Repeated measures : impact of some variable on the participant's evolution


New Member

This may seem like a basic question but I can't manage to find a clear answer online, so I would like to ask it here to be able to discuss it.

My problem is simple : I have a set of participants belonging to two groups (experimental and control), which have filled in a survey at two moments in time (before and after a treatment for the experimental group, before and after doing nothing for the control group).

In order to test the effect of the treatment, I simply do a repeated measures ANOVA with one between subject factor (the group) and one within subject factor (the time / the treatment). So far so good.

Now, I also measured another variable for the experimental group before the treatment, lets call it X, which is a continuous variable. And I want to see if the effect of the treatment depends on X. That is, if it is for example more effective for high values of X, and less effective for low values. This clearly looks to me like some kind of interaction, but I'm unclear how to test it in SPSS.

I suppose I should include X as a covariate (after centering it)? And then, what? Where in the outputs can I see what I'm looking for? Is it the interaction between the covariate and the within subject factor (time*X) ?

I'm unsure because I understand covariate analysis as controlling for the effect of the covariate while investigating another effect ; I'm not sure if this is also the right approach for the kind of interaction I'm interested in.

Thanks in advance for your insights! :)
Hi Jero,

Yes, include X as a covariate.

The part that might be tricky, though, is you only have values of X for the treatment group (if I'm understanding it correctly).

If X is missing for the control group and you include X in your repeated measures model, SPSS will drop everyone in the control group.

So you may have to do it as a separate analysis just for the treatment group.