Editor’s Note: Today’s post is by John G. Dove with Rebecca Kennison, Ginny Steel, Lea Delos Reyes, and Jo Dutilloy. John Dove is the founder of publishing consultancy Paloma & Associates. Rebecca Kennison is the Executive Director of K|N Consultants. Ginny Steel is the Norman and Armena Powell University Librarian at UCLA. Lea Delos Reyes is a Librarian Specialist at the International Rice Research Institute in Laguna, Philippines. Jo Dutilloy is a Masters in Information and Library Sciences student at Simmons College. This post is published under a CC BY 4.0 Attribution license.

Transparency has been a key topic in recent discussions of open access (OA) at conferences such as ALA Midwinter, ACRL, R2R, SSP, ALPSP, and COASP. The expectation is that a more open landscape might shed light on costs and prices, and even bring market pressures to bear on prices to more generally align with costs and values. Take Plan S, for example; its 6th of 10 principles is all about transparency. And on the day before COASP a consortium of publishers and research organizations met to discuss what a transparent transition to OA might look like.

transparent soap bubble

Ashley Farley, Associate Program Officer for Knowledge and Research Services at the Bill & Melinda Gates Foundation, as a panelist in the ACRL-SPARC Forum, mentioned how her role at the Foundation had provided her with a glimpse into the actual charges that the Foundation pays for Article Processing Charges (APCs) in fully-OA journals and for hybrid-OA fees in subscription journals. She looked forward to a much greater transparency into these charges, which might help bring about market pressure on the rates charged for opening the literature.

We therefore asked Ashley if she could provide us with their invoice data so that we could analyze it and publish our results, and she agreed.

The dataset Ashley provided us (covering the period from August 1, 2016, to March 31, 2019) includes:

  • 3,268 invoices for articles in peer-reviewed journals
  • 720 journals
  • 90 publishers

In total the Foundation paid $9,002,225 to these publishers to ensure results of all Foundation research was disseminated with a CC BY license with no embargos — an average cost for “open” of $2,755 per article.

We think we’ve uncovered some interesting trends in this data, but our main objective is to share the research dataset we have developed, along with the Foundation’s original invoicing data. That way anyone can use these for further research.

The Bill & Melinda Gates Foundation’s OA Policy

The Bill & Melinda Gates Foundation Open Access Policy was established in 2015, and is backed up by a service platform, Chronos, co-developed by the Foundation and Danish company ISSRC ApS. Jennifer Hansen, former Senior Officer and Open Access Policy lead at the Foundation, is the co-founder of Chronos, and also provided us with information for our analysis.

Since August 2016, researchers at organizations receiving funding from the Foundation have been encouraged to process their research articles through the Chronos system, which then submits the articles on their behalf and makes payments directly to the publishers. This allows researchers to focus on their research while the Foundation handles the administration of invoices and payments. It also means that there is a single point of contact for all these invoices for thousands of researchers around the world.

When the Foundation established its OA policy, there was a common misconception that it either required or strongly encouraged authors to publish only in fully-OA journals. In fact, three quarters of the journals in this dataset are subscription journals with a hybrid option and 59% of the invoices are for such hybrid fees. At just about the time that authors would have submitted articles for publication at the tail end of the stream of invoices represented in this dataset, the Foundation endorsed Plan S (November 5, 2018). One of the 10 principles of Plan S states explicitly that “The Funders do not support the ‘hybrid’ model of publishing. It will take some time for the hybrid exclusion to take full effect, but it will certainly be interesting to see how quickly this changes the set of researchers and their publishing choices.

One thing is immediately clear when you examine the subscription and OA journals separately and look for year-on-year differences: In 2016, only 22% of authors were choosing to publish in fully-OA journals; in 2019, 50% have done so.

chart showing author choice of hybrid versus fully OA journal

Both Jennifer and Ashley say that the Foundation’s OA policy office and the Chronos system itself have left the choice of journal entirely up to the author, as long as the journal complies with the Foundation’s requirement that the article will be available under a CC BY license with no embargo. This is borne out by the fact that, as of 17 August, 2019, Chronos has a database of 26,099 journals, both subscription and fully-OA, compared with Sherpa / Romeo, which listed 31,122 journals in its database on that date.

Invitation to further research

  • What evidence is there of inflation by journals or publishers in hybrid-fees?
  • Is there a difference in hybrid-fees between for-profit and non-profit publishers?
  • Does comparing prices and volume of hybrid-fee payments with retail prices of a journal’s subscription cost reveal evidence of “double dipping”?

Foundation’s dataset: https://doi.org/10.25346/S6/EEFYIP.

Our research dataset: https://doi.org/10.25346/S6/VODY5G

Open Access Journals and their APCs as evidenced by invoices

We decided to focus our analysis on the fully-OA journals. It is important to recognize upfront that any set of invoices is a distorted view of the whole landscape. The disciplines covered by journals in this dataset reflect the research agenda of the Bill & Melinda Gates Foundation, which is primarily (though not exclusively) focused on funding in the biological and medical sciences. Invoices represent actual payments rather than list prices, meaning that they include any discounts and waivers applied by the publisher. Individual articles that don’t require payment of an APC (such as invited papers) or journals whose APC is currently subsidized by the publisher or other sources do not generate an invoice. All these exceptions distort some of the average price calculations for journals and publishers. There is also a category of “diamond” or “platinum” OA journals which do not charge an APC; again, these are not included in this dataset. Thus, it doesn’t represent articles published across the entire landscape of peer-reviewed, fully OA journals.

We looked at three questions which this dataset might teach us about fully OA journals:

  1. How do APCs for fully-OA journals from the traditional subscription publishers compare to APCs charged by fully-OA journals from OA-only publishers?
  2. When choosing a fully-OA journal, do authors show any preference for fully-OA journals from the traditional publishers over those published by OA-only publishers?
  3. How do APCs charged by for-profit publishers compare with those charged by non-profit publishers?

Here are our takeaways.

  1. Traditional publishers charge significantly more for APCs than OA-only publishers

To definitively categorize journals as OA, and then publishers as either “Mixed-model” (having both subscription and OA journals) or OA-only, we consulted the Directory of Open Access Journals (DOAJ). All journals included in DOAJ were identified as definitely OA, but we also recognized that some OA journals are not in DOAJ, including a few titles from well-known publishers. We therefore double checked each publisher’s website and each journal’s website.

In two cases, large traditional publishers have acquired an OA-only publisher (Springer Nature Verlag owns BioMed Central, one of the original “born-OA” publishers, and Informa Group [Taylor & Francis] now owns Dove Medical Press). For the purpose of our analysis, we decided that these publishers currently operate on a sufficiently stand-alone basis that their pricing still reflects their “born-OA” roots. But, over time, if they are integrated into the publishing arms of their parent companies there might be good reason to include them with their parents as Mixed-model publishers.

chart showing average apc per year

In 2016 the average APC charged by traditional publishers for their OA Journals was $2,508 — $877 higher than the average APC charged by OA-only publishers. By 2019 traditional publishers’ APCs had increased to an average of $3,285 — $1,510 higher OA-only publishers.

Over these four years the Foundation is experiencing an annual rise in prices for APCs from the traditional publishers of 9.7% per year compared with an average annual rise of just 2.9% from the OA-only publishers.

Invitation to further research:

  • Which individual publishers are the biggest contributors to this difference?
  • Is there greater or lesser inflation depending on discipline? Or perhaps impact factor?
  • Do researchers make their choices based on price?
  • Would putting BMC with Springer-Nature Verlag and Dove Medical with the Informa Group change this picture significantly?

Foundation’s dataset: https://doi.org/10.25346/S6/EEFYIP.

Our research dataset:  https://doi.org/10.25346/S6/VODY5G.

  1. Researchers choosing to publish in a fully OA Journal show a strong preference for OA-only publishers’ titles

A clear majority (about 70%) of researchers publishing articles in fully OA journals have chosen OA-only publishers in each of the four calendar years of this dataset. Is there some blemish on the brand of the traditional subscription publishers? Or some glow of OA-authenticity enjoyed by the brands of the OA-Only publishers? Or are they showing price sensitivity to the big difference in prices of the APCs? More research is needed to explain these researchers’ choices of journals and publishers.

chart showing traditional versus OA-only publishers

Invitation to further research:

  • Is there a geographic or disciplinary dependency on this 70%?
  • Is the significant price differential driving the choice?
  • When choosing a fully-OA journal, are authors paying attention to Impact Factor and how does that affect these results?

Foundation’s dataset: https://doi.org/10.25346/S6/EEFYIP.

Our research dataset: https://doi.org/10.25346/S6/VODY5G.

  1. There are no significant differences in for-profit vs. non-profit publishers’ APCs for fully OA journals

This chart of the top publishers (defined as those with at least 15 articles in the dataset) shows that non-profit and for-profit publishers have among both the lowest and the highest average APCs.

Note that each individual publisher’s overall average APC could very well differ from the sample represented by these invoices to the Foundation. Only the publishers themselves know their overall averages, and in a transparent world they would be expected to provide such information.

 

Publisher Number of Articles Avg APC [in US$] For-Profit or Non-Profit
Elsevier 123 3,492 FP
BMJ Publishing Group 91 2,972 NP
Springer Nature Verlag 81 2,757 FP
Oxford University Press 18 2,681 NP
Wiley 41 2,588 FP
Frontiers Media 64 2,447 FP
BioMed Central 322 2,070 FP
PLOS 318 1,928 NP
MDPI 32 1,520 FP
F1000 Research 107 918 FP
Edinburgh University Global Health Society 42 827 NP

Invitation to Further Research

  • Categorizing all 90 publishers as for-profit or non-profit might tease out some differences that we’ve not seen.
  • More research might uncover the reasons for the substantial gap between journals at the low end of the pricing scale and those at the top. Is the pricing based on what the market will bear, gross profit margin targets, perceived value of service and quality, subsidization of some of the costs? — a combination of these? Something else? Many of the answers to these questions will have to come from publishers themselves.

Foundation’s dataset: https://doi.org/10.25346/S6/EEFYIP.

Our research dataset: https://doi.org/10.25346/S6/VODY5G.

In sum, our research is just scratching the surface of the valuable insights that can be gained from analyzing this dataset. The dataset is publicly available, and we look forward to the insights that others will find and to future reports to come.

 

Acknowledgements: The authors would like to thank Peter Suber, Lars Bjørnshauge, Impact-Hub/Boston, Anthony Watkinson, Tim Dennis, Jamie Jamieson, Jennifer Hansen, and Ashely Farley for their support in our efforts.

Discussion

23 Thoughts on "Guest Post — Transparency: What Can One Learn from a Trove of Invoices?"

I’m trying to understand the last table above listing the publisher and # of articles. The .xlsx sheet of these data had hidden rows in it. When I saved it as a .csv to get rid of the many, many hidden rows, and then ran the counts myself, I got some numbers of articles that matched and some that did not.

For instance, I got the same number of articles for F1000 Research (107), BMC (322), and PLOS (318). However, I tallied far different numbers for Elsevier (872), OUP (347), Springer Nature Verlag (203), Wiley (213), and some of the others. I also saw ACS and the American Society of Tropical Medicine and Hygiene with significant numbers (86 and 87, respectively).

Did I do something wrong? Or maybe those hidden cells affected this table?

Also, can you clarify how you treated Gates Open Research vs. F1000 Research? There seem to be some papers that went directly to F1000 Research, while most attributed to them were actually published in Gates Open Research, with F1000 Research acting as a service provider.

Thanks.

I think the F1000 question is one worth asking — if Gates is paying to have a sponsored subsection of F1000, should those costs be factored into the APCs paid for that journal (and its sub-journals)?

Gates pays an annual fees for the maintenance of the Gates Open Research platform. This cost is not reflected in the APC data set. Good question as to whether or not it should be included (I would argue not as this cost also covers the non-peer reviewed portion of the platform). Gates Open Research launched in 2016 and is not a requirement so say an author wants to publish through F1000 instead we would not say that they have to publish in Gates Open Research – so this is why there are both Gates Open Research specific articles and F1000 (timing and publication choice). I think for this analysis these were combined as the foundation does not technically view itself as a publisher as F1000 manages the editorial workflow for Gates Open Research.

Thank you, Kent, for bring these inconsistencies to light. I will check into each one of them. We excluded many rows which were out of the date range or were not peer-reviewed journals. The excluded rows were not intended to be in our published research dataset and will be explained in a revised codebook. We will make corrections in the research dataset as well, if necessary.

We did provide the data that we received from the Foundation just as we got it. So that should be the definitive source.

Thanks. Also, I should have said that I *revealed* the hidden rows. That’s a clearer description of what I think went on — revealing data maybe some counting pass or procedure missed. But do let me know when you can, as I’m curious. Thanks.

Kent,
I did not want to reply without thoroughly checking into the inconsistencies that you mention. I inadvertently had filters turned on for the spreadsheet which has the effect of hiding some rows, though it did not change any of the numbers.
The number discrepancies you mention are all due to including hybrid articles. We restricted our analysis to just articles published in Fully OA journals, so article counts we report are limited to invoices for APCs on the non-hybrid articles.
You’ll notice that all the publishers where you confirmed the article count are OA-Only publishers so they would have no hybrid articles to distort the count you got of all articles (hybrid and fully-OA).
If anyone is using the spreadsheet to focus on our analysis, you should filter out the hybrid articles by filtering column N “Derived: OA-or-Not” for “Y”. Then all the publisher numbers will match yours.
American Chemical Society only have 4 articles from Fully-OA journals, so they were left off the table. We arbitrarily presented only publishers who had at least 15 articles in the data to avoid making inferences from very small sample sizes.
If American Chemical Society had been included in the table its average APC for those four articles is just $810. That is an interesting number, but I don’t know how typical that is across all of ACS’s portfolio of Fully-OA journals.
The American Society of Tropical Medicine and Hygiene has only hybrid articles among the Foundations invoices, so they would not appear in any of the data we report on.
Thank you again for pointing out the filtering left in place on the spreadsheet. I will put up a new deposit and a better Code book next week.

Thanks. Interesting that you made this distinction at the journal level, when the invoices were at the article level. But thanks for clearing that up.

I’m interested in exploring the jump from 22% full OA in 2016 to 50% full OA in 2019 (as presented in the first figure — it would be helpful if the figures were numbered). Such a finding is really, really significant!

Is the data source for the Elsevier number in the last figure the same source that was used for the first figure? Filtering by publisher Elsevier using the Gates data returns 946 records rather than the 123 in the last figure though I might be misunderstanding what the last figure is presenting.

Ahh, I see the Paloma dataset has nearly the same number of publications. Could the authors specify how the data was calculated in the first figure? Is it based on ‘Foundation_Payment_Date’ and ‘Derived_Publisher_Type’? Again, I’m just trying to dig in to the super interesting increase in Gates authors choosing full OA journals. Many thanks for processing all of this data and making it available!

Yes, Linda. I found that to be interesting data and encouraging. I assumed that it would have been the result of some excellent promotion by the Foundation’s OA Policy team as to the benefits of OA publishing. But as we reported there was no promotion of fully OA journal choices by the Foundation.

Just by chance, Lea (our co-author who did the vast majority of our web-research) is a library specialist at a research institute in the Philippines which received some funding from the Bill & Melinda Gates Foundation. She confirmed that the Foundation did not explicitly encourage choosing fully-OA journals over hybrid journals, but just having received Foundation funding had their researchers look into Open Access and then develop and pass their own OA policy for their institution.

The Elsevier numbers you report from the Foundation includes a set of rows which we excluded. We wanted the time covered to end on a boundary of a quarter (31 March 2019) and so any records after that date were excluded.

There are no publisher specific numbers reported in Chart 1 [I leave your suggestion of numbering the Charts up to our Scholar Kitchen Chef, David Crotty] but any Elsevier hybrid articles would contribute to the bars on the left of each year, and their fully-OA article counts would contribute to the bars on the right of each year.

Very helpful, John.

For Chart 1, things got muddled for me because I was mistakenly using ‘Derived Publisher Type’ (mixed model or OA only) instead of the ‘Derived OA or Not OA’ (a code for the journal itself, not the publisher), which is indeed split at 50/50 for 2019 (!). From my observational experience as a librarian, researchers know which journals in their fields are full OA, but the fact that (Gates) authors are increasingly choosing to publish in full OA journals is a leap I wouldn’t have bet on.

Again, many thanks to you and your team for your analysis.

Some comments on the fly, both in reply to this article but also in relation to broader questions.
1. “Do researchers make their choices based on price?” Research adducing this was broached in a prior article on TSK. Unless t and p expectations are changed, which will take a very long time, we can expect the drag of this inelasticity to thwart downward pressure on pricing (APC’s or toll-access publishing).
2. I perhaps missed it, but have not seen discussion of the opportunity costs that Plan S schemes impose on limited library budgets, in terms of diverting library funds to APCs. There is a comparable need for discussion of this question: what opportunity costs will these schemes impose on use of those funds for other purposes? (notably other library products and services.)
3.The Plan S approach presuppose that tax dollars should be used to pay for scholarly publishing. Should it? It strikes me that there is a philosophical concern that could easily turn into a practical concern. In an ideal world there is a robust demarcation of the scholarly production process on the one hand, and government on the other. Persons on either side of the political aisle can and should hold this position, imo. The point is to avoid political manipulation of the knowledge distribution system. A counterargument is that we already have government funding of research, which in effect addresses an economic externality. (I.e., federal funding addresses the question: what entities are going to pay for research in the absence of economic incentives to do so?) So, advocates of Plan S might argue, isn’t it a small step to use federal grant funding to fund APCs…so what’s the problem with using them for this purpose? The responsio is that it’s not at all clear that there is any analogous philosophical justification for using federal funds to reform or change the scholarly publishing system. Is it the role of the federal government to bankroll methods of scholarly knowledge distribution when other options have not been tried? More starkly, is it really the role of the federal government to address dys-functionalities in the scholarly publishing ecosystem, when a broader range of institutionally based possible solutions have not yet been attempted? (The argument is a philosophically subsidiarist one.)
4. Concerning these other solutions: why not a twofold initiative by universities and their consortial representatives consisting of (a.) an attempt to educate faculty about the very real problems of the glut of so much sub-par literature (albeit peer reviewed) being published, thereby swamping researchers with too much information but also burdening researchers with expectations to peer review; (b.) a concerted and coordinated effort to contract the number of journals, as a way to exert downward pressures on prices whether for toll-access or APCs. (Concerning (a.), allow me to repeat the argument that consortia and librarians across the land are now committed to changing researcher behaviors by promoting OA; on pains of inconsistency, why shouldn’t they advocate for contraction of the journal market just on the basis of the need to reduce the research glut, rewarded as it is by the t and p system?)
Apologies for some repetition of earlier themes. I just haven’t seen plausible replies as yet to these counterarguments. One has the distinct since that the fascination with Plan S , “transformative” agreements, and “immediate OA” now crowd out careful consideration of counterarguments and exploration of alternative approaches. Nothing in this world of any value is “immediate”, particularly when dealing with so complex a system as scholarly publishing. [views as usual are my own]

Brian,
Thank you for setting out a good set of diverse questions. I’ve responded to some where I have something new to say in response below.
I also want to respond to your first point: I acknowledge that research has been done to try to find what the parameters are of author choice of journals. I also think that it is worth refreshing that research. As we reported in this post there are two choices authors are making: one, what journal will I publish in; and two, what publisher am I going to choose. It’s an oft repeated maxim that “academics know the journals and librarians know the publishers”. That’s why the result in the third bar-chart stands out. In choosing a fully-OA journal there seems to be a really strong preference to choose a journal from an OA-Only publisher.

So we think someone with good polling skills [not my strength] could do a survey of a sample of the specific authors represented in this dataset to tease out what went into their choices. The presentation at the Charleston Conference this week by the new CEO of Elsevier, Kumsal Bayazit, [https://2019charlestonlibraryconference.sched.com/event/UXqo/keynote-plenary-collaborating-to-support-the-research-community-the-next-chapter?iframe=no&w=100%&sidebar=yes&bg=no.] included a slide suggesting that “Open Access” was near the bottom of the list of criteria by which authors choose where to publish. We’ve reported here that authors are choosing journals from OA-only publishers 70 to 30 over the fully-OA journals of traditional publishers. So it’s either because of the large and growing difference in costs of APCs from these two categories of publishers, or it has something to do with the relative authenticity of what almost every publisher says these days, “We fully support Open Access”. Or it could be something about being a Bill & Melinda Gates Foundation funded research project that influences this choice. Asking a random sample of these specific authors might tease apart these explanations.

A p.s.:
doesn’t a contraction of the journal space provide at least as much potential to drive down prices for APCs and toll-access publishing, as any (speculative) role that one or another OA framework would play in exerting downward prices? After all, a contraction of this space would immediately reflect reduced demand, which affects prices, at least on the premises of textbook microeconomics (with possible unusual exceptions). To wit, I’m not clear on what downward pressure price mechanisms are supposed to eventuate in downward pricing in the Plan S/transformative type model. What are the details of how this is supposed to happen?

Lots of food for thought, Brian. I’m not an author of the SK post — just a (U.S.) librarian and rare engager in these discussions. Are your questions about the lack of Plan S discussions limited to the U.S.? I wasn’t at Charleston, but in your liblicense post you state that Plan S is advancing in the U.S. Do tell! Based on your discussions there, do you have a sense for whether signatories will be at the individual institution / state system / regional or national consortia or association (ACRL, Solinet, CIRLA) and/or federal level?

I’ve recently taken online courses with several European librarians where I was reminded that they’ve been discussing many of the principles addressed in Plan S for several years. And as you know, many have already implemented some of the principles through transformative agreements at the national level. The Efficiency and Standards for Article Charges/ESAC site includes links to the transformative agreements where you can see how Plan S-like principles have been implemented (https://esac-initiative.org/about/transformative-agreements/agreement-registry/ and guidelines at https://esac-initiative.org/guidelines/). The OAD also has direct links to APC funding policies to get a better sense of how Plan S signatories are already funding (obv may change when Plan S is implemented): http://oad.simmons.edu/oadwiki/OA_publication_funds

As for contraction of the journal space, I’m definitely out of the loop here! Is this something proposed as a viable solution to driving down prices? Yes, there are sooooo many journals, but I’m curious if this means a contraction of actual amount content given the increasing number of articles and data, especially if it’s agreed that science is enhanced by publication of a broader set of findings (not just novel results) and the importance of providing a voice for LMIC researchers and publics. I think of the rich and diverse entertainment content space where kids born in the time of Hulu, Netflix, and Apple TV couldn’t begin to fathom the days of 3 networks :).

Hi Linda, great questions. I’ve used up enough bandwith here so will be brief. I wrote some about the contraction issue in a preprint (about preprints) and intend to expand that and also develop a critique there of the plan S type scheme plus reflect on pricing dynamics more. As for plan S’s reach on US shores, a TSK writer should help to “calibrate” it plus report on SSP conference discussions about it. Like you, I can only address these topics intermittently, but they are fascinating esp given that I have graduate degrees in economics and also philosophy, both of which provide perspectives that give me lots of concerns.

I wonder if the authors would clarify the Foundation’s position on hybrid journals. They state:

“At just about the time that authors would have submitted articles for publication at the tail end of the stream of invoices represented in this dataset, the Foundation endorsed Plan S (November 5, 2018). One of the 10 principles of Plan S states explicitly that “The Funders do not support the ‘hybrid’ model of publishing”. It will take some time for the hybrid exclusion to take full effect, but it will certainly be interesting to see how quickly this changes the set of researchers and their publishing choices.”

Two paragraphs later, that information is contradicted by:

“Both Jennifer and Ashley say that the Foundation’s OA policy office and the Chronos system itself have left the choice of journal entirely up to the author, as long as the journal complies with the Foundation’s requirement that the article will be available under a CC BY license with no embargo.”

Are researchers funded by the Foundation allowed to publish in hybrid journals or not? Similar contradictory statements have been made by other funders, leaving researchers uncertain of where they are allowed to publish.

I did not intend to make any assertion about if, how, or when the Foundation would implement the Plan S principle that disparages the hybrid-option. My point about the timing is that one might reasonably assume that having the Foundation join the Plan S Coalition on November of 2018 was not likely to influence enough authors to change their intention to submit a paper, submit it, have it go through peer-review and be accepted all before 31 March of 2019, so it was unlikely to change the Q1 2019 numbers. The data, however, is there for you to check, and the payment dates are specific, so you could find out if there was a dramatic bump up caused the by the November 8th announcement.

Perhaps Ashley can let us know how the Foundation is planning to deal with hybrid going forward.

Who and how is going to pay for all the stuff being published? Entire university budgets will be consumed by OA charges. If we settle on an average fee of say $2500 and multiply that by 2.5 million (number of new articles per year) it is a lot of money even to Bill Gates and it will loom even larger to some government bureaucrat when s/he starts looking for a place to cut money!

” According to research from the University of Ottawa, in 2009 we passed the 50 million mark in terms of the total number of science papers published since 1665, and approximately 2.5 million new scientific papers are published each year.”

“What’s driving this publication explosion?
At its most basic level, we’ve seen a substantial increase in the total number of academic journals. As of 2014 there were approximately 28,100 active scholarly peer-reviewed journals. Add to this the increasing number of predatory or fake scientific journals, which produce high volumes of poor-quality research, and you have a veritable jungle of journals to wade through.”

“Another key factor is the sheer number of publishing scientists worldwide, which is increasing at a rate of approximately 4-5% per year.”
http://blog.cdnsciencepub.com/21st-century-science-overload/

Brian’s arguments are very interesting in that he challenges the basic premise of the OA movement! A movement that is not being challenged in the light of economic reality. A very interesting statistic would be from where does authorship come and in regards to OA, who is paying, and who is providing the funds? If it is the university library are their budgets being increased to cover the costs of publishing and who in the university is suffering because of the diversion of funds to the few who are publishing? Will the library need to increase headcount and will that be added onto the OA publishing fee? If funding is coming from some agency or foundation who is handling the transaction because there is a fee in that too. Thus, the fee the author pays is but part of the true cost of publishing the article.
Regarding cutting the number of journals. That would have worked in a non internet world. To think, Plos One has surpassed 200,000 articles since its founding. Thus, I am not sure cutting is a solution to the problem. Thus, the solution to the publish or perish for free access dilemma will only be addressed when someone says: I am unwilling to provide the funds for Sally or Dave to publish their article!

I’m replying to the points that Brian and Harvey raise about who is going to pay. This dataset shows an example of a foundation funding some significant amount of research (not as a government agency). It is one example of how funding of APCs can be done. One argument for APCs in areas of funded research is that many funders already do include support for dissemination of the research that they support. This includes things like travel and participation in scientific meetings and expenses independent of the publication of results in peer-reviewed journals. But the latter can reasonably be included in the broader category of dissemination.

The Bill & Melinda Gates Foundations OA policy, as I understand it now, was set up in order to lift as much as possible any administrative or financial burden on authors or their institutions to cover the costs of reporting the results of the research in peer-reviewed journals.

Ashley tells me that all of this $9M cost for hybrid fees and APCs came out of central fund, not the individual grants. We did not report on growth of this fund, but the dataset clearly shows significant year-on-year growth. Annualizing the two partial years (2016 and 2019) I get: 2016–$810,966; 2017–$3,141,235; 2018–$4,001,091; and 2019: $5,825,664. This growing amount is a significant savings for any of the funding that libraries are using to support for APCs for their researchers.

To respond to the comments about Plan S as to funding of the APCs, the fourth principle of Plan S [https://www.coalition-s.org/principles-and-implementation/]. addresses funding an suggests one of the alternatives is that APCs are funded by the funding organizations. [But certainly not individual researchers.] The Bill & Melinda Gates Foundation has chosen this path. If most funders choose that same path this might mean a significant lifting of any concern about the effect on library budgets.

Perhaps someone from the Plan S Coalition could weight in on this.

Hi John,
Many thanks for getting back.
I have yet to fully develop my arguments about the role of non-profit (non-governmental) organizations in funding APCs.
But the concerns in a listserv message I posted this morning, in the wake of this week’s Charleston conference, extend in my view as well to non-profit funding of plan S schemes and, more generally, the overall scheme itself and efforts to perpetuate it pose very many questions, regardless how they are funded:
https://list.indiana.edu/sympa/arc/chminf-l/2019-11/msg00005.html
This was cross-posted to a few other listservs.
I should note that the other Socratically posed questions in that posting are independent of the question about researcher sensitivity to pricing for APCs that is so intriguing. This latter issue will assuredly be a continuing area of research. I’d like to spend time on it, but have a primary commitment to the folks in the six subject areas I serve at a university and so can only work on these issues intermittently and un-systematically. But it is a very intriguing issue and one about which a lot more will appear. A particular view on the answer to the question about the responsiveness of researchers to APC pricing, however, has already been broached by one U Cal administrators and it is not clear that arguments to the contrary have been fully engaged.

Harvey, hanks for the comment. It is indeed a tall order to contract the number of journals and your comment is very sound about the barriers that the predominance of the electronic format imposes on contracting the market. I’m under no illusion that contraction may never happen, if t and p assessment criteria are not changed. They, and grant funding agencies that reward number of publications and miles long cv’s, are important drivers of journal demand and extremely difficult to reform.
In my view, a far more interesting data collection effort than, say, all the efforts to collect data about APC’s would to conduct surveys about the needs of researchers, with respect to whether they think the glut of research is now hampering their efforts and promoting over-specialization. Currently ongoing “natural experiments” (a term I’m not using with precision) in contracting the journal space could provide very interesting survey data along these lines. See http://listserv.crl.edu/wa.exe?A2=ind1910&L=LIBLICENSE-L&P=70057
U Cal is now ideally suited to conduct such surveys and then to educate faculty about the problem of research glut that properly designed surveys will no doubt reveal. Perhaps research of this kind is already underway in Europe in those contexts in which journal big deals were temporarily shut off? Related to this, a systematic study of efforts by the library profession to promote OA publication among their faculties will surely challenge any argument that librarians do not regard one of their goals as one of changing faculty perceptions about how the latter should (prescriptively) publish.

Leave a Comment