Hi, I solved using SAS Studio that has neighbour option in PROC MCMC. I'd need to weight the single area (cod_pro) and the neighbors (cod_pro_adj) with their Observedevents . How could I do? follow my sintax without the weight ods select none; proc mcmc data=step1 seed=615926 nbi=10000 nmc=50000 thin =10 plots=none outpost=/*OUT.*/step1post; parms tau_b 0.5 tau_h 0.2 provar 0.5; parms alpha 0; prior tau: ~ gamma(0.5, is=0.0005); prior alpha ~ general(0); prior provar ~ gamma(0.001, is=0.001); random h ~ n(0, prec=tau_h/**poptime*/) s=cod_pro; random b ~ normalcar(neighbors=cod_pro_ad, num=num, prec=tau_b*num/*/poptime ObservedEvents*/) s=cod_pro; mu=alpha + b + h; model stdrate ~ normal(mu, sd=provar); run; I've attached the db with few rows. Many thanks Alessandra
... View more