s-news
[Top] [All Lists]

information from frailty models

To: s-news@lists.biostat.wustl.edu
Subject: information from frailty models
From: Jan Box-Steffensmeier <jboxstef+@osu.edu>
Date: Tue, 25 Jun 2002 16:01:58 -0400
Cc: sdeboef@psu.edu
Hi,

We are running monte carlos on the cox frailty models. We want some values
returned to us, such as the variance of the random effect, which we can get
by requesting theta. We would also like to get the statistical
significance returned (re: the p-value of the Chi-Square stat). Here is
our monte carlo code and an example of what is produced. How do we get the
p-value returned? We'd be very grateful for your help. Below is the relevant portion of our S+ script.

agfit.frailty <- coxph(Surv(sim1[,2],sim1[,3],sim1[,4]) ~ sim1[,5] + frailty.gamma(sim1[,1]))
                beta<-agfit.frailty$coef
#               naivese<-sqrt(diag(agfit.frailty$naive.var))
#There is no robust =true option so only one variance
                se<-sqrt(diag(agfit.frailty$var))
                frailty<-agfit.frailty$frail
                theta<-agfit.frailty$frail
                tmp<-cbind(i,beta,se,frailty,theta)
                write(tmp,"agfrailty.txt",5,T)
                #write.table(tmpf, "frailtb", sep = ",",T)

Thanks, Jan








Janet M. Box-Steffensmeier, Associate Professor
Ohio State University
2140 Derby Hall, 154 N. Oval Mall
Columbus, OH  43210-1373
Phone:  (614) 292-9642
Fax:  (614) 292-1146
E-mail:  jboxstef+@osu.edu
Web page:  http://psweb.sbs.ohio-state.edu/faculty/jbox/jboxstef.htm

"Yes, spanking a child stops a child from doing whatever it is
you don't want done.  Throwing a stick of dynamite into your
flower bed will stop the weeds from growing awhile." - Anthony J. Lapray


<Prev in Thread] Current Thread [Next in Thread>
  • information from frailty models, Jan Box-Steffensmeier <=