Overall racial bias estimates can be decomposed from in-group bias regression coefficients

Evidence

This is significant because it demonstrates that the effects of overt racial bias (β1 in Equation 1) can be recovered from the coefficients of the in-group bias regression (Equation 3). To make this clear, assume that Y stands for sentence length, measured in months, and assume that δ3 is estimated as -6 from Equation (3). This means that there is in-group bias in sentencing and that defendants receive sentences that are six months shorter if they are matched by a judge of their own race. This important finding can be broken down into its components. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 768

Categorization

• DEFINE: Explains the statistical link between Equation 1’s overall bias parameter and Equation 3’s in-group bias decomposition.
• corroborated: 2 independent passages support this claim — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law: A Definitive Look at…, 768