Major Universities Are Using Race as a "High Impact Predictor" of Student Success – The Markup
algorithmic-biaseducationpredictive-analyticsraceai-ethics
Abstraction: EAB Navigate advising software uses race to predict student dropout risk
Key points:
- EAB's Navigate software, used by 500+ universities, incorporates race as a "high-impact predictor" (accounts for >5% of score variance) at 4 of 7 examined schools
- Black students labeled high-risk at 2.3–3.9x the rate of white students at UMass Amherst, UW-Milwaukee, and Texas A&M
- Georgia State University—the cited success model—does NOT include race in its algorithm and invests heavily in advisers (350:1 ratio vs 700:1 elsewhere)
- Several university administrators were unaware race was included or that large racial disparities existed in scores
- Navigate's Major Explorer function may steer minority students away from STEM by surfacing lower-risk alternative majors
- Critics warn models trained on historic data encode existing discriminatory patterns; school administrators have little training on score interpretation
Connections: Eab · The Markup · Algorithmic Bias · Predictive Analytics · AI Ethics