ABSTRACT
In thischapter,ourconcerniswithmultiplecomparisonprocedures(MCPs)thatinvolve comparisons among the group means. Recall from Chapter 11the one-factor analysis of variance (ANOVA) where the means from two ormore samples were compared. What do
we do if the omnibus F test leads us to reject H0? First, consider the situation where there are only two samples (e.g.,assessing the effectiveness oftwo types ofmedication), and H0 has already been rejected inthe omnibus test. Why was H0 rejected? The answer should be obvious. Those two sample means must besignicantly different asthere isnoother way that the omnibus H0 could have been rejected (e.g., one type of medication is signicantly more effective than the other based on an inspection of the means).
