Permutation test for the multivariate coefficient of variation in factorial designs

New inference methods for the multivariate coefficient of variation and its reciprocal, standardized mean, are presented. While there various testing procedures both parameters in univariate case, it is less known how to do setting appropriately. There some existing but they rely on restrictive assumptions underlying distributions. We tackle this problem by applying Wald-type statistics context general, potentially heteroscedastic factorial designs. In addition k-sample higher-way layouts can be incorporated into framework allowing discussion main interaction effects. The resulting shown asymptotically valid under null hypothesis consistent general alternatives. To improve finite sample performance, we suggest permutation versions tests show that tests’ asymptotic properties transferred them. An exhaustive simulation study compares new tests, their counterparts methods. further analyze differences between conduct two illustrative real data examples.