Towards Equalised Odds as Fairness Metric in Academic Performance Prediction

TL;DR

This paper evaluates various fairness notions for academic performance prediction systems and identifies equalised odds as the most suitable fairness metric considering the task's nature and long-term benefits.

Contribution

The study analyzes fairness guidelines and advocates for equalised odds as the optimal fairness measure in academic performance prediction tasks.

Findings

Equalised odds aligns well with APP's worldview.

Applying equalised odds can lead to long-term population benefits.

Literature review supports equalised odds as the most appropriate fairness notion.

Abstract

The literature for fairness-aware machine learning knows a plethora of different fairness notions. It is however wellknown, that it is impossible to satisfy all of them, as certain notions contradict each other. In this paper, we take a closer look at academic performance prediction (APP) systems and try to distil which fairness notions suit this task most. For this, we scan recent literature proposing guidelines as to which fairness notion to use and apply these guidelines onto APP. Our findings suggest equalised odds as most suitable notion for APP, based on APP's WYSIWYG worldview as well as potential long-term improvements for the population.