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: Computing predicted probabilities in multi-level model (xtmelogit)


From   Philip Ender <ender97@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Computing predicted probabilities in multi-level model (xtmelogit)
Date   Fri, 22 Oct 2010 10:18:41 -0700

"Michael.Stuetzer" <Michael.Stuetzer@uni-jena.de> had a question on
computing predicted probabilities from -xtmelogit-.  In particular, he
had a question on how to handle an interaction term in the model.
First off, the method he selected from the ATS Stat Consulting FAQ
page,

http://www.ats.ucla.edu/stat/stata/faq/xtmelogit_prob.htm

is meant to produce predicted probabilities that include both fixed
and random effects.  If the predicted probabilities using just the
fixed portion of the model are acceptable then the process is much
easier using -margins- with the predict(mu fixedonly).

If I understand things correctly, the variable,
int_ownmge_bipwachstum, is the interaction of the categorical
variable, Zownmge, and the continuous variable, Zbipwachstum_l1t5.  To
get the correct probabilities the interaction term needs to change
when Zownmge changes.  When Zownmge equals zero the interaction should
also equal zero.  When Zownmge equals one the interaction value equals
Zbipwachstum_l1t5.

So hopefully I understand your model correctly.

Phil
-- 
Phil Ender
UCLA Statistical Consulting Group
*
*   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