Texas A&M Drops "Race" from Student Risk Algorithm Following Markup Investigation – The Markup

Abstraction: University drops race variable from student dropout risk algorithm after bias investigation

Key points:

• EAB's Navigate software (used by 500+ schools) assigned Black students high-risk scores at double to quadruple the rate of White peers, based on The Markup's investigation
• Race was explicitly used as a "high impact predictor" in several schools' predictive models without administrators' awareness
• Texas A&M paused risk scores and asked EAB to rebuild models excluding race as a variable following the reporting
• The Major Explorer feature steered high-risk students toward "less risky" majors, compounding racial disparities in STEM access
• Experts warn removing race explicitly doesn't eliminate proxy discrimination — zip code, high school, and family income can still encode race in models
• Faculty and advisers were rarely told how risk scores were calculated or trained to interpret them

Connections: Eab · The Markup · Algorithmic Bias · Predictive Analytics · AI Fairness

Source: https://themarkup.org/news/2021/03/30/texas-am-drops-race-from-student-risk-algorithm-following-markup-investigation