|
Activity Number:
|
216
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #305266 |
|
Title:
|
Imputation of Nominal Variables Using Gaussian-Based Routines
|
|
Author(s):
|
Recai M. Yucel*+
|
|
Companies:
|
State University of New York at Albany
|
|
Address:
|
Department of Epidemiology and Biostatistics, Rensselaer, NY, 12144,
|
|
Keywords:
|
nominal data imputation ; missing data ; rounding ; multiple imputation ; missing data software ; categorical incomplete data
|
|
Abstract:
|
The multivariate normal (MVN) distribution is arguably the most popular parametric model used in imputation and is available in commonly used software packages (e.g. SAS PROC MI). When the incompletely-observed variables include nominal variables, practitioners often apply techniques such as creating a distinct ``missing" category or disregarding the nominal variable from the imputation process, both of which may lead to biased results. In this work, we propose practical rounding rules to be used with the existing MVN-based imputation methods, allowing practitioners to obtain usable imputation with small biases. These rules are calibrated in the sense that values re-imputed for observed data have distributions similar to those of the observed data. A simulation study demonstrating the advantages of this approach is presented.
|
