Re: st: Last estimates not found after bs4rw

[Saidé Salazar <saide.salazar@gmail.com>]

Thank you Jeff for all your help and explanation. I am using Stata 13,
and as you correctly assumed, I need the -logit- results to do
something else. Actually, I want to perform the following steps:

1. Estimate logit
2. Estimate predicted probabilitites for out of sample observations
3. Obtain the mean of predicted probabilities
4. Repeat the process for each rweight.

So far, my do-file looks like this:

use myfile, clear
program define myprog
   syntax [pw iw] [if] [in] , first(varlist numeric) second(string)
third(string)
   preserve
   logit `first' if sample=="`second'"
   predict pre if sample=="`third'"
   mean pre
   restore
end
foreach y in y1 y2 y3{
   bs4rw, rw(wt*): myprog [pw=wt], first(`y' x1 x2 x3) second(sampleA)
third(sampleB)
}


The result I obtain is the mean table, but with missing values in all
columns except -Estimated Coef.-. It looks like Stata is using the
same sample every time, not the rweights.

I have tried with -svy- prefix, but I get "las estimates not found".
Here is the do-file:

use myfile, clear
svyset idstd [pw=wt],vce(bootstrap) bsrweight(wt*)
program define myprog
   syntax [pw iw] [if] [in] , first(varlist numeric) second(string)
third(string)
   preserve
   logit `first' if sample=="`second'"
   predict pre if sample=="`third'"
   mean pre
   restore
end
foreach y in y1 y2 y3{
   svy: myprog, first(`y' x1 x2 x3) second(sampleA) third(sampleB)
}


I would deeply thank any advice.

Regards,

Aranzazu
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/