I have a spread of data with 3 groups (A, B, and C). I originally normalized the averages of groups B, and C to the average of group A and propagated the error associated with those averages. However, when I try to run a nonparametric ANOVA (my data is non-Gaussian), I am not able to without the data spread!
It seems my only alternative is to normalize individuals within groups B and C to the average of group A. However, now I have no error to propagate from groups B and C, and am worried about the inaccuracy of this method along with the increase in false-positive rates!
What do I do? Is there a way to propagate error in this situation?
Thanks for your help!! :)