Computational models of the social play an increasingly important role in political processes, from consulting to algorithmic governance. The dissertation project investigates the relation between the knowledge production practices underlying such models and the governance practices that make use of them. In doing so it aims to understand how different forms of modeling the social translate into different forms of governing.

Besides a reconstruction of the history of modeling in politics and the development of a practice theoretical approach to modeling and the contestation of its various forms of expertise, the thesis empirically examines three fields of knowledge production: the computational social sciences, the field of fairness, accountability and transparency in machine learning as well as modeling in the context of the Covid-19 pandemic. Relying on qualitative interviews, document analysis and network analysis, the dynamics and performative effects of these fields are studied from a mixed-methods perspective. On the basis of the gained insights the thesis reflects on possible social theoretical and political responses to the emerging structural role of computational models in the digital society.