as long as its linear, there is no problem, yes. it would be consistent
with the normal imputation model.
Gary
On Wed, 17 Nov 2004, Gina M Yannitell Reinhardt wrote:
Dear Amelia Help,
I know that I should not use variables in an analysis that were not used in
the imputation phase. That being said, is it possible to generate new
variables based on linear transformations of the imputed variables, and use
the newly generated variables in the analysis? For example, if I impute
values for var1 and var2, can I create var3=var1+var2 and use only var3 in the
analysis, without compromising the integrity of the dataset?
Thank you,
Gina
Gina Yannitell Reinhardt
Political Science Doctoral Candidate
Washington University, St. Louis
One Brookings Drive, Box 1063
St. Louis, MO 63130
phone (314)935-5654
fax (314)935-5856
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia
-
Amelia mailing list served by Harvard-MIT Data Center
[Un]Subscribe/View Archive:
http://lists.hmdc.harvard.edu/?info=amelia