MANOVA: TYPE I ERROR RATES OF SUDOKU SQUARE DESIGN’S MODELS

Type I error is the rejection of true null hypothesis in an experiment, the probability of incurring type I error is sometime referred to as significant level, this type of error does affect sensitivity or robustness of test statistics especially in multivariate analysis. This study proposed procedure for the estimation of type I error rate of multivariate analysis of variance of Sudoku square design models. It compares type I error rates of four multivariate test statistics namely Wilk’s Lambda, Pillai, Lawley-Hotelling and Roy’s largest root tests using simulated Sudoku square data. The study also considered the robustness of the test statistics used within the 95% confidence interval.