Correspondence analysis

Most survey data are nominally scaled. The scaling may be either dichotomous or categorical. Unlike metric- or ratio-scaled variables, nominal variables often have no directional intensity. Researchers frequently encounter nominal scaled data, such as “Do you intend to visit this destination?” (yes/no). In this case, correspondence analysis is an appropriate analytical tool.

This method is an exploratory multivariate data analysis technique for the graphical display of contingency tables (cross-tabs). It portrays sets of data points in a joint space where the axes are principal components (contributors to explained variance) identified in the analysis. The rows and columns of the input table are scaled in corresponding units so that each can be displayed graphically in the same low-dimensional space. The resulting (two-dimensional) map provides insight into similarities and differences within rows (a column) compared to a column (rows). While relationships between rows and columns can...