No, I wasn't referring or alluding to
this FAQ, but it clearly points you
in the right direction.
Suppose for example you
. use http://www.stata-press.com/data/r8/nmihs.dta
. svymean age
Survey mean estimation
pweight: finwgt Number of obs = 9953
Strata: stratan Number of strata = 6
PSU: <observations> Number of PSUs = 9953
Population size = 3898922
------------------------------------------------------------------------------
Mean | Estimate Std. Err. [95% Conf. Interval] Deff
---------+--------------------------------------------------------------------
age | 26.28999 .0775855 26.1379 26.44207 1.920533
------------------------------------------------------------------------------
It would seem that you can then knit your own standardized score by
. gen stdage = (age - _b[age]) / ( _se[age]*sqrt($S_E_nobs/S_E_deff[1,1]))
However, on -svy-, never believe an answer unless experts on -svy- agree.
Nick
[email protected]
anju parthan
> Thanks. I am not sure I understood your comment.
>
> Are you referring to the explanation given by William
> Sribney in the following link
>
> http://www.stata.com/support/faqs/stat/supweight.html
> --- Nick Cox <[email protected]> wrote:
>
> > Neither -zscore- nor -center- supports
> > pweights.
> >
> > This seems to be the first mention in this
> > thread of survey data. Is the concept of
> > standard deviation well defined for survey
> > data? If not, the whole idea of a standardized
> > variable would appear not defined in turn.
> >
*
* 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/
