I need to compare means of 4 groups (2 categorical predictors and 1 continuous outcome variable). The homogeneity of variance assumption is tenable. However, due to the non-normal distribution and unequal sample sizes (different number of people in each group), I need to use a robust independent factorial ANOVA.
I’ve done the test and post hoc tests. Now I need to calculate the effect size. Field et al.’s (2012) manual suggests omega squared and there is a code for it (p. 542).
However, in the code, there is a place where you are supposed to enter the number of participants in one group since the code seems to assume equal number in each group, which is not the case with my data.
I tried to find whether I can use omega squared then or whether there is any other test that I need to use instead. Unfortunately, I had no luck so any piece of advice would be highly appreciated.
