MISTRESS is a method for imputing missing categorical data. The method was published in S. van Buuren and J.L.A. van Rijckevorsel (1992). Imputation of missing categorical data by maximizing internal consistency. Psychometrika, 57, 567-580. pdf
Abstract of the paper
This paper suggests a method to supplant missing categorical data by "reasonable" replacements. These replacements will maximize the consistency of the completed data as measured by Guttman's squared correlation ratio. The text outlines a solution of the optimization problem, describes the relationships with the relevant psychometric theory, and studies some properties of the method in detail. The main result is that the average correlation should be at least 0.50 before the method becomes practical. At that point, the technique gives reasonable results up to 10-15% missing data.
A SAS/IML Program: Mistress 1.17
The SAS/IML program, documentation, examples, and test code can be downloaded here.