I am looking for a little help on determining the correct way to run a power analysis on a within-subjects repeated measures design (i.e., no between subjects variable) with combined contrasts. Originally, I was under the impression that I could run a power analysis on a repeated measures design with planned contrasts and not worry about the fact that we are combining the variables because everything is within subject. However, I have since realized that this is not the case because, for a three factor Repeated (Within) design, there will be 7 error terms owing to combination of main effects and interactions. Depending on how you want to form contrasts (and power or effect sizes) we pool different SS error and their associated df terms...
A SS Error(A)
B SS Error(B)
C SS Error(C)
and so on...
For example, we are looking at how different apps impact quality of life measures. One of the planned contrasts is to compare manufacturers (2). However, we would also like to collapse across manufacturers (i.e., combining the planned contrasts) to look at app use versus no app (baseline).
This website has been so helpful for me in my research, and I have been struggling for some weeks now to figure this out/get a straight answer out of some wonderful statisticians. Perhaps I am asking the wrong question, but hopefully someone here can help me. Any advice would be much appreciated. Thank you in advance!