We have to just pick one outcome variable for each experiment though, or it's going to get too complicated.