# Debiasing classifiers: is reality at variance with expectation?

arXiv: https://arxiv.org/abs/2011.02407 · HTML: https://mlsanity.com/paper/2011.02407/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Nov 2020.

## Built on by (most cited first, 10 of 10)

- [Multi-Objective Hyperparameter Optimization in Machine Learning -- An Overview](https://mlsanity.com/paper/2206.07438/) (2022)
- [Can Fairness be Automated? Guidelines and Opportunities for Fairness-aware AutoML](https://mlsanity.com/paper/2303.08485/) (2023)
- [Fair Enough: Searching for Sufficient Measures of Fairness](https://mlsanity.com/paper/2110.13029/) (2021)
- [One Model Many Scores: Using Multiverse Analysis to Prevent Fairness Hacking and Evaluate the Influence of Model Design Decisions](https://mlsanity.com/paper/2308.16681/) (2023)
- [fairmodels: A Flexible Tool For Bias Detection, Visualization, And Mitigation](https://mlsanity.com/paper/2104.00507/) (2021)
- [X Hacking: The Threat of Misguided AutoML](https://mlsanity.com/paper/2401.08513/) (2024)
- [The Intersectionality Problem for Algorithmic Fairness](https://mlsanity.com/paper/2411.02569/) (2024)
- [Beyond Fairness Metrics: Roadblocks and Challenges for Ethical AI in Practice](https://mlsanity.com/paper/2108.06217/) (2021)
- [Bias Begins with Data: The FairGround Corpus for Robust and Reproducible Research on Algorithmic Fairness](https://mlsanity.com/paper/2510.22363/) (2025)
- [Addressing Missing Sources with Adversarial Support-Matching](https://mlsanity.com/paper/2203.13154/) (2022)

## Abstract (arXiv)

We present an empirical study of debiasing methods for classifiers, showing that debiasers often fail in practice to generalize out-of-sample, and can in fact make fairness worse rather than better. A rigorous evaluation of the debiasing treatment effect requires extensive cross-validation beyond what is usually done. We demonstrate that this phenomenon can be explained as a consequence of bias-variance trade-off, with an increase in variance necessitated by imposing a fairness constraint. Follow-up experiments validate the theoretical prediction that the estimation variance depends strongly on the base rates of the protected class. Considering fairness--performance trade-offs justifies the counterintuitive notion that partial debiasing can actually yield better results in practice on out-of-sample data.