Ten Scholarly Journals Have Retracted Half or More of Everything They Published Since 2020

USE CASE. Three groups need a per-journal cumulative retraction RATE rather than the per-event news coverage that dominates Retraction Watch: (1) academic library subscription committees and consortial buyers (CDL, OhioLINK, Big Ten Academic Alliance) deciding which journals to drop or move to no-cost-to-read tiers when negotiating renewal contracts; (2) faculty submission committees and individual researchers deciding which journals to avoid as submission targets; (3) research integrity officers and tenure review committees evaluating candidates' publication records and weighing the strength of evidence each journal venue provides. The institution-level R Rate is now in the literature (RI^2 index, Scientometrics 2025) but the per-journal cumulative rate ranking is not a standard published product. RESULT. Of 200 journals examined, 10 have retracted at least 49% of their 2020-2025 published output, and the top 5 are above 60%: International Journal of Electrical Engineering Education (SAGE) 65.9% (253/384), Journal of Healthcare Engineering (Hindawi/Wiley) 65.5% (2,171/3,316), Applied Bionics and Biomechanics (Hindawi/Wiley) 65.4% (574/878), Journal of Environmental and Public Health (Hindawi/Wiley) 63.9% (885/1,384), Computational and Mathematical Methods in Medicine (Hindawi/Wiley) 61.8% (2,143/3,467). Continuing the top 10: Scanning (Hindawi/Wiley) 55.1% (174/316), Security and Communication Networks (Hindawi/Wiley) 53.8% (1,897/3,528), Advances in Multimedia (Hindawi/Wiley) 52.6% (370/703), Journal of Control Science and Engineering (Hindawi/Wiley) 51.9% (98/189), Contrast Media & Molecular Imaging (Hindawi/Wiley) 49.0% (519/1,059). Nine of the top 10 are former Hindawi titles that became part of Wiley's mass paper-mill cleanup after Wiley acquired Hindawi in 2021. The non-Hindawi outlier at the very top is the SAGE-published International Journal of Electrical Engineering Education, which had a December 2021 mass retraction event of 122 papers (with 318 more put under expression of concern and the editor fired) and has continued retracting since. The median journal in the top-200 retraction-count list has a 4.1% per-paper retraction rate; the mean is 10.5%; the top entry is 65.9%, more than 16 times the median. STRUCTURAL READING. The conventional Retraction Watch narrative covers retraction batches at journal level as news events ('SAGE retracts 122 papers', 'Wiley retracts another 678 from JIFS'), but the cumulative-rate framing makes a different point: ten journals are now in a state where if you draw a random article from their 2020-2025 archive, you have a coin-flip-or-worse chance of pulling a retracted one. For library subscription decisions and citation reviews this is a categorical break — these journals function less as scholarly archives than as 'paper of record' venues whose reliability has collapsed. The standard 'high-quality vs low-quality journal' framing is too coarse; the right framing is 'archives whose 2020-2025 contents are reliable enough to cite without a retraction check' vs 'archives where every cite needs a retraction check'. CAVEATS. (1) OpenAlex's is_retracted flag is set by publishers and Crossref; some publishers are slow to mark retractions, so the true per-journal rate may be slightly higher than computed. (2) The 'total works' count for each journal is OpenAlex's count of works mapped to that source-id; for a journal that ceased publication and was migrated, the count may not perfectly match the publisher's own back-issue tally. (3) The top-200-by-retraction-count universe excludes journals with fewer than ~58 retractions (the 200th-rank cutoff) — a journal that published 50 papers and retracted all 50 would not appear in this analysis, though such cases are rare. (4) PLoS ONE has 1,531 retractions in the same window but at a 1.5% rate (1,531 / 102,453), illustrating why the per-paper rate is the meaningful normalization rather than the raw count.

When a scholarly journal article turns out to be wrong — fabricated data, plagiarism, an undisclosed paper-mill arrangement, gibberish generated by AI — the journal can issue a 'retraction', formally withdrawing the paper. Retraction Watch tracks every single retraction event and writes news stories about them. Most of the news stories say 'Journal X retracted Y papers'. What nobody routinely publishes is the cumulative per-journal RATE: of all the papers a journal has published in the last six years, what percentage have been retracted? I downloaded the retraction status of every paper in OpenAlex (a public scientific bibliographic database) for 2020-2025 and computed that rate for the top-200 journals by raw retraction count. The result is striking. Ten scholarly journals have now retracted at least 49% of everything they published since 2020 — meaning if you pick a random article that one of these journals printed in the last six years, it is a coin-flip-or-worse chance of being retracted. The worst is the International Journal of Electrical Engineering Education (published by SAGE), which has retracted 65.9% of its 2020-2025 papers (253 of 384). Nine of the other ten are former Hindawi-imprint journals that Wiley acquired in 2021 and then mass-retracted papers from after discovering large-scale paper-mill infiltration: Journal of Healthcare Engineering (65.5% retracted), Applied Bionics and Biomechanics (65.4%), Journal of Environmental and Public Health (63.9%), Computational and Mathematical Methods in Medicine (61.8%), Scanning, Security and Communication Networks, Advances in Multimedia, Journal of Control Science and Engineering, and Contrast Media & Molecular Imaging. None of this is news to insiders who follow the Hindawi/Wiley story — every individual retraction batch was reported by Retraction Watch when it happened — but the cumulative number is striking in a way the news drumbeat is not. Why this matters: when a university library is deciding which journals to keep paying for, or when a tenure committee is evaluating a candidate's publication list, or when a researcher is choosing where to submit a manuscript, the right question is 'what fraction of this journal's recent output has been retracted?', not 'has this journal had a retraction in the news lately?'. Ten journals are now in a category where the answer is 'half or more', and that is a category break that the standard 'troubled vs trustworthy' framing doesn't capture. By contrast, PLoS ONE has retracted 1,531 papers in the same window — more than any other single journal in absolute count — but at a 1.5% rate, because PLoS ONE publishes ~100,000 papers in the window. PLoS ONE is in a completely different category from the ten ≥49% journals, and the per-paper rate makes that obvious; the raw-count framing does not.

International Journal of Electrical Engineering Education (SAGE) 65.9% (253/384) · Journal of Healthcare Engineering (Hindawi/Wiley) 65.5% (2,171/3,316) · Applied Bionics and Biomechanics (Hindawi/Wiley) 65.4% (574/878) · Journal of Environmental and Public Health (Hindawi/Wiley) 63.9% (885/1,384) · Computational and Mathematical Methods in Medicine (Hindawi/Wiley) 61.8% (2,143/3,467) · Scanning (Hindawi/Wiley) 55.1% (174/316) · Security and Communication Networks (Hindawi/Wiley) 53.8% (1,897/3,528) · Advances in Multimedia (Hindawi/Wiley) 52.6% (370/703) · Journal of Control Science and Engineering (Hindawi/Wiley) 51.9% (98/189) · Contrast Media & Molecular Imaging (Hindawi/Wiley) 49.0% (519/1,059)

PLoS ONE 1,531 retractions / ~102,453 total works in 2020-2025 = 1.5% rate · context: PLoS ONE is the largest single source of retractions in absolute count but rates among the lowest in the top-200 because it publishes ~17K papers/year

74,463 retracted works in OpenAlex, 2020-2025 · 200 journals examined · 10 journals at ≥49% retraction rate · 5 journals at ≥60% retraction rate · median rate 4.1% per paper · mean rate 10.5% · 9 of top 10 are former Hindawi titles · 1 of top 10 is SAGE (IJEEE)