THREE BAYESIAN TEST INDEXES FOR NORMAL DISTRIBUTIONS

Abstract

In clinical trials, data that are consistent with a normal distribution are often regarded as candidates for primary evaluation variables. In such cases, t-tests are frequently used to compare different groups of data. However, the repeated use of t-tests leads to problems with multiplicity. This study proposes three new indexes that are based on (Kawasaki and Miyaoka, 2012; 2013). These indexes can be used to evaluate the superiority, non-inferiority and equivalency of population means for normal distributions. These new indexes are constructed based on the Bayesian framework and can be used to prevent multiplicity issues. We apply these three new indexes to actual data in order to demonstrate their usefulness.