Inequality orderings, normalized stochastic dominance, and statistical inference

This article derives large-sample properties and provides asymptotically distribution-free statistical inference procedures for an alternative approach to inequality orderings-normalized stochastic dominance (NSD). NSD is a straightforward extension of standard dominance techniques. As such, it can be generalized to higher orders of dominance, thereby providing a more powerful technique for ranking distributions. We illustrate the NSD method by applying it to Current Population Survey data for 1970-1990.