Big Data's Disparate Impact¶
Barocas, S. (2016). Big Data's Disparate Impact. California Law Review, 104(3), 671-732.
Cited by¶
5 citations across 5 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Equity
- The abstraction generalizes beyond formal law to any institutional context where case-by-case judgment and remedial discretion must operate alongside rule systems: algorithmic exception-handling, software engineering's case-specific patches, organizational review of policy-rule enforcement, ethical committees adjudicating borderline cases, and dispute-resolution processes where rigid rules would fail the underlying purpose of fairness, a substrate-extension that Barocas and Selbst (2016) develop in their analysis of disparate impact in algorithmic systems.
This sourceShows how data-driven decision systems produce discriminatory outcomes through training data, feature selection, and proxies, and argues solving this requires reexamining 'discrimination' and 'fairness'.
- The abstraction generalizes beyond formal law to any institutional context where case-by-case judgment and remedial discretion must operate alongside rule systems: algorithmic exception-handling, software engineering's case-specific patches, organizational review of policy-rule enforcement, ethical committees adjudicating borderline cases, and dispute-resolution processes where rigid rules would fail the underlying purpose of fairness, a substrate-extension that Barocas and Selbst (2016) develop in their analysis of disparate impact in algorithmic systems.
- Fairness
- This reframing allows organizations to audit systems (hiring algorithms, resource allocation, performance evaluation) against explicit fairness metrics, identify misalignment, and adjust, as Barocas and Selbst (2016) document for "big data's disparate impact" and the auditing of algorithmic decision systems.
This sourceShows how data-mining/algorithmic systems produce discriminatory outcomes through training data, feature selection, and proxies, and argues for auditing such systems and rethinking 'discrimination' and 'fairness.'
- This reframing allows organizations to audit systems (hiring algorithms, resource allocation, performance evaluation) against explicit fairness metrics, identify misalignment, and adjust, as Barocas and Selbst (2016) document for "big data's disparate impact" and the auditing of algorithmic decision systems.
Mechanisms¶
- Causal Feature Review Panel
- The danger it addresses — a feature that predicts a protected class rather than the target — is the core of proxy discrimination.
This sourceExplains that seemingly neutral data features can proxy for protected characteristics and reproduce discrimination without explicit use of class membership.
- The danger it addresses — a feature that predicts a protected class rather than the target — is the core of proxy discrimination.
- High-Risk Targeting List
- Its failure modes are the ones the archetype flags for risk scoring specifically: a score can encode bias when a feature is a proxy for a protected group, producing disparate impact even with no intent, and a self-reinforcing loop can manufacture the very risk profile it claims to measure.
This sourceExplains how data-mining features can act as proxies for protected-class membership and produce disparate impact without discriminatory intent.
- Its failure modes are the ones the archetype flags for risk scoring specifically: a score can encode bias when a feature is a proxy for a protected group, producing disparate impact even with no intent, and a self-reinforcing loop can manufacture the very risk profile it claims to measure.
- Risk Score Proxy Metric
- Its failure mode is uniquely ethical: a proxy can encode and launder the very attributes it must not use
This sourceShows how neutral-seeming variables can proxy protected-class membership and yield systematically less favorable outcomes behind an apparently objective model.
- Its failure mode is uniquely ethical: a proxy can encode and launder the very attributes it must not use
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Not recorded. Neither this nor any other of the 5 citations of this work carries a recorded support check.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Links previously used in the corpus¶
Before the registry existed this work was also linked 1 other way.
Registry ID ref:0aba451de0f5 · see in the full table