Graphical Methods for Categorical Data

The graphical displays shown here are implemented in SAS/IML software whose combination of matrix operations, built-in Statistical methods for categorical data, such as loglinear models functions for contingency table analysis, and graphics provide a and logistic regression, represent discrete analogs of the analysis of convenient environment for graphical display for multiway variance and regression methods for continuous response variables. categorical data (Friendly 1991a; 1992). However, while graphical display techniques are common adjuncts to analysis of variance and regression, methods for plotting contingency table data are not as widely used. This paper provides a brief introduction to graphical methods Several schemes for representing contingency tables graphically are that are useful for understanding the pattern of association among based on the fact that when the row and column variables are categorical variables. These methods can be helpful both for data independent, the estimated expected frequencies, e ij , are products of exploration and for communicating results to others. The methods the row and column totals (divided by the grand total). Then, each described include association plots for two-way tables, mosaic cell can be represented by a rectangle whose area shows the cell displays for multiway tables, correspondence analysis and effect frequency, f ij , or deviation from independence. plots for logit models.