RE: st: Creating time-dependent covariates to replicate the tvc option for stcrreg

["Travis Coan" <tcoan@law.harvard.edu>]

Great, this helps a lot. Thanks again Bobby.

Travis


-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Roberto G.
Gutierrez, StataCorp
Sent: Monday, April 05, 2010 2:24 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: Creating time-dependent covariates to replicate the tvc
option for stcrreg

Travis Coan <tcoan@law.harvard.edu> writes:

> I have one quick follow-up question. When using the -tvc()- option you
> cannot produce the CIF using stcurve. Would you recommend producing
the CIF
> in this situation -- i.e., when you've violated the proportionality
> assumption -- using the predict and the basecif option? Or am I
missing
> something here?

I would not recommend producing the CIF curve with -stcurve- after using
the
-tvc()- option, which is why it is not supported.  You really want to
think of
using -tvc()- as for testing the proportional hazards assumption, that
is, for
modeling time-varying coefficients.  When your coefficients vary with
time,
the concepts of baseline and of modifying baseline for a specific
covariate
pattern become muddled.

As such, if you decide you want to predict the baseline CIF function
with
-predict, basecif-, this is perfectly fine but you want to be careful
how you
transform this function to be the CIF for a specified covariate pattern.
When
you use -tvc()-, coefficients change with time, and you need to account
for
this in your calculations.

--Bobby
rgutierrez@stata.com
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/