Rationality and politics of algorithms. Will the promise of big data survive the dynamics of public decision making?

H.G.van der Voort and A.J.Klievink (TU Delft, the Netherlands), M.Arnaboldi (Politecnico di Milano, Italy) and A.J.Meijer (Utrecht University, the Netherlands) published in Government Information Quarterly Volume 36 Issue 1 (January 2019)

Highlights

• A framework respecting both a rational and a political view on decision-making and respecting the perspectives of both data analysts and decision makers.
• We found that data analysts are politically significant in all stages. They have a crucial role in constructing information, facilitating framing and adjusting information to political preferences. However, powers of data analysts are limited. Data analysts do their work in the context of predefined policies. Decision makers sometimes just neglect or bypass big data analysis.
• Mere functional support for big data to decision making or mere warnings for the politics of big data may miss an important aspect of big data. An important additional question is whether the powers of data analysts and decision makers are well-balanced and how checks and balances could be organized.

Abstract

Big data promises to transform public decision-making for the better by making it more responsive to actual needs and policy effects.

However, much recent work on big data in public decision-making assumes a rational view of decision-making, which has been much criticized in the public administration debate.

In this paper, we apply this view, and a more political one, to the context of big data and offer a qualitative study. We question the impact of big data on decision-making, realizing that big data – including its new methods and functions – must inevitably encounter existing political and managerial institutions.

By studying two illustrative cases of big data use processes, we explore how these two worlds meet. Specifically, we look at the interaction between data analysts and decision makers.

In this we distinguish between a rational view and a political view, and between an information logic and a decision logic. We find that big data provides ample opportunities for both analysts and decision makers to do a better job, but this doesn't necessarily imply better decision-making, because big data also provides opportunities for actors to pursue their own interests.

Big data enables both data analysts and decision makers to act as autonomous agents rather than as links in a functional chain. Therefore, big data's impact cannot be interpreted only in terms of its functional promise; it must also be acknowledged as a phenomenon set to impact our policy-making institutions, including their legitimacy.

Full text (HTML)

---