The Big Picture
What We Did
Researchers analyzed data from 117,071 adult probationers in Texas, released from probation between 2000 and 2010. They investigated the predictive accuracy and error distribution of the Wisconsin Risk Needs Assessment instrument across different racial groups (Black, Hispanic, and White). The study utilized logistic regression models to examine predictive bias, including interaction effects between race, risk scores, and offense severity.
What We Found
1
Predictive Equity and Error Distribution:
– The instrument showed a significant predictive error for racial groups, with under-classification more common among Black probationers and over-classification more prevalent among White probationers.
– The risk scores were generally more accurate for “higher risk” probationers, while “lower risk” individuals, particularly those who were Black, experienced greater predictive inequities.
2
Predictive Bias:
– Black probationers were more likely to be inaccurately classified as lower risk when compared to White and Hispanic probationers at equivalent risk levels.
– Hispanic probationers showed no significant difference from White probationers at lower risk scores, but predictive accuracy declined for higher scores.
3
Impact of Offense Severity:
Why This Matters
The findings highlight significant racial and ethnic disparities in the predictive accuracy of widely used probation risk assessment tools. These disparities can lead to unequal treatment in the probation system, with minorities facing higher risks of incorrect classification and, consequently, inappropriate levels of supervision and intervention. Understanding these biases is crucial for developing fairer correctional practices and policies that minimize predictive errors and racial inequities.
What Can We Do?
1
Refine Risk Assessment Tools:
2
Improve Training and Oversight:
3
Further Research:
About the Study
This study provides crucial insights into the racial and ethnic predictive inequities of the Wisconsin Risk Needs Assessment instrument, underscoring the need for critical evaluation and reform of risk assessment practices in the criminal justice system. By highlighting the differential impacts on Black, Hispanic, and White probationers, it calls for targeted efforts to address these disparities in correctional settings. The full APA citation is: Henderson, H., Tanana, M., Bourgeois, J. W., & Adams, A. T. (2015). Psychometric racial and ethnic predictive inequities. Journal of Black Studies, 46(5), 1-20. https://doi.org/10.1177/0021934715574096