Abstract:Each year, selective American colleges sort through tens of thousands of applications to identify a first-year class that displays both academic merit and diversity. In the 2023-2024 admissions cycle, these colleges faced unprecedented challenges. First, the number of applications has been steadily growing. Second, test-optional policies that have remained in place since the COVID-19 pandemic limit access to key information historically predictive of academic success. Most recently, longstanding debates over affirmative action culminated in the Supreme Court banning race-conscious admissions. Colleges have explored machine learning (ML) models to address the issues of scale and missing test scores, often via ranking algorithms intended to focus on 'top' applicants. However, the Court's ruling will force changes to these models, which were able to consider race as a factor in ranking. There is currently a poor understanding of how these mandated changes will shape applicant ranking algorithms, and, by extension, admitted classes. We seek to address this by quantifying the impact of different admission policies on the applications prioritized for review. We show that removing race data from a developed applicant ranking algorithm reduces the diversity of the top-ranked pool without meaningfully increasing the academic merit of that pool. We contextualize this impact by showing that excluding data on applicant race has a greater impact than excluding other potentially informative variables like intended majors. Finally, we measure the impact of policy change on individuals by comparing the arbitrariness in applicant rank attributable to policy change to the arbitrariness attributable to randomness. We find that any given policy has a high degree of arbitrariness and that removing race data from the ranking algorithm increases arbitrariness in outcomes for most applicants.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the impact on the prioritization of applicants in the American college admission process due to policy changes (especially the abolition of affirmative action) and inherent variability. Specifically, the research focuses on the following two main issues: 1. **How do policy changes affect the overall composition of the freshman class in colleges?** - The research explored how colleges are affected when choosing the freshman class after the consideration of race is removed. In particular, whether this change will affect the academic level and diversity of applicants who are given priority for review. 2. **How do policy changes affect individual applicants?** - The research analyzed the ranking changes of each applicant under different policies and whether these changes are random. By comparing the ranking results under different policies, the researchers attempted to quantify the impact of policy changes on individual applicants. To answer these questions, the researchers used four years of admission data and developed machine - learning models to simulate the rankings of applicants under different policies. The main contributions of the research are: - **Predicting the specific impact of the admission policy after the removal of the consideration of race on the academic level and diversity of college admissions.** - **Providing background information by comparing with changes in other hypothetical policies to better understand the impact of policy changes.** - **Proposing a method beyond group fairness assessment to understand the relative impact of policy changes on individual applicants.** The research found that the policy of removing the consideration of race did not significantly improve the academic level of top - tier applicants, but instead significantly reduced diversity. Moreover, the negative impact of excluding race data on diversity is greater than excluding other variables (such as expected major). At the individual level, the research found that any given policy is highly random, that is, at most only 9% of applicants are always in the top 20% under different policies. In addition, the ranking changes of individual applicants under different policies are greater than those under the same policy. Finally, the ranking algorithm that removes the consideration of race increased the randomness for most applicants (including Asian, Latino, white, and multi - ethnic applicants).

Algorithms for College Admissions Decision Support: Impacts of Policy Change and Inherent Variability

The Disparate Impacts of College Admissions Policies on Asian American Applicants

Towards Data-Driven Affirmative Action Policies under Uncertainty

Affirmative Action Policies for Top-k Candidates Selection, With an Application to the Design of Policies for University Admissions

Trading off performance and human oversight in algorithmic policy: evidence from Danish college admissions

Bias Analysis of AI Models for Undergraduate Student Admissions

Inside the Black Box: Detecting and Mitigating Algorithmic Bias across Racialized Groups in College Student-Success Prediction

Assessing the Potential Impact of a Nationwide Class-Based Affirmative Action System

Dropping Standardized Testing for Admissions Trades Off Information and Access

Explicit vs. statistical targeting in affirmative action: Theory and evidence from Chicago's exam schools

Augmenting Holistic Review in University Admission using Natural Language Processing for Essays and Recommendation Letters

The Disparate Impact of Uncertainty: Affirmative Action vs. Affirmative Information

Reducing the Filtering Effect in Public School Admissions: A Bias-aware Analysis for Targeted Interventions

Test-Optional Admissions

Impact of Masking Academic Metrics on Enrolling a Broadly Diverse and Mission-Aligned Student Body

Test-optional Policies: Overcoming Strategic Behavior and Informational Gaps

An empirical evaluation of Chinese college admissions reforms through a natural experiment

Evaluation of Fairness Trade-offs in Predicting Student Success

Admission Prediction in Undergraduate Applications: an Interpretable Deep Learning Approach

An algorithm for removing sensitive information: application to race-independent recidivism prediction