Date:         Sun, 2 Nov 2008 18:32:00 -0800
Reply-To:     Jeff <jeffrey.m.allard@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Jeff <jeffrey.m.allard@GMAIL.COM>
Organization: http://groups.google.com
Subject:      Chi Square to Rank Variables
Comments: To: sas-l@uga.edu
Content-Type: text/plain; charset=ISO-8859-1

This is a question not focused on SAS per say but will be executed
with SAS.

Does anyone know for data mining (classification specifically)
exercises, when a chi square is used to rank which potential predictor
variables are expected to be useful in predicting the target variable,
is the chi square statistic or the p value best used? Note that
variables will have different number of levels (degree of freedom)
which causes the question about comparing the chi square statistics.
Thanks!