Comparing three groups that have been analysed by Friedman's ANOVA

#1
Dear all,

I have been struggling with the following problem and would be extremely grateful for some help. :confused:

I have three groups of participants. For each participant I have six measurements for my experimental manipulation.

I want to know whether there is an interaction of group with my experimental manipulation. If my data were parametric, I would simply run a mixed-ANOVA with my within-subject factor (6 levels) and my between-subject factor group (3 levels). Then I would look for an interaction. However, I have very non-normally distributed data which I have not been able to normalise via transformations.

I therefore analysed the three groups separately using Friedman's ANOVAs.

As predicted, the Friedman's ANOVA for Group A is highly significant with a chi-square > 20, whereas for Group C is it is non-significant with a chi-square around 2.

Of course this is not enough: I need to show if the effect is significantly STRONGER in Group A compared to C.

Is it possible to directy compare the Chi-square values? Or could one compare effect sizes directly between groups? Or how else would one compare between groups for such a design?

Best wishes
Tulip