A more powerful familywise error control procedure for evaluating mean equivalence. Issue 11 (1st November 2020)

• Record Type: Journal Article
• Authors: Davidson, Heather; Cribbie, Robert A.
• Is Part Of: Communications in statistics. Volume 49:Issue 11(2020)
• Page Start: 2914
• Page End: 2929
• Publication Date: 2020-11-01
• Subjects: Equivalence testing -- Familywise error -- Stepwise multiple comparison procedures; Mathematical statistics -- Periodicals; Mathematical statistics -- Data processing -- Periodicals; Digital computer simulation -- Periodicals; 519.5
• DOI: 10.1080/03610918.2018.1530783
• Languages: English
• ISSNs: 0361-0918

Abstract

Abstract: When one wishes to show no meaningful differences among group means, equivalence tests should be used, as a nonsignificant test of mean difference does not provide evidence supporting equivalence. This research proposes two modified stepwise procedures for controlling the familywise Type I error rate, based on the Bonferroni-type correction of k 2 /4 (where k is the number of groups to be compared) proposed by Caffo, Lauzon and Rohmel (2013 Correction to "easy multiplicity control in equivalence testing using two One-Sided tests. The American Statistician 67 (2):115–6) Bonferroni-type correction of k 2 /4 (where k is the number of groups to be compared). Using a Monte Carlo simulation method, we show that adopting a stepwise procedure increases power, while maintaining the familywise error rate at or below α . Implications for applied research and directions for future study are discussed.