LIKELIHOOD RATIO CRITERIA OF PARAMETERS IN GROWTH CURVE MODEL FOR MULTIVARIATE ELLIPTICALLY CONTOURED DISTRIBUTIONS

The maximum likelihood estimates of parameters in growth curve model under normal distribution have been studied by many authors. In this paper, the maximum likelihood estimates of parameters in the growth curve model under two classes of matrix elliptically contoured distributions are discussed. Likelihood ratio criteria are obtained for many null hypotheses, which have the same form as that of the normal case. The testing statistics that are invariant in the classes of matrix elliptically contoured distributions are also studied.