Quantile-preserving spread
提出分位数保持展形这一分布函数的偏序关系，适用于序数数据，给出五种等价刻画并比较其与不同变异性概念的关系。
Quantile-preserving spread is a partial ordering of distribution functions which is appropriate for dealing with ordinal data. This partial ordering is a natural extension of stochastic ordering to cases where the dispersion of the random variables under study is of interest. This paper provides five equivalent characterizations of quantile-preserving spread, studies its consequences and compares it to different notions of increasing variability.