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

Re: st: interpretation of exponentiated coefficients

From   Richard Williams <[email protected]>
To   [email protected], <[email protected]>
Subject   Re: st: interpretation of exponentiated coefficients
Date   Fri, 01 Jul 2005 06:53:10 -0500

At 10:48 AM 7/1/2005 +0100, Barbara A Willey wrote:
I wonder if anyone may be able to help/comment...

Due to low prevalence of my outcome I used a complementary log log
transformation of my data.

Regression can still be carried out (cloglog command) but this give
coefficients not odds ratios.
A display of odds ratios and 95%CI can be obtained using the exponentiated
coefficients (ereturn display, eform ("OR")).

However, I am not sure if this can be interpreted in the same way as odds
ratios, Stata help also described it as a hazard ratio.

Can anyone advise on how to interpret results in this format?
I can't answer your specific question, but can only note that there is a history of questions like this and this isn't the first time there has been confusion over such matters. There was a somewhat related discussion a few months ago; in particular see Roberto G. Gutierrez's comment at

Different Stata programs use different terms (odds ratio, relative risk ratio, etc.) for the same mathematical transformation (exponentiating the coefficient). This has to do with the proper interpretation of the transformation (although, as Roberto points out, other software packages just always call it "odds ratio" or generically "exp(b)").

In the case of cloglog - I notice that it doesn't include an option, under any name, for exponentiating the coefficients. In such cases, you wonder if (a) it simply isn't appropriate, or (b) nobody got around to programming it.

Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: [email protected]
WWW (personal):
WWW (department):

* For searches and help try:

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