Joint equivalence of means and variances of two populations.

Clinical equivalence has almost exclusively been treated from the view of average equivalence. However, as in bioequivalence, there has been recent interest in more general definitions of clinical equivalence. In particular, Bauer and Bauer (1) have investigated the possibility of the use of a pair of tests for the equality of means and variances taking account of the multiplicity issues involved. In bioequivalence, a number of authors have considered a Bayesian approach; in this article we generalize these previous Bayesian approaches to this more general problem. We investigate the operating characteristics of a decision procedure based on the posterior probability that the parameters lie within a prespecified region of equivalence.