Yet another intersting thread -

David Cassell wrote (in part) <<< The fact that R is free is both good and bad. It also means that you are on your own when using hand-crufted code from someone else. There is R code to do mixed models, but it is nowhere close to what SAS users expect from PROC MIXED. I wouldn't do sophisticated mixed models in R. Yet. But then, how long did we SAS users have to wait before PROC MIXED was ready for us? (Hint: it came along so late that it had the original experimental version of ODS incorporated in it.) >>>

Two comments

1) I have found that many of the developers of packages in R are extremely accessible on the R-help list. In addition, that list (like this one) is loaded with knowledgable and helpful people. Also, part of the 'style' of R is that functions have extensive help files

2) I am certainly no expert in mixed models, but I will need to become more expert - they are becoming more and more necessary in my work.

I was curious as to what SAS can do that R cannot, particularly if you load the packages for lme and nlme developed by Pinheiro and Bates into R

I am going to have to devote some considerable time to learning either PROC MIXED and related things, or the R packages, or possibly the HLM program - if one is much better than the others, that's the one I'll learn

Peter