Racial Bias in Using USMLE Step 1 Scores to Grant Internal Medicine Residency Interviews
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
PURPOSE: To determine whether the United States Medical Licensing Examination (USMLE) Step 1 score, commonly used in screening residency applicants for interviews, eliminates a greater proportion of African-American applicants from the interview process at an internal medicine residency program. METHOD: A survey of internal medicine residency programs was performed to determine the prevalence of using USMLE Step 1 scores to grant interviews. A cohort of applicants was analyzed by constructing a database of USMLE Step 1 scores from the Electronic Residency Application Service (ERAS) database of applications from U.S., Canadian, and osteopathic medical schools to one residency program in 2000. Each applicant was classified as African American or non-African American. Rejection rates were then calculated for each five-point increment from a hypothetical threshold rejection score of <180 to <215. RESULTS: Responses were received from 259 residency programs (69%), and 92% used the USMLE Step 1 score in deciding which applicants to interview. A cohort of 626 non-African-American and 47 African-American applicants was analyzed. The proportion of applicants below each incremental threshold score was significantly higher for African-American applicants (p <.05 at each level). Depending on the threshold score used, an African-American applicant was three to six times less likely to be offered an interview. CONCLUSIONS: When USMLE Step 1 scores are used to screen applicants for a residency interview, a significantly greater proportion of African-American students will be refused an interview.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it