While the MI package will automatically combine estimates and standard
errors from analysis models across datasets, the formulas are not
themselves very complicated (I hope Ken Scheve doesn't take that as any
knock against the usefulness of those procedures).
Run your analysis, or compute your means, or whatever parameters you are
interested in, in each dataset, then average these to get your quantity of
interest. For the standard errors in these estimates, the formula has
a second term. You need to average the standard errors, and also add
uncertainty if the different datasets are predicting different model
parameters. The formula is in:
http://gking.harvard.edu/files/evil.pdf
in equation (3). It is the same formula used in model averaging, only
here the model weights are equal, since each imputed dataset is equally
valid.
regards,
James Honaker
On Sun, 30 Nov 2003, paul and penny wrote:
Hi all,
I do not have a licence for stata to combine the imputations for amelia. I have thought
about using an aggregate function in SPSS to generate means and standard deviations for
cases across all the five files (My gut instinct is that this is way too simple). I have
also thought about using NORM to combine the five file parameters (Again seems to simple).
Any advice on how to combine the amelia data files without stata would be much
appreciated.
Kind regards Paul
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia