Send message to Author

Paper short abstract

This paper examines how algorithms used to moderate access to higher education construct comparability between students’ grades. It shows how state-led harmonisation processes align heterogeneous evaluations, while generating tensions between statistical consistency and situated fairness.

Paper long abstract

Determining who gets into higher education (HE) is one of the fundamental prerogatives of the modern state. However, the growing number of applicants to HE means that schools have to compare a raising pool of students each year. Over the past decade, various states have introduced automated systems that rely on standardised indicators to assist with the evaluation of candidates. This paper examines how the making of such systems lead to specific forms of governance by inscribing values through the making of the metrics they use to evaluate.

Drawing on innovative approaches to crack open those algorithms, I trace the making of metrics in two cases especially relevant to this debate. The first is the OfQual grading algorithm deployed in England during the COVID-19 crisis to grade students. Whilst it was initially made to ensure overall grade comparison over time, its cancellation due to public backlash led to a substantial grade inflation. The second case is the Parcoursup algorithm in France, which is used to match students with HE institutions through algorithmic processes.

Through this comparative analysis, I explore how automated systems build on a variety of metrics at their disposition to produce standardised and comparable variables at a national scale. The controversies surrounding OfQual and Parcoursup highlight the limits and contestation of such harmonisation processes. By opening the black boxes of these algorithms, this paper demonstrates how such algorithms produce by states are imbued with specific set of values that are discussed by limited numbers of actors.