Christie --
What "natural language" hypotheses do you have in mind when
you say
"test for interactions between the DV (Dependent
Variable) and covariates"?
It is very rare to find a "natural language"
hypothesis that involves statements about
"interactions between the DV and covariates".
Usually there are interests in
"interactions
among the predictor
attributes".
If you start with an ASSUMED MODEL and state your hypotheses
in terms of parameters of your
ASSUMED MODEL , then impose your restrictions on the
ASSUMED MODEL to
obtain your RESTRICTED MODEL and compare the 2 models via an
F-test you can do what
your hypotheses imply.
I will be interested to hear about what your hypothesis
"really is" (in natural language) and in
terms of parameters of your ASSUMED
MODEL.
.
-- Joe
**********************************
Joe H. Ward, Jr.
167 East Arrowhead Dr.
San Antonio, TX 78228-2402
Phone: 210-433-6575
Fax: 210-433-2828
Email: [EMAIL PROTECTED]
http://www.northside.isd.tenet.edu/healthww/biostatistics/wardindex
==============================
Health Careers High School
4646 Hamilton Wolfe Road
San Antonio, TX 78229
*********************************
Joe H. Ward, Jr.
167 East Arrowhead Dr.
San Antonio, TX 78228-2402
Phone: 210-433-6575
Fax: 210-433-2828
Email: [EMAIL PROTECTED]
http://www.northside.isd.tenet.edu/healthww/biostatistics/wardindex
==============================
Health Careers High School
4646 Hamilton Wolfe Road
San Antonio, TX 78229
*********************************
----- Original Message -----
From: "Christie Befort" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Monday, September 01, 2003 12:54 PM
Subject: Interaction terms in ANCOVA
> univariate ANCOVA model. The SPSS (version 10) GLM window does not
> allow to test for this type of interaction; it only lists the factors
> and covariates, but not the DV, from which to build the model. Is it
> possible to build in this term using syntax? Or do I have to create a
> separate variable for the interaction term and then enter that as the
> DV? Thanks,
> Christie
> .
> .
> =================================================================
> Instructions for joining and leaving this list, remarks about the
> problem of INAPPROPRIATE MESSAGES, and archives are available at:
> . http://jse.stat.ncsu.edu/ .
> =================================================================
