Date:         Wed, 5 Mar 1997 14:55:56 GMT
Reply-To:     David Nichols <nichols@SPSS.COM>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         David Nichols <nichols@SPSS.COM>
Organization: SPSS, Inc.
Subject:      Re: Collin Maslow Stat

In article <970305082739_1216230670@emout10.mail.aol.com>,
Barry Dickinson  <JBARR@AOL.COM> wrote:
>Is there a way to compute the C(sub p), Collin Maslow Statistic for
>determining which predictor variables should be in a regression equation.  I
>know you can do stepwise regression but I would like to compute this
>statistic.  Also, is there a way to calculate Allen and Cady's PRESS
>statistic (predicted error sum of squares)?
>
>Thanks in advance,
>Barry Dickinson
>Drexel University
>Doctoral Student
 
The Cp statistic is one of the four measures you get in REGRESSION
when you specify the SEL keyword on the STATISTICS subcommand (it's
not supported in the dialog boxes in current releases, so you have to
either write or paste syntax and add it if you're using a GUI version
of SPSS). REGRESSION won't explicitly give the PRESS statistic, but
it's fairly easy to compute: save the deleted residuals, square them,
and compute the sum of those squared values.
 
--
-----------------------------------------------------------------------------
David Nichols             Senior Support Statistician              SPSS, Inc.
Phone: (312) 329-3684     Internet:  nichols@spss.com     Fax: (312) 329-3668
-----------------------------------------------------------------------------