topic Re: Proc reg in Statistical Procedures

Proc reg

Elkridge_SAS — Mon, 05 Apr 2010 01:28:56 GMT

I am trying to do backward/forward/stepwise regression while also calculating Cp's . However, instead of using the p-value (sle 0.05 for example) as the cut-off, I am supposed to use an F-value. Can someone help me with the syntax to carry this out ? Thanks

Re: Proc reg

sbb — Mon, 05 Apr 2010 02:09:15 GMT

Suggest checking out the SAS support http://support.sas.com/ website for SAS-hosted DOC and supplemental technical reference/conference materials.

Scott Barry
SBBWorks, Inc.

Suggested Google advanced search argument, this topic/post:
+"proc reg" +"f value" site:sas.com

Re: Proc reg

plf515 — Mon, 05 Apr 2010 03:04:02 GMT

You can do this using PROC GLMSELECT, see the documentation for that PROC. But it's not a good way to do things. See my paper with David Cassell: Stopping Stepwise: Why stepwise procedures are bad and what you should use instead. Available in various places, most recently

www.nesug.org/Proceedings/nesug09/sa/sa01.pdf

Peter

Re: Proc reg

Elkridge_SAS — Mon, 05 Apr 2010 09:42:50 GMT

Thanks Peter. I am trying to do this for a homework so I really am constrained in terms of choice of approach. I looked at the paper and I don't think it mentions how to use the F- value to stop the regression process. For instance, there is a part where it mentions the available stop criteria and there does not seem to be any mention of using F. I just want to be able to say, if F=4, remove/enter data. I'd like to know if SAS can do this. Thanks a lot for your time and for responding.