[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: re: heteroskedasticity questions
Am I correct in saying that -estat imtest- is White's general test,
or something like that? If I don't have any strong theory about the
form of hetero, would it make sense just to do -estat hettest- and
-estat imtest- routinely?
With the white option it is. The White test is just a special case of
Breusch-Pagan with a particular choice of -varlist-. -imtest- without
the -white- option is a broader test, due to Cameron and Trivedi,
which examines 2d-4th moments of the error distribution as I
understand it. White, Breusch-Pagan, -robvar-, Goldfled-Quandt all
just consider 2d moment.
Kit Baum, Boston College Economics
An Introduction to Modern Econometrics Using Stata:
* For searches and help try: