Problem Solving in Judgment Under Uncertainty

This article presents a problem-solving approach to judgment under uncertainty. According to this approach, the application of statistical rules and nonstatistical rules depends on the same general problem-solving factors. The experiments reported herein used base-rate problems to study four such factors: the prior and immediate activation of the rules, their relation to the goal, and their applicability to the givens of the problem. All of the experiments used purely statistical base rates and highly representative case information. Supporting the prior activation assumption, Experiment 1 found that prior use of base rates enhanced subsequent use of other base rates. Consistent with the immediate activation assumption, Experiment 2 showed that the use of base rates increased with their immediate salience. Experiments 3 and 4 demonstrated that the application of inferential rules is goal driven. Specifically, base-rate and case information were used to the extent that they led to satisfactory solutions. Finally, Experiments 5 and 6 demonstrated subjects' sensitivity to restrictions on the application of the relevant statistical and nonstatistical rules. In these experiments, subjects' reliance on base rates increased with the perceived randomness of the process whereby the case was selected from the population and with the perceived unreliability of the source of case information.