Multiple imputation MICE

WinMICE V0.1 is a Windows program for creating multiple imputations under the linear model. Download installs the program, and contains documentation and examples (July 2005). Has a subset of MICE features.

MICE V1.0 Old S-Plus library (mice.zip, 16/06/00, 76KB)

Documentation (manual.pdf, 16/06/00, 222KB), use Acrobat Reader 4 (AR 3 has some problems).

MICE V1.0 for R under Unix, converted by Peter Malewski (10/08/00, 35Kb) + neat function list.

MICE V1.1.0 for R under Windows. Package created by Jason Turner (25 Feb 2004).

MICE						Any bugs found? Please let me know.

What is MICE?
MICE stands for Multivariate Imputation by Chained Equations. We have written a software library for multiple imputation using S-Plus V4.5 and higher for Windows, and R 1.8 and higher for Windows. The library assists in performing the steps required in a full multiple imputation analysis. There is also an implementation in STATA. Specific features of MICE V1.0 include: columnwise specification of the imputation model; arbitrary patterns of missing data; transformations and index variables; subset selection of predictors; supports standard `lm` and `glm` complete-data methods; automated pooling using the Barnard-Rubin adjustment; callable user-written imputation functions; online help files.

Downloads
WinMICE V0.1 is a Windows program for creating multiple imputations under the linear model. Download installs the program, and contains documentation and examples (July 2005). Has a subset of MICE features. MICE V1.0 Old S-Plus library (mice.zip, 16/06/00, 76KB) MICE for V1.13.1 for R 2.1.0 by John Fox (April 26, 2005): mice.1.13.1.zip (Windows), mice-manual.pdf, mice.1.13.1.tar.gz S-Plus and R V1.13 source code. For those creating their own libraries and packages (25 Feb 2004). Documentation (manual.pdf, 16/06/00, 222KB), use Acrobat Reader 4 (AR 3 has some problems). MICE V1.0 for R under Unix, converted by Peter Malewski (10/08/00, 35Kb) + neat function list. MICE V1.1.0 for R under Windows. Package created by Jason Turner (25 Feb 2004).


Other relevant reports
OUDSHOORN CGM, VAN BUUREN S & VAN RIJCKEVORSEL. Flexible multiple imputation by chained equations of the AVO-95 Survey. Leiden: TNO Prevention and Health. Report PG/VGZ/99.045, 1999.[pdf]
VAN BUUREN S & OUDSHOORN CGM. Flexible multivariate imputation by MICE. Leiden: TNO Preventie en Gezondheid, TNO/VGZ/PG 99.054, 1999.[pdf]
Note: The syntax used in the above two reports is based on unpublished versions of the MICE-library. The appropriate syntax is given in de MICE V1.0 documentation.
VAN BUUREN S, BRAND JPL, GROOTHUIS-OUDSHOORN CGM, RUBIN DB. Fully Conditional Specification in Multivariate Imputation. Journal of Statistical Computation and Simulation, in press. Simulation study on the MICE algorithm. [pdf]

Poster
This poster provides a condense overview of the most important aspects of MICE (.pdf, 106KB).

Presentation
Our Munich presentation explains how we impute multivariate categorical data.



DISCLAIMER: THE SOFTWARE IS PROVIDED AS IS, WITHOUT REPRESENTATION AS TO ITS FITNESS FOR ANY PURPOSE, AND WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. TNO SHALL NOT BE LIABLE FOR ANY DAMAGES, INCLUDING SPECIAL, INDIRECT, INCIDENTAL, OR CONSEQUENTIAL DAMAGES, WITH RESPECT TO ANY CLAIM ARISING OUT OF OR IN CONNECTION WITH THE USE OF THE SOFTWARE, EVEN IF IT HAS BEEN OR IS HEREAFTER ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.