Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

st: Logistic regression interpretation -- vs. M-H


From   "Dr. Bill Westman" <webicky@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Logistic regression interpretation -- vs. M-H
Date   Thu, 23 Sep 2010 09:09:12 -0700

helpful and generous with your time.  To other listers who replied --
thank you as well for your excellent guidance

I am wondering whether Mantel-Haenszel. is more or less appropriate in
2x2x2 situation with binary variables which spans across groups than
logistic regression.  I am reading Agresti's work which discusses this
now.  I believe epitab M-H is producing this test (at least I think it
is).  The results are quite different in terms of statistical
significance.  I am not sure if you or others have an opinion of the
differences in these models.

 I hope this last question isn't a nuisance, but there are several
Stata list posts on this issue and I was not able to come away with a
meaningful concluston

Thanks for reading...

Bill

On Thu, Sep 23, 2010 at 2:50 AM, Maarten buis <maartenbuis@yahoo.co.uk> wrote:
> --- On Wed, 22/9/10, Dr. Bill Westman wrote:
>> Also - those two di statements at show the exact same odds
>> - was this intentional?
>
> It is important to make a distinction between odds and odds-ratios.
> The former is a measure of the likelihood of an event, the latter
> is a measure of how this likelihood differs across groups.
>
> So in the example below we first compute the odds of union membership
> separately for married and non-married people. Remember that the odds
> is the expected number of union members per non-union member, so if
> we observe 10 union members and 20 non-union members then we expect
> to see 10/20=.5 union member for every non-union member.
>
> After that we look at how the odds of union membership differs between
> married and non-married women, by computing the ratio.
>
> After that I show that that is exactly what -logit- does.
>
> *--------------------------- begin example --------------------------
> sysuse nlsw88, clear
> tab married union
> di "the odds of union membership for married women is: " 280/942
> // so there are .30 union members for every non-union member when
> // the women is married
>
> di "the odds of union membership for non-married women is: " 181/475
> // so there are .38 union members of every non-union member when the
> // women is not married
>
> di "the odds of union membership for married women is " ///
>   (280/942)/(181/475)  " times the odds of union membership" ///
>   " for non-married women"
> // so mariage reduces the odds of memebership with 22%
>
> // this odds ratio can be recovered with -logit-.
> logit union married, or
>
> // the odds can be recovered using -margins.
> margins , exp(exp(xb())) over(married)
> *---------------------- end example ---------------------------------
>
> Hope this helps,
> Maarten
>
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://www.maartenbuis.nl
> --------------------------
>
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index