Standardisation and Statistics

Walter J. Radermacher (2021)

The Survey Statistician, 84, pp. 24-31. Open access

Abstract. An infinite amount of data and a variety of ways to use it promise an individualisation of statistics; the tools, algorithms and (artificial) intelligence of data science are available to everyone. The age of statistics is being replaced by the data era. There is a promise in this, similar to that for the internet as a whole, of an equal and thus democratically organised basis of technology and access to quantitative information, and perhaps even knowledge. The article deals with the question of whether and to what extent this individualisation is beneficial or detrimental to the democratic discourse. For this objective, the concept of information quality is introduced and used, which is oriented towards the principle that statistical products should be fit for purpose. It is shown that individualisation and standardisation are two alternative approaches to statistics, the choice of which ultimately depends on the application and its intended function. As a language for public discourse, standardisation is of vital importance.

Download link