One kind of imputation involved substituting the record of another person or an entire household: 5.8 million people required such whole person imputation in 2000, amounting to 2.1 percent of the household population count. (In 1990, only 1.9 million people, or 0.8 percent of the household population, were imputed in this way.) Whole person imputations in 2000 included: cases for which there was no information about the number of people living at that address or their characteristics (0.4% of the household population); cases for which household size was known but not the characteristics of the members (0.8% of the household population); and cases for which no information was provided for the individual, although other household members had reported data (0.9% of the household population).

Editing and imputation rates for missing values for individual shortform content items, such as age, race, sex, and housing tenure, were low—ranging from 1.1 percent to 4.3 percent. (These rates exclude wholly imputed people.) In many instances, it was possible to fill in an answer from other information for the person or household, so that rates of hot-deck imputation for short-form items were lower still. Information about editing and imputation rates for long-form content items is not yet available.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement