[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: xtlogit is slow

From   Maarten buis <>
Subject   RE: st: xtlogit is slow
Date   Tue, 8 Jul 2008 12:05:29 +0100 (BST)

--- "Mitchell F. Berman" <> wrote:
> Good thought on the xtlogit issue, but no, the if clause is not the 
> problem.  Takes closer to an hour, I think, for the entire data set,
> but I've only done this once.  I can tell that it is running slowly,
> because each iteration takes many minutes.
> I sent this issue in to Stata Tech Support.  I'm hoping they
> investigate it.

I remember a talk by Bobby Gutierrez (from StataCorp) at the 2007
Nordic and Baltic Stata Users Group meeting where he discussed the
related -xtmelogit- command. He made the following remark about the
speed: The point with these models is that the maximum likelihood
solution requires integration over the random effects, which can be
very computationally intensive in non-linear models, like logistic
regression. Shortcuts exist, for instance in the form of PQL and MQL,
but these can be severely biased. For that reason StataCorp decided for
-xtmelogit- to integrate over the marginal effects using adaptive
Gaussian quadratures. Basically preferring to be slow but correct
instead of fast but wrong. Something similar may be the case with 
-xtlogit-. You can find that talk here:

-- Maarten

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Not happy with your email address?.
Get the one you really want - millions of new email addresses available now at Yahoo!
*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index