Date:         Fri, 25 Oct 2002 17:50:15 -0500
Reply-To:     "Simon, Steve, PhD" <ssimon@cmh.edu>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         "Simon, Steve, PhD" <ssimon@cmh.edu>
Subject:      Re: combining Logistic and Linear Models
Comments: To: Sherry Downing-Alfonso <SDowningAlfonso@gaic.com>
Content-Type: text/plain

Sherry Downing-Alfonso writes:

> I am working on a model to predict a measure which is
> present only for about 2% of the population I am
> evaluating.  I've taken those 2%, as well as a
> representative sample of similar size from the remaining
> 98%, and analyzed using logisitic regression (event, y/n).
> Next, working with the group for which the event is present
> and using linear regression to predict the measure of
> interest.  How can I combine these 2 models?

I'm not sure you need to combine them. Just report the two models as they
currently stand. If you do want to combine them, you need to use something
like a Tobit regression which is a rather specialized approach not available
in SPSS.

Steve Simon, ssimon@cmh.edu, Standard Disclaimer.
The STATS web page has moved to
http://www.childrens-mercy.org/stats.