The GLMMOD Procedure

The GLMMOD procedure constructs the design matrix for a general linear model; it essentially constitutes the model-building front end for the GLM procedure. You can use the GLMMOD procedure in conjunction with other SAS/STAT software regression procedures or with SAS/IML software to obtain specialized analyses for general linear models that you cannot obtain with the GLM procedure.

While some of the regression procedures in SAS/STAT software provide for general linear effects modeling with classification variables and interaction or polynomial effects, many others do not. For such procedures, you must specify the model directly in terms of distinct variables. For example, if you want to use the REG procedure to fit a polynomial model, you must first create the crossproduct and power terms as new variables, usually in a DATA step. Alternatively, you can use the GLMMOD procedure to create a data set that contains the design matrix for a model as specified using the effects modeling facilities of the GLM procedure.

Note that the TRANSREG procedure provides alternative methods to construct design matrices for full-rank and less-than-full-rank models, polynomials, and splines. See Chapter 117: The TRANSREG Procedure, for more information.