Categorical Data Analysis Using The SAS® System, 2nd Edition

8.3. Using the CLASS Statement

In the previous example, PROC LOGISTIC used the values of the explanatory variables to construct the model matrix. These values were already coded as 0s and 1s. However, often your SAS data set contains a response variable or explanatory variables that have character values. Or, your model may involve a number of explanatory variables and/or interaction terms so that constructing and managing the required terms as variables becomes a chore. The LOGISTIC procedure handles character-valued response variables by creating ordered values based on the alphabetical order of the response variable values. In addition, PROC LOGISTIC now includes a CLASS statement and allows GLM-like model specification so that constructing ...