you're not doing anything wrong; zelig/arima isn't set up to deal with
multiply imputed data sets. you'll have to do the combining "by hand"
i'm
afraid.
Gary
---
http://gking.harvard.edu
On Wed, Jul 21, 2010 at 11:49 AM, Brad Epperly <epperly(a)u.washington.edu>wrote;wrote:
I'm using Amelia for multiple imputation, and
after the creation of the
Amelia object and m data sets, I want to use Zelig to automatically combine
the results of my AR models.
I have no problem getting it to work for LS models, when using summary and
subset I can get either the combined model outputs or the output from
individual imputed data sets.
Using model="arima", however, this is not working. Instead, I always get m
identical summaries, and this error:
Error in tmp[, 1] : incorrect number of dimensions
This occurs on different imputed data sets, so I'm not quite sure what I'm
doing wrong. I can write (inelegant) code to do the averaging myself, but
it's not easily portable. Any help would be much appreciated.
Brad Epperly
PhD Candidate
Department of Political Science
University of Washington
Box 353530
Seattle, WA 98195
epperly(a)u.washington.edu