PaperlayerView exact revision on arXiv

Source-linked AI summary

The strain on scientific publishing

Mark A. Hanson, Pablo Gómez Barreiro, Paolo Crosetto, Dan Brockington

arXiv:2309.15884v2cs.DL

TL;DR

Scientific publishing is strained by exponential article growth and limited capacity to vet the expanding literature. The paper combines five publisher-practice metrics from indexed data, web scrapes, publisher requests, and public materials, finding that strain is not strictly tied to one business model and that MDPI was an outlier across metrics. The authors conclude that greater transparency and coordinated action are needed, while noting that rejection-rate trends—not raw rates—are the more defensible comparison.

  • Problem

    Exponential growth in peer-reviewed articles has overwhelmed scientists and outpaced the capacity to vet new work, creating strain on scientific publishing.

  • Method

    The paper compares five metrics of publisher practice using indexed Scopus and Web of Science data, web scraping, publisher requests, and publicly available materials.

  • Results

    A comparative analysis finds strain is not strictly tied to one publisher business model, while MDPI was an outlier across every metric, often by wide margins.

  • Takeaways & Limitations

    The metrics identify areas needing transparency and support actions by publishers, researchers, and funders addressing strain in scientific publishing.

  • Takeaways & Limitations

    Because rejection-rate methodologies could not be standardized, raw rejection-rate comparisons carry caveats, whereas within-group changes are assumed more robust.

Abstract

from arXiv · show

Scientists are increasingly overwhelmed by the volume of articles being published. Total articles indexed in Scopus and Web of Science have grown exponentially in recent years; in 2022 the article total was approximately ~47% higher than in 2016, which has outpaced the limited growth - if any - in the number of practising scientists. Thus, publication workload per scientist (writing, reviewing, editing) has increased dramatically. We define this problem as the strain on scientific publishing. To analyse this strain, we present five data-driven metrics showing publisher growth, processing times, and citation behaviours. We draw these data from web scrapes, requests for data from publishers, and material that is freely available through publisher websites. Our findings are based on millions of papers produced by leading academic publishers. We find specific groups have disproportionately grown in their articles published per year, contributing to this strain. Some publishers enabled this growth by adopting a strategy of hosting special issues, which publish articles with reduced turnaround times. Given pressures on researchers to publish or perish to be competitive for funding applications, this strain was likely amplified by these offers to publish more articles. We also observed widespread year-over-year inflation of journal impact factors coinciding with this strain, which risks confusing quality signals. Such exponential growth cannot be sustained. The metrics we define here should enable this evolving conversation to reach actionable solutions to address the strain on scientific publishing.

2. Introduction

Scientific publishing faces strain because article growth has outpaced the capacity to vet new work, while publisher and researcher incentives encourage continued output. The paper introduces five metrics to analyse this strain and finds it is not strictly tied to one business model.

  • Exponential article growth has outpaced researcher training, leaving editors struggling to recruit reviewers and scientists overwhelmed by new literature.
  • Growth can broaden access and reduce systemic bias, but becomes strain when it compromises rigorous information vetting.
  • The study analyses five metrics: article and special-issue growth, turnaround times, rejection rates, and citation-based “impact inflation.”
  • Publisher business models include subscription revenue and gold open access article-processing charges, with either for-profit or not-for-profit ownership.
  • A comparative multi-metric analysis finds strain is not strictly tied to one publisher business model, although some behaviours associate with specific gold open access publishers.
  • Publishers, funders, and researchers jointly shape the system through incentives tied to output, quality signals, and publication prestige.

Data collection

The study combines indexed publication data with web-scraped, publisher-provided, and publicly available information to compare publisher practices. Web-scraping constraints limited several analyses to selected publishers and articles.

  • Web-scraped analyses of special-issue shares, turnaround times, and rejection rates covered a selected mix of major established and gold open access publishers.
  • Publisher labels were aggregated by corporate branding while subsidiary brands remained separate when their behaviour differed meaningfully from parent companies.
  • Historical article, citation, SJR, and reference data came from Scimago, while Impact Factor data covered 16,174 journals but only subsets per publisher.
  • Rejection-rate data came from publisher reports, direct requests, and web scraping, whereas turnaround times were calculated from submission to formal acceptance using publisher corpora or article pages.
  • Special-issue classification required manual curation, and scraping scripts were tailored to publisher website and metadata formats before standardizing comparisons.

Global researcher statistics

The study combines OECD PhD counts with complementary China and India data and UNESCO researcher-density measures to characterize the global researcher population.

  • OECD data supplied PhD graduate numbers for 37 countries through 2020, supplemented by China and India data available through 2019.
  • UNESCO researchers-per-million data were used and projected to 2022 with a linear regression model.

Rejection rates

Rejection rates were calculated from accepted, rejected, and total submissions, but publisher-specific definitions make raw comparisons difficult. The authors therefore assume internal methodological consistency when comparing within-group trends.

  • Rejection rates were defined from accepted, rejected, and total submissions according to the data available for each publisher.
  • Publisher practices, including revision handling and acceptance-rate reporting, introduce cryptic factors that complicate direct comparisons.
  • Frontiers used 1 - (accepted articles / total submissions), while MDPI totals combined accepted and rejected articles and Hindawi reported publisher-defined acceptance rates.
  • Because methodologies could not be standardized, the analysis assumes publishers maintained internally consistent methods across years.
  • Comparing changes in rejection rates within groups is treated as more robust than comparing raw rejection-rate levels.

Impact Inflation

The paper distinguishes Impact Factor from SJR and introduces “Impact Inflation” as a ratio applicable to both metrics.

  • Impact Factor is the mean total citations per article published within the previous two years.
  • SJR uses a citation-network approach that weights same-field journal citations more heavily and limits self-citation or citation-cartel contributions.

Data Availability

The authors could not publicly release their web-scraped data and scripts because of copyright concerns, but shared supplementary materials and developed a web app.

  • Copyright concerns prevented public release of the web-scraped data and scripts.
  • The supplementary materials include README files explaining how readers can assemble the data.
  • The authors developed an R Shiny web app hosted on the project website.

A few publishers disproportionately contribute to total article growth

Article output increased sharply from 2016 to 2022, with a few publishers contributing disproportionately through different growth patterns.

  • ~897k more indexed articles were published annually in 2022 (~2.82m) than in 2016 (~1.92m), representing ~5.6% year-on-year exponential growth.
  • MDPI (~27%), Elsevier (~16%), Frontiers (~11%), Springer (~9.5%), and Wiley (~7.0%) contributed over 70% of the increase since 2016.
  • MDPI and Frontiers became disproportionately high producers of articles per journal, while Elsevier’s output per journal increased only slightly.
  • Elsevier and Springer increased total output by distributing articles across more journals, whereas MDPI and Frontiers increased publications per journal.

Growth in articles published through “special issues”

Special issues became a major publishing route for some publishers between 2016 and 2022, especially Hindawi, Frontiers, and MDPI. Their prevalence was publisher-dependent rather than an inherent feature of open access.

  • Special issues are invited by journals or editors and commonly delegate responsibilities to guest editors.
  • Between 2016 and 2022, the proportion of special-issue articles grew drastically for Hindawi, Frontiers, and MDPI.
  • Comparable changes were not observed for gold open access publishers BMC and PLOS, while Nature, Springer, and Wiley published small proportions of special issues.
  • The strain associated with special issues was publisher-dependent, not a direct consequence of open access publishing or article processing charges alone.
  • Frontiers, Hindawi, and especially MDPI published a majority of their sampled articles through special issues, including recent increases.

Decreasing mean, increasing homogeneity of turnaround times

Turnaround times varied substantially across publishers, while publishers using more special issues showed faster and increasingly homogeneous acceptance timelines. These patterns suggest special issue articles are typically processed more rapidly and consistently than normal articles.

  • ~37 days was MDPI’s 2022 average turnaround time, compared with 72 days for Frontiers and 83 days for Hindawi.MDPI had maintained this level since ~2018, while Frontiers and Hindawi also reduced mean turnaround times between 2020 and 2022.
  • >130 days characterized turnaround times for other publishers in the dataset, which increased slightly between 2016 and 2022.
  • Turnaround-time variance declined for Hindawi, Frontiers, and especially MDPI, making acceptance timelines increasingly homogeneous.This pattern occurred despite expected variation by article quality and research field.
  • Special issue articles were processed more rapidly than normal articles for most publishers, independently of the proportions of special issue and normal articles.Mean-turnaround-time decreases aligned with inflection points in special issue growth for Hindawi, Frontiers, and MDPI.
  • Special issue articles were typically accepted more rapidly and within more homogeneous timeframes than normal articles.The study presents this as a previously undescribed pattern.

Journal rejection rates and trends are publisher-specific

Rejection-rate patterns were largely publisher-specific, with no general relationship to publisher growth, journal size, or citations per document. The analysis also examined citation-based impact inflation, finding especially high values for MDPI and Hindawi and linking MDPI’s anomaly to self-citation.

  • No clear trend connected rejection-rate evolution with publisher growth, journal size, or citations per document.Among younger journals, 2022 rejection rates differed by publisher but not journal size; citations were not a strong predictor.
  • The publisher itself best predicted rejection rates, and rejection-rate patterns were largely publisher-specific over time and among similarly aged journals.
  • Within the limited publisher subset with data, rejection rates declined as the share of special issue articles increased.
  • Impact inflation is the ratio of Cites-per-doc to SJR, contrasting a citation-count proxy with a network-adjusted prestige metric.SJR limits the prestige earned from a single source, so disproportionate self-citation can raise Cites-per-doc without raising SJR.
  • Impact inflation was significantly higher for MDPI and Hindawi than for all other publishers in 2022, while Frontiers remained comparable to other publishers.
  • MDPI journals nearly universally increased within-journal self-citation rates during the study period, explaining its anomalous impact inflation.MDPI had significantly different self-citation rates from most compared publishers, whereas Hindawi’s rates were more comparable despite high impact inflation.

6. Discussion

The strain on scientific publishing reflects rapid article growth and interacting incentives among researchers, publishers, and funders. Comparative metrics reveal distinct publisher strategies, widespread impact inflation, and the need for targeted transparency and governance responses.

  • Scale and drivers: Nearly one million indexed articles were added annually over the last six years, increasing writing, reviewing, and publication-fee burdens.The analysis focuses strictly on indexed articles, not total articles.
  • Scale and drivers: Researchers’ publish-or-perish incentives and publishers’ commercial and reputational interests jointly sustain article-output growth.Funders shape these incentives by rewarding researchers who demonstrate productivity.
  • Publisher strategies: Article growth is widespread, but most indexed-article growth comes from a few publishing houses using two broad strategies.Older houses combine mild growth in journals and articles per journal, while some gold open access publishers rely heavily on special issues.
  • Publisher strategies: Special issues are systematically handled differently from normal submissions, with lower rejection rates and shorter, more homogeneous turnaround times.The paper presents impact inflation (IF/SJR) as a publicly assessable litmus test for citation gaming.
  • Publisher strategies: MDPI was an outlier across every measured metric, including +1080% indexed-article growth, 88% special-issue articles, and 37-day turnaround times.It also had −8 percentage-point rejection-rate change, 5.4 impact inflation, and 9.5% within-journal mean self-citation.
  • Scope and responses: The study’s scope excludes how editorial independence is maintained, although that balance is identified as important for scientific integrity and authority.Rejection rates and turnaround-time statistics appear largely independent of the other assayed metrics.
  • Scope and responses: Transparency is limited because basic publishing data are difficult to collect, especially without standardized rejection-rate reporting.The authors recommend transparency about special issues, turnaround times, rejection rates, revenue, and operating costs.
  • Scope and responses: Funders could mitigate publish-or-perish pressure through policies such as narrative CVs, while researchers and ethics bodies could discourage special-issue overuse.The authors argue that regulation should target specific publishers and behaviors rather than business models.

Supplementary figures and tables

The supplementary analyses document publisher growth, special-issue use, turnaround-time patterns, and citation-based impact inflation, while highlighting important dataset and metric caveats.

  • Publisher growth: Between 2013-2022, Elsevier, Springer, Taylor & Francis, MDPI, and Nature noticeably expanded their indexed journal portfolios.The dataset includes journals indexed in both Scopus and Web of Science and treats subsidiaries with separate licensed names independently.
  • Publisher growth: Journals publishing over one paper per day increased sharply while those publishing under one paper per week declined, indicating greater concentration into megajournals.The supplementary analysis also reports a slight decline in the number of journals publishing one paper per day or less.
  • Processing times: Across publishers, special-issue articles generally had lower turnaround times than regular articles, with Springer the only reported exception.The figure notes that special-issue organisation varies across journals and publishers, which may explain differences.
  • Caveats: The analyses are constrained by nested publisher ownership, selective publisher coverage, and rejection-rate measures drawn from nonstandardised, voluntarily reported data.Rejection rates also reflect editorial capacity, journal scope, and submissions rather than quality or rigour alone.
  • Citation behaviour: MDPI’s lower SJR relative to its Impact Factor reflects citation patterns involving elevated within-journal self-citation and extensive citation within the MDPI network.The three flagship journals reported within-journal self-citation rates of 9.4%, 11.8%, and 15.3% in 2022.
  • Citation behaviour: Similar article-per-journal growth did not produce the same impact-inflation pattern for Frontiers, which had more diverse citations and lower impact-inflation scores than MDPI.The comparison indicates that article growth alone is not an intrinsic explanation for the disparity.
  • Metric interpretation: Year-over-year Impact Factor increases can reflect inflation from expanding article and citation totals, so absolute values change meaning over time even when relative ranks remain comparable.The authors suggest reporting relative rank or percentile rather than absolute Impact Factor values; impact inflation is likewise proposed for comparison through relative ranks such as quartiles.

Themes in Publisher Motivation

Publisher reports frame growth around author service, visibility, quality assurance, and expanding publication capacity. Their stated motivations span business growth, open-access expansion, research dissemination, and researcher experience.

  • Quality assurance and dissemination: Elsevier presents its role as validating, improving, and disseminating research through more than 2,900 journals and over 1.5 million reviewers.It reported publishing more than 630,000 articles in 2023 from almost 3 million submissions.
  • Researcher experience: Publishers describe growth through expanded journals, author relationships, open-access services, and systems intended to support quality control.Examples include Scopus AI for navigating growing workloads and AIRA-assisted desk review before peer review.
  • Researcher experience: Frontiers emphasizes researcher experience, reporting that 92% of authors and 88% of reviewers rated their experience excellent or good.The figures came from more than 150,000 pieces of feedback collected during 2022.
  • Publisher growth and author service: MDPI attributes publication growth to author-focused processes, service quality, and attention to authors’ needs.Its 2022 report also highlights surpassing one million total articles, supported by reviewers, editorial board members, and staff.
  • Publisher growth and author service: MDPI reported more than 50% average annual growth in open-access articles over five years, while remaining the largest open-access publisher by article count.
  • Business growth and open access: Springer Nature linked revenue growth to Nature Portfolio journals, open-access expansion, new titles, and strong submission levels.Its fully open-access portfolio comprised around 600 journals.
Loading 2309.15884v2…