The purpose of this study is to present and discuss statistical issues as they arise in the design and analysis of gingivitis clinical trials demonstrating either superiority or equivalence. The problems of designing and interpreting data that are obtained from an equivalence trial are being investigated. The basic issues considered include the scale of measurement, the type of statistical test procedure employed, whether to use the difference or ratio of group means, and the type of statistical rule for demonstrating either superiority or equivalence. Their impact on the design and analysis of gingivitis studies focusing on superiority or equivalence was investigated. The best procedures for comparing study groups consisted of the use of either the student t-test (on original measurement scale) or the randomization test; there was little difference when using the ratio of means rather than differences; and there were large differences in effect by choice of statistical rule selected for demonstrating superiority or equivalence.
