Statistical Laws of Confidence Versus Behavioral Response: How Individuals Respond to Public Management Decisions Under Uncertainty
通过实验研究个体能否主观运用推断统计中预测区间和置信区间的生成、使用与解释，发现人们能理解区间准确性与宽度的权衡，但无法根据经验信息校准区间大小，表明人类判断需要外部校准来有效纳入不确定性。
Rational policy analysis confronts the problem of uncertainty directly through various methods of quantification. This paper considers the extent to which individuals are able to subjectively apply the basic tenets of inferential statistics surrounding the generation, use, and interpretation of prediction and confidence intervals. In a small-group experiment, subjects successfully internalized the tradeoff between an interval's accuracy and its sensitivity or width. Unfortunately, subjects could not calibrate the magnitude of the interval based on empirical information about underlying uncertainty (i.e., variance) in the phenomena being estimated. The results, consistent with much of the cognitive psychological research, suggest human judgment needs external calibration to successfully incorporate uncertainty into decisionmaking.