I recieved answers from Brian Ripley and James McBroom
Brian Ripley has written software for fitting GLMMs using penalized quasi
likelihood. It's in the latest version of the MASS library, and is called
glmmPQL (for S+6 and R). It builds on nlme.
Gordon Smyth has reglm for S, also PQL (by another name), simpler.
Jose Pinheiro has GLME for S in alpha/beta, also PQL and similar in effect
to glmmPQL but more integrated with nlme.
GlmmGibbs by Myles and Clayton uses MCMC, currently R only.
Jim Lindsey has a glmm function for R for a single random intercept only,
using simple numerical integration.
Prof Ripley has heard of (but not seen) other functions, e.g. by Doug
Clarkson of
Insightful.
I found the MASS version easiest to track down.
Thanks to those who answered ... if anyone else knows of other software,
please let me know.
-------------------------------------------------------------------
Dr Ann Cowling
Head, Statistical Consulting Unit
The Graduate School
John Dedman Mathematical Sciences Building, Room G10
Australian National University
Phone: (02) 6125 4792
-------------------------------------------------------------------
|