Dear Allstat,
Here is the summary of replies I had for the question below. Thank you to
everyone who responded.
>Dear Allstat
>
>I have repeated measures data ( 4 obs per person) in which the dependent
>variable is ordinal(1,2,3,4,5,6) where 1=very mild, 6=very severe. I wish
>to fit a random effects model, but cannot find any standard software
>packages that will fit a model to ordered categorical repeated
>measurements.
>Does anyone know of any software that can do this?
>
>Laurence
1) You can in MLwiN
2) In reference to your post on Allstat, you could try SABRE:
http://www.cas.lancs.ac.uk/software/sabre3.1/sabre.html
3) You can analyse such data by Generalized Estimating Equations (GEE) in
PROC
GENMOD of SAS. The following WEB page may help
http://support.sas.com/rnd/app/da/new/dagee.html
I have done some further research myself and have found that the 'gllamm'
command written in STATA syntax can also deal with these types of models.
The following website has comprehensive details on this command:
http://www.gllamm.org/
_________________________________________________________________
Find a cheaper internet access deal - choose one to suit you.
http://www.msn.co.uk/internetaccess
|