Estimating Means of a Set of Measures from Small Samples by Using Multivariate Composite Estimating Concept—The Strong Heart Study

In this study, we describe a method employing the multivariate composite estimating (MCE) concept to estimate the means of a set of measurements based on small samples from different groups in a population as well as the grand means of the measurements for the population. The advantages of these estimators over simple averages used in the literature are that they have lower mean squared errors, and they improve on simple averages by taking account of variations and correlations among the set of measurements as well as differences among groups. An application to the Strong Heart Study baseline exam data is provided.