Household Finance and Consumption Network

1. There is inconsistent practice and some uncertainty on the terms “editing” and “imputation”. In this document we will refer to data editing as the procedure used to correct or reconcile contradictory information, whereas imputation is the procedure used to deal with missing data. The glossary of the United Nations Statistical and Economic Commission for Europe (UNECE) states that data editing is “the activity aimed at detecting and correcting errors (logical inconsistencies) in data”, whereas data imputation is the “substitution of estimated values for missing or inconsistent data items (fields). The substituted values are intended to create a data record that does not fail edits”. 2. According to UNECE (2000), up to 40% of costs in a typical survey may be due to statistical data editing, although this figure should be lower given enough attention to the interviewing and the questionnaire design. This document addresses first the question of data editing, before describing imputation and recommending the use of multiple imputation. Specific organisations for data editing and imputation, as well as common tools are discussed at the end.