Date:         Mon, 10 Oct 2011 07:08:09 -0400
Reply-To:     Philippe Cogneau <philippecogneau@SKYNET.BE>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Philippe Cogneau <philippecogneau@SKYNET.BE>
Subject:      Two questions about PROC PHREG

Hi,

I've two questions about the outputs of PROC PHREG:
- when I request for the OUTPUT with SURVIVAL, I receive "the survivor
function estimate at tau(j) where tau(j) is the observation time" according
to the documentation (p 4574). But which time is it? The time of the event
for the observation (j)? And then, for censored observations, is it the time
of censoring? Or is it possible to request for the value of the function
estimate at a series of selected times?
- I would like to build a model using half of the observations, and check
its validity using the overall C by application of the model on the other
half of the observations.
With PROC LOGISTIC, I can do that easily using the option "SCORE". Here,
this is not possible immediately, if I'm not wrong. I can get estimation of
the regression coefficients, using the OUTEST option; but which formula do I
have to apply using those regression coefficients and the covariates of my
test sample?

Thanks to everybody that will answer me.


Philippe