MANCOVA Model For Heteroscedasticity And Unbalance Cell Size: Framework

Abstract

Multivariate analysis of covariance models are widely used in human societies, principally in health, ecology, environmental science and psychological perspective. As such, the researchers face the task of seeking for ways to improve the performance of these models to respond to the requirements which are treating. This paper proposes an extension of two way MANOVA to two way MANCOVA model under effect one covariate factor for heteroscedastic and unbalanced cell sizes. A type III sums squares of an influence is computed that adjusted for all other influences for suggested two way MANCOVA model, no matter in which order they are included. As classical tests could dangerously biased when homogeneity violates the assumption of cell covariance matrices, the suggested model implemented the modified tests obtained by Zhang & Xiao (2012) to test the hypothesis of model