I am looking for formal statistical tests of the robustness of regression
coefficient estimates. What I am currently doing is to plot how the
t-statistics of estimated coefficients change as I drop one observation at a
time from the sample.
In my experiment, this is not equivalent to testing for outliers because I
have three dimensions in my panel (countries, industries and time). I first
eliminate outliers by means of formal tests based on the three dimensions
(ijt). Then I wish to test for robustness of estimation results relative
only to the country and industry dimensions (ij).
Can anyone help me on this?
OECD Department of Economics
Structural Policies Analysis Division
2, rue Andr� Pascal, 75016 Paris, France
Tel. (33) 1 45 24 87 30
Fax (33) 1 45 24 13 47
E-mail [email protected]
*
* 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/
