Nested Repeated Measures Model Sufficiency and Estimation
-
-
Abstract
A nested repeated measures model occurs in analysis of variance when a particular individual has the same number of subindividuals and each subindividual receives every pair of treatment levels. Let Yi = (Yilll,...,Yimrc)’ be the vector of observations on the ith individual. It is assumed that the Yi are independently normally distributed with mean μi and common covariance ∑>0. The simplifying assumptions that all measurements have the same variance σ2, every pair of measurements that comes from different subindividuals but the same individual has covariance σ2ρ1; and every pair of measurements that comes from the same individual and same subindividual with (i) different columns and different rows treatments, (ii) the same column but different rows treatments, (iii) different columns but the same row treatments have σ2ρ2,σ2ρ3, σ2ρ4 as their respective covariances is made. We assume that the design is given and use a coordinate-free approach to derive a complete sufficient statistic, minimum variance unbiased estimators (MVUE) and maximum likelihood estimators (MLE) for the nested repeated measures model.
-
-