 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: Standard Error on a Survival median, weighted data
Median is the "standard difficult case" in survey inference. Since the
underlying estimating function is non-smooth, the really standard
approach of Taylor series linearization does not work. One popular
approach is cdf inversion: you first get the interval around 1/2 of
your distribution function (in i.i.d. case, something like z times
sqrt( one half * one half / sample size); in more complex cases,
appropriately more different), and then your interval on x axis by
inverting the cdf for the endpoints of that interval. This is known
as Woodruff's method. I don't think I've seen an implementation in
Stata though.
See doi:10.1214/aos/1176347993, doi:10.1093/biomet/73.3.597.
On 12/17/08, [email protected] <[email protected]> wrote:
> Hello
>
> I have weighted data but need confidence intervals on the median survival time. I can pick out the median and get it to come out in sts list (see below) but the confidence intervals are suppressed because the data is weighted. Does anyone know how can I get confidence intervals on the median survival with weighted data without assuming a parametric model
>
> Thanks
>
> Milly
>
> stset time [pw=weight],fail(fail)
> stsum
> local median=r(p50)
> sts list, at(14 `median' 20)
>
>
> *
> * 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/
>
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.
*
* 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/