Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: st: Odds ratio

From   "Garth Rauscher" <>
To   <>
Subject   RE: st: Odds ratio
Date   Sat, 10 Apr 2010 17:05:42 -0500

In my discipline (epidemiology) we have relied heavily on odds ratios simply
out of habit and convenience (logistic regression) when our interest is
almost never in the OR itself, but rather in the risk or prevalence ratio or
difference. Because of past help I have received on this list I have learned
how to convert model-based predictions from logistic regression into risk
ratios or differences using marginal standardization with bootstrapped
confidence intervals. The syntax is fairly straghtforward. I don't know what
the implication would be if predictions were generated from a random
intercept or gee logistic model. Does anyone have additional thoughts about
this that might help to answer Rosie's original problem?

Garth Rauscher
UIC School of Public Health


-----Original Message-----
[] On Behalf Of Lachenbruch,
Sent: Friday, April 09, 2010 11:24 AM
To: ''
Subject: RE: st: Odds ratio

I almost agree - ORs are tough to interpret for non-Epidemiologists and
Statisticians.  However, logistic regression is designed to estimate
log-odds ratios.  I'd report both so both of you can feel 'happy' - of
course, if the difference in probability of event is tiny, a large OR
doesn't really indicate a big issue


Peter A. Lachenbruch
Department of Public Health
Oregon State University
Corvallis, OR 97330
Phone: 541-737-3832
FAX: 541-737-4001

-----Original Message-----
[] On Behalf Of E. Paul Wileyto
Sent: Thursday, April 08, 2010 12:17 PM
Subject: Re: st: Odds ratio

The problem is that vastly different sets of numbers can give you the 
same odds ratio...  same odds ratio, with different variances.  Effect 
size (in Cohenesque d speak) is best obtained from fractions and and 
changes in fractions.  Your effect size can come from the log of the 
odds-ratio, but the variance will be determined by the actual 
proportions involved in calculating the OR.

It doesn't sound like the reviewer is asking for much.  Would it hurt to 
give the proportions?

You can actually generate those effect size numbers (d)  if you report 
an Odds Ratio with CI95 and a sample size, but that is more convoluted.


Rosie Chen wrote:
> Hello, dear all,
> I have a question regarding a reviewer's comment on my use of odds ratio
in interpreting the results of a logistic regression, and would appreciate
it very much if you can provide any insight or any references for responding
to the comment. 
> The reviewer commented that all results are expressed in terms of odds
ratios which makes it very difficult to assess the magnitude of the effect.
Probabilities and changes in probabilities would be much easier to
interpret. My impression is that, although it is true that predicted
probabilities might be easier to understand, odds ratios have been used
extensively in research when we interpret results from logit models. 
> Do you have any suggestions regarding how to respond to this comment, or
do you have any statistics textbooks in your mind that recommend odds ratio
as a standard approach reporting results from logistic models?
> Thank you very much in advance!
> Rosie
> *
> *   For searches and help try:
> *
> *
> *

E. Paul Wileyto, Ph.D.
Assistant Professor of Biostatistics
Tobacco Use Research Center
School of Medicine, U. of Pennsylvania
3535 Market Street, Suite 4100
Philadelphia, PA  19104-3309

Fax: 215-746-7140 

*   For searches and help try:

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index