ABSTRACT
Adaptive designs are usually applied when changes of the experimental design after an interim analysis are intended. Common examples are modification of the sample size, but also changes of the setup of the trial, the test procedure, or the inclusion criteria. When multiple hypotheses are to be tested, the same modifications are possible. The question arises as to whether it would even be possible to change the position or the structure of the hypotheses themselves, given a multiple test problem. The result could be a change of the weights or of the hierarchy of the hypotheses. Of course, this type of action has been applied for a long time in practice, though without fixed decision rules. From the statistical point of view, this has been strongly advised against, since the result is nearly always an excessive number of falsely positive decisions. Therefore, it is not surprising that a positive result could often not be repeated.
