A group sequential procedure for all-pairwise comparisons of k treatments based on the range statistic.

In this paper, a group sequential procedure for all-private comparisons of the means of k independent normal populations with a common known variance is proposed. A repeated range test is defined and its critical points are tabulated. The power function is studied and minimum group size needed to achieve a desirable power level is provided. The test can also be converted into a set of repeated simultaneous confidence intervals from which all-pairwise comparisons can be made. An example of the application of this new interference method is given.