Predictive models can suggest case outcomes under counterfactual race-neutral processing.

Evidence

We describe a project to develop a predictive model that can render prosecutorial decision-making more equitable and resistant to racial bias. The project involves building a model that, based on initial case information that a prosecutor receives, suggests how the case would resolve if the defendant were treated race-neutrally. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 1210

Categorization

• SEED: An article could profile how counterfactual simulation — treating the defendant ‘as if’ race-neutral — reimagines algorithmic fairness as a thought experiment rather than a reweighting.
• XREF: connects to: Missing counterfactuals make outcome-based evaluation of street-level programs impossible — Michael Lipsky, Street-Level Bureaucracy: Dilemmas of…, 390