Statalist


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

st: RE: running a backward stepwise multivariate analysis on my specific dataset


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: running a backward stepwise multivariate analysis on my specific dataset
Date   Wed, 3 Dec 2008 18:01:13 -0000

It seems to me that you already found out how to run a stepwise
analysis. So what is the question? Finding out more about the odds ratio
seems best addressed by further study of logit/logistic modelling texts
and papers in your field. 

On a point of detail: I think that these days logit models of the kind
discussed here wouldn't qualify as "multivariate". The meaning has
shifted from the classic "many variables" to the modern "many responses"
-- and you have one response. 

See Frank Harrell's book "Regression modelling strategies" (Springer,
New York 2001) for a trenchant critique of the stepwise strategy. 

Nick 
n.j.cox@durham.ac.uk 

Scott Gilmore

I have a dataset of N=50 patients and 20 variables.  13 are categorical
and 7 are continuous variables.  I have no missing values.  My outcome
is one of the variables and it is a binary outcome, 1= yes disease, 0=
no disease.
 
I want to run an analysis for multivariate predictors of disease.  Using
logistic regression and doing a backwards stepwise multivariate
analysis.
 
I used the commands:
sw logistic var1 var2 var3 var4 var5, pr(0.5)
 
where var1=the outcome variable (binary)
var2= categorical (1,2,3,4,5)
var3= continuous
var4= categorical
and so forth.
 
I am not sure how to interpret the Odds Ratio and don't quite understand
what I am getting.
 
Can someone please explain to me how to run a backward stepwise
multivariate analysis on my specific dataset.

*
*   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   |   What's new   |   Site index