Narrowly tailored actuarial models for affirmative action in higher education

Previous research discusses the superiority of actuarial models over clinical models in a number of areas related to accuracy and consistency in decision making. The current project sought to develop an actuarial candidate selection model for affirmative action in higher education that would achieve the goal of diversity by assigning points for a number of diversity-related characteristics in addition to standard academic admission criteria. Two experiments showed that participants who used actuarial models selected applicants with more markers of academic success and greater diversity using factors favored by recent U.S. Supreme Court cases. The second experiment showed that an unweighted actuarial model also helped decision makers select more minority student candidates and that it produced higher ratings of procedural fairness. The article discusses how an actuarial model might pass Constitutional muster in light of recent Supreme Court cases.