Fitting Linear Mixed Models with SAS

Part of the Springer Series in Statistics book series (SSS)

In Chapters 5 and 6, estimation and inference on all parameters in the marginal model (5.1) were discussed. Chapter 7 considered inference for the random effects in the hierarchical model (3.8). At present, among the most flexible commercially available statistical packages is the SAS procedure PROC MIXED (SAS 1992, 1996, 1997). In this chapter, we will therefore explain in full detail how all previously discussed inferences can be obtained with this procedure, using SAS Release 6.12 (SAS 1997). Although this may seem anomalous to many, given the availability of Version 7.0 and higher, it has to be noted that Version 7.0 (SAS 1999) was not available on a commercial basis in 1999, for example, in Europe. For a thorough description of PROC MIXED in SAS Version 7.0, we refer to the on-line manual (SAS 1999). Further, some of the important changes in comparison to Version 6.12 are summarized in Appendix A.

In this chapter, our original model (3.10) for the prostate data will be used as a guiding example. In Section 8.2, the program for fitting the model will be presented, together with some available options. It is by no means our intention to give a full overview of all available statements and options. Instead, we restrict to those statements and options that are, in our experience, most frequently used in the analysis of longitudinal data. When fitting mixed models in other contexts, other statements or options may be more appropriate. We refer to the SAS manuals (SAS 1992, 1996, 1997) and to Littell et al. (1996) for more details on the procedure MIXED and for a variety of examples in other contexts.


Random Effect Stochastic Process Statistical Theory Linear Mixed Model Longitudinal Data 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer Verlag New York, LLC 2009

Personalised recommendations