A Framework for Evaluating Intersectional Fairness in Algorithmic Decision Making

• In Event: Differential Item Functioning II

Abstract

This study proposes an extension of a recently proposed framework for algorithmic fairness, known as “differential algorithmic functioning” (DAF), to include the concept of intersectionality. The proposed intersectional DAF framework provides theoretical and technical definitions, as well as practical tools and guidelines for assessing intersectional fairness in modern algorithms.

Authors

• Youmi Suk, Teachers College Columbia University

• Kyung (Chris) T. Han, GMAC