‘Predatory’ open accessa longitudinal study of article volumes and market characteristics

For most of the twentieth century, the prestige of a scientific journal was inseparable from scarcity. Print space was finite, and editorial boards were selective. If your paper made it through peer review and onto the page, that physicality itself carried a signal: this was vetted and this mattered. Then the internet dissolved that constraint overnight, and in the gap it left, something unexpected grew. Not just open access science — something else entirely, wearing open access like a costume. That's the phenomenon Cenyu Shen and Bo-Christer Björk set out to measure in a 2015 longitudinal study — the first serious attempt to put hard numbers on predatory publishing. To understand what they found, you need the basic landscape. Legitimate open access publishing works through a few distinct routes. Gold open access means articles are freely available the moment they're published, often funded by article processing charges, or APCs, paid by the authors or their institutions. Green open access means authors deposit manuscript versions in repositories, bypassing journal paywalls. Hybrid journals sit in between, keeping a subscription model while allowing individual articles to go open for a fee. The APC-funded gold route grew fast, and megajournals like PLOS ONE — publishing around thirty thousand articles per year — showed the model could work at scale by judging papers on scientific rigor rather than perceived novelty. But the APC model has a structural vulnerability. If you collect a processing charge regardless of quality, you get paid whether or not the peer review was real. Jeffrey Beall, a librarian at the University of Colorado, noticed publishers exploiting exactly this gap — collecting fees, publishing fast, doing little to no actual review, and spamming academics with solicitation emails. He coined the term "predatory publishers" and built a public index of journals and publishers that met his criteria. That list became the starting point for Shen and Björk's study. The methodological challenge was real. You can't build a new global registry of predatory journals from scratch with limited resources, and the phenomenon actively resists enumeration — many journals exist as placeholder websites with no content at all. Shen and Björk downloaded Beall's two lists in September 2014 and identified nine hundred sixty-six publishers operating eleven thousand eight hundred seventy-three journals. Manually auditing that many journals was impossible, so they designed a stratified sampling approach, dividing publishers into four groups by portfolio size: those with one hundred or more journals, those with ten to ninety-nine, those with two to nine, and single-journal publishers. From that frame, they randomly selected two hundred ninety publishers and then sampled journals within each group, arriving at a working sample of six hundred thirteen journals. For each one, they manually collected subject field, country of publisher, article volumes from 2010 through 2014, and APC. For a subset, they also sampled individual articles to track author affiliations and publication timelines. The results are explicitly described as rough estimates — but transparent, reproducible, and far more systematic than anything that had come before. What those estimates revealed is a market in full sprint. In 2010, predatory journals published an estimated fifty-three thousand articles. By 2014, that number had climbed to roughly four hundred twenty thousand, across about eight thousand active journals. That's nearly an eightfold increase in four years. The number of active journals itself grew from around one thousand eight hundred in 2010 to eight thousand by 2014. And the shape of the market shifted as it grew. Early on, the largest publishers — those with portfolios of one hundred or more journals — dominated total article counts and remained the biggest players through 2011, but by 2012 the mid-sized tier, publishers with ten to ninety-nine journals, had taken over market leadership and held it. The mega-publishers didn't shrink; the middle just grew faster around them. One structural feature stands out: the prevalence of empty journals. Forty-six percent of journals from the largest publisher portfolios had never published a single article. The rate dropped to twenty-three percent in the mid-sized tier, eighteen percent in the small tier, and only two percent among single-journal publishers. The big players were running what amounted to shelf companies — journals that existed online as placeholders, perhaps to establish credibility before activating, or perhaps simply as speculative infrastructure. Average articles per active journal rose from about thirty in 2010 to fifty-three by 2012, before leveling off, indicating that as the market matured, the journals that did publish were publishing more. The disciplinary breakdown of those articles is striking. In 2014, general-scope journals accounted for an estimated one hundred sixty-two thousand articles, engineering journals for about ninety-seven thousand, and biomedicine for around seventy thousand. That concentration in engineering and medicine — fields where publication pressure and institutional evaluation criteria are intense — points directly to the demand side of this market. And that's where the picture gets genuinely complicated. Shen and Björk sampled two hundred sixty-two corresponding authors and found that roughly three-quarters came from Asia and Africa. About thirty-five percent were from India alone, eight percent from Nigeria, and six percent from the United States. This isn't a story about a few bad actors. It's a story about structural pressure. The paper cites qualitative research showing that academics at Nigerian universities reported difficulty getting into Western journals while their institutions required international publication for advancement. Serbian institutional rules mandated ISI-indexed publications for appointments and doctoral degrees. The pattern, as Shen and Björk put it, is "highly contained to just a few countries, where the academic evaluation practices strongly favor international publication, but without further quality checks." Against that backdrop, the economics of predatory journals make uncomfortable sense. Shen and Björk report two APC figures: an unweighted average per journal of three hundred four US dollars, and a weighted average per article of one hundred seventy-eight dollars. The gap between those two numbers is itself a finding. Authors were systematically choosing lower-priced journals, driving per-article costs down even as the market grew. In the mid-sized publisher tier, the per-journal APC averaged two hundred thirty-nine dollars but fell to one hundred four dollars when weighted by actual article counts. Compare that to legitimate open access journals indexed by the Directory of Open Access Journals, or DOAJ, where average APCs ran around nine hundred to one thousand dollars, and universities in the UK and Germany were paying roughly one thousand two hundred to one thousand three hundred dollars on average. Predatory journals were offering something at roughly one-sixth the price. Speed compounded the appeal. Publication delay averaged three point six months across journals, but the median — which Shen and Björk flag as the more meaningful figure — was two point seven months. Typically, two to three months from submission to publication. For a researcher under a promotion deadline, or in a country where one international publication changes a career trajectory, that combination of low cost and fast turnaround is a real and quantifiable value proposition. Here is where Shen and Björk land their hardest finding. The total volume of articles published in journals on Beall's list in 2014 — an estimated four hundred nineteen thousand, with a standard error of about ninety-one thousand — is now of the same order of magnitude as the total output of legitimate open access journals indexed by DOAJ. Laakso and Björk had estimated DOAJ journal output at three hundred forty thousand articles in 2011; extrapolated forward, that puts legitimate open access output at roughly half a million in 2014. Predatory journals were producing about eighty to eighty-five percent of that volume. And average output per predatory journal — about fifty articles per year — actually slightly exceeded the roughly forty articles per year per DOAJ journal in 2009. The shadow industry wasn't a footnote. It had become structurally comparable to the real one. Only seven point eight percent of journals from Beall's list appeared in DOAJ, meaning the vast majority existed entirely outside quality-control frameworks. That's the damage. Not just to individual researchers who might cite a poorly-reviewed paper without knowing it, but to open access as a project. Every predatory journal scandal — every sting operation, every fake manuscript accepted, every spam wave — generates ammunition for critics of open access publishing in general, even though the APC-funded gold model and the predatory model are functionally distinct. Shen and Björk don't argue for abandoning open access. The solution they point toward is stronger identification infrastructure: DOAJ's tightened criteria after 2014, membership standards from the Open Access Scholarly Publishers Association, funder requirements that mandate publication in verified journals, and APC waivers that give researchers in lower-income countries legitimate options. The problem, their data make clear, is not open access. It's that the market for publishing pressure is real, and without quality gates, something will always fill it. The internet didn't just open up science. It opened up everything — including the space for industries that look like science. What Shen and Björk gave us was a first clear look at how large that space had become. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.