Dear list, I have a problem regarding choosing the distribution family for a mixed model. I want to compare my response variable y with treatment A,B,C and D. In each of the treatments the distribution of y is differently. Is there a way to implement this in a mixed model? Kind regards, Maarten