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 on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: compare effect size between dummys and metrics variables in logistic regression


From   Morten Hesse <mh@crf.au.dk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: compare effect size between dummys and metrics variables in logistic regression
Date   Mon, 27 Sep 2010 12:59:49 +0200

Hi
I am a newbie, but I get confused about this. It may well be that Michael is right. But if I conduct a logstic regression, and then write:

test [continuous covariate]=[1.dichomotomous covariate]

Then I get a chi-square statistic (with one df) and a p-value. It looks just like any other result of "test":

 ( 1)  [dep var]cont_covar - [dep var]1.dic_covar = 0

           chi2(  1) =    4.28
         Prob > chi2 =    0.0387

In my understanding, this asks the question "is the difference between the value of the coefficient when dic_covar=1 significantly different from the coefficient of cont_covar.

If this does not make sense, then why does STATA do it? Is this not what Joerg would want to use?

Greetings
Morten

Den 27-09-2010 09:21, Michael N. Mitchell skrev:
Greetings Joerg

If you were doing an OLS regression, I would say that you could report the partial (or semi-partial) correlation associated with each of the variables. However, as you note, you are using binary logistic regression. While you could technically create z scores for the binary variables to try and put them on the same scale as the z-transformed variables, those results would not make much sense. You could take the z-transformed variables and convert them into 0/1 variables by cutting them into two arbitrary groups, but that would not make much sense. My thoughts are that there is not much you can do to compare the "apples" and "oranges" here in terms of the effect size.

Best regards,

Michael N. Mitchell
Data Management Using Stata - http://www.stata.com/bookstore/dmus.html A Visual Guide to Stata Graphics - http://www.stata.com/bookstore/vgsg.html
Stata tidbit of the week         - http://www.MichaelNormanMitchell.com



On 2010-09-26 11.23 AM, Jörg Eulenberger wrote:


Dear Statalisters,
a have an more statistical Question. I want to calculate an binary logistic regression. I have all metric variables z-transformed (mean = 0, std=1) to compare the effect size between the independent variable. But I have also dummys in my Regressionmodell. What can i do to compare the effect size of the dummy's with the effect size of the metric
independent variables? Or is that completely impossible?
Thanks,
Joerg
*
* 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/
*
*   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