henhao
Sign in
Scientific Communication & EducationbioRxiv
Preprint

Updated science-wide author databases of standardized citation indicators including retraction data

John P. A. IoannidisiD, Angelo Maria PezzulloiD, Antonio CristianoiD, Stefania BocciaiD, Jeroen BaasiD

Abstract

Citation metrics are widely used in research appraisal, but they provide incomplete views of scientists impact and research track record. Other indicators of research practices should be linked to citation data. We have updated a Scopus-based database of highly-cited scientists (top-2% in each scientific subfield according to a composite citation indicator) to incorporate retraction data. Using data from the Retraction Watch database (RWDB), retraction records were linked to Scopus citation data. Of 55,237 items in RWDB as of August 15, 2024, we excluded non-retractions, retractions clearly not due to any author error, retractions where the paper had been republished, and items not linkable to Scopus records. Eventually 39,468 eligible retractions were linked to Scopus. Among 217,097 top-cited scientists in career-long impact and 223,152 in single recent year (2023) impact, 7,083 (3.3%) and 8,747 (4.0%), respectively, had at least one retraction. Scientists with retracted publications had younger publication age, higher self-citation rates, and larger publication volume than those without any retracted publications. Retractions were more common in the life sciences and rare or nonexistent in several other disciplines. In several developing countries, very high proportions of top-cited scientists had retractions (highest in Senegal (66.7%), Ecuador (28.6%) and Pakistan (27.8%) in career-long citation impact lists). Variability in retraction rates across fields and countries suggests differences in research practices, scrutiny, and ease of retraction. Addition of retraction data enhances the granularity of top-cited scientists profiles, aiding in responsible research evaluation. However, caution is needed when interpreting retractions, as they do not always signify misconduct; further analysis on a case-by-case basis is essential. The database should hopefully provide a resource for meta-research and deeper insights into scientific practices.

PDFDOIData
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
30
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 1 (17 Sept 2024) · CC BY 4.0

Linking 39,468 eligible Retraction Watch retractions to Scopus shows 3.3% of 217,097 career-long and 4.0% of 223,152 single-year (2023) top-2% cited scientists have at least one retraction.

Citation metrics alone give a partial view of a scientist's record, so this annual update of the authors' Scopus-based top-2% cited scientist databases attaches per-scientist counts of retracted papers and of citations to and from retracted papers. Clarivate's Web of Science highly-cited list now omits anyone with a retraction; here those scientists stay listed and interpretation is left to case-by-case assessment.

Key results

  1. ShownAbout 3 to 4% of top-2% cited scientists have at least one eligible retraction, and the number of retractions per author follows a power-law distribution.

    7,083 of 217,097 (3.3%) career-long and 8,747 of 223,152 (4.0%) 2023 single-year top-cited scientists had at least one retraction; 1,710 (0.8%) and 2,150 (1.0%) had two or more. Based on 39,468 RWDB retractions linked to Scopus (38,364 by DOI, 1,104 by title and year); Figure 1 shows the power-law distribution. · Abstract; METHODS AND RESULTS; Figure legends

  2. ShownScientists with retractions publish many more papers than those without. Their other differences (younger publication age, more self-citation, more co-authorship by h/hm ratio, slightly better rank) are small or modest. Retractions are most frequent at the very top of the ranking.

    Mann-Whitney U p<0.001 for all indicators in the career-long dataset and for all except publication age and absolute subfield rank in the 2023 dataset (Table 1). Among the 1,000 scientists with the highest composite indicator, 13.8% (career-long) and 11.1% (single-year) had at least one retraction. · METHODS AND RESULTS; Abstract

  3. ShownRetractions among top-cited scientists are concentrated in the life sciences; several fields have few or none.

    Across 20 Science-Metrix fields the share ranges from 0% to 5.6%, highest in Clinical Medicine and Biomedical Research (4.9 to 5.6%). Among 174 subfields (2023 single-year list), Complementary & Alternative Medicine 10.5%, Oncology & Carcinogenesis 9.9% and Pharmacology & Pharmacy 9.4% lead. · METHODS AND RESULTS; Abstract

  4. ShownThe share of top-cited scientists with retractions varies widely by country. The highest rates are in several developing countries that have few top-cited scientists.

    Among the 20 countries hosting most top-cited authors: India 9.2% career-long / 8.6% single-year, China 8.2% / 6.7%, Taiwan 5.2% / 5.7%, versus Israel 1.7% / 2.0%, Belgium 2.1% / 2.1%, Finland 2.2% / 2.2%. Smaller countries: Senegal 66.7%, Ecuador 28.6%, Pakistan 27.8% (career-long); Kyrgyzstan 50%, Senegal 41.7% (single-year). · METHODS AND RESULTS; Abstract

  5. ShownCitations to and from retracted papers are now recorded for each scientist. They are large for a few individuals but make up a small fraction of nearly all top-cited scientists' total citations.

    Among scientists with a retraction, citations to their retracted papers range 0 to 7,491 (median 25, IQR 6 to 80) career-long and 0 to 832 (median 1, IQR 0 to 4) in 2023; 114 exceed 1,000, and for 230 (0.1%) and 260 (0.1%) they are over 5% of all citations. Only 5 scientists received over 1,000 citations from retracted papers; for 14 and 7 these exceed 5%. · METHODS AND RESULTS; DISCUSSION

Methods, data and reproducibility

Retraction Watch entries downloaded from Crossref were filtered to retractions not solely due to publisher error, retract-and-replace or obsolescence. These were linked to Scopus by DOI, or by title plus year (+/- 1 year). Counts were attached to each scientist in the top-2% composite-citation-indicator lists (Science-Metrix 20 fields, 174 subfields) and summarized descriptively with Mann-Whitney U tests.

  • Retraction Watch database entries retrieved via Crossref on August 15, 2024 (50,457 retractions; 47,964 after exclusions) · n = 55,237 entries; 39,468 eligible retractions linked to Scopus
  • Scopus-based top-2% cited scientists by career-long impact (includes 2,789 in the overall top 100,000 but not the top 2% of their subfield) · n = 217,097 scientists
  • Scopus-based top-2% cited scientists by single recent year (2023) impact (includes 6,325 in the overall top 100,000 but not the top 2% of their subfield) · n = 223,152 scientists
Methods & tools
Scopus, Retraction Watch database (RWDB), Crossref, Science-Metrix classification, composite citation indicator, Mann-Whitney U test
Data & accessions
https://api.labs.crossref.org/data/retractionwatchhttps://elsevier.digitalcommonsdata.com/datasets/kmyvjk3xmd/2
Reproducibility gaps
No code is given for the RWDB-to-Scopus linkage or for the indicator calculation. The composite indicator method is described only by reference to earlier iterations, and the text does not clearly say where the updated release can be downloaded.
Limitations (6)
  • The roughly 4% prevalence is conservative: it misses retractions in venues that Scopus does not cover and retractions that could not be linked. · stated by the authors
  • Excluding no-fault retractions does not ensure that every included retraction reflects author error or misconduct, and not every co-author may be responsible. Any use at the level of individual scientists needs case-by-case review of each retraction. · stated by the authors
  • Merging the data may introduce errors. Scopus author IDs and affiliations contain some mistakes, and because names can be similar or duplicated, wrongly attributing a retraction to a scientist cannot be fully ruled out. · stated by the authors
  • The databases make no judgement on the ethics or fairness of any retraction, and some retractions may still be contested or in legal proceedings. · stated by the authors
  • The analyses are descriptive and cross-sectional. Explanations for field and country differences (scrutiny, incentives, ease of detection) are not tested, and the most extreme country rates (e.g. Senegal 66.7%) come from countries with few top-cited authors, whose denominators are not given in the main text. · noted in this summary
  • Some reported numbers do not agree with each other: the text gives 39,950 retracted papers but 39,468 linked retractions, and it reports a median of 0 with an IQR of 2 to 5 for citations received from retracted papers. · noted in this summary

Who should read:Read this if you use or build bibliometric indicators for research evaluation, or want a ready-linked dataset of retractions and citation metrics for meta-research on research integrity across fields and countries.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISCandal-Pedr '20Toma '21Boccia '26Ioannidis '26Bhatt '21Kacem '20
79% semantic matchScientific Communication & Education
C. Candal-Pedreira, A. Ruano-Ravina, E. Fernandez+3 more·bioRxiv·Aug 11, 2020