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 on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: RE: xtabond2 for 2 DVs -- xtivreg2


From   Austin Nichols <austinnichols@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: xtabond2 for 2 DVs -- xtivreg2
Date   Thu, 16 Feb 2012 10:12:39 -0500

Lisa Marie Yarnell <lisayarnell@yahoo.com>:
See also
http://www.ssc.ca/survey/documents/SSC1996_P_Kott.pdf
which says (p43):
"The use of [IV in the presence of measurement error]
was suggested to me by Wayne Fuller."  ....
"The randomization-based theory supporting its use,
in general, has no documentation as far as I know."

On Thu, Feb 16, 2012 at 10:03 AM, Austin Nichols
<austinnichols@gmail.com> wrote:
> Lisa Marie Yarnell <lisayarnell@yahoo.com>:
> Note that pweights+cluster is all the survey machinery you need, since
> you are doing superpopulation inference.
> I.e. you would be computing SE for your IV models even if you had the
> whole population, not a sample.
> You can restrict to a subpop with an -if- qualifier in your case.
> If you have a stratified survey, you will not quite get the right SE,
> so not quite the right superpopulation inference, but the theory for
> IV superpopulation inference is not developed yet.
> See also http://www.jstor.org/stable/3182811 and more recent papers
> that cite that paper.
>
> On Thu, Feb 16, 2012 at 8:49 AM, Lisa Marie Yarnell
> <lisayarnell@yahoo.com> wrote:
>> Many thanks, Mark, for clarifying this!
>>
>> All best,
>> Lisa
>>
>> --- On Thu, 2/16/12, Schaffer, Mark E <M.E.Schaffer@hw.ac.uk> wrote:
>>
>>> From: Schaffer, Mark E <M.E.Schaffer@hw.ac.uk>
>>> Subject: st: RE: xtabond2 for 2 DVs -- xtivreg2
>>> To: statalist@hsphsun2.harvard.edu
>>> Date: Thursday, February 16, 2012, 7:26 AM
>>> Lisa,
>>>
>>> I'm afraid -xtivreg2- doesn't support any -svy- features,
>>> including subpop weights.
>>>
>>> It is true that -xtivreg2- does "support" pweights, but only
>>> in the sense that it allows them.  They're equivalent
>>> to aweights+robust.  How you interpret the results of
>>> estimating with pweights is up to you.
>>>
>>> HTH,
>>> Mark

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


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