On Feb 3, 2004, at 2:33 AM, Kalina wrote:
I am not able to use the dwstat or bgtest tests in Stata 7 with a panel
data. Is there a fix for this problem or other tests I can use
instead?
Yes (sort of). 'findit panelauto' will locate a set of minor
modifications to official commands that I wrote. These routines will
work on ONE timeseries of a panel: that is, you may analyze the
residuals of unit i alone (presumably with some sort of loop over i).
This is NOT a "panel durbin watson test', if you consider such a test
one that gives you a single verdict on the panel, but one that removes
the annoying restriction that many of Stata's official time-series
commands (including unit root tests) will not work on a panel unless
you drop all but one timeseries of the panel.
Also please note that 'bgtest' and 'durbinh' (Baum and Wiggins, STB-55)
are now implemented in official Stata as 'bgodfrey' and 'durbina'. The
latter routines are to be preferred, since you can consult StataCorp
tech support for assistance with them.
KIt
*
* 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/
