ProPublica used 2014 recidivism outcomes to test the risk model’s accuracy and bias

Categorization

• corroborated: 2 independent passages support this claim
• XREF: connects to: ProPublica tested COMPAS recidivism predictions against actual reoffense outcomes for bias

Related

• ProPublica tested COMPAS recidivism predictions against actual reoffense outcomes for bias (similarity 0.85)
• Science only recently began predicting parolee behavior after release (similarity 0.79)