MétaCan
Menu
← Back to cohort
Record W7118987466 · doi:10.1504/ijsmarttl.2025.150926

ChatGPT assists in identifying and recommending research study designs to support study protocol writing

2025· article· en· W7118987466 on OpenAlexaff
Gary KK. Low, Osamudiamen Favour Omosumwen, Sudarshan Subedi, Sam Froze Jiee, Wan Hsuan Lee, Sirjana Devkota, Selvanaayagam Shanmuganathan, Zelda Doyle

Bibliographic record

VenueInternational Journal of Smart Technology and Learning · 2025
Typearticle
Languageen
FieldMedicine
TopicArtificial Intelligence in Healthcare and Education
Canadian institutionsAlberta Health Services
Fundersnot available
KeywordsProtocol (science)Research designDesign studyCohen's kappaProtocol designScientific writingScheme (mathematics)Kappa

Abstract

fetched live from OpenAlex

This study aimed to validate ChatGPT by identifying the types of study designs in the published scientific literature and recommending the types of research study designs. ChatGPT Version 3.5 was asked to classify the published literature into different study designs. ChatGPT answers were then compared to the blinded independent reviewers' findings. The ChatGPT was then used to provide recommendations for study designs with the research titles from 333 unpublished study protocols. A total of 463 articles were used for validation. The Kappa coefficient was 84.45% agreement for the published literature. The agreement between ChatGPT and reviewers based on whole study protocol was 56.2%. ChatGPT may assist in classifying the types of study design with reasonably good accuracy if an abstract or full text is provided. ChatGPT may also be useful in recommending study designs to a novice researcher but with oversight from experts.

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.

How this classification was reachedexpand

Full frame machine prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.

metaresearch head score (Codex)0.562
metaresearch head score (Gemma)0.769
Version: metacan-v3-hybrid-931329e0061cValidation status: machine_predicted_unvalidated
Candidate categoriesMetaresearch
Consensus categoriesMetaresearch
DomainCandidate signal: Methods · Consensus signal: Methods
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.438
Threshold uncertainty score0.541

Distilled classifier scores by category (both heads)

CategoryCodexGemma
Metaresearch0.5620.769
Meta-epidemiology (narrow)0.0030.004
Meta-epidemiology (broad)0.0050.005
Bibliometrics0.0300.015
Science and technology studies0.0060.005
Scholarly communication0.0080.015
Open science0.0050.012
Research integrity0.0040.008
Insufficient payload (model declined to judge)0.0620.027

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.

Opus teacher head0.305
GPT teacher head0.579
Teacher spread0.274 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; the direct Gemma label and the distilled Codex classifier agree on what is shown here.

Study designSimulation or modeling
DomainMethods
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations0
Published2025
Admission routes1
Has abstractyes

Explore more

Same venueInternational Journal of Smart Technology and Learning→Same topicArtificial Intelligence in Healthcare and Education→French-language works237,207→