ABSTRACT
The contribution relies on the French institutionalist approach of economics of convention (EC). This approach was developed in the analysis of statistics, categorization, and quantification and links the issue of data generation to the question of the common good. EC focuses on how data quality is grounded in (measurement) conventions and how data serves social collectives for the pursuing of a common good. EC can be regarded as a neopragmatist approach, which points to the plurality of data worlds and the upcoming tensions between conventions as different institutional logics, how to generate, evaluate, interpret and use data in situations. Big data and algorithms bring in new tensions in situations because of its (mostly) opaque character of data generation. The emerging issues at stake are how data quality and its link to the common good can be evaluated by civic and collective actors. More and more realms of society are governed by private companies, whose data generation, collection and analysis are based on privately owned and invisible algorithms. As the relevance of digital platforms and social media grows, so does the power of governance that such entities have on society and social life. A key to understanding this influence is the ability of digital actors to regulate effectively, be it through (automated) enforcement of rules or their law-making potential. This contribution mobilizes the convention-based notion of data worlds and the critical capacities of their actors to explain conflicts and coordinative action in different areas of algorithmic regulation. The approach, which is based on situations of uncertainty, will be exemplarily applied to different areas of content moderation as well as payment, public administration and legal tech services.
