gms | German Medical Science

In many experiments in the life sciences several endpoints, potentially measured on different scales, are recorded per subject. Classical MANOVA models assume normally distributed errors and homogeneity of the covariance matrices, two assumptions that are often not met in practice. In particular, if the observations are not even metric, such applications are no longer possible since means no longer provide adequate effect measures. To this end, several rank-based methods have been proposed for nonparametric MANOVA and repeated measures designs. Null hypotheses are either formulated in terms of distribution functions or in terms of meaningful Mann-Whitney-type effects. The latter has the additional advantage of providing multiplicity-adjusted p-values and simultaneous confidence intervals for subsequent post-hoc tests. Resampling procedures improve the small sample behavior of the proposed methods.

The methods are implemented in an R package, which allows for different factorial designs with nested and crossed factors and comes with a plotting routine and a graphical user interface. Moreover, methods for post-hoc tests are implemented as well. We present the main functionalities of the package and use it to analyze a practical data set.

The authors declare that they have no competing interests.

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