CoChat

Performance of ChatGPT on the MCAT: The Road to Personalized and Equitable Premedical Learning

Open access
Bommineni, Vikas L, Bhagwagar, Sanaea, Balcarcel, Daniel, Davatzikos, Christos, Boyer, Donald · 2023 · medRxiv
PreprintCited by 54

The DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.

Abstract

Despite an increasingly diverse population, an unmet demand for undergraduates from underrepresented racial and ethnic minority (URM) backgrounds exists in the field of medicine as a result of financial hurdles and insufficient educational support faced by URM students in the premedical journey. With the capacity to provide highly individualized and accessible no- or low-cost dynamic instruction, large language models (LLMs) and their chatbot derivatives are posed to change this dynamic and subsequently help shape a more diverse future physician workforce. While studies have established the passing performance and insightful explanations of one of the most accurate LLM-powered chatbots to date—Chat Generative Pre-trained Transformer (ChatGPT)—on standardized exams such as medical licensing exams, the role of ChatGPT in premedical education remains unknown. We evaluated the performance of ChatGPT on the Medical College Admission Test (MCAT), a standardized 230-question multiple choice exam that assesses a broad range of competencies in the natural, physical, social, and behavioral sciences as well as critical analysis and reasoning. Depending on its visual item response strategy, ChatGPT performed at or above the median performance of 276,779 student test takers on the MCAT. Additionally, ChatGPT-generated answers demonstrated both a high level of agreement with the official answer key as well as insight into its explanations. Based on these promising results, we anticipate two primary applications of ChatGPT and future LLM iterations in premedical education: firstly, such models could provide free or low-cost access to personalized and insightful explanations of MCAT competency-related questions to help students from all socioeconomic and URM backgrounds. Secondly, these models could be used to generate additional test questions by test-makers or for targeted preparation by pre-medical students. These applications of ChatGPT in premedical education could be an invaluable, innovative path forward to increase diversity and improve equity among premedical students.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1101/2023.03.05.23286533

Cites

9 of 9 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 9 references

Cited by

See all 54

Metadata

Authors
Bommineni, V. L., Bhagwagar, S., Balcarcel, D., Davatzikos, C., & Boyer, D.
Venue
medRxiv
Publisher
openRxiv
Published
2023-03-07
OpenAlex
W4323352552
Cited by
54
References
9
Type
posted content
Language
en
Open access
green

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Performance of ChatGPT on the MCAT: The Road to Personalized and Equitable Premedical Learning. https://app.cochat.ai/papers/10.1101/2023.03.05.23286533
The paper
Bommineni, V. L., Bhagwagar, S., Balcarcel, D., Davatzikos, C., & Boyer, D. (2023). Performance of ChatGPT on the MCAT: The Road to Personalized and Equitable Premedical Learning. medRxiv. https://doi.org/10.1101/2023.03.05.23286533

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.