A clustering algorithm for university admissions

Abstract

In 2003 The Supreme Court declared that all government funded universities, which choose to consider race in their admissions processes, must utilize a holistic process. A holistic process includes a thorough evaluation of all aspects of each applicant. For larger universities this type of admissions process would be very taxing. A computer scientist from Auburn University created an algorithm, Applications Quest, to handle large quantities of applications in a way that would evaluate applicants holistically with a computational tool. Applications Quest utilizes the Euclidean distance measure, Similarity matrices, Divisive Clustering, and Random Selection. This algorithm produces a diverse admittance class for a university. In this research we simulate this algorithm and run tests with hypothetical Rice University data. Ultimately, we are left with the following question: Can a computational use of arbitrary difference account for human qualities that define certain social phenomena, such as underrepresentation in higher education?