Representativity is one of the most frequently invoked terms in communication of statistics. Despite its central role in survey and official statistics, the term is often used without a clear formal framework, leading to questionable interpretations of statistical output. Further, in the presence of representative statistics, high quality output is assumed by consumers of statistics up to policy decisions. This urges the definition of a clear framework of statistical quality. Additionally, the use of the statistics has to be considered carefully. Are we interested in data driven information, do we need a careful assessment of evidence, or are we interested in legislation or law court decisions.

The classical framework of quality in (survey) statistics is addressed e.g. in the European Statistics Code of Practice or similar frameworks or in the total survey error concept. However, in the age of big data and AI, the classical concepts are in question since uncertainty and accuracy assessment has to be carefully redefined.

The session will focus on new requirements for quality and uncertainty for surveys and statistics while considering the aim of use. Additionally, the communication of new accuracy concepts will be discussed.