I am computing a 3x3 mixed anova with time (day0, day1, day2) as the within subjects factor and stress-eating response (less than, more than, normal) as the between subject factor. Both the main effects and interaction are significant. My hypothesis is that there will be a difference between day0 and day1, and between day0 and day2 for the "normal" stress-eating response group only, and not for the other two groups. To evaluate such differences, can I do paired samples t-tests? For ex : one paired samples t-test for the "normal" group to show a difference between day 0 and day 1 and another comparing day 1 and day 2 for the same group. And then do similar t-tests for the other two groups and show there is not too much of a difference?
Hello Alisha. Use the EMMEANS sub-command for your interaction term with the COMPARE keyword. See an example in this TechNote:
https://www-01.ibm.com/support/docview.wss?uid=swg21475404
You can use the GUI dialog to generate a first draft of the code you'll need (if you exit via PASTE), but it will not generate the exact code you need. You'll have to do some editing in the syntax window.
HTH.
- Badges
- Science topic
- Similar topics
- Computer Science
- Computing