|Date: ||Fri, 7 Nov 2008 06:32:13 -0500|
|Reply-To: ||Peter Flom <firstname.lastname@example.org>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||Peter Flom <peterflomconsulting@MINDSPRING.COM>|
|Subject: ||Re: Stepwise selection in proc logistic|
|Content-Type: ||text/plain; charset=UTF-8|
"cat.." <cat.b41@GMAIL.COM> wrote
>I'm trying to understand in detail, step by step, how the stepwise
>selection process works and I study the following SAS example:
Mind if I ask why?
>But I'm confused about one thing: The first set of covariates is: cell
>smear infil li blast temp.
>At step 0, only the intercept is entered, and, before entering the
>step 1, a table is displayed that shows the analysis of effects
>eligible for entry. Which includes, at step 0: cell smear infil li
>Is this table supposed to display the Chi-square score of significance
>of the estimate of each effect when fitting the model remis = cell
>smear infil li blast temp ?
>In that case, why does the code
>proc logistic data=Remission descending outest=betas covout;
> model remiss=cell smear infil li blast temp;
>yield chi-square scores that are completely different !!
Because that will include *all* the independent variables, and all the effects are controlled for the other IVs.
You'd have to try, e.g
proc logistic data=Remission descending outest=betas covout;
and compare that to the first run result for cell
Peter L. Flom, PhD
www DOT peterflom DOT com