Abstract
Background: Socioeconomic differences in matriculant composition may emerge because fewer people apply, applicants are accepted at a lower rate, acceptees matriculate at a lower rate, or these processes operate together. We compared the AAMC Education-Occupation (EO) classification, first-generation college student status, and AMCAS fee assistance use within a common stage-decomposition framework. The primary question was where visible national aggregate differences were concentrated and how much each stage contributed to the overall application-to-matriculation difference. Methods: We analyzed publicly available AAMC FACTS aggregate counts of applicants, acceptees, and matriculants. The primary EO analysis compared the lower EO group (EO1/2, reflecting lower parental education and/or nonprofessional occupational categories) with the higher EO group (EO3/4/5, reflecting degree-holding parents in administrative, managerial, or professional occupations) among records with known EO status. We also analyzed first-generation college students and AMCAS fee assistance users. Available periods were 2018–2019 to 2025–2026, 2020–2021 to 2025–2026, and 2022–2023 to 2025–2026, respectively. We calculated acceptance, matriculation-yield, and applicant-to-matriculant pipeline rates; stage-specific changes in representation; descriptive ordinary least-squares trends; and descriptive stage-equalization decompositions. Results: In 2025–2026, the lower EO group, first-generation college students, and AMCAS fee assistance users had lower acceptance and applicant-to-matriculant pipeline rates than their comparison groups. Acceptance-rate gaps were 11.7, -11.2, and -7.3 percentage points, respectively; the corresponding pipeline-rate gaps were -11.6, -11.2, and -22 percentage points, whereas post-acceptance matriculation-yield gaps were only -1.0 to -0.4 percentage points. Two complementary descriptive calculations quantified the stage contrast. In a direct representation decomposition, the applicant-to-acceptee stage accounted for 96.6%-97.9% of the decline from applicant share to matriculant share. In a separate rate-equalization calculation, replacing the focal-group acceptance rate with the reference-group rate reduced the pipeline-rate gaps by 95.7%-97.6%. Conclusions: National AAMC aggregate data indicate that the main visible differences across three socioeconomic markers were concentrated between application and acceptance, whereas national acceptee-to-matriculant yield differences were small. The findings identify the application-to-acceptance stage as the priority for individual- and institution-level research on the mechanisms underlying these national patterns and their implications for admissions policy and downstream outcomes.
Funding Information
  • the Fundamental Research Funds for the Central Universities (CCNU26XJ009)