Can Generative Pre-trained Transformers (GPT) Pass Assessments in Higher Education Programming Courses?
Open accessAbstract
We evaluated the capability of generative pre-trained transformers (GPT), to pass assessments in introductory and intermediate Python programming courses at the postsecondary level. Discussions of potential uses (e.g., exercise generation, code explanation) and misuses (e.g., cheating) of this emerging technology in programming education have intensified, but to date there has not been a rigorous analysis of the models' capabilities in the realistic context of a full-fledged programming course with diverse set of assessment instruments. We evaluated GPT on three Python courses that employ assessments ranging from simple multiple-choice questions (no code involved) to complex programming projects with code bases distributed into multiple files (599 exercises overall). Further, we studied if and how successfully GPT models leverage feedback provided by an auto-grader. We found that the current models are not capable of passing the full spectrum of assessments typically involved in a Python programming course (<70% on even entry-level modules). Yet, it is clear that a straightforward application of these easily accessible models could enable a learner to obtain a non-trivial portion of the overall available score (>55%) in introductory and intermediate courses alike. While the models exhibit remarkable capabilities, including correcting solutions based on auto-grader's feedback, some limitations exist (e.g., poor handling of exercises requiring complex chains of reasoning steps). These findings can be leveraged by instructors wishing to adapt their assessments so that GPT becomes a valuable assistant for a learner as opposed to an end-to-end solution.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
A straightforward application of these models could enable a learner to obtain a non-trivial portion of the overall available score (over 55%) in introductory and intermediate courses alike.
“but a straightforward application of these models could enable a learner to obtain a non-trivial portion of the overall available score (over 55%) in introductory and intermediate courses alike.”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
There appeared to be a difference in success rate between MCQs that contained a code snippet and those that did not.
“and that there appeared to be a difference in success rate between MCQs that contained a code snippet and those that did not [45, 46].”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
Prior findings [45, 46] no longer hold.
“Overall, the findings reported in prior work [45, 46] no longer hold.”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
The then current GPT models were not capable of passing the full spectrum of assessments typically involved in a Python programming course (below 70% on even entry-level modules).
“They found that the then current GPT models were not capable of passing the full spectrum of assessments typically involved in a Python programming course (below 70% on even entry-level modules);”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
An important limitation of the GPT models was their apparent struggle with activities that required multi-hop reasoning.
“They also observed that an important limitation of the GPT models was their apparent struggle with activities that required multi-hop reasoning,”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
Findings reported in [46] appear challenged by GPT-4, as well as findings in [7, 10, 11, 38].
“Hence, the findings reported in [46] appear to be challenged by the more capable GPT-4 model as well as the findings reported in [7, 10, 11, 38].”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
The ability of GPT-4 to utilize feedback seems to be even stronger than the ability of the GPT-3.5 model evaluated in [38, 46].
“Hence, the ability of GPT-4 to utilize feedback seems to be even stronger than the ability of the GPT-3.5 model evaluated in [38, 46].”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
Strengths and limitations reported in prior work [7, 10, 11, 38, 46] still hold to a certain degree.
“Hence, the strengths and limitations reported in prior work [7, 10, 11, 38, 46] still hold to a certain degree.”
From Thrilled by Your Progress! Large Language Models (GPT-4) No Longer Struggle to Pass Assessments in Higher Education Programming Courses (2023) · cites this paper as [46]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
The citation lists haven’t been fetched yet
References 18 · Cited by 109
- Authors
- Savelka, J., Agarwal, A., Bogart, C., Song, Y., & Sakr, M.
- Affiliations
- Carnegie Mellon University, Pittsburgh, PA, USA
- Carnegie Mellon University (US)
- Venue
- Proceedings of the 2023 Conference on Innovation and Technology in Computer Science Education V. 1
- Publisher
- ACM
- Published
- 2023-06-29
- Pages
- 117-123
- OpenAlex
- W4382653968
- Cited by
- 109
- References
- 18
- Citations by year
- 2026: 10 · 2025: 35 · 2024: 50 · 2023: 14
- FWCI
- 32.96
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- ACM Conference on Innovation and Technology in Computer Science Education (ITiCSE) (publishedVersion, cc-by)
- Research Publications (Maastricht University) (submittedVersion, cc-by)
- Funders
- National Science Foundation (2040490)
- University of Pittsburgh
- Topics
- Physical Sciences › Computer Science › Computer Science Applications › Teaching and Learning Programming
- Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Keywords
- generative pre-trained transformer, GPT, programming education, Python programming course, automated assessment, automated grading, code generation, cheating, multiple-choice questions, programming projects, feedback use, higher education
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
