Well first of all, what kind of standard errors do you expect to get
from tabulations? One way to get the standard errors on your
proportions would be to generate the dummy variables corresponding to
the entries of the table (with -xi- or through -egen, group()- ) and
then take -svy brr : mean- of those. Generally, -brr- expects
something like the coefficient estimate vector, and -mean- does
provide that.
On 1/18/07, Alan Diener <[email protected]> wrote:
(thanks for the suggestion Maarten - your example did work for me)
I now realize that the problem occurs when using BRR and bootstrap weights.
The data set which I am using does not have strata or PSU info. I also
tried it using nhanes2brr and I was not able to obtain the SEs or CIs.
I have tried to use both of the following commands and some variants:
. svy: tab var1 var2, row se ci
. svy brr: tab var1 var2, row se ci
Neither method has produced the standard errors or confidence intervals. I
am able to get the estimates by using the proportion command and subpop,
but the output is not as clear as that produced by the tab command.
--
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/
