07-13-2011 08:40 AM
Has anyone had an opportunity to beta PROC HPMIXED in 9.3? In particular, I am curious about the analysis of repeated measures models. Looking through the 9.3 documentation, it seems the possible covariance types are limited in the REPEATED statement (pretty much AR(1), unstructured, and VC). I did see an example of fitting a repeated measures through the RANDOM statement, which offers CS and CSH as covariance types.
Any comments will be greatly appreciated!