[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: RE: xtlogit with a fractional response

From   "Verkuilen, Jay" <>
To   <>
Subject   st: RE: RE: xtlogit with a fractional response
Date   Thu, 28 Feb 2008 15:08:43 -0500

Garry Anderson wrote:

>Thanks to Maarten, Nick and Jay for their suggestions and historical
perspective. Regards, Garry<

If GEE does the trick for you to handle the dependence, try the
quasi-likelihood trick. Population-averaged coefficients are, by their
very nature, pushed towards 0 compared to random effects models. You may
get good results using betafit by  Nick Cox, et al., and employing
clustered robust standard errors. 

If you want, feel free to get in touch. I can give you some SAS or MCMC
(winBUGS) code that does what you want. I'm simply not familiar enough
with Stata programming (and lack the time to work on it until summer,
probably) to write the beta regression equivalent of xtlogit's random
effects program. Or drop Mike Smithson a line, he's at ANU.... 


-----Original Message-----
[] On Behalf Of Garry
Sent: Wednesday, February 27, 2008 6:00 PM
Subject: st: xtlogit with a fractional response

Dear Statalist,

I was attempting to analyse a fractional response with -xtlogit propn x1
,re i(id)-. However, it seems that all nonmissing values other than zero
are regarded as a positive outcome. This is not the situation with -glm
proportion xvars ,fam(bin) link(logit) robust-

This is mentioned in the FAQ 'How do you estimate a model when the
dependent variable is a proportion?' at

'glm has since been enhanced specifically to deal with fractional
response data.'

I was wondering if there is a way that xtlogit can be used with
fractional response panel data? 
(Other options include glm and the cluster robust variance, or xtgee.)

Best wishes, Garry
Garry Anderson
School of Veterinary Science
University of Melbourne
250 Princes Highway    Ph  03 9731 2221
WERRIBEE    3030       Fax  03 9731 2388

*   For searches and help try:

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index