Date:         Fri, 10 Nov 2006 10:32:22 -0800
Reply-To:     David L Cassell <davidlcassell@MSN.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         David L Cassell <davidlcassell@MSN.COM>
Subject:      Re: Keeping Demographic Variables in Behaviour Scorecard
              Development
In-Reply-To:  <1163125141.827134.12900@f16g2000cwb.googlegroups.com>
Content-Type: text/plain; format=flowed

ranjan.rao@GMAIL.COM wrote:
>Hi All
>
>I am currently developing a behaviour scorecard for credit card
>portfolio. Please guide me if it is acceptable to keep demographic
>variables in model of a behaviour scorecard.
>
>If we just go by definition it is score based upon its performance with
>  the bank so demographic variables shouldnt be there in the behaviour
>score .
>However considering that it is used for purposes other than just
>measuring performance like for campaign management we should not
>refrain from using any information that improves my prediction of a
>potential good customer.
>
>Although this is not a forum to specifically discuss this issue but
>since the expertise of our esteemed group members goes far beyond just
>using SAS.Thats why I am still taking this up .
>Please excuse me if its irrelevant to the group.
>
>Thanking you in advance
>Regards

There are ethical issues in here as well as technical ones.  In part,
it depends on what you mean by 'demographic variables'.  If you're
characterizing different rates by, say, whether it is a 'blue' state
vs. a 'red' state that has entirely different implications than the
case where you have 9-digit zip codes and you're implicitly using
them as surrogates for poverty, race, etc.

David
--
David L. Cassell
mathematical statistician
Design Pathways
3115 NW Norwood Pl.
Corvallis OR 97330

_________________________________________________________________
Add a Yahoo! contact to Windows Live Messenger for a chance to win a free
trip!
http://www.imagine-windowslive.com/minisites/yahoo/default.aspx?locale=en-us&hmtagline