We've recently used -mi impute mvn-.
(1) You can fit the full MVN model via
. mi impute mvn var1 var2 var3 var4
(2) You can fit the conditional MVN model via
. mi impute mvn var1 var2 = var3 var4
We had ~10,000 observations, with 5 outcomes (repeated measures over
time) with arbitrary missingness and 8 covariates (with no missingness).
But our covariates included dichotomous and categorical variables and we
also wanted some interactions in there.
So, for us, (1) was really not practical given the large dimension of
the covariance matrix. So we did (2) with
. mi impute mvn y1 y2 y3 y4 y5 = x1 x2 ...
(we also needed to define manually the indicator terms for categorical
variables).
Hope this is helpful.
CD
On 10/4/2011 3:09 PM, Paul von Hippel wrote:
Is there a Stata command that implements the multivariate normal
imputation model developed by Schafer (1997)? This is a model that
assumes the joint distribution of the complete and incomplete variables
is multivariate normal. Implementations include NORM for S-Plus/R and
PROC MI in SAS.
I hoped that this model was implemented in Stata's /mi impute mvn/
command in Stata, but it appears that /mi impute mvn/ imputes via
regression -- i.e., specifying the conditional distribution of the
incomplete variables, rather than the joint distribution of all the
variables.
Many thanks for any pointers.
--
Best wishes,
Paul von Hippel
Assistant Professor
LBJ School of Public Affairs
Sid Richardson Hall 3.251
University of Texas, Austin
2315 Red River, Box Y
Austin, TX 78712
mobile, preferred (614) 282-8963
office (512) 232-3650
--
The documents accompanying this transmission may contain confidential
health or business information. This information is intended for the
use of the individual or entity named above. If you have received
this information in error, please notify the sender immediately and
arrange for the return or destruction of these documents.
Constantine Daskalakis, ScD
Associate Professor,
Thomas Jefferson University, Division of Biostatistics
1015 Chestnut St., Suite M100, Philadelphia, PA 19107
Tel: 215-955-5695
Fax: 215-503-3804
Email: [email protected]
Webpage: http://www.jefferson.edu/clinpharm/biostatistics/
