Isolating racial bias requires cases identical except for race

Evidence

To identify the “true” and direct impact of the defendant’s race, ethnicity, or gender on the outcome of a case, one needs to account for every aspect of a case that may be correlated with the race, ethnicity, or gender of the defendant and which may also impact the outcome of the case. Thus, researchers aim to find two groups of cases that are identical in all possible dimensions (e.g., the aggravating and mitigating circumstances, the criminal history of the defendant, the attributes of the prosecutor, the characteristics of the defense attorney, and the income and education level of the defendant), but that differ only in the race of the defendant. If the resolution of these cases is different (e.g., the defendants in the first group are convicted more frequently in comparison to the defendants in the other group), then the disparity can be attributed to the racial differences between the defendants. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 732

Categorization

• DEFINE: A clean statement of the ceteris-paribus identification ideal for bias research — the counterfactual standard implicit in most social-science causal claims.
• XREF: connects to: Geographic quotas can function as covert racial discrimination that is nearly impossible to prove — Henry M. Hart, Jr. and Albert M. Sacks, The Legal Process: Basic Problems in…, 1704; the MOC “The False Compass: How Racial Categories Reorder Bodies, Minds, and the Self”