[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
RE: st: baseline function estimates for stratified estimation
Roberto G. Gutierrez answered Christian Kaiser
> > I did some Cox prop. hazard estimations using the strata
> option. According
> > to the reference manual, estimates of the baseline
> functions for each
> > stratum should have been stored. However, when I type
> stcox xvars, strata
> > (svars) basechazard(bch) basesurv(bsv) I only get the two
> vectors bch and
> > bsv. As far as I understand I should get one of these
> vectors for each value
> > of my svars. How can I retreive the single vectors?
> The estimated baseline functions are always generated as
> single variables.
> When you stratify, the variable containing the estimated
> baseline function
> is "stratified" as well. The estimated function values for
> each stratum are
> contained within the one generated variable and are
> organized to mirror the
> organization of your stratification variable.
> So, for example, if your variable -svars- consists of the
> values 1, 2, and
> 3, then
> . list bsv _t if svars == 1
> will list the estimated baseline survival function for the
> "first" strata,
> . list bsv _t if svars == 2
> for the second strata, etc.
> If this seems a bit sparse on information recall that the
> estimated cumulative
> hazard and survival function are step functions where the
> steps occur only at
> the observed failure times. Thus, even when you stratify
> you only need a
> single variable to contain all the required information.
As a footnote to Bobby's posting, check out -separate-.
. separate bsv_t, by(svars)
would give you a separate variable for each distinct stratum,
should you want them.
* For searches and help try: