SELECTIVE EDITING – NEW APPROACHES AT THE STATISTICAL OFFICE OF THE REPUBLIC OF SLOVENIA Cover Image

SELECTIVE EDITING – NEW APPROACHES AT THE STATISTICAL OFFICE OF THE REPUBLIC OF SLOVENIA
SELECTIVE EDITING – NEW APPROACHES AT THE STATISTICAL OFFICE OF THE REPUBLIC OF SLOVENIA

Author(s): Rudi Seljak, Nejc Jevšnik, Jerneja Pikelj
Subject(s): Methodology and research technology, ICT Information and Communications Technologies
Published by: Ekonomski fakultet u Sarajevu
Keywords: selective editing; error detection; score function; statistical software;

Summary/Abstract: Data editing has always been considered as one of the most demanding and consequently most costly parts of the statistical process. Therefore, statistical offices are putting a lot of effort into activities that would reduce costs and decrease the workload caused by this part of the process. Implementation of a selective editing approach is certainly one of the most commonly used means to achieve this goal. The entire editing process should be designed in such a way that it focuses on the mistakes that actually affect the estimates. Selective editing approach aims to search for significant errors, accepting that the final data set still contains a number of errors with no effect on the final estimates. Its main purpose is therefore to reduce the amount of changes in the microdata, maintaining the acceptable quality of statistical results. This not only leads to a reduction of costs, but also to improved overall quality of the data. The Statistical Office of the Republic of Slovenia has already been using the selective editing approach for some time. The problem of the current practice is that it is based on the so-called system of key respondents. Key respondents are the most influential units that are pre-selected to be treated differently in the data collection and data editing phases. Recently we have been trying to introduce a different approach, which would be based on the posterior information obtained in the data collection phase and on the appropriate implementation of the score function idea. With this approach, the number of errors that require manual attention is reduced.

  • Issue Year: 2017
  • Issue No: 35
  • Page Range: 168-180
  • Page Count: 13
  • Language: English
Toggle Accessibility Mode