Enumeration errors of census imputation

Abstract Nonresponse to census enumeration is unavoidable despite the effort and resource committed. Imputation for the census nonrespondents so as to create a complete census database is common practice for preparing the numerous census outputs. Many countries have used administrative records for census imputation or are investigating such uses, in order to benefit from imputing genuine data of the census nonrespondents, which would not be possible when the imputed records are taken from the census respondents. But there is a gap in the literature regarding the assessment of the associated enumeration errors. In this paper, we develop methods for estimating the enumeration errors induced by census imputation and illustrate their uses in the context of Australian Census 2016.