Chapter 24: The CORRESP Procedure

Overview

The CORRESP procedure performs simple and multiple correspondence analysis. You can use correspondence analysis to find a low-dimensional graphical representation of the rows and columns of a crosstabulation or contingency table. Each row and column is represented by a point in a plot determined from the cell frequencies. PROC CORRESP can also compute coordinates for supplementary rows and columns . Experimental graphics are now available with the CORRESP procedure. For more information, see the ODS Graphics section on page 1109.

PROC CORRESP can read two kinds of input: raw categorical responses on two or more classification variables , and a two-way contingency table. The correspondence analysis results can be output and displayed with the %PLOTIT macro.

Background

Correspondence analysis is a popular data analysis method in France and Japan. In France, correspondence analysis was developed under the strong influence of Jean-Paul Benz cri; in Japan, it was developed under Chikio Hayashi. The name correspondence analysis is a translation of the French analyse des correspondances . The technique apparently has many independent beginnings (for example, Richardson and Kuder 1933; Hirshfeld 1935; Horst 1935; Fisher 1940; Guttman 1941; Burt 1950; Hayashi 1950). It has had many other names , including optimal scaling, reciprocal averaging, optimal scoring, and appropriate scoring in the United States; quantification method in Japan; homogeneity analysis in the Netherlands; dual scaling in Canada; and scalogram analysis in Israel.

Correspondence analysis is described in more detail in French in Benz cri (1973) and Lebart, Morineau, and Tabard (1977). In Japanese, the subject is described in Komazawa (1982), Nishisato (1982), and Kobayashi (1981). In English, correspondence analysis is described in Lebart, Morineau, and Warwick (1984), Greenacre (1984), Nishisato (1980), Tenenhaus and Young (1985); Gifi (1990); Greenacre and Hastie (1987); and many other sources. Hoffman and Franke (1986) offer a short, introductory treatment using examples from the field of market research.