Re: st: mi estimate using gllamm and pweight

[Maarten Buis <maartenlbuis@gmail.com>]

On Tue, May 31, 2011 at 3:30 PM, Wike, Traci wrote:
> I am running a multi-level model with multiply imputed data and using propensity score weights. Because xtmixed does not allow pweights, I am using gllamm. Yet, when I try to run the model using mi estimate: (to account for the imputed data), I receive a message saying that mi estimate is not supported with gllamm. Does anyone have suggestions for how to handle this?

So you are trying to combine propensity scores with a multilevel model
with multiple imputation? If I were in a positive mood, I would say
that that is very brave. I am not, so I will say that I would not try
to combine that many techniques. Who knows what the properties of such
a combined estimator are? Each of them are reasonable approximations
of the thing they want to measure, but that does not mean that the
same is true for the combined estimator. At the very least you are
collecting an awful lot of assumptions with such a combined model.

Hope this is not too depressing,
Maarten

--------------------------
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/