Spotfire Statistica® Correspondence Analysis

Correspondence analysis is used to uncover relationships and create segments; soccer moms, Pepsi taco consumers, etc. In other words, it is an exploratory technique designed to analyze simple two-way and multi-way tables containing some measure of correspondence between the rows and columns. The results are similar in nature to those produced by Factor Analysis techniques.

Multiple correspondence analysis is a simple correspondence analysis carried out on an indicator (or design) matrix with cases as rows and categories of variables as columns. The user can directly specify a Burt table as input for the analysis. Or the module can automatically create a Burt table from grouping variables (e.g., you included a variable Gender, with the two possible values Male and Female). 

The module will compute various tables, including the table of row percentages, column percentages, total percentages, expected values, observed minus expected values, standardized deviates, contributions to the Chi-square values, generalized eigenvalues, and eigenvectors. For each dimension and row or column point, the program will compute the inertia, quality, and cosine² values.

For a description of correspondence analysis computational details read the classic text by Greenacre (1984). These methods were originally developed primarily in France by Jean-Paul Benzécri in the early 1960s and 1970s (e.g., see Benzécri, 1973; see also Lebart, Morineau, and Tabard, 1977).  Other English references are Carrol, Green, and Schaffer, 1986; Hoffman and Franke, 1986. See Greenacre (1984, p. 280, Example 9.6) for a comparison between correspondence analysis and factor analysis (principal components).