Testing a covariance matrix structure in a mixed model with no empty cells (Q1118948)

Scheffé's mixed two-way classification model for balanced as well as unbalanced data is considered. The covariance structure associated with this model's random effects is quite general. Under the usual assumptions of independence of the random effects and homogeneity of their respective variances, the covariance structure assumes a particular form. A multivariate test is presented to check the validity of this form. A numerical example is given to illustrate the implementation of the proposed test.