Sociological Aspects of the Statistical Research Process: Toward a Sociology of Public Statistics

Włodzimierz Okrasa (2020):

Polish Sociological Review 3, pp. 323-344

Abstract. The statistical process is a multifaceted, socially immersed research process. Statistics as a researchactivity is—and statistics as a product are—influenced by a variety of factors, covariates, and participants/stakeholdersin data generation, analysis, dissemination and use. The aim of this paper is to identify and systematizekey elements of the socio-cognitive status of statistics and of relations between statistics and society. To this end,a conceptual meta-model of the statistical process, involving an interwoven “logic of method” and “logic of action,”provides a framework for the exploration of statistics from the perspective of sociology. Given that social,cultural, and political conditions affect the quality of statistical products and the status of statistical institutions,sociological study seems especially well suited to dealing with the complexity of the issues involved. To be successful,however, advances in the sociology of statistics are also desirable.

Extract. “Statistics, perceived in substantive or content terms, is usually defined as a body of knowledge consisting of registered and presented quantitative data.2 As a scholarly discipline, it is supplemented by algorithms for data compilation and analysis, including the choice of indicators, the design of models (“statistical” and “econometric”), methods of data generation, the construction of databases, and so forth. One version of such an approach was developed, for example, by the pioneer of the French school of the sociology of statistics, Alain Desrosières (2011: 59). […] Within such a multi-dimensional approach, sociologization emerges as an appropriate strategy for analyzing the statistics-society relationship, and the statistical process as a primary social process. Instruments for such an implementation of sociologization are provided by the sociology of statistics, which has developed so far primarily in relation to official statistics, as demonstrated by Desrosières (1998). Sociologization also means taking into account the historical context, which interacts with virtually every aspect of statistical processes. It includes what is sometimes referred to as the “Desrosières standard” (e.g., Barnes 2000),5 that is, analysis of the interaction between scientific description, coding, categorizing, measuring, and analyzing on the one hand, and the administrative and political area of activity, such as decision making, intervention, and improvement, on the other hand (Desrosières 1998).” (pp. 325/326-327)

Link to the review