I'm having troubles interpreting and comparing effect sizes in factorial designs based on partial eta squared values. E.g. in a 3x3-within-subjects design with:
- η²part, A = .2
- η²part, B = .4
- η²part, AxB = .1,
can it be concluded that:
Effect of Factor A is half (or twice) as strong on DV as Effect of Factor B (respectively Effect of Interaction)? How can this results be reported? I'm having cases where Factor A accounts for more percentage of total variance (in terms of generalized η²), so its kind of hard to comprehend for me.
I appreciate any help on how to get that together. Thank you very much on this.
