[R] : Re: coxme frailty model standard errors?

[Terry Therneau]

-- begin included message --
Hi, but why we do the difference : ltemp <- 2 * diff(tfit
$loglik[1:2]) ??

Where I can find information about Integrate Likelihooh and null like
lihood??

--- end inclusion ---

 1. Basic statistical fact: 2 * difference in loglik between two nested
models = distributed as a chi-square distribution.   For coxme loglik[1]
= likelihood from a null model (all coefs =0).  

 2. In a random effects model the random terms are integrated out of the
likelihood. This is a large topic and it's not easy to give a simple
reference other than 'read one of the class textbooks on mixed effects
models'.