In the context of business surveys at National Statistical Institutes (NSIs), imputation of missing values is often complicated by the fact that the data should conform to a large number of edit rules. In this module, we consider two basic approaches to obtain imputations that satisfy edit rules. Under the first approach, the edits are incorporated directly in the imputation model, so that all imputations are automatically consistent. Unfortunately, this can lead to a very complex model. Therefore, in practice, another approach is often used, in which the missing values are first imputed without taking the edits into account. In a subsequent step, the initial imputations are then minimally adjusted to become consistent with the edits.
To read the entire document, please access the pdf file (link under "Related Documents" on the right-hand-side of this page).
Your feedback is appreciated. Please send your remarks, suggestions for improvement, etc. to email@example.com.