st: New article comparing software for GLMMs with binary responses

[Roger Harbord <rmharbord@gmail.com>]

Interesting-looking article just published online in Statistics in
Medicine comparing approaches, algorithms and software for fitting
generalized linear mixed-effects models with binary responses:

Zhang,Hui; Lu,Naiji; Feng,Changyong; Thurston,Sally W.; Xia,Yinglin;
Zhu,Liang; Tu,Xin M.
On fitting generalized linear mixed-effects models for binary
responses using different statistical packages. Statistics in Medicine
2011 <http://dx.doi.org/10.1002/sim.4265>

They compare two SAS procedures and three R packages, but ignore Stata
completely -- I feel offended on Stata's behalf. Extending to Stata
might make a worthwhile short student project if someone on this list
is responsible for a suitable course. I'm aware that -xtmelogit- and
-gllamm- sometimes give noticeably different results due to subtle
differences in their approaches and algorithms, but I'm never sure
exactly what the differences are. It would be helpful to list these
and compare results to those from SAS and R given in this paper. I
imagine Statistics in Medicine might be willing to publish as a
letter.

Roger.

-- 
Roger Harbord
http://www.epi.bris.ac.uk/staff/rharbord.htm
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/