Design and Analysis of Experiments with more than two Treatments – a Review

Summary

In this paper a review over 8 problems of judging a treatments (levels of a factor) are given. Not only the 3 types of multiple comparisons with comparisonwise and experimentwise risks of first and second kind but also comparisons with a standard (comparisonwise and experimentwise risk of first kind) and selection rules are discussed. In the case of selection the problem of selecting the t best out of a populations and the problem of selecting r out of a populations so, that the t best are included in the r selected ones are handled. To each problem hints for determining the minimal sample size and for the analysis are given.