Suppressor variables

[Larry Lutsky]

I am running a stepwise regression to determine which of 8 scale scores
predict overall satisfaction at a college.  All the independent variables
are postively correlated with the dependent measure and with each other.
However on variable (one of the 3 signifiacnt ones) had a negative beta
weight.  This variable was correlated .45 with the dependent measure and
had correlations between .45 and .79 with the other 7 independent
variables.  I know a change in sign of a beta weight indicates net
suppresion.

My question is how do I interpret the negative beta weight?  Do I just
present the results and indicate that this variable explains a significant
amount of variance because it suppressed the error variance in the other
variables?

[Ornelas, Fermin]

Make sure you run collinearity diagnostics on your regression. Often if
a variable does not have the expected sign it is an indication of high
correlation among the predictor variables. The value of .79 suggest that
the information provided by one of your predictors overlaps with the
information already contained in another predictor. As you may be aware
this has an impact on statistical inferences but not on prediction.

Fermin Ornelas, Ph.D.
Management Analyst III, AZ DES
Tel: [phone]
E-mail: [hidden email]


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Larry Lutsky
Sent: Tuesday, May 15, 2007 9:53 AM
To: [hidden email]
Subject: Suppressor variables

I am running a stepwise regression to determine which of 8 scale scores
predict overall satisfaction at a college.  All the independent
variables
are postively correlated with the dependent measure and with each other.
However on variable (one of the 3 signifiacnt ones) had a negative beta
weight.  This variable was correlated .45 with the dependent measure and
had correlations between .45 and .79 with the other 7 independent
variables.  I know a change in sign of a beta weight indicates net
suppresion.

My question is how do I interpret the negative beta weight?  Do I just
present the results and indicate that this variable explains a
significant
amount of variance because it suppressed the error variance in the other
variables?

NOTICE: This e-mail (and any attachments) may contain PRIVILEGED OR
CONFIDENTIAL information and is intended only for the use of the
specific
individual(s) to whom it is addressed.  It may contain information that
is
privileged and confidential under state and federal law.  This
information
may be used or disclosed only in accordance with law, and you may be
subject to penalties under law for improper use or further disclosure of

the information in this e-mail and its attachments. If you have received

this e-mail in error, please immediately notify the person named above
by
reply e-mail, and then delete the original e-mail.  Thank you.