[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: Which random effect estimators use Gauss-Hermite
The manual points out that after running xtlogit with random effects, one
should use quadchck (though I don't see why this isn't just the default in
xtlogit). This is because the quadrature method of computing the log
likelihood and the derivatives may give unstable estimates. This makes
Rather cryptically the manual (version 7.0) also says in the 'quadchck'
entry "Some random-effects estimators in Stata use Gauss-Hermite
My questions are:
1. Which estimators do and which don't use G-H quadrature?
2. Or, if #1 is too much to answer, what does xtnbreg use?
3. How can we check for 'stability' with whatever other methods are used?
4. If Stata does use different optimization methods for different
estimators, what determines which algorithm is chosen?
5. And how about making (perhaps a non-noisy) quadchck a default for those
xt family members that need it?
Alex Cowell, Ph.D.
Behavioral Health Economics Program
3040 Cornwallis Road PHONE:
PO Box 12194 FAX:
Research Triangle Park E-MAIL:
North Carolina 27709-2194
* For searches and help try: