Thanks for the quick response, Thomas. I am not sure that would be the reason they are different in my case since I am using a balanced data set. My understanding (albeit it is limited) is that regardless of the type of sums of squares used, in a balanced design they should lead to the same result. Thanks for trying to help.
Thank you Thomas for putting together this function. I like how you incorporated adjustments for sphericity and output all the significant post-hoc comparisons.
I am trying to run a 3-way repeated measures anova. I believe I have inputted the data and the levels correctly based on your example, however, the results do not match up to what I get using this other function also in File Exchange (RMAOV33). In addition, when I run the analysis in R I get the same results as the RMAOV33.
If you take one of your data sets and run RMAOV33 do you get the same results or are there any reasons why they would be different?