Analysis of Association of Categorical Variables by Numerical Scores and Graphical Representation

摘要

An expository treatment of correspondence analysis (CA) and homogeneity analysis (HA), two data analytic techniques which describe association among categorical variables in terms of numerical scores assigned to the categories is presented. These scores are analogous to loadings of numerical variables on principal components. Thus CA and HA allow graphical representations of the data. Rules for interpretation of such representations are given and applied to a real data set.