Weighted sequential hot desk imputation macros

Abstract

Item nonresponse occurs when questions from an otherwise completed survey questionnaire are not answered. Since the population estimates formed by ignoring missing data are often biased, nonresponse adjustment procedures to reduce this bias should be considered.
Hot deck imputation is a commonly used nonresponse adjustment procedure that replaces missing data with available survey data. Conventional hot deck methods do not, however, consider the weighted distribution of the data in the imputation process. Weighted sequential hot deck imputation (Cox, 1980) replicates the weighted distribution of the available data in the imputed data by using the sample weights, or inverse selection probabilities, of item respondents and nonrespondents.