The DOI matches its Crossref record; 32 of the 32 references it cites were found in a registry or an index.
Predicting retracted research: a dataset and machine learning approaches
Open accessAbstract
Background: Retractions undermine the scientific record’s reliability and can lead to the continued propagation of flawed research. This study aimed to (1) create a dataset aggregating retraction information with bibliographic metadata, (2) train and evaluate various machine learning approaches to predict article retractions, and (3) assess each feature’s contribution to feature-based classifier performance using ablation studies. Methods: An open-access dataset was developed by combining information from the Retraction Watch database and the OpenAlex API. Using a case-controlled design, retracted research articles were paired with non-retracted articles published in the same period. Traditional feature-based classifiers and models leveraging contextual language representations were then trained and evaluated. Model performance was assessed using accuracy, precision, recall, and the F1-score. Results: The Llama 3.2 base model achieved the highest overall accuracy. The Random Forest classifier achieved a precision of 0.687 for identifying non-retracted articles, while the Llama 3.2 base model reached a precision of 0.683 for identifying retracted articles. Traditional feature-based classifiers generally outperformed most contextual language models, except for the Llama 3.2 base model, which showed competitive performance across several metrics. Conclusions: Although no single model excelled across all metrics, our findings indicate that machine learning techniques can effectively support the identification of retracted research. These results provide a foundation for developing automated tools to assist publishers and reviewers in detecting potentially problematic publications. Further research should focus on refining these models and investigating additional features to improve predictive performance. Trial registration: Not applicable.
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.
Integrity signals
Registry
Cites
Statistics
Figures




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.
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In the context of the increasing number of research retractions, few AI platforms have been developed to screen articles for potential retraction, which can also be used in the peer-review process to evaluate the possibility of future retraction.37”
From Artificial Intelligence in Detecting Statistical Errors: Implications for Authors, Reviewers, and Editors (2025) · cites this paper as [37]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“On the other side, using retracted articles to train AI to help predict future or possible retractions has also been studied.”
From Artificial Intelligence in Detecting Statistical Errors: Implications for Authors, Reviewers, and Editors (2025) · cites this paper as [37]
The cited source says what the claim attributes to it.
In the citing paper
“In a study by Fletcher and Stevenson,37 who studied the precision of the AI platform Llama 3.2 Base Model, achieving an accuracy of 68%.”
From Artificial Intelligence in Detecting Statistical Errors: Implications for Authors, Reviewers, and Editors (2025) · cites this paper as [37]
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
1 retracted
Retracted Science and the Retraction Index
Fang, F. C., & Casadevall, A. (2011) · Infection and Immunity · doi:10.1128/iai.05661-11 · Cited by 405 · cited 1×
“…this issue from several angles: classifying the reasons why already withdrawn papers were retracted [17], analysing characteristics common among articles deemed to warrant retraction [18, 19], and highlighting systemic difficulties in retracting papers even when significant issues are known [20].”
RetractedThe probabilistic relevance framework: BM25 and beyond
Robertson, S., & Zaragoza, H. (2009) · Foundations and Trends® in Information Retrieval · doi:10.1561/1500000019 · Cited by 3,374 · cited 1×
“All other text features were represented using term weights produced by Best Matching 25 [29] with k = 2 , b = 0.3 and no maximum vocabulary length specified.”
InfluentialMisconduct accounts for the majority of retracted scientific publications
Fang, F. C., Steen, R. G., & Casadevall, A. (2012) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.1212247109 · Cited by 1,232 · cited 1×
“For example, Steen [2] attributed 73.5% of Pub-Med retractions to errors, whereas Fang et al. [16] later, using additional sources, reclassified 15.9% of this dataset from errors to fraud, suggesting that clarifying a paper’s retraction status is more straightforward than classifying its cause.”
InfluentialFairness in machine learning: A survey
Caton, S., & Haas, C. (2024) · ACM Computing Surveys · doi:10.1145/3616865 · Cited by 646 · cited 1×
“Predictive models, such as the ones described here, can introduce bias thereby raising potential fairness issues [32].”
Fairness in Machine Learning: A Survey
Caton, S., & Haas, C. (2020) · arXiv · doi:10.48550/arxiv.2010.04053 · Cited by 646 · cited 1×
“Predictive models, such as the ones described here, can introduce bias thereby raising potential fairness issues [32].”
A Survey on Text Classification: From Traditional to Deep Learning
Li, Q., Peng, H., Li, J., Xia, C., Yang, R., Sun, L., Yu, P. S., & He, L. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3495162 · Cited by 494 · cited 1×
“Feature-based classifiers were selected that have demonstrated effectiveness for text-classification tasks: Gradient Boosting, SVM, XGBoost, Random Forest, MLP and Decision Trees [26].”
Retractions in the scientific literature: is the incidence of research fraud increasing?
Steen, R. G. (2011) · Journal of Medical Ethics · doi:10.1136/jme.2010.040923 · Cited by 405 · cited 1×
“Retracting scientific articles is essential for safeguarding the integrity of the research record, but the growing number of retractions also reveals weaknesses in peer review and editorial oversight [1, 2].”
More than 10,000 research papers were retracted in 2023 — a new record
Van Noorden, R. (2023) · Nature · doi:10.1038/d41586-023-03974-8 · Cited by 351
Priem, J. R., Piwowar, H., Barrett, J., Demes, K., McCafferty, N., Meyer, C., Orr, R., & Portenoy, J. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2205.01833 · Cited by 297 · cited 1×
“These resources were combined to create a single dataset suitable for predicting article retractions [22].”
Model Assessment and Selection
Hastie, T., Tibshirani, R., & Friedman, J. (2009) · Springer Series in Statistics · doi:10.1007/978-0-387-84858-7_7 · Cited by 265 · cited 1×
“Group sizes were chosen with nested split using standard machine learning splits [25].”
Research works printed but not linked: 1
- Bai Y, Kadavath S, Kundu S, Askell A, Kernion J, Jones A, et al. Constitutional AI: Harmlessness from AI Feedback. 2022. https://arxiv.org/abs/2212.08073. Accessed 23 Jan 2025. · arXiv
Cited by
Scancar, B., Byrne, J. A., Causeur, D., & Barnett, A. G. (2026) · BMJ · doi:10.1136/bmj-2025-087581 · Cited by 38
“Also, Bless and colleagues 24 and other studies 25 26 have shown that machine learning methods can predict retractions and paper mill products from text using data from Retraction Watch—a non-profit group that records paper retractions 27 —in a cross disciplinary context rather than within a…”
Alnaimat, F., AlSamhori, A. R. F., El Sharu, H., Othman, L., Oralbek, A., & Zimba, O. (2025) · Journal of Korean Medical Science · doi:10.3346/jkms.2025.40.e342 · Cited by 4
“37 Table 1 provides a concise overview of the mentioned AI tools.”
InfluentialDeng, X., Wang, X., & Stevenson, M. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744614 · Cited by 10
Self-citationResearch Retractions: Scoping Review with ☸️SAIMSARA
SAIMSARA (2026) · SAIMSARA Journal · doi:10.62487/saimsaraada9ec85
Hussain, A., Ahmad, A., Rahman, A., & Ahmad, K. (2026) · Serials Review · doi:10.1080/00987913.2026.2663516
“Fletcher and Stevenson (2025) developed predictive models using metadata and classifiers, such as random forests, to identify potential retractions, achieving high precision.”
Metadata
- Authors
- Fletcher, A. H. A., & Stevenson, M.
- Affiliations
- University of Sheffield (GB)
- Corresponding author
- Fletcher, Aaron H. A.
- Venue
- Research Integrity and Peer Review
- Published
- 2025-06-11
- Volume
- 10
- Issue
- 1
- Pages
- 9-9
- ISSN
- 2058-8615 (ISSN-L)
- PMC
- PMC12153192
- OpenAlex
- W4411162931
- Cited by
- 9
- References
- 32
- Citations by year
- 2026: 7 · 2025: 2
- FWCI
- 9.31
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Research Integrity and Peer Review (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- White Rose Research Online (University of Leeds, The University of Sheffield, University of York) (acceptedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- UK Research and Innovation (EP/S023062/1)
- Topics
- Social Sciences › Education › Academic integrity and plagiarism
- Social Sciences › Library and Information Sciences › scientometrics and bibliometrics research
- Social Sciences › Library and Information Sciences › Academic Publishing and Open Access
- Keywords
- retraction, Retraction Watch, bibliographic metadata, Llama 3.2, contextual language models, feature-based classifiers, random forest, scientific integrity
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: Predicting retracted research: a dataset and machine learning approaches. https://app.cochat.ai/papers/10.1186/s41073-025-00168-w
- The paper
- Fletcher, A. H. A., & Stevenson, M. (2025). Predicting retracted research: a dataset and machine learning approaches. Research Integrity and Peer Review. https://doi.org/10.1186/s41073-025-00168-w
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.