Mathematical Notions vs. Human Perception of Fairness: A Descriptive Approach to Fairness for Machine Learning

TL;DR

This paper investigates which mathematical fairness notions align best with lay people's perceptions across different scenarios, finding that simple demographic parity often matches societal fairness intuitions better than complex definitions.

Contribution

It introduces a descriptive approach to identify societal perceptions of fairness, challenging the focus on mathematical tradeoffs by emphasizing human-centered fairness evaluation.

Findings

Demographic parity aligns closely with lay perceptions in multiple scenarios

Explicit explanations of complex fairness notions do not significantly change perceptions

Simple fairness definitions are often preferred over complex ones by non-experts

Abstract

Fairness for Machine Learning has received considerable attention, recently. Various mathematical formulations of fairness have been proposed, and it has been shown that it is impossible to satisfy all of them simultaneously. The literature so far has dealt with these impossibility results by quantifying the tradeoffs between different formulations of fairness. Our work takes a different perspective on this issue. Rather than requiring all notions of fairness to (partially) hold at the same time, we ask which one of them is the most appropriate given the societal domain in which the decision-making model is to be deployed. We take a descriptive approach and set out to identify the notion of fairness that best captures \emph{lay people's perception of fairness}. We run adaptive experiments designed to pinpoint the most compatible notion of fairness with each participant's choices through a small number of tests. Perhaps surprisingly, we find that the most simplistic mathematical definition of fairness---namely, demographic parity---most closely matches people's idea of fairness in two distinct application scenarios. This conclusion remains intact even when we explicitly tell the participants about the alternative, more complicated definitions of fairness, and we reduce the cognitive burden of evaluating those notions for them. Our findings have important implications for the Fair ML literature and the discourse on formalizing algorithmic fairness.