SPSS: in regression test interaction between a continuous and a within subjects var

#1
I have a SPSS analysis question regarding how to use regression to test for a moderator in a 2 x 2 design when the 2 x 2 design has both a between and within subjects variable. This is the design: 2 (gender, between subjects) x 2 (chance of a happy marriage or chance of divorce, within subjects). This interaction is significant (using a mixed design ANOVA). Now I want to test whether religiosity (keeping it as a continuous variable) moderates the significant 2 x 2 interaction.
How do I do that in a regression? I would typically center the variables (gender, risk_happy, risk_divorce, religiosity) and then enter all the centered variables along with all the interactions. But I need to somehow tell SPSS that one of the variables is measured as a within subjects variable. How do I tell SPSS that a new variable “risk” is a within subjects variable? I need to do a risk x gender x religiosity interaction but because risk is within subjects it has to be specified somehow).
Obviously I could just do a medium split on religiosity and include it in an ANOVA but it is clearly better to keep the moderator variable continuous.
 
#2
Hi Helwegm,

You do it in General Linear Model-->Repeated Measures.

The first window that pops up will ask about your within-subjects variable. Specify your two levels of Marriage type.

Then in the next window, put in your two marriage variables for the within-subjects factor, Gender as a between subject factor, and Religiosity as a covariate.

If you click on Model, then choose custom, you can create the 3 way interaction.

Karen
 
#3
Hi Karen
Following your directions, it is possible to get the 2-way interactions between the variables but it is not possible (even if you select "3 way interaction") to get any output that shows the 3-way interaction between Risk, Gender, and Religiousity.
Thoughts?
Marie