Multiple testing in 2^k factorial experiments
by
Jane Chang
Department of ASOR, Bowling Green State University
Coauthors: Jason Hsu (Department of Statistics, The Ohio State University)

In 2^k factorial experiments a total of 2^k-1 hypotheses on the factor effects are simultaneously tested. However most of the literatures have ignored the problem of multiple testing. To address the multiplicity aspect Familywise Error Rate and False Discovery Rate are reviewed and a new concept of error rate is introduced. Also statistical test procedures that control these error rates for 2^k factorial experiments are described.

Date received: August 30, 2003