Statalist


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

st: question: random-coefficient models with bernoulli-distributed coefficients


From   László Sándor <sandorl@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: question: random-coefficient models with bernoulli-distributed coefficients
Date   Sun, 4 Oct 2009 15:04:54 -0400

Dear all,

I have been trying to get my head around a seemingly intuitive
problem, but I am puzzled that I don't really find relevant treatment
(nor implemented tools) in the literature. Perhaps I am missing the
obvious, but I would be grateful if you could comment on what this
problem looks like and how I could estimate it:

I want to allow different treatment effects by clusters, but only want
to allow two different treatment effects. (These could be: compliers
and non-compliers in the execution of the experiment.) This seems to a
simple, straightforward case of random-coefficient (also called
growth-curve) models within the larger class of multilevel models,
only restricting the distribution of the coefficients differently than
being random. However, after checking the -xtmixed- and -xtgee-
commands, or the treatment of -gllamm- in "Multilevel and Longitudinal
Modeling Using Stata" (2ed) from Stata Press, I have not found
anything useful.

I can also imagine that this problem is close to "latent class" models.

Any help will be highly appreciated.

Thank you,

Laszllo
*
*   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   |   What's new   |   Site index