Stat Corner
Statistical Programmer and Consultant
robinh@uoregon.edu
Get Acquainted with PROC GENMOD
Extracting information contained in the data is one important component of a branch of statistics called "Generalized Linear Models," which includes a wide variety of statistical models. The choices are flexible enough to allow the type of data to define what particular model should be chosen for analysis (in contrast to attempting to apply one model to analyze many types of data). These range from familiar regression techniques to relatively unknown methods for categorical data expressed as counts. All of these models contain a variance/covariance matrix of the parameters, which indicates how good information (that is, the amount and quality of data) produces good estimates of the population parameters.
Many books and journal articles have been written on the broad topic encompassing the SAS procedure called GENMOD (among others). To find out more about how PROC GENMOD works and to see examples of the wide variety of models available within it, go to http://www.uoregon.edu/~robinh/genmod_sas.html