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

Re: st: negative binomial model - transformation

From   Willard Manning <>
Subject   Re: st: negative binomial model - transformation
Date   Tue, 25 Mar 2003 17:05:17 -0600


If you use the default link function (log) for the negative binomial, all you have to do is exponentiate the x beta hat or to make adjusted predictions in the usual way. The negative binomial is estimating the log(E(y|x)), which is more appropriate for mean comparisons than the E(log(y|x)) that OLS on log(y is estimating.


At 05:09 PM 3/25/2003 -0500, you wrote:

I am running a negative binomial model simulating the effect of changes in right-hand side variables on the predicted count of the outcome. There is a well documented transformation problem with log OLS models. Are there any transformation issues associated with analyzing predicted counts from the negative binomial model? If so, do you have any do files or concrete suggestions for handling this problem?

Thank you,
Maureen Stewart

* For searches and help try:
Willard G. Manning
Dept. of Health Studies -- MC 2007
The University of Chicago
5841 So. Maryland Ave.
Chicago, IL  60637

Phone:  773-834-1971 (direct)
        773-702-2453 (department)

Fax:    773-702-1979


*   For searches and help try:

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