Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: question re: bootstrapping joint tests


From   john.bluedorn@economics.oxford.ac.uk
To   statalist@hsphsun2.harvard.edu
Subject   st: question re: bootstrapping joint tests
Date   Thu, 20 Jul 2006 15:57:22 +0100 (BST)

Hello,

I was hoping to do a joint hypothesis test using the bootstrap (similar to
doing a bootstrapped t-test). However, it appears that Stata does not have
any "_variables" (system variables) with which to capture the relevant
covariances for each replication, and thus calculate the joint test
statistic for each replication. This is eminently feasible for a t-test
(if it weren't already available), using something like:

bootstrap tstat = _b[X1]/_se[X1]: regress Y X1

Does anyone know if it is possible/how to generate a bootstrap sample of
joint test statistics (which use covariances), which could then be used to
calculate critical values (e.g., 2.5 and 97.5 percentiles)?

I've tried using e(V) to get to the covariances in each replication, but
Stata doesn't appear to like referencing matrices when it's doing the
bootstrap. Thanks in advance for your suggestions.

Best,
John

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
John Bluedorn, Research Fellow
Dept. of Economics, Univ. of Oxford
Manor Road Building
Manor Road
Oxford OX1 3UQ, United Kingdom
Email:  john.bluedorn@economics.ox.ac.uk
Office:  +44 (0) 1865 281 483
Fax:  +44 (0) 1865 271 094
Mobile:  +44 (0) 7771 535 765
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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   |   What's new   |   Site index