[R] output of variance estimate of random effect from a gamma frailty model using Coxph in R

[Thomas Lumley]

On Thu, 21 Jul 2005 ghwei at umich.edu wrote:

> Hi,
>
> I have a question about the output for variance of random effect from a gamma
> frailty model using coxph in R. Is it the vairance of frailties themselves or
> variance of log frailties? Thanks.
>

For a Gamma frailty model it is the variance of the Gamma distribution, so 
the variance of the frailties.  For Gaussian frailty it will be the log 
frailties, though.


 	-thomas