I conducted a MANOVA in SPSS with two factors:
- Factor A (2 levels)
- Factor B (2 levels), and I have 14 replications for each group.
There are 14 dependent variables. After the MANOVA showed significant results, I attempted to perform post hoc tests. However, SPSS gave the following warning:
“Post hoc tests are not performed for A in split file B=H because there are fewer than three groups.”
Can I use post hoc tests in this case, and how should I proceed with further analysis?
I'm not sure, but in your case, post hoc tests are not directly applicable after a significant MANOVA for factors with ONLY TWO levels.
Post hoc tests are designed to perform pairwise comparisons between multiple groups (i.e., three or more levels). If I understood well, since your factors (A and B) have only two levels each, pairwise comparisons are unnecessary because there is only one pair per factor.
The warning you encountered indicates that post hoc tests cannot be run because your design does not meet this requirement of having at least three levels per factor.
But maybe someone else has a better idea.
Miklós Gyorgyovich is correct. With 2 levels a factor can only support one contrast which is equivalent to the F test of the factor testing the difference between the levels. If you ran this as a t test then the t^2 for the pairwise contrast equals F for the main effect.
Generally I'd avoid MANOVA anyway e.g., see
http://psychologicalstatistics.blogspot.com/2021/08/i-will-not-ever-never-run-manova.html
Wow! Thanks for the blog and the article! Very interesting what you say about MANOVA, I will think about it. The other article titles are interesting too, I will read them, even though I am a sociologist. :)
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python