Nonparametric homogeneity test based on ridit reliability functional

The paper provides nonparametric test procedures for comparing K(>2) unknown univariate populations, in which the tests are formulated by using consistent estimators of the ridit reliability functionals (see, for example, Bandyopadhyay and De, 2011) for comparing more than two populations. The tests are asymptotically distribution free and can be used for data on both discrete and continuous random variables. An extensive numerical study is performed to compare the proposed test with the nonparametric tests, obtained from Konietschke et al. (2012), in terms of type I error rate and power. A data study illustrates the use of such tests.