Roy and Gnanadesikan  showed that inference for a general multivariate variance components model may be carried out using the standard multivariateF distribution under certain condtions. It is shown in this note that the theory of zonal polynomials, and their extension by the author to invariant polynomials in two matrix arguments, provide a concise approach to the derivation of these conditions. Relevant distributions are also derived for the general case.
Latent Root Invariant Polynomial Wishart Distribution Matrix Argument Zonal Polynomial
This is a preview of subscription content, log in to check access.
Chakravorti, S. (1968). On the analysis of variance test in multivariate variance components model,Calcutta Statist. Ass. Bull.,17, 57–78.MathSciNetCrossRefGoogle Scholar
Davis, A. W. (1979). Invariant polynomials with two matrix arguments extending the zonal polynomials: applications to multivariate distribution theory,Ann. Inst. Statist. Math.,31, A, 465–485.MathSciNetCrossRefGoogle Scholar
James, A. T. (1964). Distributions of matrix variates and latent roots derived from normal samples,Ann. Math. Statist,35, 475–501.MathSciNetCrossRefGoogle Scholar
Roy, S. N. and Cobb, W. (1960). Mixed model variance analysis with normal error and possibly non-normal other random effects, I. The univariate case, II. The multivariate case,Ann. Math. Statist.,31, 939–957, 958–968.MathSciNetCrossRefGoogle Scholar
Roy, S. N. and Gnanadesikan, R. (1959). Some contributions to ANOVA in one or more dimensions: II,Ann. Math. Statist,30, 318–339.MathSciNetCrossRefGoogle Scholar