The Kruskal–Wallis test is a nonparametric test to decide whether k independent samples are from different populations. Different samples almost always show variation regarding their sample values. This might be a result of chance (i.e., sampling error) if the samples are drawn from the same population, or it might be a result of a genuine population difference (e.g., as a result of a different treatment of the samples). Usually the decision between these alternatives is calculated by a one-way analysis of variance (ANOVA). But in cases where the conditions of an ANOVA are not fulfilled the Kruskal–Wallis test is an alternative approach because it is a nonparametric method; that is, it does not rely on the assumption that the data are drawn from a probability ...
Looks like you do not have access to this content.