Split File for Mixed Measures

#1
This is a rookie question, so I apologize in advance -

I am doing a mixed ANOVA with 2 levels of within groups (Time) and 2 levels of between groups (Intervention status - control or intervention). I got significant main effects as well as a significant interaction effect. Given that there are only two levels, I don't think I need to post hoc the main effects, right? So I am trying to run a post hoc for the interaction effect.

In order to do this:

I split the file (separating by group intervention), then
I compared means with a paired sample t-test.

That's where I ran into problems - there was still only 1 intervention group and I didn't have a non-intervention group to pair it with.

What am I doing wrong?

--- Newbie stats guy
 
#2
Sounds like what you're doing is right. I'm assuming you're using SPSS. After you split file and run the paired t test, you should have two different outputs: one for the intervention and one for the control group. Since your interaction was significant, these two t tests should be different. Also, since the interaction was significant, there is usually no need to interpret the main effects.