UMASS Department of: Biostatistics and Epidemiology 691F

Lecture 15
Additional Example for DATA _NULL_;

In many studies, we will be interested in summarizing analyses controlling for different variables. In such settings, it is nice to have a table that gives the coefficients of each variable in the different models, to see how control for other variables affects the variable in the list. We illustrate this for a relatively complicated study of seasonal variation in total cholesterol.

A small sample of data for the study are given in the file aa1ax.sd2. These data are copied to the c:\temp directory, and read into the SAS program LEC15P13.SAS. This program fits mixed models to the data, controlling for various covariates. As the models are fit, an output data set is created that contains coefficients and P-values from the models. These coefficients are accumulated in another data set.

Once all models have been fit, the resulting data set with the coefficients is printed in a Table using as DATA _NULL_; resulting in this output.