[R] duration analysis

Dimitri Szerman dimitrijoe at ipea.gov.br
Mon Mar 6 18:17:33 CET 2006


It seems the list faced some problems during the weekend, so I am re-sending
this message.

I am trying to estimate the effects of covariates on the hazard function, rather
than on the survival. I know this is actually the same thing. For example, using
the survival package, and doing:

>> myfit <- survreg( Surv(time, event) ~ mymodel )

all I have to do to get the quantities of my interest is

>> -myfit$coefficients/myfit$scale

The standard erros are easily worked out, as the absolute z-statistics are the same.

Ok, so I can get easily what I want from survreg(). But I'd like to get it even
more directly. Does anyone know some function to do it?

Thank you,
Dimitri Szerman

More information about the R-help mailing list