Imputation for cross-classified random effects model

[scott.raynaud]

I’m interested in running an APC model via cross classified random effects on some NHIS data.  My interest is in some of the health outcome and health care access data controlling for several covariates.  I need a way of handling the missing data (I think these are MAR) in the data set that is congenial with the substantive model, at least to the extent that reviewers will accept it.  IVEware is always a possibility, but I’m not sure if it will pass muster in terms of congeniality.  I’ve seen Bayesean approaches in SAS (http://www.hindawi.com/journals/jam/2014/368791/) and Mplus (https://www.statmodel.com/download/Imputations7.pdf, see example 11.5) that seem to be equivalent to treating all the missing data as a single group and allowing them to correlate freely, however, this too does not seem to preserve the hierarchical structure of the data.  Mplus talks about modeling covariates, but there is very little documentation on how to do this.  The only way I could see how to write the probability statements for such an approach is to use the data itself.  Is that even legitimate?  So the question is what approach and software is best.  Thoughts?

[scott.raynaud]

Let's simplify this.  What do you think about using IVEware for imputation when the substantive model is a cross classified random effects  model?  Is that congenial enough?

[Trivellore Raghunathan]

​Yes, as long as the variables defining the cluster (associated with random effects) have been included in the imputation model. 

Raghu ​

--
You received this message because you are subscribed to the Google Groups "IVEware" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iveware+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Trivellore Raghunathan (Raghu)

Director, Survey Research Center

Research Professor, Institute for Social Research

Professor, Department of Biostatistics

University of Michigan

Phone: (734)-764-8365

Fax: (734)-763-9831

www.src.isr.umich.edu

www-personal.umich.edu/~teraghu

"A good life is filled with selfless actions full of compassion knowing well that we are all one"