Using SAS for Nonparametric Statistics

Definition A nonparametric procedure is a statistical procedure that has certain desirable properties that hold under relatively mild assumptions regarding the underlying population(s) from which the data are obtained – Hollander and Wolfe. The definition is not only vague but it is also extremely inclusive. “Statistical procedure” can refer to a test or an estimator. There are a number of new methodologies that can be seen as nonparametric. This discussion will be limited to traditional rank based analyses. In this presentation we will exploit the idea that if the data are replaced by ranks then we can often apply the usual computations to the ranks and arrive at a test that is equivalent to an established nonparametric procedure.