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: xtmixed not producing LR results - how to compare models?


From   Maarten Buis <maartenlbuis@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: xtmixed not producing LR results - how to compare models?
Date   Tue, 18 Oct 2011 08:58:34 +0200

On Mon, Oct 17, 2011 at 9:40 PM, adlynch  wrote:
> I am using a 3 level model (time within individuals within cities) in xtmixed
> with 10 multiply imputed datasets in Stata 12. <snip>
> As a result, I am in particular having trouble assessing the importance of
> random effects.  I have no way of assessing whether adding random effects at
> level2 improves model fit.    Normally, I would run the lrtest command to
> compare nested models.

you should not use -lrtest- for that. The problem is that you want to
test a parameter at the boundary of its parameter space, the absolute
minimum value a variance can take is 0. Such tests tend to have weird
sampling distributions, requiring specialized commands. See for
example: (Gutierrez et al. 2001)  I don't think this has been
implemented for multiply imputed datasets.

Hope this helps,
Maarten

Gutierrez, R., S. Carter, and D. M. Drukker. 2001. On boundary-value
likelihood-ratio tests. Stata Technical Bulletin 60: 15–18.
<http://www.stata.com/products/stb/journals/stb60.pdf>

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   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