Re: st: stcox: weighted regression changes the # of observations

[Michael McCulloch <mm@pinest.org>]

**RE-SENDING BECAUSE OF DISPLAY PROBLEM**

>  > *Note: I'm using inverse probability of treatment weighting, with the
> >  following command immediately preceding -stcox-:
> >  stset datedied [iweight=wt_unst_II]
> -bootstrap- is not meant for weighted data.   Currently -stcox- with the
> -vce(bootstrap)- option does not complain when weights have been -stset, but
> it should (we will fix this soon).
Dear Statalist members,
In my prior posting, Jeff Pitblado helpfully pointed out that 
-bootstrap- is not meant for weighted data. I'd like to ask advice on 
alternate strategies.

I wish to do Cox survival analysis, with weights (inverse probability 
of exposure), and with bootstrapping of the standard errors. Each 
observation has its own weight value (as below):

    weight 	  exposed      datetx    datedied 
     1.08515   yes   01feb1992   01sep2001 
     1.08515   yes   01jul1992   01jan1996 
     1.08515   yes   01apr1990   01jan1994 
    12.13202    no   01jun1991   01apr1993 

My question:
Because Stata's -bootstrap- is not meant for weighted data, could the 
following sequence work:?
	1. -stset- the data
	2. estimate weight AFTER -stset-
	3. run the bootstrapped -stcox- with weight included as a 
covariate, for example:
		bootstrap, reps(1000): stcox tcm weight

Thank you kindly,
Michael









--


Michael McCulloch, LAc, MPH
University of California at Berkeley
School of Public Health, Division of Epidemiology

Research office:
124 Pine St
San Anselmo, CA 94960-2674
tel	415-407-1357
fax	415-485-1065
email	mm@pinest.org
web	www.pinestreetfoundation.org
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/