The problem of multiplicities arises in almost all aspects of scientific studies. When multiple hypotheses are tested in a study, one must adjust the significance level of individual tests to maintain an overall probability error rate. The reasons for this are explored in a non-technical way. Techniques for dealing with multiplicities are described and motivated and guidelines are given, with implications for both study design and publication.
No takes yet. Share an insight, caveat, or question.
William F. Rosenberger (1996) studied this question.