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 panel data xttobit


From   Stas Kolenikov <skolenik@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Bootstrapping panel data xttobit
Date   Wed, 23 Jun 2010 06:52:28 -0400

-bootstrap- resamples the whole cluster with this syntax. If companies
are nested within the clusters, then you are good to go. If you have a
relatively low number of clusters, your bootstrap will suffer small
sample bias: it is off by a factor of (n-1)/n where n is the number of
clusters. So with 20 or so clusters (when asymptotic inference is
problematic per se), you will surely get something quite a bit off the
target.

To see what exactly is going on, you can shorten your sample to say 5
clusters and 15 firms, so that it is tractable, and -bootstrap,
noisily: tab firm_ID year- or even -bootstrap, noisily: list firm_ID
year clustervar newclustervar-.

On Wed, Jun 23, 2010 at 3:38 AM, Catharina Klepsch
<catharina.klepsch@gmx.de> wrote:
> Hi together,
>
> I have a problem with the bootstrap command, and there are not a lot
> information referring to panel data problems..
>
> I am not sure if I get the right bootstrap standard errors if I have a panel
> data set (data are clustered for different companies)
> If I use the following command:
>
>
> Bootstrap, reps(500) seed (123) saving(ergbenis.dta)cluster(clustervar)
> idcluster(newclustervar): xttobit y x i(newclustervar), ul() ll()
>
>
> I just want to be sure that for each estimation of the bootstrap  the sample
> includes all datas for the company which was drawn (as there are sometimes
> different time periods for each company  available).  I am not quite sure if
> the above mentioned bootstrap guarantees this. Because I also did not find
> out when to use the group option?
>
> So if anyone has some experience with bootstrapping panel data this would
> help us a lot.
>
> Otherwise perhaps someone knows if it is possible to get more information
> about the drawn sample using bootstrap (which companies were chosen at the
> first replication which companies for the second rep.) to check the results.

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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