Interactions using regression

Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 1
Joined: Thu Sep 24, 2015 1:36 am

Interactions using regression

Postby lmay91 » Thu Sep 24, 2015 1:43 am

Hi all,

I'm examining main and interaction effects using regression in SPSS, but have hit a dead end and would love some help. My variables are as follows:
- Outcome: disordered eating (continuous variable)
- Predictors: 1) peer relationships (continuous variable); 2) gene 1 (categorical variable - either 2 or 3 categories); 3) gene 2 (categorical - either 2 or 3 categories)
- Controlling for: socioeconomic status (continuous), ethnicity (continuous)

I'm looking at main effects of peer relationships, gene 1, and gene 2 on eating. I'm then looking at gene x environment interaction effects for gene 1 x peer relationships, gene 2 x peer relationships. Also looking at gene x gene interactions for gene 1 x gene 2. Finally, I'm looking at a 3-way interaction for gene 1 x gene 2 x peer relationships.

Basically, regressions have indicated that there are significant interactions, but I can't figure out where that significant effect is coming from. For example, is it from the 9R/9R, 9R/10R, or 10R/10R category/group of gene 1, etc. I've tried so many things from dummy coding to entering each combination separately, but no luck so far. Does anyone know how to tease this out??

I also need to plot the interactions on a scatterplot/line graph, controlling for socioeconomic status and ethnicity. Again, any hints?

Thank you SO much in advance.

Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm

Re: Interactions using regression

Postby GerineL » Mon Nov 09, 2015 10:15 am

figuring out which gene it comes from indeed is a matter of using dummies.
Plots I usually do myself in excel using the regression output, but there is some manuals online on how to do it in spss, for instance: ... ted-values

Who is online

Users browsing this forum: No registered users and 2 guests