The validity of the test of individual equivalence ratios

Abstract

We evaluate the validity of the test of individual equivalence ratios (tier), a term coined by Anderson & Hauck (1990). The test was also proposed in Wellek (1989). It is proved to be a valid test under a class of unimodal symmetric distributions; this includes normal distributions. Since most bioequivalence studies involve data which are normal, or at least unimodal and symmetric, the test is typically valid. It is, however, not always valid, as shown by counter examples.