Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: interactions for xtmelogit models
From
[email protected] (Roberto G. Gutierrez, StataCorp)
To
[email protected]
Subject
Re: st: interactions for xtmelogit models
Date
Tue, 23 Feb 2010 14:12:19 -0600
Dale Hardy <[email protected]> asks:
> A few months age I asked you a question about xemelogit interaction and you
> send me the link below.
> On page 15/36 of the presentation by Gutierrez entitled, Â?Tricks of the
> trade: getting the most out of xtmixedÂ?, is a model for xtmixed
> interaction.
> My question is: Why isn't age not in the model? Age is with the interaction
> terms but not as a single term. Would that make the model not hierarchally
> well-formed?
The variable -age- is in the model, it's just that it is coded as two
interactions of complimentary binary variables with age, rather than as one
main effect and one dummy interaction. That is, instead of the standard
. xtmixed ... age ageXgirl ...
You have
. xtmixed ... ageXboy ageXgirl ...
The model is the same in both, but the latter coding is used to match what was
used in the random part of the model, where such a coding makes interpretation
easier -- you have one set of variance components for boys and one for girls.
--Bobby
[email protected]
*
* 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/