Statistical Approaches to the Fat Tail Problem for Dichotomous Choice Contingent Valuation
针对二分选择数据估计的累积分布函数右尾过厚导致支付意愿均值高估的问题，提出一种新的统计方法，在截断和标准化方法之外给出最佳猜测估计，但该估计因外推而变异性高。
If a cumulative density function estimated from dichotomous choice data has an unrealistically fat right-hand tail, mean willingness to pay (WTP) will be overestimated. Truncating the range of integration results in a lower-bound estimate of the true mean WTP. A normalization procedure proposed by Boyle, Welsh, and Bishop (1988) unnecessarily depresses the estimate further. A new statistical approach is presented that allows a best-guess estimate of mean WTP. Because this estimate involves extrapolation beyond the range of the data, it tends to be highly variable.