In repeated measures within subject design, the SPSS calculates pairwise comparisons for each factor values. For example the factors are medication A (1, 2, 3), medication B (1, 2, 3, 4) and time (1, 2, 3, 4, 5). SPSS will calculate pairwise comparisons for each factor (1 vs 2; 1 vs 3; 1 vs 4; 2 vs 3; 3 vs 4) and the same for other factors. But it does not calculate for example: Medication A1 x Medication B1 x Time 1 (and so on).
When all the interaction effects are significant, what is the best way to compare pairwise comparisons between factors level? Shall I calculate it by paired t test? and how to correct alpha level? Is there anyway to calculate it in SPSS without requiring to paired t-test?
I'm not real familiar with SPSS, but there should be an option to use EMMEANS as a post-hoc analysis. This is what you want to use. You don't want to use t tests. I imagine EMMEANS can handle interactions, but you may need to use script and not the gui menus.
This is another reason to avoid SPSS. You could run your model as a straight regression which is much easier and then just use Bonferroni or your favorite alternative.. The regression form of your model is discussed at this link:
https://www4.stat.ncsu.edu/~davidian/st732/notes.html
you can also find it here:
https://www.amazon.com/Design-Analysis-Experiments-Douglas-Montgomery/dp/1119589061/ref=sr_1_1?keywords=Montgomery+Design+and+Analysis+of+Experiments&qid=1566752991&s=books&sr=1-1
For my favorite methods see the attached. Best, D. Booth
OOPS Mea Culpa I missed something. If you are talking about interaction like A1xA2 then these are what you are interested in. There's an old saying that in the presence of interaction main effects are meaningless. Thus are you sure 1. your model is set up correctly? and 2. Do you have a factorial design-this is the only way you can study interaction. See the attached paper. Best, David Booth
Hi Carmelo. You could modify the sintax in "paste" option when running the main ANOVA and choose Bonferroni, for example, for correcctions. In that option you could command to do all comparissons for significant interactions, that I guess is what you want to do
- Badges
- Science topic