Abstract
Quality in official statistics in one aspect relates to how the statistical procedures are designed and implemented in order to assure an adequate statistical quality, and to achieve this in a cost-effective way. Another equally important aspect relates to methods and techniques for assessing the level of quality actually arrived at. This paper considers both these aspects through a system for quality assurance and assessment developed for the annual Census of Agriculture in Sweden, which is arranged as a permanent panel study. The system depends on a matrix approach: the census variables form one dimension and the operations (data collection, data entry, etc.) form the other. In principle each combination of these two dimensions has been analysed with regard to error sources and error mechanisms as well as possible methods to prevent, detect, correct and measure the errors.
Get full access to this article
View all access options for this article.