Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Use of offset variables with discrete-choice models

From   [email protected]
To   [email protected]
Subject   Re: st: Use of offset variables with discrete-choice models
Date   Wed, 19 Jan 2005 13:36:57 EST

You can use offsets for other reasons than for rates. For instance, you can  
use an offset as a constraint to perform a likelihood ratio test when  
comparing the coefficient you may have for a particular model with  some comparative 
value. Suppose that you have a logistic regression  coefficient of 0.15 on 
variable x and you wish to test, using the  likelihood ratio test, if it is 
significantly equal to 0.20. You can set  variable x as an offset with value 0.20 
and use it as the comparison model for  the LL ratio test. 
This use of offsets, as well as others, can be found in Hardin & Hilbe,  
"Generalized Linear Models and Extensions" (2001)  Stata Press. 
I should also mention that offsets have been used with dose reponse cloglog  
models for some time. 
Joe Hilbe
Dear all,

I understand why one might wish to use an offset  when
estimating a Poisson regression (i.e., to model rates
as opposed to  counts).  However, can anyone tell me
why one would use an offset when  fitting a logistic or
complementary log-log model?  Particularly in  the
context of longitudinal (panel) data?  I have gone
through all of  the Stata reference manuals I own
(which, unfortunately, do not include the  User's
manual) and have also tried to find the answer online.
Any feedback  would be appreciated.



James  W. Shaw, Ph.D., Pharm.D., M.P.H.
Post-Doctoral Fellow
Tobacco Control  Research Branch
Behavioral Research Program
Division of Cancer Control and  Population Sciences
National Cancer Institute
Tel.:   215-852-3045
Fax:   215-542-7150

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index