Skip to content
DraconiticResearch
Search Sources API
181,686 studies31 sourcesUpdated 40 min ago
PreprintCC BY 4.0arXiv

Technical Report: Causal ANOVA Applied to COMPAS

  • Yiming Xiong
  • Zijun Gao
Open at sourcePDFDOICite

Abstract

This report applies Causal ANOVA to the COMPAS recidivism prediction dataset to study how demographic variables contribute to variation in two-year recidivism outcomes. Using a directed acyclic graph (DAG) to encode causal assumptions, we estimate counterfactual explainability scores for Race, Sex, and Age Category. We then examine whether the explanatory roles of Race and Sex change after conditioning on age group. The results suggest that Age Category explains the largest share of variation in the full causal model. After conditioning on age, Sex explainability is highest among younger defendants and declines with age, while Race explainability increases with age and is most pronounced among older defendants. As a complementary analysis, we use logistic regression models to interpret the demographic patterns suggested by the Causal ANOVA results. We provide a Python package of Causal ANOVA at https://github.com/devonxiong/causal-anova.

Related

  1. KAYROS: An Anytime and Exact Open-Source Solver for Duration-Minimization Time-Dependent Vehicle Routing. A Technical Report and a Case Study in Human-AI Engineering
    Florian Rascoussier, Romain Billot, Lina Fahed +1
    arXivPreprintCC BY 4.0Full textOptimisation & controlarXiv
  2. TabPFN-3.5: Technical Report
    Benjamin Jäger, Nick Erickson, Léo Grinsztajn +44
    arXivPreprintarXiv licenceFull textStatisticsarXiv
  3. TabPFN-3: Technical Report
    Léo Grinsztajn, Klemens Flöge, Oscar Key +38
    arXivPreprintarXiv licenceFull text14 citedStatisticsarXiv
  4. Weight-adapted ANOVA kernels for high-dimensional periodic approximation
    Zhengjie Sun
    arXivPreprintarXiv licenceFull textNumerical analysis & scientific computingProbabilityarXiv
  5. Worst-Case Completion of Tensors with Approximately Few ANOVA Terms
    Simon Foucart, Jingchun Shao
    arXivPreprintCC BY 4.0Full textNumerical analysis & scientific computingarXiv

Cite

BibTeXRIS
BibTeX
@misc{xiong2026technical,
  title = {{Technical Report: Causal ANOVA Applied to COMPAS}},
  author = {Yiming Xiong and Zijun Gao},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.30110},
  eprint = {2610.30110},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.30110},
  note = {CC BY 4.0},
}
RIS
TY  - UNPB
TI  - Technical Report: Causal ANOVA Applied to COMPAS
AU  - Yiming Xiong
AU  - Zijun Gao
PY  - 2026
DA  - 2026/10/21
PB  - arXiv
DO  - 10.48550/arxiv.2610.30110
UR  - https://arxiv.org/abs/2610.30110
AB  - This report applies Causal ANOVA to the COMPAS recidivism prediction dataset to study how demographic variables contribute to variation in two-year recidivism outcomes. Using a directed acyclic graph (DAG) to encode causal assumptions, we estimate counterfactual explainability scores for Race, Sex, and Age Category. We then examine whether the explanatory roles of Race and Sex change after conditioning on age group. The results suggest that Age Category explains the largest share of variation in the full causal model. After conditioning on age, Sex explainability is highest among younger defendants and declines with age, while Race explainability increases with age and is most pronounced among older defendants. As a complementary analysis, we use logistic regression models to interpret the demographic patterns suggested by the Causal ANOVA results. We provide a Python package of Causal ANOVA at https://github.com/devonxiong/causal-anova.
ER  -