Edifying Presentation of Risk Estimates: Not as Easy as It Seems
本文指出，在不确定性下提出风险建议时，若忽视不确定性会陷入四个常见陷阱，并以EPA对致癌空气污染物年超额死亡人数的估计为例，揭示其呈现方式的缺陷。
In the preceding Insight piece, Gray and Graham (GG Finkel, 1990]. Recommendations made in ignorance of uncertainty are vulnerable to four common pitfalls: (1) they do not allow for the weighing of the probabilities and the consequences of error; (2) they do not permit a reliable comparison of the costs and benefits of a portfolio of decisions, so that priorities can be established; (3) they fail to communicate to decision makers and the public the range of control options that would be compatible with different assessments of the true state-of-nature; and (4) they preclude the opportunity for identifying research initiatives that might reduce uncertainty and thereby reduce the probability or the impact of being caught by surprise. In selecting for analysis EPA's estimate of the annual excess death toll from carcinogenic air pollutants, G&G reveal the flaws in a particularly important risk number that EPA presents in a particularly clumsy fashion. As they indicate, it is easy for the reader of Cancer Risk from Outdoor Exposure to Air Toxics to infer that the interval 1734-2597 is meant to subsume the total