Richard,
Thanks for alerting me to -tsfill-. You'd have had every right to blast me
for _not_ knowing about it properly. After using it, this is what
happened:
. tsset id betime
panel variable: id, 1 to 3457
time variable: betime, 1 to 102
. tsfill
. newey becon clmargin, lag(1)
betime is not regularly spaced
r(198);
. newey becon clmargin, lag(1) force
no observations
r(2000);
So -newey- _still_ wouldn't run after -tsfill- (and that's without the
-full- option)! Maybe it's simply not meant to.
I never got to -set trace on-, because Justina has cleverly provided the
solution: bascially, trick -newey- by switching off -iid-. I should have
known this: I do the same thing when running -prais-.
It has occurred to me that no particular regression may be ultimately
suitable for this dataset, given that we're talking about British
by-elections. In one by-election, I could be assessing the Conservative
vote in Glasgow Govan, the next in South West Surrey: both completely
different constituencies with completely levels of Tory support. What is
really needed, therefore, is to have the previous general election vote as
the lag. I know little about the mechanics of -reg3- and -ivreg-: would it
suitable here? I don't know.
Anyway, thanks for your help, though!
C.
> I don't know much about this, but I can note that -newey2- includes the
> command
> -tsfill, full-. Ergo, I suggest reading up on the -tsfill- command and
> figuring out if your data are violating any of its requirements.
>
> Also, if you are not sure exactly where a program is dying, it can help to
> specify -set trace on- before running it.
CLIVE NICHOLAS |t: 0(44)191 222 5969
Politics |e: [email protected]
Newcastle University |http://www.ncl.ac.uk/geps
*
* 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/
