Predicting Inclusion: Institutional Correlates of Transgender-Inclusive Policies in U.S. Public Universities using Machine Learning Analysis

• In Event: AERA Roundtable Session Saturday 3:45 pm - Gold 4
  In Roundtable Session: RT2: Quantitative Advances and Approaches to Research on Inclusion, Equity, Educational Disparities and Student Success

Abstract

Amid more than 900 anti‑trans bills, U.S. public universities vary widely in supportive policies. Drawing on neo‑institutional, queer, and QuantQueer perspectives, we merged a 2025 audit of transgender‑inclusive health‑insurance, housing, and athletics policies at 94 land‑grant institutions with 4,975 IPEDS variables (2023). Random‑forest models in R (v 4.3.2), validated by 220‑fold cross‑validation (accuracy =.79–.82; κ =.68–.72), pinpointed key predictors. Surgery‑inclusive insurance clustered at open‑access campuses graduating many multiracial women and logging few residency‑unknown students, while selective, low‑diversity schools offered none. Gender‑inclusive housing appeared at well‑resourced universities enrolling more NHOPI men; varsity trans participation thrived where Asian four‑year and Hispanic five‑year graduation counts diverged. Findings highlight how diversity, selectivity, and resources jointly shape policy adoption and showcase QuantQueer Machine Learning as a tool for equity research.

Authors

• Amanda J. Davis Simpfenderfer, College of William & Mary

• Ailidh Wallace, College of William & Mary

• Travis Heath Olson, Michigan State University

• Obed Amoakoh Boateng, Virginia Commonwealth University

• Kamden Strunk, Virginia Commonwealth University

• Jason C. Garvey, University of Vermont

• Mario I Suárez, Utah State University