Imputation of Establishment Survey Data

Résumé du livre

Nonresponse to establishment surveys is an ongoing problem for agencies responsible for the collection and dissemination of economic data. This is not to say that nonresponse is of no concern to social survey statisticians. For economic surveys, however, the data are often highly skewed and quantitative in nature, which poses unique challenges in dealing with the problem of missing data. As nonrespondents may not behave in the same manner as the respondents in terms of the characteristics of interest, special care must be taken to avoid potential biases that can result from inappropriate treatment of missing or inconsistent data. While imputation is but one method of dealing with incomplete data, its use is widespread. In this chapter, we discuss the reasons for and against using imputation techniques.