On 8/10/07, Janelle Knox <[email protected]> wrote:
> I am trying to create a sample weight for a dataset, which will
> correct for variations in gender, age, etc from population means.
> Does anyone know how to do this, or where I can find information for
> setting up a sample weight.... pweight=?
Yet another route to take is to use Stata's post-stratification
adjustments, as it looks more like what you are doing there. The
[pweights] are supposed to be the inverse probabilities of selection,
and are treated as such by Stata's -svy- procedures... leading to
Horvitz-Thompson estimators, etc. The post-stratification applies to
the already collected data with known design features, and works by an
additional ex-post re-weighting of the data, so that the re-weighted
sample margins equate the known population ones -- which seems to be
exactly your problem. So read up [SVY] post-stratification to see how
it is done; rather than matching the means and such, Stata's
implementation of post-stratification requires categorical information
on the total groups sizes. My understanding is that if you produce
weights through post-stratification procedures, and start using them
as probability weights, you would get incorrect inference, although I
am not 100% positive on that -- I have not studied that area of survey
statistics. Here's a couple of articles on the general topic of
post-stratification to help you out:
http://www.citeulike.org/user/ctacmo/tag/poststratification.
--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: Please do not reply to my Gmail address as I don't check
it regularly.
*
* 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/
