The Effectiveness of Hot-deck Procedures in Small Samples.

The hot-deck method is a simple and useful technique to impute missing values in a data matrix. It is primarily a record matching technique in which an incomplete record is compared with a complete record having similar characteristics (Ernst, 1978; Rander, 1978). The missing field in the incomplete record is imputed from the value which appears on the corresponding variable in the complete record. The hot-deck method has been used successfully in the past by the American and Canadian Census Bureaus, the Social Security Administration, and the Internal Revenue Service. However, application of this procedure to date has been limited to large data sets. Because of the lower cost and the simplicity with which hot-de~k methods may be used in real life situations it seemed advantageous to determine their usefulness when applied to small samples. The present study was an attempt to provide this information.