A Regression Model for Multilevel Homogeneity Analysis

One of the basic techniques that analyzes categorical data is homogeneity analysis. The technique originated in the work of Guttman [17] as a method of scale construction using reciprocal averaging. Burt [5] described homogeneity analysis as a principal components analysis of qualitative data. Hayashi [19] stressed homogeneity analysis as one possible way of quantifying categories. It should be noted that earlier work by Hirschfeld [21] and Fisher [13] concentrated on the bivariate case (analysis of contingency tables). Extensive reviews on the history of homogeneity analysis can be found in de Leeuw [8], Nishisato [26], Benz ecri [3], Tenenhaus and Young [28] and Gifi [14].