st: -mixlogit- module available from SSC

["Arne Risa Hole" <arnehole@gmail.com>]

Thanks to Kit Baum the -mixlogit- module for fitting mixed logit
models using maximum simulated likelihood is now available from the
SSC archive.

This is an update of -mixlogit- as published in SJ7-3 (st0133). The
new features include options for specifying weights (including
sampling weights) and for obtaining robust and cluster-robust standard
errors. The estimation speed has also been improved by using
analytical instead of numerical derivatives when maximising the
simulated log-likelihood function. This change has the side-effect of
producing somewhat different estimation results from the old version
of -mixlogit- with some datasets/ model specifications. The new
-numerical- option may be used to replicate estimation results
produced with the old version but this should only be used for this
purpose as it causes the module to run very slowly.

To install type -ssc describe mixlogit- from within Stata and follow
the instructions.

Comments and bug reports are welcome.

Arne
*
*   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/