Statalist The Stata Listserver


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

Re: st: random effects logit


From   "Rodrigo A. Alfaro" <raalfaroa@gmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   Re: st: random effects logit
Date   Thu, 24 May 2007 17:17:07 -0400

Xiaodong Chen wrote:
Hi all,
I run random effects logit model using both STATA and LIMDEP, but the
results are quite different. Does anyone have any idea what's the
likelihood function of random effects logit used in STATA program?

Thanks,
Richard wrote:
Sometimes differences are due to differences between algebraically equivalent parameterizations. How do the fits of the models compare? If they are different as well, I suspect you haven't specified equivalent models like you think you have. You might post the output from both.
Totally possible to get different results. The likelihood is in the manual.
In addition to Richard's suggestions, RE models need to deal with the computation of an integral, which implies (1) define the number of points to compute the quadrature and (2) get the integration method itself. This could explain the differences.

More important, no matter which method you will use you need to check
that the quadrature is accurate. This is recommended in the manuals for
all the RE model. I don't know about LIMDEP.
R



*
* 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/




© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index