there are ways of combining sets of parameters
rather than a single parameter, but in almost all cases the ultimate
quantity of interest is still a single parameter. so whatever you were
going to do with the covariance matrix is what I'd consider the
ultimate quantity of interest. then you can use the usual combining
rules described in our article and in the manual.
Gary
---
http://gking.harvard.edu
On 03/29/2009 08:46 AM, Simon Wigley wrote:
Hello
How (if at all) can we create a single covariance matrix by combining
the m imputed data sets?
Many thanks in advance
Simon Wigley/ Arzu Akkoyunlu
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive: http://lists.gking.harvard.edu/?info=amelia
More info about Amelia: http://gking.harvard.edu/amelia