Skip to content
Prev 154291 / 398500 Next

Variance-covariance matrix

Laura Bonnett wrote:
You need to read up on vectorized arithmetic, (and don't use fix() to
define functions, x1 <- function(....) {....} will do, and what's with
the unused arm() function??).

However, if I get your data structure correctly, what I meant was that
you could just look at

fit3 <- coxph(Surv(Withtime,Wcens)~Treat, data=armb,method="breslow")

(possibly factor(Treat) instead). If VPS is your control, I think you'll
even get the desired coefficient directly from the output because the
group comparisons will be against LTG.