Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Bootstrapping with sampling weights?


From   Michael Boehm <michael.boehm1@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Bootstrapping with sampling weights?
Date   Wed, 3 Oct 2012 10:51:13 +0100

Dear Stas,

Thanks a lot for this. I'll read your paper and see whether I'll use
the methods described there or rather do a different correction to my
standard errors than the bootstrap.

Michael

On Tue, Oct 2, 2012 at 10:54 PM, Stas Kolenikov <skolenik@gmail.com> wrote:
> Michael,
>
> (1) read my SJ paper, http://stata-journal.com/article.html?article=st0187
>
> (2) yes, if you have independent samples in those periods 20 years
> apart, you can treat them as (meta) strata. If you have stratification
> information for your data, you should cross it with year:
>
> egen metastrata = group( year strata)
>
> I really doubt that your data are "independent". If this is a personal
> interview survey, this is simply a very wrong assumption to make that
> would compress your standard errors to an unknown extent. You would
> need to figure out what the sample designs for both studies were.
>
> --
> -- Stas Kolenikov, PhD, PStat (SSC)  ::  http://stas.kolenikov.name
> -- Senior Survey Statistician, Abt SRBI  ::  work email kolenikovs at
> srbi dot com
> -- Opinions stated in this email are mine only, and do not reflect the
> position of my employer
>
> On Tue, Oct 2, 2012 at 11:59 AM, Michael Boehm <michael.boehm1@gmail.com> wrote:
>> Hi statalisters,
>>
>> I have survey data with sampling weights (assume the observations
>> themselves are independent) from two different periods twenty years
>> apart. I want to obtain bootstrapped standard errors for a pooled
>> regression of the kind
>>
>> regress logwage constant constant_period1 college college_period1
>> [pweight=weight_attr_afqt_hrs], noconstant hascons
>>
>> where college_period1 is college interacted with a dummy for the later
>> period, i.e. I want to know whether the return to college increased
>> over twenty years. I have two questions:
>>
>> (1) The vce(bootstrap) option is not allowed with sampling weights.
>> How do I correctly employ the bootstrap when different observations
>> represent different numbers of individuals in the population?
>>
>> (2) Since the dummy period1 represents a different population, should
>> I draw my bootstrap samples from the two periods independently? If
>> yes, does this amount to the strata(period1) option?
>>
>> Thanks,
>> Michael
>> *
>> *   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/
> *
> *   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/
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index