PaperlayerView exact revision on arXiv

Source-linked AI summary

Tracing high-profile attention to questionable research as a case for funder due diligence

Federica Silvi, Leslie D. McIntosh

arXiv:2608.30613v1cs.DLcs.SI

TL;DR

Paper mills have expanded the sale of authorship, raising questions about how questionable research affects the scientific record and funding. Using a known authorship-for-sale network, the study examines its influence on policy and related outputs and finds that linked research had informed these domains while authors had benefited from legitimate grants.

  • Problem

    Paper mills have advertised over 20,000 authorship positions for sale, raising questions about the consequences of this expansion for research integrity.

  • Method

    The study analyzes a known authorship-for-sale network and examines who appears in authors’ networks and how those networks behaved.

  • Results

    PNN-linked research had informed policy documents, patents, and clinical guidelines for years before exposure, while its authors had already benefited from legitimate public grants.

  • Takeaways & Limitations

    Funders occupy a distinct position in the chain of research-integrity responsibility examined by the paper.

  • Takeaways & Limitations

    The study analyzes one documented network and a small fraction of activity at this scale, rather than a representative sample of paper-mill activity.

Abstract

from arXiv · show

When questionable research has entered the scientific record, it stands the same chance as other literature of shaping national and international policy, clinical guidance, and research and development activities. This study uses a known authorship-for-sale network identified in August 2022 to examine whether questionable research influences policy, patents, and clinical guidelines, and whether its authors continue to secure funding and publish after exposure. Across nearly 2,000 publications in our dataset, 57 were cited in policy documents, 12 in clinical guidelines, and 480 in patents. Nearly a quarter of these publications are linked to one or more grants: funding that could otherwise have supported rigorous, ethical science. Among the 278 authors we traced, publishing and funding continued well after the PNN's public exposure in 2022: over 90% continued publishing, and 23% were linked to a grant. As our findings show, the participants of paper mills and authorship-for-sale networks can still inform policy and clinical practice and support their authors' career progression even after the questionable practices behind them are exposed. We present a case for funders to consider authorship and network structure as part of a holistic due diligence process, alongside metrics such as citation counts and attention data.

Introduction

The paper frames questionable research as a growing integrity problem whose visibility can influence policy, practice, funding, and careers. It argues that evaluating authors and their networks requires more than conventional attention and citation metrics.

  • Scale of questionable research: Paper-mill activity has grown sharply since 2016, when the first paper mill-linked retraction was recorded.
  • Scale of questionable research: Over 20,000 authorship positions were advertised for sale by paper mills across multiple countries.
  • Why it matters: Questionable research can erode public trust, enable disinformation, and divert scarce funding from rigorous, ethical science.
  • Limits of current evaluation: Career advancement and funding decisions rely largely on visibility and output measures that can conflict with making reliable science available.
  • Limits of current evaluation: Citation metrics are narrow, can misapply journal averages to individual articles, and can be gamed through citation cartels or coercive citation.
  • Signals of influence: Policy, patent, and clinical-guideline mentions indicate that research has entered evidence bases or commercial and technological processes.
  • Study focus: The study examines a known authorship-for-sale network to test questionable research attention, funding, and post-exposure publication patterns.

2. RQ2: Are its authors receiving legitimate grant funding?

The second research question asks whether authors linked to the exposed network continued to publish and receive funding after its questionable practices became public.

  • Post-exposure activity: RQ3 asks whether authors continue publishing and receiving funding after their network or questionable practices have been exposed.
  • Post-exposure activity: The question focuses on activity after exposure rather than only on the network's earlier participation in questionable research.
  • Post-exposure activity: The wording treats publication and funding as separate post-exposure outcomes to be examined.

Methods

The study is a forensic scientometric case study of the Pharmakon Neuroscience Network, combining publication, attention, grant, and retraction data to examine three research questions. It uses dated data snapshots, manual verification for policy and guideline mentions, and linked records for grants and authors.

  • Study design: The study examines the Pharmakon Neuroscience Network, an authorship-for-sale network publicly identified in August 2022 and previously characterized as spanning over 300 authors across 40 countries.
  • Data sources: Data sources include Dimensions, Altmetric, Policy Commons, Retraction Watch, and PubMed's E-utilities API.
  • Dataset construction: RQ1 combines publications linked to 16 PNN authors or naming the network with Altmetric-tracked policy and guideline mentions and patent counts.
  • Dataset construction: RQ2 examines grant records linked to the combined publication population, including grants receiving specific scrutiny in a 2025 investigation.
  • Dataset construction: RQ3 uses 278 authors derived from outputs linked to the PNN at exposure to test whether patterns from a smaller sample generalize across the wider network.
  • Temporal design: The 24 August 2022 public exposure date defines pre-exposure and post-exposure periods for publications, mentions, grants, and retractions.
  • Mention curation: Policy and guideline mentions were manually checked and deduplicated, whereas high-volume patent mentions were not subject to the same manual verification.

Results

Across 1,975 publications linked to the network, questionable research received attention in policy documents, clinical guidelines, and patents, while authors continued publishing and securing grants after exposure.

  • Real-world attention: 57 publications received 76 policy mentions, 12 received 22 clinical guideline mentions, and 480 were mentioned 2,623 times in patents.The mentions came from 27 policy organisations and 18 clinical-guideline organisations.
  • Real-world attention: Policy citations were concentrated among intergovernmental and government organisations, with the FAO the largest single source.The passage identifies the FAO as the largest source, followed by the Publications Office of the European Union, OECD, and CADTH.
  • Real-world attention: Guideline-cited research reached practice recommendations across 13 distinct clinical areas.Citations spanned internationally recognised medical societies and national and specialty societies.
  • Attention in practice: Case studies show network-linked publications entering high-stakes policy and clinical contexts, including EU regulation and recommendations from major medical associations.One paper cited in an EU technical report could form the legal basis for regulation on waste-derived fertilizers; another clinical guideline was cited by major medical associations as recently as 2024.
  • Grant funding: 450 of 1,975 publications, or 22.8%, had at least one grant linked in Dimensions.Among the 450 grant-linked publications, 129 were directly linked to the 2025 investigation, representing 28.7% of grant-funded publications in the dataset.
  • Publishing and funding after exposure: Of 278 analysed authors, 70, or 25.1%, received or were linked to at least one grant after exposure, and none stopped publishing.The authors were credited on 12,508 papers published after the network's exposure; 22 authors published nothing after exposure, while one author published 567 papers.
  • Publishing and funding after exposure: Public exposure did not meaningfully disrupt correction: only one of 66 publications cited in policy documents or clinical guidelines had been retracted.The same pattern held for the 98 Altmetric-tracked Pharmakon papers, of which only one was retracted after exposure.

Discussion

The discussion argues that fragmented integrity checks allow questionable research to gain attention, funding, and legitimacy. It proposes that funders examine authorship and network structure alongside conventional indicators, while noting important study and data limitations.

  • Distributed integrity risk: Paper-mill activity operates through distributed enterprises and compromised peer-review networks, so individual journals may see only part of the pattern.The discussion describes activity across competing publishers and notes that peer review can be compromised by the same networks it is intended to police.
  • Downstream reach: Questionable research can reach policy, clinical guidelines, and patents, creating systemic risks when its underlying research or authorship is insufficiently scrutinised.Evaluation systems treat societal impact as evidence of research value, while misconduct-detection systems have not kept pace with downstream uptake.
  • Verification constraints: Grant-stage due diligence is constrained when grant and ethics identifiers are not consistently recorded in structured, auditable form.Without consistent identifiers, funders and institutions cannot verify which acknowledged outputs warrant closer scrutiny.
  • Funder exposure: Funding creates both allocation and reputational risks when grants support unreliable work or researchers implicated in producing it.Public and philanthropic funding is finite, and funder acknowledgements can tie an organisation's name to work that later proves unreliable or fraudulent.
  • Study scope: The PNN case is a single documented network and an in-depth analysis of only a small fraction of its potential output, not a representative sample of paper-mill activity.The study also notes that only one publisher was directly notified about one author's problematic practices and that Altmetric coverage has limits.
  • Persistence after exposure: PNN-linked research informed policy, patents, and clinical guidelines before exposure, while funding and attention continued after public exposure.The discussion states that neither pattern was disrupted by exposure through retractions and misconduct findings.
  • Limits of conventional indicators: Citation counts and attention metrics cannot establish whether attention is warranted, because they can be slow, misapplied, manipulated, or disconnected from authorship legitimacy.The proposed shift is from measuring whether attention was received to examining what stands behind it.
  • Proposed due diligence: Funders are positioned to add scrutiny of authorship and network behaviour to existing due diligence rather than relying on grant history, citations, or policy uptake alone.The discussion supports integrity commitments, grant-contract clauses, prompt breach reporting, and collaboration between funders, institutions, and publishers.

Declaration of generative AI use

The authors used Claude for coding, data exploration, and language editing, while independently validating the resulting code and verifying AI-assisted content.

  • Claude was used for coding and data exploration across the project.
  • The resulting code was independently validated by Sam Scribner.
  • Claude also assisted with language editing of the manuscript text.
  • The authors verified the originality and accuracy of AI-assisted content and reviewed Claude’s terms of use.
  • The authors accepted responsibility for the manuscript’s integrity, including reference accuracy.

Data availability statement

The study’s data, code, and methodological materials are available through specified repositories or upon reasonable request, with author identifiers de-identified but potentially re-identifiable.

  • The curated Altmetric mention-level dataset is available from the corresponding author upon reasonable request.
  • Dimensions publication and grant metadata, retraction-status determinations, and supplementary methodological detail are available at Figshare.
  • Author names were replaced with anonymised labels, but retained Dimensions IDs can resolve to public researcher profiles.
  • Python code for constructing datasets and producing reported results is available at Figshare.
  • The Retraction Watch Database snapshot analysed in the study is a third-party dataset available at retractiondatabase.org.

Data usage statement

The study used data from five named sources and states that each source was accessed and used according to its applicable terms and policies.

  • The underlying data came from Dimensions, Altmetric, Policy Commons, the Retraction Watch Database, and PubMed.
  • The authors state that each source was accessed and used in accordance with its platform terms and data usage policies.

Funding

The authors report that the research received no funding.

  • No funding was obtained for the research reported in the article.

Declaration of interest statement

The authors disclose employment ties to Digital Science, whose commercial Altmetric and Dimensions platforms underpin the study, and an unpaid advisory role with VeriMe.

  • Digital Science: FS and LDM are employees of Digital Science, which owns the Altmetric and Dimensions platforms used in the study.The platforms were accessed through employee licenses rather than a research grant or external subscription.
  • Digital Science: Digital Science had no role in the study’s design, data interpretation, or publication decision.The research was conducted independently of commercial interests.
  • VeriMe: LDM serves as VeriMe’s unpaid, volunteer Pre-Launch Advisor, providing subject-matter expertise, context, advice, and network access.The advisory role concerns VeriMe’s products and services.

Figure captions

The figure captions describe datasets tracing policy attention, grant linkages, and post-exposure publication and funding activity among PNN-linked authors.

  • Dataset overview: The study traces attention and funding activity using two datasets summarized in Figure 1.The figure provides an overview of the datasets underlying the analyses.
  • Investigation-authors dataset: The investigation-authors dataset contains 16 authors and 1,975 publications with Altmetric-tracked attention.Grant-linkage identification narrows this publication set through Dimensions records and prior investigation links.
  • Post-exposure activity: The Pharmakon papers dataset covers 278 authors and 12,508 post-exposure publications recorded in Dimensions.Post-exposure publication counts use full researcher profiles, while both author populations are identified through Altmetric-tracked outputs.
  • Policy attention: Policy-attention outputs are organized by source category across six categories and 27 organisations.The organisations accounted for 76 total policy mentions, with categories assigned using Policy Commons taxonomy and official websites.
  • Policy attention: A bar chart ranks the 27 citing organisations by policy mentions in descending order.The chart reports 76 total mentions and shows the source-category breakdown in Figure 1.
  • Grant linkage: Grant-linkage charts track publications from the full 1,975-publication set through successive identification stages and separate attention types.The stages include any Dimensions-recorded grant linkage and grants connected to a prior 2025 investigation.
Loading 2608.30613v1…