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: A question about identification strategies


From   Nils Braakmann <nilsbraakmann@googlemail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: A question about identification strategies
Date   Thu, 1 Apr 2010 10:18:40 +0200

Dear Charlie,

i'd say it depends on the question. A situation where I could imagine
that this arguments makes sense: Say, you're interested in evaluating
a training program where unemployed individuals are assigned to the
program by some case worker based solely on information contained in
some files (i.e., there's no personal contact between the case worker
and the unemployed) and you have access to these files. In these
cases, a claim that assingment is essentially random conditional on
the information contained in the files might have some merit as the
assignment mechanism should depend only on the information available
to you. However, identification would still require the assumption
that the remaining fluctuations in treatment assignment are
essentially random, which might be still be violated if, for instance,
treatments are assinged on a first come, first served base so that
individuals who become unemployed at the beginning of the month are
more likely to get treated (and this time difference matters for the
outcomes).

Hope this helps,
Nils

P.S.: I guess, Angrist (1998, "Estimating the Labor Market Impact of
Voluntary Military Service Using Social Security Data on Military
Applicants", Econometrica 66(2), pp. 249-288) uses a version of this
argument in his first strategy. If I come across other references, I
let you know.

On Thu, Apr 1, 2010 at 4:20 AM, C Engelbrecht <cngelbrecht@gmail.com> wrote:
> I read in a couple of working papers where the authors claim that they
> "obtain all information presented to the agent", therefore they should
> be able to claim causality using a simple OLS. I certainly sympathize
> because IVs are hard to come by, but is this a valid argument? Have
> you seen similar arguments made in published papers?
>
> Many thanks for any reference you can point me to.
>
> Charlie
> *
> *   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/
>

*
*   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