Skip to content

Manusights/retracted-citation-study

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Retracted-citation rate in recent biomedical papers

License: CC BY 4.0 Data version

How often does a recently published biomedical paper cite research that has since been retracted, corrected, or placed under an expression of concern?

We checked 12,402 recent (2025) papers across 54 major biomedical journals, covering 744,770 individual citations (435,108 unique cited DOIs), against the Retraction Watch Database and Crossref retraction notices.

Headline results

Measure Rate Count 95% CI (Wilson)
Papers citing a retracted, EoC, or corrected reference 7.1% 875 / 12,402 6.6 to 7.5%
Papers citing a retracted or EoC reference (higher severity) 1.4% 170 / 12,402 1.2 to 1.6%

About 1 in 73 recent papers cites a retracted or expression-of-concern reference.

We lead with the 1.4% figure rather than the headline 7.1%, because "correction" is the lowest-severity flag and the most common one: 783 of the flagged references were corrections, against 162 retractions and 25 expressions of concern. An erratum is not misconduct.

The gradient by field is the real finding

The corpus average hides a sharp split. Cell and cancer biology run far above it; clinical medicine sits near the floor.

Full 54-journal table (click to expand)
Journal Sampled Papers with flag Rate
Nature Reviews Immunology 42 9 21.4%
Nature Reviews Drug Discovery 30 5 16.7%
Cancer Cell 128 21 16.4%
Cell 305 47 15.4%
Cell Reports 306 45 14.7%
Cancer Research 291 42 14.4%
Journal of Experimental Medicine 158 20 12.7%
Diabetes 182 23 12.6%
Immunity 178 22 12.4%
American Journal of Human Genetics 166 19 11.4%
Nature Immunology 183 20 10.9%
Neuron 232 25 10.8%
Molecular Cell 274 27 9.9%
Developmental Cell 205 20 9.8%
Science Translational Medicine 255 24 9.4%
Nature Biotechnology 163 15 9.2%
eLife 289 26 9.0%
Nature Genetics 264 22 8.3%
Nature 294 24 8.2%
Biochemical and Biophysical Research Communications 297 24 8.1%
Journal of Immunology 306 23 7.5%
Journal of Cell Biology 209 15 7.2%
Nature Communications 303 21 6.9%
Journal of Biological Chemistry 306 21 6.9%
PLOS Genetics 295 20 6.8%
Nature Neuroscience 208 14 6.7%
PLOS Biology 308 19 6.2%
Nature Methods 227 14 6.2%
Proceedings of the National Academy of Sciences 299 18 6.0%
Nucleic Acids Research 302 18 6.0%
BMC Bioinformatics 270 16 5.9%
Nature Cell Biology 172 10 5.8%
Journal of Virology 297 16 5.4%
BMJ 148 7 4.7%
Science Advances 307 14 4.6%
Journal of Bacteriology 198 9 4.5%
Gastroenterology 179 8 4.5%
Blood 296 13 4.4%
Brain 302 13 4.3%
PLOS Medicine 140 6 4.3%
Nature Medicine 306 13 4.2%
Science 286 12 4.2%
Molecular Biology and Evolution 268 11 4.1%
The Lancet 244 10 4.1%
Circulation 253 10 4.0%
Bioinformatics 172 6 3.5%
The Journal of Neuroscience 300 10 3.3%
PLOS ONE 300 9 3.0%
JAMA 180 5 2.8%
Annals of Internal Medicine 162 4 2.5%
New England Journal of Medicine 298 7 2.3%
The Lancet Oncology 158 2 1.3%
Journal of the American Society of Echocardiography 83 1 1.2%
The Lancet Neurology 78 0 0.0%

Machine-readable copies in data/. The 54 rows sum to 12,402 papers sampled and 875 flagged.

If you work in cell biology, oncology, or immunology, your bibliography's base rate is roughly five times a clinical journal's.

What this does not mean

A paper is not wrong because one of its references was later retracted. Retractions frequently land after the citing paper was written, and the citing authors usually had no way to know.

The narrower and more useful point: at submission time, nobody ran a check that would have caught it, and that check now takes seconds.

Data

File Contents
data/retracted-citation-rate-by-journal-2026-06-04.csv Per-journal table: journal, papers sampled, papers with a flag, flag rate. Methodology in the header comments.
data/retracted-citation-rate-by-journal-2026-06-04.json Same data plus corpus-level summary and study metadata.

Row-level data (every citing/cited DOI pair) is available on request: erik@manusights.com

Method in brief

Full detail in METHODOLOGY.md.

  • Sample: 12,402 recent (2025) journal-articles drawn across 54 major biomedical journals via the Crossref /works API by ISSN, restricted to papers with at least 10 resolvable reference DOIs (4,969 thinner candidates excluded).
  • Check: all 744,770 reference citations matched against the live Retraction Watch dataset (over 61,000 notices) and Crossref update-to[] retraction notices.
  • Severity precedence: retracted > expression of concern > correction.
  • Intervals: Wilson 95%.

Known limitations, stated plainly

  1. This is a conservative lower bound. The hijacked-journal clone-domain check was not run in this batch, though production applies it.
  2. Flag probability scales with reference-list length. Excluding thin-reference papers pushes the rate up. Read 7.1% as the rate among papers with substantive bibliographies, not among all 2025 papers.
  3. The 54-journal panel is a sample of major journals, not a random draw of all biomedical literature. The per-field gradient is the honest way to read it.
  4. Review articles score high partly because they cite more. The figures are unadjusted for reference count.

How to cite

Manusights. (2026). How often do recent papers cite retracted research? A 12,402-paper study. https://manusights.com/tools/reference-integrity/retracted-citation-study

@misc{manusights2026retracted,
  title        = {How Often Do Recent Papers Cite Retracted Research? A 12,402-Paper Study},
  author       = {{Manusights}},
  year         = {2026},
  howpublished = {\url{https://manusights.com/tools/reference-integrity/retracted-citation-study}},
  note         = {Data study. 12,402 recent biomedical papers across 54 journals;
                  built on Crossref and the Retraction Watch Database (ISSN 2692-4579).}
}

If you cite this study, please also cite the underlying sources: the Retraction Watch Database (ISSN 2692-4579) and Crossref. Neither is ours; the analysis is.

License

CC BY 4.0. Reuse freely with attribution to Manusights, the Retraction Watch Database, and Crossref.

Related

About

Retracted-citation rate in recent biomedical papers: 12,402 papers across 54 journals, 744,770 citations checked. 7.1% cite a retracted/EoC/corrected reference (95% CI 6.6-7.5). CC BY 4.0. Study: https://manusights.com/tools/reference-integrity/retracted-citation-study

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors