Statalist The Stata Listserver


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

st: 2 stage variance correction in panel structure


From   "Jake Kendall" <kendall_jake@hotmail.com>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: 2 stage variance correction in panel structure
Date   Wed, 5 Apr 2006 14:09:04 -0700

Dear Statalist-ers

I am attempting a 2-stage estimation with both stages having a panel
structure. I would like to do a poisson regression, random effects
[xtpoisson ...,re], for the first regression and the more typical
linear random effects [xtreg ...,re] for the second stage using the
predicted results of the first stage as regressors. I am willing to
consider doing a pooled poisson in the first stage if that makes
things simpler.

My question is how to do the variance correction for a panel
structure. I am aware of the papers by Murphy and Topel (1985) and by
James Hardin (200?) but I am having difficulty in calculating the
corrections for either method in the panel structure. I wonder if
anyone out there has done this before in a panel or can give me a
pointer. I believe I have to get the individual observation likelihood function for xtreg ...,re and then the derivatives that I would need to
calculate the C and R matrices. If anyone has the proper formulas
or can help with a pointer on the convoluted vector calculus, I am having trouble especially with the second stage.

Also, if there is an alternative way
to correct the variance when doing a two stage estimation, I am all
ears.
Any help is appreciated.


Jake
*
* 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