Recommended Practices for Editing and Imputation for Cross-Sectional Business Surveys

Abstract

The EDIMBUS project developed a manual for editing and imputation in cross-sectional business surveys. Survey data always contains erroneous and missing data. Editing and imputation cover the phases which deal with erroneous and missing data and thus are crucial for the quality of the data provided by Eurostat and the National Statistical Institutes of the European Statistical System. Eurostat entrusted the project partners to developed a guide for the design, the implementation, the testing and the documentation of editing and imputation strategies. The manual should present the best current methods for statistical data preparation. More importantly, it should give advise to survey managers how to approach the difficult task of developing a system for statistical data preparation which takes into account the relation between different phases of statistical data preparation, the resources needed and the effect on outcomes.