arbiter
JournalsAccount

Building Nondiscriminatory Algorithms in Selected Data

David Arnold et al.

American Economic Review: Insights2025https://doi.org/10.1257/aeri.20240249article
AJG 3ABDC A*
Weight
0.41

Abstract

We develop new quasi-experimental tools to understand algorithmic discrimination and build non-discriminatory algorithms when the outcome of interest is only selectively observed. We first show that algorithmic discrimination arises when the available algorithmic inputs are systematically different for individuals with the same objective potential outcomes. We then show how algorithmic discrimination can be eliminated by measuring and purging these conditional input disparities. Leveraging the quasi-random assignment of bail judges in New York City, we find that our new algorithms not only eliminate algorithmic discrimination but also generate more accurate predictions by correcting for the selective observability of misconduct outcomes.

2 citations

Open via your library →

Cite this paper

https://doi.org/https://doi.org/10.1257/aeri.20240249

Or copy a formatted citation

@article{david2025,
  title        = {{Building Nondiscriminatory Algorithms in Selected Data}},
  author       = {David Arnold et al.},
  journal      = {American Economic Review: Insights},
  year         = {2025},
  doi          = {https://doi.org/https://doi.org/10.1257/aeri.20240249},
}

Paste directly into BibTeX, Zotero, or your reference manager.

Flag this paper

Building Nondiscriminatory Algorithms in Selected Data

Flags are reviewed by the Arbiter methodology team within 5 business days.

Evidence weight

0.41

Balanced mode · F 0.40 / M 0.15 / V 0.05 / R 0.40

F · citation impact0.25 × 0.4 = 0.10
M · momentum0.55 × 0.15 = 0.08
V · venue signal0.50 × 0.05 = 0.03
R · text relevance †0.50 × 0.4 = 0.20

† Text relevance is estimated at 0.50 on the detail page — for your query’s actual relevance score, open this paper from a search result.