Estimating Variance Components for Two-Way Disproportionate Data with Missing Cells by the Method of Unweighted Mean

Abstract

The method of unweighted means, which is a simple analysis based on cell means, has been limited to designs where every cell has at least one observation. Here, the method of unweighted means is examined with regard to estimating the variance components for a two-way classification with unequal numbers of observations in the cells and with one or more cells having no observations. In particular, two estimating procedures based upon the method of unweighted means are developed to estimate the variance components for a design with missing cells. These two unweighted means procedures are then examined numerically