Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Pfohl, Stephen R., Harris, Natalie, Nagpal, Chirag, Madras, David, Mhasawade, Vishwali, Salaudeen, Olawale, Dieng, Awa, Sequeira, Shannon, Arciniegas, Santiago, Sung, Lillian, Ezeanochie, Nnamdi, Cole-Lewis, Heather, Heller, Katherine, Koyejo, Sanmi, D'Amour, Alexander
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918168662900736
author Pfohl, Stephen R.
Harris, Natalie
Nagpal, Chirag
Madras, David
Mhasawade, Vishwali
Salaudeen, Olawale
Dieng, Awa
Sequeira, Shannon
Arciniegas, Santiago
Sung, Lillian
Ezeanochie, Nnamdi
Cole-Lewis, Heather
Heller, Katherine
Koyejo, Sanmi
D'Amour, Alexander
author_facet Pfohl, Stephen R.
Harris, Natalie
Nagpal, Chirag
Madras, David
Mhasawade, Vishwali
Salaudeen, Olawale
Dieng, Awa
Sequeira, Shannon
Arciniegas, Santiago
Sung, Lillian
Ezeanochie, Nnamdi
Cole-Lewis, Heather
Heller, Katherine
Koyejo, Sanmi
D'Amour, Alexander
contents Disaggregated evaluation across subgroups is critical for assessing the fairness of machine learning models, but its uncritical use can mislead practitioners. We show that equal performance across subgroups is an unreliable measure of fairness when data are representative of the relevant populations but reflective of real-world disparities. Furthermore, when data are not representative due to selection bias, both disaggregated evaluation and alternative approaches based on conditional independence testing may be invalid without explicit assumptions regarding the bias mechanism. We use causal graphical models to characterize fairness properties and metric stability across subgroups under different data generating processes. Our framework suggests complementing disaggregated evaluations with explicit causal assumptions and analysis to control for confounding and distribution shift, including conditional independence testing and weighted performance estimation. These findings have broad implications for how practitioners design and interpret model assessments given the ubiquity of disaggregated evaluation.
format Preprint
id arxiv_https___arxiv_org_abs_2506_04193
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness
Pfohl, Stephen R.
Harris, Natalie
Nagpal, Chirag
Madras, David
Mhasawade, Vishwali
Salaudeen, Olawale
Dieng, Awa
Sequeira, Shannon
Arciniegas, Santiago
Sung, Lillian
Ezeanochie, Nnamdi
Cole-Lewis, Heather
Heller, Katherine
Koyejo, Sanmi
D'Amour, Alexander
Machine Learning
Computers and Society
Disaggregated evaluation across subgroups is critical for assessing the fairness of machine learning models, but its uncritical use can mislead practitioners. We show that equal performance across subgroups is an unreliable measure of fairness when data are representative of the relevant populations but reflective of real-world disparities. Furthermore, when data are not representative due to selection bias, both disaggregated evaluation and alternative approaches based on conditional independence testing may be invalid without explicit assumptions regarding the bias mechanism. We use causal graphical models to characterize fairness properties and metric stability across subgroups under different data generating processes. Our framework suggests complementing disaggregated evaluations with explicit causal assumptions and analysis to control for confounding and distribution shift, including conditional independence testing and weighted performance estimation. These findings have broad implications for how practitioners design and interpret model assessments given the ubiquity of disaggregated evaluation.
title Understanding challenges to the interpretation of disaggregated evaluations of algorithmic fairness
topic Machine Learning
Computers and Society
url https://arxiv.org/abs/2506.04193