In the GMANOVA model or equivalent growth curve model, shrinkage effec
ts on the MLE (maximum likelihood estimator) are considered under an i
nvariant risk matrix. We first study the fundamental structure of the
problem through which we decompose the estimation problem into some co
nditional problems and then demonstrate some classes of double shrinka
ge minimax estimators which uniformly dominate the MLE in the matrix r
isk. (C) 1996 Academic Press, Inc.