Statalist The Stata Listserver


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

Re: st: mprobit and ivlogit?


From   "Clive Nicholas" <Clive.Nicholas@newcastle.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: mprobit and ivlogit?
Date   Fri, 1 Sep 2006 08:26:15 +0100 (BST)

Richard Gates replied to Nishant Dass:

[...]

> -mprobit- is slow when compared to -mlogit- since it uses quadrature to
> approximate the multivariate normal distribution.  Moreover, it only
> computes
> the first order derivatives of the log likelihood analytically (in -ml-
> terms
> it has a d1 likelihood evaluator) so the second derivatives are
> approximated
> numerically.  All compuations are done in ado an Mata.  In comparison,
> -mlogit-
> is implemented in C and computes the second derivatives analytically.
>
> If you run the document example, the sysdns3 data, -mprobit- is about 900
> times
> slower than -mlogit-.  My guess is you have a large data set with many
> alternatives so Stata's version of -mprobit- may not be a feasible routine
> to
> use, or you need a faster computer.  Note, that -mprobit- is using Newton-
> Raphson (-ml, technique(nr)-) and you can get a faster feedback in the
> iteration
> history if you use technique(bfgs), but the overall computation time will
> probably be the same.

I'd like to add to Rich's valuable tips on getting the best out of
-mprobit- by tackling Nishant's comment that, given his data and model, he
makes no distinction between -mlogit- and -mprobit- on 'theoretical'
grounds.

Both -mlogit- and -mprobit- assume that the categories of a model's
dependent variable are conceptually distinct from each other, which can of
course be most effectively identified by running an 'independence of
irrelevant alternatives' (IIA) test such as -mlogtest, hausman base-.
(Although such tests are not without their critics.) Both also assume that
the errors associated with the DV categories are uncorrelated. Both assume
that the data are Normally distributed.

Indeed, given all the above, and given Nishant's complaint that -mprobit-
is much slower than -mlogit- (it certainly is), I can't think of a single
good reason why _anyone_ would want to use -mprobit-, especially as it has
no analogy to odds ratios -mlogit-, which can be extremely useful.

My advice would be to kick -mprobit- into touch and stick with -mlogit- if
possible.

CLIVE NICHOLAS        |t: 0(044)7903 397793
Politics              |e: clive.nicholas@ncl.ac.uk
Newcastle University  |http://www.ncl.ac.uk/geps

Whereever you go and whatever you do, just remember this. No matter how
many like you, admire you, love you or adore you, the number of people
turning up to your funeral will be largely determined by local weather
conditions.

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index