[The matched-pairs differences-W-test for evaluating clinical treatment effects in matched-pair samples (author's transl)].

Starting from the U-test (see Buck 1975), for 2 independent samples of differences (e.g. pre-post differences), a corresponding test for 2 dependent samples of differences of pre- and post-measurements of 2 treatments (e.g. verum vs. placebo) is described (matched-pairs differences-W-test). Clinically relevant applications (verum vs. placebo) of the matched-pairs differences-W-test are mentioned and discussed as to their restrictions.