--- "Claudia L. Rangel" <[email protected]> wrote:
> You mention that you are using imputed data by means of the ICE
> command. If your dataset is of a multilevel nature, then ICE is not
> appropriate, since it only imputes with flat data structures
Say your second level is country then if you have enough complete cases
in each country you can perform separate imputations in each country.
This way your imputation model is more general then your substantive
model (multilevel model) and there is no problem.
Note the word complete in complete cases: -ice- won't complain if you
have more complete and incomplete cases then parameters in the
imputation model but it won't converge if the number of complete cases
is less then the number of parameters. This can be tricky since -ice-
gives you imputations without checking if it has convergenced. (I've
learned the hard way...) I am not critizising Patrick Royston (author
of -ice-) here. It is in the nature of the algorithm (a Gibbs sampler)
that declaring convergence is much more an art then it is in for
instance the algorithms usually used to maximize likelihood functions.
Hope this helps,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
___________________________________________________________
What kind of emailer are you? Find out today - get a free analysis of your email personality. Take the quiz at the Yahoo! Mail Championship.
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
