Thank you to everyone who replied to my query. Here was the original
question
“Dear Allstat,
I am wondering if anyone knows if it is possible to perform multi-level
modelling in S-Plus. I do have MLWin which seems to be the package to use,
but if there is a way of doing it in S-Plus (version 6), then it might be
helpful also. Replies to me please, i will summarize the responses and post
them later.
Thanking you in advance,
Mark Kelly,
UWCM.”
The responses could be categorized as follows:
1. those that told me that multi-level modelling has many other names,
including mixed-effects modelling.
2. those that told me to use the “lme” and “nlme” functions in S
3. those that suggested a website or mailing list
4. and those that recommended a text book
See below for further details on each of these.
1.
'Multi-level modelling' is MLWin-speak. The same concept appears in various
guises as random effect modelling, hierarchical modelling mixed-effects
modelling. As concerns the latter there is a book by Pinheiro and Bates,
Mixed-Effects Models in S and S-Plus, Springer, 2000 but you will not find
that SPlus does everything that MLWin does and vice versa. Proc NLMixed is
the SAS procedure for this sort of thing and various procedures in GenStat
cover the same topic and no doubt many other packages cover it too.
2.
I think you can use the function nlme. See also the book "Modern
Applied Statistics with S-plus" (Venables and Ripley) (Section 6.11)
3. WEBSITE
suggest you look at http://multilevel.ioe.ac.uk/softrev/index.html
"The Centre for Multilevel Modelling aims to provide a comprehensive set of
reviews of packages for multilevel modelling. The first set of reviews is
now nearing completion and they may be accessed by clicking on the
appropriate package name at the left. This will take you to a brief
introduction from where you can download the full review in PDF format. The
remaining reviews will be provided as they become available. "
This has a review for 'R' which "can be regarded as an Open Source
implementation of the S language which in turn underlies the S-Plus
software.". So the Centre obviously think you can use it [and hence S-plus]
for multi-level modelling. So suggest you ask them how.
MAILING LIST
There is a Multilevel modelling mailing list at JISCmail that might
help you with your question and any further problems you might have once you
start your modelling: please see
http://www.jiscmail.ac.uk/lists/multilevel.html
If you search the archives for "S-PLUS" you'll find some entries, including
one by Douglas Bates <[log in to unmask]> who states that S-PLUS uses the
NLME (nonlinear mixed-effects) library (see http://nlme.stat.wisc.edu/).
4.
The book on this topic;
Pinhiero, J. C. and
Bates, D. M. (2000). Mixed-Effects Models in S and S-plus. New York:
Springer.
http://cm.bell-labs.com/cm/ms/departments/sia/project/nlme/MEMSS/
_________________________________________________________________
Stay in touch with absent friends - get MSN Messenger
http://www.msn.co.uk/messenger
|