MANOVA PROCEDURES WIYH RANDOMIZED COMPLETE BLOCK DESIGN : A CASE STUDY OF SELECTED VARIETIES OF SUGARCANE

Abstract

ABSTRACT Multivariate analysis if variance may be adopted in any specified experimental designs involving two or more response variables. Further, we may exploit multivariate analysis of variance technique when experimental materials ie either homogenous of heterogeneous. For instance when experimental materials is satisfactorily homogenous we have completely randomized design. This paper therefore presents four MANOVA procedures for analysing designed experiments namely Hotelling’s T-square, Wilk.s Lambda, Pillai-Bartlett and Roy’s characteristic root. Each of these four procedures involves the solution of a determinantal equation to obtain the characteristic roots of the sums of squares and sums of products for the appropriate sources of variation be it of the systematic and/or random type. A data set of fifty four varieties of sugarcane obtained from a study conducted at National Cereal Research Institute, Baddegi was used to illustrate the utility of these four MANOVA procedures. The design used for the experiment is of the randomized complete block type in that the two periods of harvest constitute two distinct blocks.