The assessment of risk, however, is a cognitively difficult task (Huber [1974, p. 434]), and often the results may be systematically biased (Tversky and Kahneman [1974]). It seems that there is a general problem of underutilizing or ignoring normatively