Risk scores, label bias, and everything but the kitchen sink
Science Advances · 29.03.2024 · 10 (13)
DOI: 10.1126/sciadv.adi8411 ↗Kurz gefasst
Fachartikel in der Zeitschrift Science Advances, erschienen am 29.03.2024. Verfasst von Michael Zanger-Tishler und 2 weiteren Autor:innen. Beteiligte Einrichtung: Sociology and Social Policy, Harvard University, Cambridge, MA 02138, USA. Die Arbeit wurde bisher 22-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
In designing risk assessment algorithms, many scholars promote a “kitchen sink” approach, reasoning that more information yields more accurate predictions. We show, however, that this rationale often fails when algorithms are trained to predict a proxy of the true outcome, for instance, predicting arrest as a proxy for criminal behavior. With this “label bias,” one should exclude a feature if its correlation with the proxy and its correlation with the true outcome have opposite signs, conditional on the other model features. This criterion is often satisfied when a feature is weakly correlated with the true outcome, and, additionally, that feature and the true outcome are both direct causes of the proxy outcome. For example, criminal behavior and geography may be weakly correlated and, due to patterns of police deployment, direct causes of one’s arrest record—suggesting that excluding geography in criminal risk assessment will weaken an algorithm’s performance in predicting arrest but will improve its capacity to predict actual crime.
Autor:innen und Institutionen
- Sociology and Social Policy, Harvard University, Cambridge, MA 02138, USA.
- Stanford Law School, Stanford, CA 94305, USA.
- Harvard Kennedy School, Cambridge, MA 02138, USA.
ORCID-Profile: Michael Zanger-Tishler, Julian Nyarko, Sharad Goel
Zitieren
APA
Zanger-Tishler, M., Nyarko, J., & Goel, S. (2024). Risk scores, label bias, and everything but the kitchen sink. Science Advances, 10(13). https://doi.org/10.1126/sciadv.adi8411
Vancouver
Zanger-Tishler M, Nyarko J, Goel S. Risk scores, label bias, and everything but the kitchen sink. Science Advances. 2024;10(13). doi:10.1126/sciadv.adi8411
BibTeX
@article{ZangerTishler2024,
author = {Zanger-Tishler, Michael and Nyarko, Julian and Goel, Sharad},
title = {Risk scores, label bias, and everything but the kitchen sink},
journal = {Science Advances},
year = {2024},
volume = {10},
number = {13},
doi = {10.1126/sciadv.adi8411}
}
EUVIMED