[Erik Ruzek]

Clyde Schechter I am trying to interpret the random slope coefficient for the group#time interaction. I guess it is telling you whether there is variation in outcome growth within groups? That's not something I've seen included often. I would suggest that Jeffrey consider using likelihood ratio testing (lrtest in Stata) to determine whether a model with the group#time random slope fits the data better than the model that just includes the time random slope with the cross-level interaction in the fixed effect only.

[Clyde Schechter]

Yes, it would estimate variation in outcome group within groups. Your suggestion of a likelihood ratio test for that is a good one.

[Jeffrey Chan]

Thank you very much Erik and Clyde! I will take all of your advice into account