[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: overidentification test
This is a repost of a message sent last week. I am hoping
that if I rephrase my question I will get a positive answer,
but donít worry If I donít I wonít repost it again.
I am interested in doing an overidentification test for a
tobit with instrumental variables. I have one endogenous
variable and three excluded instruments. If I were using
ivreg (ivreg2) I could use the command overid to do what I
want but I have a lot of zeros in my left hand side variable
and prefer to use tobit.
I have searched the statalist archives (and google) and I
donít think there is a code that can do what I want
automatically for tobit. My question is, do any of you have
any suggestions on how to test for the validity of
instruments using tobit? I already did an informal test of
the relevance of the instruments by testing their
significance in a ďfirst-stepĒ regression.
* For searches and help try: