Statalist


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

st: countfit and exposure


From   "Biernbaum, Lee" <Lee.Biernbaum@volpe.dot.gov>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: countfit and exposure
Date   Thu, 2 Aug 2007 14:41:45 -0400

Franklin

1) For the uninitiated out there, -countfit- is a command that is part
of J Scott Long and Jeremy Freese's invaluable package, -SPOST-,
designed to help those doing regression in the world of GLM (logit,
probit, oprobit, poisson, negbin, etc). I don't have their website on
hand, but -findit spost- should be sufficient. Note the package name
varies with version of Stata.

2) -countfit- does not support the -exposure- suffix. 

What you can do is 

gen ln_exposure = ln(exposure)

and then put ln_exposure in the -countfit- arguments. This will not be
perfect as it does not constrain the coefficient to be equal to 1 (I've
checked Long & Freese's book, and it does not note support for the
-constraint- option, but you could try), but certainly does a better job
than leaving thee variable out altogether. Moreover, if the estimated
coefficient is terribly different than 1, you should probably be
thinking carefully about not using the exposure constraint in the first
place.

If you do not have it already, I highly recommend Long and Freese's
Regression Models for Categorical Dependent Variables (Stata Press) as
both a great intro to Stata and a detailed manual to the -spost-
commands. Be sure to pick up the new 2nd edition if you run Stata 9 or
higher.

Lee Biernbaum


------------------------------

Date: Wed, 1 Aug 2007 09:36:31 -0400
From: "Franklin, Frank" <frfrankl@jhsph.edu>
Subject: st: Countfit and Exposure 

Hello Fellow Listeners,

I have a question regarding the use of the command countfit...how does
the command manage "exposure", is this built in the command or must one
add it manually as you would when running a count model outside of the
countfit command.  If so, what is the best way to do this for the
command?

I have tried many different approaches and I can not seem to get
countfit to replicate the same outcomes as when I compare the PMR, NBRM,
ZIP and ZINB separately.

Thanks
Franklin

*
*   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/



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