Listy Biometryczne - Biometrical Letters Vol. 35(2), 1998, 127-132
|
OPTIMALITY OF SOME EXPERIMENTAL DESIGNS UNDER MIXED EFFECTS LINEAR MODEL Augustyn Markiewicz Department of Mathematical and Statistical Methods, Agricultural University of Poznan, Wojska Polskiego 28, 60-637 Poznan, Poland |
A design which is universally optimal for estimating a given set of parameters under the fixed effects model is also universally optimal for estimating a reduced set of parameters under the related mixed effects model. The aim of the paper is to study conditions under which a design universally optimal in the model without random effects (i.e. a model in which variances of random effects are zero) is also universally optimal in the mixed effects model. Using the Kiefer ordering some sufficient conditions are established.
mixed effects model, Schur complement, G-majorization, Kiefer ordering, Kiefer optimality, optimal experimental designs, row-column designs, repeated measurements designs.
