gms | German Medical Science

In multivariate data analysis, the underline research questions usually refer to specific endpoint-wise localizations of significant differences across different treatment groups. Most existing nonparametric methods for the analysis of such data, however, can only be used to test global null hypotheses, i.e., detecting if in any of the group-endpoint combinations significant differences exist. Furthermore, existing nonparametric methods test hypotheses formulated in terms of the distribution functions of the data and thus assume identical covariance matrices across the groups. In this study, we derive pseudo-rank based tests to test hypotheses formulated in terms of purely nonparametric treatment effects. Thus, the new approaches can be used for testing the global null hypothesis as well as for performing multiple comparisons and for the computation of compatible simultaneous confidence intervals. The small-sample performance of the procedures in a simulation study indicates that the proposed procedures control the family-wise error rate quite accurately and have a comparable power compared to rank-based MANOVA-type tests. A real data example illustrates the application of the proposed tests.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.