Non-symmetric correspondence analysis and biplot representation: Comparing differences in market share distribution

Abstract

In this paper Non-Symmetric Correspondence Analysis (NSCA, Lauro and D’Ambra, 1984; D’Ambra and Lauro, 1989, 1992) is proposed as a useful technique for evaluating contingency table with a dependence structure, in particular within the context of comparing market share differences. Technical aspects of the method are discussed with a view towards application, giving special attention to the biplot representation ofNSCA as compared to the symmetric graphical display. Two examples dealing with canned food and the car market, respectively, are used to illustrate the usefulness of the technique and its kind of representation.