-glm- (and -betafit-) will allow the cluster option.
HTH,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z434
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
-----Original Message-----
From: [email protected] [mailto:[email protected]]On Behalf Of Cameron Gillies
Sent: donderdag 28 september 2006 18:03
To: [email protected]
Subject: st: GLLAMM and a dependent variable that is a proportion
Greetings,
I have data with samples nested within individuals that I suspect would best
be analyzed with GLLAMM where I can include a random intercept and one or
more random coefficients. The difficulty is that the dependent variable is
a proportion (0 to 1). Is anyone aware if GLLAMM handles this appropriately
using family(binomial) and link(logit) or is it treating non-zeros as ones?
glm does handle this correctly, but cannot include random effects. I have
run the data using GLLAMM, but it is not clear to me from the results how
the non-zero values are being handled.
I am trying to avoid doing a logit transformation (see
http://www.stata.com/support/faqs/stat/logit.html).
Thanks very much for any insight you may have!
Regards,
Cam
Cam Gillies
Biological Sciences
University of Alberta
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
