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]

st: Postestimation after regress, robust


From   Dana Chandler <dchandler@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Postestimation after regress, robust
Date   Sat, 13 Mar 2010 13:04:27 -0600

Hi fellow statalisters -

I am looking to get a sense of whether my data has any influential
observations in a panel regression I am running. Ideally, I would want
to predict and graph studentized residuals, cook's distance, and the
hat matrix - predict, rstudent, cooksd, hat.

However, the regression has several nonstandard options on it:
aweights(), robust(), cluster(). I think robust is the one that is
causing stata to return that I cannot run the postestimation commands.

I'm not sure whether the issue is that it's theoretically dubious to
talk about those measures after doing a robust estimation, whether
there is simply an additional package I can download somewhere, or
whether I would have to program something myself.

Another workaround might be to just try regressing with the aweights
and predicting the measures of influence (which stata DOES report
after reg y x [aw=N]). Since "robust" and "cluster" don't affect
parameter estimates, this might be sufficient for finding out
observations that are affecting parameter values. Does anyone think
that would yield a substantially different answer?

I'm also open to other ways of finding observations in my panel that
are influencing the regression.

Thanks,
Dana
*
*   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