Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: RE: Stata 12 Announcement


From   "William Gould, StataCorp LP" <wgould@stata.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: Stata 12 Announcement
Date   Mon, 27 Jun 2011 10:24:01 -0500

Richard Williams <richardwilliams.ndu@gmail.com> asked, 

> will -sem- be a faster or more powerful alternative to -gllamm-, at 
> least for some problems?

These two commands obviously address different problems.  There is
some overlap in their features.  When there is, -sem- is faster than
-gllamm- because it benefits from using analytical first and second
derivatives that are specific to the problem at hand.

I have no test results comparing -gllamm- and -sem- directly.  We did
some tests, however, comparing -xtmixed- and -sem-, and on one of
those, I know the timings.  It was a growth-curve model with a small
number of measurements per subject, and heterskedastic variances.
Estimation took 100 seconds using -xtmixed-, and 1 second using -sem-.

On a personal note, I have used -sem- to fit many models and it 
results have always appeared instantly. 

Richard also asked whether -sem- would be "more powerful" than
-gllamm-.  The answer in yes in the sense that the amount of
post-estimation diagnostics available after -sem- is staggering and
because they are tailor-made for SEM models but, model-wise, is there
anything you can't estimate using -gllamm-?

-- Bill
wgould@stata.com
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index