Some distribution-free k-sample rank tests of homogeneity against ordered alternatives

Abstract

A problem which occurs frequently in statistical analysis is that of deciding whether several samples should be regarded as coming from the same population. This problem, usually referred to as the k-sample problem, when expressed formally is stated as follows: Let Xii,j=1,···,mi,i=1,···,k, be a set of independent random variables and let Fi(x) be the probability distribution function of Xii . The set of admissible hypotheses designates that each Fi belongs to some class of distribution functions Ω.