topic Re: How to specify the slentry and slstay in stepwise or backward regression? in Statistical Procedures https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101308#M5290 <HTML><HEAD></HEAD><BODY><P>Guidance #1. Don't do backward/forward/stepwise regression.</P><P>Guidance #2. See guidance number 1.</P><P></P><P>There is an abundant number of posts and papers regarding the dangers of this method.&nbsp; The resulting models have poor predictive ability and the regression parameters range from moderately to severely biased.&nbsp; See the NESUG paper by Flom and Cassell: "Stopping stepwise: Why stepwise and similar selection methods are bad, and what you should use"&nbsp; <A href="http://www.nesug.org/proceedings/nesug07/sa/sa07.pdf">http://www.nesug.org/proceedings/nesug07/sa/sa07.pdf</A></P><P></P><P>Still if you must deliberately use a poor tool, stick to the default values.</P><P></P><P>Steve Denham</P><P></P><P><BR /> </P></BODY></HTML> Mon, 17 Dec 2012 13:05:16 GMT SteveDenham 2012-12-17T13:05:16Z How to specify the slentry and slstay in stepwise or backward regression? https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101307#M5289 <HTML><HEAD></HEAD><BODY><P>Hi,</P><P>I have never run the stepwise /backward regression before. I am wondering on what basis I can decide the values for slentry (sig level to allow the variable in the model)&nbsp; and slstay (sig level to keep the variable in the model).</P><P>Any guidance/help will be appreciated.</P><P>Thanks much,</P><P>Ashwini</P></BODY></HTML> Mon, 17 Dec 2012 01:23:43 GMT https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101307#M5289 Ashwini_uci 2012-12-17T01:23:43Z Re: How to specify the slentry and slstay in stepwise or backward regression? https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101308#M5290 <HTML><HEAD></HEAD><BODY><P>Guidance #1. Don't do backward/forward/stepwise regression.</P><P>Guidance #2. See guidance number 1.</P><P></P><P>There is an abundant number of posts and papers regarding the dangers of this method.&nbsp; The resulting models have poor predictive ability and the regression parameters range from moderately to severely biased.&nbsp; See the NESUG paper by Flom and Cassell: "Stopping stepwise: Why stepwise and similar selection methods are bad, and what you should use"&nbsp; <A href="http://www.nesug.org/proceedings/nesug07/sa/sa07.pdf">http://www.nesug.org/proceedings/nesug07/sa/sa07.pdf</A></P><P></P><P>Still if you must deliberately use a poor tool, stick to the default values.</P><P></P><P>Steve Denham</P><P></P><P><BR /> </P></BODY></HTML> Mon, 17 Dec 2012 13:05:16 GMT https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101308#M5290 SteveDenham 2012-12-17T13:05:16Z Re: How to specify the slentry and slstay in stepwise or backward regression? https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101309#M5291 <HTML><HEAD></HEAD><BODY><P>Thanks Steve for advising me on this! It's great to know about the latest guidelines on SAS/STAT procedures.</P><P></P><P>Best,</P><P>Ashwini</P></BODY></HTML> Mon, 17 Dec 2012 22:47:49 GMT https://communities.sas.com/t5/Statistical-Procedures/How-to-specify-the-slentry-and-slstay-in-stepwise-or-backward/m-p/101309#M5291 Ashwini_uci 2012-12-17T22:47:49Z