Date:         Fri, 11 Feb 2000 09:09:25 GMT
Reply-To:     Znarf Akfak <znarf_akfak@MY-DEJA.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Znarf Akfak <znarf_akfak@MY-DEJA.COM>
Organization: Deja.com - Before you buy.
Subject:      Re: Standardized Coefficients

Posted on behalf of Michael.

In article  "abdu.elnagheeb" <abdu.elnagheeb@BANKOFAMERICA.COM> wrote:

> I may have seen on this list and somewhere else a discussion of "using
> standardized coefficients" from linear regression to rank independent"
> variables in terms of importance. Some have concern about that. Any
> ideas or references?  Please, respond to me as I don't have access to
> the list now. I will summarize responses and post.

A major problem, when the predictors are correlated, is that this type
"ranking" depends upon the order in which variables are entered into
the model.

Michael F.


Sent via Deja.com http://www.deja.com/
Before you buy.