Dear R users, Suppose that I generate the following : library(survival) set.seed(123) n=200 x=rbinom(n,size=1,prob=0.5) v=rep(rgamma(n/2,shape=1,scale=1),2) w=rweibull(n,shape=1,scale=1) b=-log(2) t=exp(-x*b-log(v)+log(w)) c=rep(1,n) id=rep(seq(1:(n/2)),2) group=rep(1:(n/10),10) fit2=coxph(Surv(t,c)~x+frailty(id,dist="gamma",sparce=T,method="em")) Using the command "fit2$frail" I get the id level frailty terms , Some of the frailty terms are negative , so Survival package gives not v but log(v) Am I wrong ? best regards, D.Lalountas University of Patras __________________________________________________ [[alternative HTML version deleted]]