Date:         Wed, 31 Oct 2007 08:16:34 -0400
Reply-To:     Henry Crawford <hcrawford@LAKERESEARCH.COM>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Henry Crawford <hcrawford@LAKERESEARCH.COM>
Subject:      Re: SPSS errors when running a regression

SPSS unfortunately tends to be correct when giving this error -- so where
is our mistake?

Either status_new has only one value for all cases in the file or, in case
you've run the frequency and found this not to be true, it only has one
value FOR THE CASES THAT SPSS KEEPS FOR THE REGRESSION.  With pairwise
deletion, any cases with missing values FOR ANY of the variables (dependent
or independent) will be dropped from the equation.

I think it's very likely you have at least one variable in your regression
with missing values perfectly correlated with one of the two values of
status_new, leaving status_new with only one value among cases included..

-Hank


On Tue, 30 Oct 2007 12:18:12 -0400, Alina Sheyman <alinashe@GMAIL.COM>
wrote:

>Hi all,
> I'm running a liner regression in SPSS and am getting the following error
>message -
>"For models with dependent variable Tenure of Drops (Month), the following
>variables are constants or have missing correlations: status_new. They will
>be deleted from the analysis."
>
>The only thing that's special about this particular variable is that it's
>dichotomous with values of either 0 or 1. I've used variables like this in
>building regressions in the past and have never encountered any errors. Any
>idea what exactly might be causing this error to occur?
>
>thank you

=====================
To manage your subscription to SPSSX-L, send a message to
LISTSERV@LISTSERV.UGA.EDU (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD