Date:         Thu, 17 Nov 2005 08:52:29 -0800
Reply-To:     Yifan Lu <ylu@ibiweb.org>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Yifan Lu <ylu@ibiweb.org>
Subject:      Re: Multicollinearity and Regression
Comments: To: Sunish George <sunish.george@SURREY.AC.UK>
In-Reply-To:  <200511161708.jAGGs1GE008405@mailgw.cc.uga.edu>
Content-Type: text/plain; charset="iso-8859-1"

Hi Sunish,

I would detect multicollinearity in multiple linear regression and
delete/combine those variables causing the problem, then rerun the model in
multiple regression.
You can use the /statistics=defaults tol to request the display of
"tolerance" and "VIF" values for each predictor as a check for
multicollinearity.

Yifan
-----Original Message-----
From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of
Sunish George
Sent: Wednesday, November 16, 2005 9:09 AM
To: SPSSX-L@LISTSERV.UGA.EDU
Subject: Multicollinearity and Regression

Hi all

I have some subjective data obtained from a test. I tried to predict these
results by using multiple linear regression with the help of different
parameters that I have developed. When I use all of them in my model, I get
a high correlation and less error of prediction. But I cannot accept the
result due to the serious multicollinearity problem. I am in search of an
alternative regression method offered by SPSS that can handle the data with
multicollinearity problem. Can anyone suggest such a method?

Many thanks

Sunish