LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (July 2001)Back to main SPSSX-L pageJoin or leave SPSSX-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Wed, 11 Jul 2001 11:33:50 -0300
Reply-To:     Dan Waschbusch <dan.waschbusch@DAL.CA>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Dan Waschbusch <dan.waschbusch@DAL.CA>
Subject:      simple interaction effects
Content-Type: text/plain; charset="iso-8859-1"

Hi,

I have a significant three way interaction (A x B x C), where each factor is between subjects with 2 levels, and I would like to decompose it using simple effects test through GLM procedure. That is, I'd like to be able to test whether there is an A x B interaction at C1 and whether there is an A x B itneraction at C2. I have been tyring to find an answer as to how to do this, and have not yet come up with a satisfactory response. Options I've encountered so far:

(1) Run new, seperate analyses examining A x B at each level of C. For example, if C = gender, then examine whether A x B is significant for males, then do the same for females. This solution is what I've found on many internet sites, but I'm concerned because it changes the number of subjects and degrees of freedom, relative to the original analysis (A x B x C) that included all factors. That's not really a simple effets test, is it?

(2) I found a "white paper" on the SPSS website that describes very clearly teh procedure, but it's written for hte old MANOVA and doesn't work for GLM (see http://www.spss.com/tech/stat/articles/simple.txt) about 1/2 down, under the heading General Between Models: Estimating Simple, Simple Main Effects, and Simple Interaction Effects")

(3) Use BMDP

(4) Use the EMEANS and COMPARE commands. I have tried this, but it only seems to give the simple effects of one factor at every level of the other factors (e.g. A at B1C1, A at B1C2, A at B2C1, A at B2C2), but it doesn't seem to test interactions (i.e. A x B at C1, A x B at C2).

Can anyone tell me how to decompose a 3-way (or higher) interaction using simple effects tests in SPSS (e.g. examine A x B at each level of C)?

Thanks, Dan Waschbusch

*************************************************** FREE SUMMER CAMP for elementary age children with and without ADHD: www.dal.ca/STP *internships for undergraduate and graduate students also available ***************************************************


Back to: Top of message | Previous page | Main SPSSX-L page