Nonparametric ridit-based methods for comparing multiple treatments under repeated measurement design

This study explores the application of ridit-based nonparametric measures in analyzing repeated measurement design involving K(≥2) treatments, where the response variables are not necessarily continuous. We introduce several asymptotically distribution-free (ADF) procedures designed to test the null hypothesis of no simple factor treatment effect. Furthermore, these procedures facilitate the assessment of treatment interchangeability under repeated application schemes. We investigate two distinct testing strategies: the usual (conventional) approach and Roy’s step-down approach, both adapted to the ridit-based framework. The efficacy of these proposed methods is rigorously evaluated through extensive simulation studies and compared against established Wald-Type and ANOVA-Type tests. The practical utility of our methodology is further demonstrated through the analysis of a real-world dataset.