Hi Dimitris, Thanks for your syntax! However, I tried it and saw that the mu-values saved with gm1 at mu probably contain the (posterior) estimate of the random effect as well, as plotting these against 'period' shows many different predicted probabilities for the same period. Ben.