Dear Amelia users and creators, I am using Amelia to multiply impute missing data in a data set. In this data set, more variables are assessed for cases than for the controls. There exists a strong correlation between this additional data and the missing values that I would like to impute. Should I incorporate these additional variables in Amelia, even if none of the controls have an existing value for these variables? It is not my intention to include these additional variables in any further statistical analyses, I just want to use this additional data to make a better estimation of the missing values. Thanks in advance, Roel Mocking
if you decide on some quantity of interest that would be output from VAR (there are a lot of possibilities, but you can do one at a time if you like), then use the combining rules described in our APSR paper. you'd have to do it by 'hand' (or write a little script) but they are quite simple rules.
Gary
---
http://gking.harvard.edu
On 08/07/2009 09:13 AM, D.J. van der Pas wrote:- 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/ameliaDear Amelia users and creators,
I am trying to fill in a ts dataset using Amalia, after which I want to do a VAR (Vector Autoregression) analysis in Stata. As I understand it Ken Scheve's MI program and Clarify can not automatically combine the results of a VAR analysis (or ARIMA). Is this correct? Is there an alternative way to automatically combine the VAR results?
Thanks a lot and kind regards,
Daphne van der Pas