The Pairwise Multiple Comparison Multiplicity Problem: An Alternative Approach to Familywise and Comparisonwise Type I Error Control

When simultaneously undertaking many tests of significance researchers are faced with the problem of how best to control the probability of committing a Type I error. The familywise approach deals directly with multiplicity problems by setting a level of significance for an entire set (family) of related hypotheses, while the comparison approach ignores the multiplicity issue by setting the rate of error on each individual contrast/test/hypothesis. A new formulation of control presented by Benjamini and Hochberg, their does not provi false discovery rate, de as stringent control as the familywise rate, but concomitant with this relaxing in stringency is an increase in sensitivity to detect effects, compared to familywise control. Type I error and power rates for four relatively powerful and easy-to-compute pairwise multiple comparison procedures were compared to the Benjamni and Hochberg technique for various one-way layouts using test statistics that do not assume variance homogeneity.