You probably meant post-stratification weights, or sampling and
post-stratification weights... Looking at -clogit- code, it seems like
it does not matter in the end whether you have [fw] or [iw] at the
important call to -_robust- that computes the proper standard errors.
So what I would practically do is to use importance weights and
-cluster-, disregarding the startification information, if any. If you
are curious enough, you can also try the resampling weights from -svy
jackknife-, just to see if your standard errors are comparable.
Omitting the weights whatsoever will lead you to bad bad biases not
only in standard error, but in point estimates, as well.
On 12/15/05, Anita Alves Pena <[email protected]> wrote:
> Dear Statalist,
> I have been struggling with what importance weights do in the clogit
> command. Clogit only seems to allow freqency weights and importance
> weights. My weights are post-sampling weights. In most processes, I use
> pweights but this doesn't seem to be an option here and there is no
> corresponmding svy command (that I know of at least). Does anyone know if
> I should be using importance weights or simply no weights at all here?
> Thanks!
> Anita Pena
> *
> * 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/
>
--
Stas Kolenikov
http://stas.kolenikov.name
*
* 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/
