Library Access Perspective
The Empire Strikes Back
Somehow that seems like an appropriate overall term for several clusters of material, including some of the proceedings and commentaries from a set of (U.K.) House of Commons Science and Technology hearings on STM publishing and a bunch of other commentaries specifically including part of an ongoing Nature debate. My sense in reading this is that open access publishing must be perturbing the STM oligarchy a lot; otherwise, they wouldn’t be so busy spreading misinformation about it.
In addition to notes here, I’ve seen other cases where representatives of commercial and scholarly publishers assert that open access journals don’t have proper peer review procedures embedded in their operating assumptions. Since this is clearly false—free scholarly journals have used proper peer review for more than 15 years, and it’s explicitly part of contemporary OA models—I have to wonder whether this is ignorance or malice.
One key question that may never be answered is critically important if you assume—as I do—that one way or another, a substantial portion of open access publication fees is likely to be diverted from library funding. The question is how many STM articles actually get published each year. That’s particularly important given the uncertain issue of how much has to be charged for each published article in order for OA journals to survive in the long run, and the wide range of such fees at present (the two dominant numbers being BioMed Central’s $500 and PLoS’ $1,500).
Some OA advocates assert 2.25 million articles a year. Elsevier claims 1.2 million articles a year. That’s a huge difference. If Elsevier’s claim that $4.5 billion is spent on STM journals is correct—that’s how Elsevier arrives at the $3,750 “cost” per published article overall—then consider two outcomes:
Ø For 2.25 million articles a year, $1,500 fees (PLoS) yield a total cost of $3.375 billion dollars. Since publication fees won’t replace all of the money spent on STM journals (quite a few journals add value beyond refereed articles in their print editions, and even OA journals charge for that added value and for print subscriptions), it’s not at all clear that there would be any overall savings.
Ø For 1.2 million articles a year, however, the total cost comes out to $1.8 billion, which would seem to assure overall savings to the community, even with lots of money spent on print extras.
Interestingly, PLoS’s own cost analysis (mentioned in the regular Library Access to Scholarship section in this issue) shows $870 per article as their total production costs, plus $20 to handle each submitted manuscript regardless of whether it’s published. I could poke at those numbers—e.g., if there are an average of six graphics per article, then why does the analysis show $138 per 11-page article for graphics layout at $12.50 per page—but never mind. (Based on PLoS’ analysis, a text-only journal should have total costs of about $565 per published article and $20 per submitted article.) PLoS arrives at a total cost of $1,070, assuming a 90% rejection rate. Note that PLoS’ model properly includes significant amounts for copy-editing and layout.
I would apologize for the length of this section—but as with other similarly long thematic sections, part of this is for the record. I want to be able to come back in two or three years and trace what’s happened without trying to return to primary materials, and I particularly want to be able to refute charges of strawmen or red herrings.
I don’t have all the unedited minutes from the hearings and certainly lack most of the position papers. I’m going to ignore substantial portions of the hearings that had to do with aspects of STM publishing other than pricing, open access, and copyright.
A good starting point may be Elsevier’s comments on evolutions in scientific, technical and medical publishing and reflections on possible implications of Open Access journals for the UK, dated February 2004 and apparently a position paper for the UK parliamentary hearings. I don’t know whether the 15-page PDF will be available at any given point: It’s apparently already appeared, disappeared, and reappeared through at least one cycle.
“The current worldwide system of Scientific, Technical and Medical (STM) publishing has evolved over hundreds of years, and we believe it serves science and medical communities well.” I’d be surprised if Elsevier said the system was broken. “Hundreds of years” hardly says much about the relatively recent dominance of commercial publishing in STM, but never mind.
One ongoing quandary is just how many peer-reviewed STM journals and articles actually exist. Elsevier claims 1.2 million articles a year published by “some 2,000 STM publishers,” the articles then used by “millions of researchers.”
Elsevier asserts that 97% of UK researchers have direct access to around 90% of Elsevier journals—and that “UK citizens have access to all Elsevier journals and articles either directly through their local libraries, or via inter-library loan agreements.”
Then comes the attack on several fronts:
Ø The OA model “risks penalising the UK because British researchers produce a disproportionately high number of articles every year.” British researchers supposedly contribute 5% of all STM articles while British spending on journals is about 3.3% of the world total—if you believe Elsevier’s figures.
Ø OA risks “undermining public trust.” The subscription model “ensures high quality, independent peer review and prevents commercial interests from influencing decisions to publish. This critical control measure would be removed in a system where the author—or indeed his/her sponsoring institution—pays.” The specific claim is that, because publication fees are only for accepted papers, OA publishers would be under continual pressure to increase output “potentially at the expense of quality.”
Ø Even the highest OA article fees “cover only about 40-60% of the estimated total costs” to publish high-quality articles—which means “to provide all the revenue to publishers that they currently make.” The equation of costs with revenues is a constant in the empire’s counterattack.
Ø Universal access requires print, since only 11% of the world’s population uses the Internet and “only 64% of UK citizens have ever been online.”
Details include descriptions of the Big Deal in the UK, a claim that the STM publishing market is “highly competitive,” with 2,000 publishers publishing 16,000 “unique learned journals,” the assertion that no STM publisher “has disproportionate power,” and a more detailed discussion of why OA models won’t work.
“A wide range of supporting evidence shows that costs exceed $3,000 per article at existing quality levels… For example, the Open Society Institute suggests Open Access publishers will need to recoup $3,750 per article published… By contrast, Science magazine estimates that it would have to charge $10,000 per article in a pay-per-article model… Similarly, the American Journal of Biological Sciences estimates that the journal BioScience would have to charge $7,000 per article. “$3,750…is in line with Elsevier’s estimated mean costs per article across the range of its some 1,800 journals.”
Elsevier also takes credit for “nurturing new areas of science” by launching 31 new journals a year. The claims that article quality would suffer from OA are repeated and expanded. The significance of existing OA journals is dismissed with the note that “ISI, the industry standard that provides key data…on the quality of research, currently measures only two out of some 500 Open Access Journals…”
Elsevier claims that OA models would increase fraud and malpractice because individual researchers would lack the resources and legal expertise to identify infringements and pursue transgressors. “Publishers, together with their journal editors, have been vigilant in identifying and taking action against issues such as multiple publication and plagiarism.”
On or around February 23, 2004 Jan Velterop issued a set of comments on Elsevier’s position paper, noting that Elsevier seems “curiously ill-informed” about OA publishing.
Regarding UK researcher access, Velterop notes that this seems to assume that all UK researchers work for academic institutions. The UK National Health Service does not have Elsevier’s Big Deal, and smaller biotech (and other science and technology) companies surely have researchers but no assurance of Elsevier access.
Velterop’s response to Elsevier’s assertion that OA would penalize the UK—and would penalize major universities to the benefit of commercial organizations and the like—includes the following:
Scientists and institutions benefit from making their published research available to a wide audience—it is by publishing influential research that institutions acquire a reputation that brings them high levels of funding and top researchers. And the cost of dissemination is tiny compared to the cost of doing the research in the first place… In the traditional environment, the less well-off institutions, which publish little research, effectively subsidize (through subscriptions) the publication costs of better-off institutions, which publish a lot.
Velterop’s response to the whole argument that OA models would undermine article quality and the peer review process is pointed, although the first sentence raises some questions about grade inflation:
If a student pays tuition fees, does that make his exam easier to pass? The overwhelming majority of the journals published by Elsevier have traditionally seen price increases proportional with the increase in their volume [and for other reasons]…As a result, they would benefit from a higher acceptance rate in the same way that they imply Open Access publishers do.
As regards OA viability, Velterop says Elsevier’s estimates are based on inefficient operation of traditional publishers. Elsevier also claimed (which I didn’t quote in detail) that their huge profits result in technological innovation as well as nurturing emergent areas of science to which Velterop responds: “It is not huge investment by a large corporation that best drives innovation in the online world. Open platforms drive innovation, as the internet has shown.” Later, he notes that OA publishing means that a publisher doesn’t have to demonstrate a commercially viable market for subscriptions in order to begin a needed new niche journal. “This allows journals to develop in new niches that would have been too small or too poor to support a traditional journal. BioMed Central has published several journals that show how the previous publishing models had failed to cover a particular area, e.g. Malaria Journal.”
Velterop mostly makes fun of Elsevier’s blather about print publishing and the non-Internet population. “It is somewhat bizarre that Elsevier imagines that the 89% of the world’s population who have never used the internet are somehow likely to have access to print copies of Elsevier journals.” In any case, as he notes, the most prominent OA publishers do offer print subscriptions, “and the logic that some people may want to pay for print has very little bearing on open access.” Taking on Elsevier’s note that “only” 64% of UK adults have ever used the internet (68% when Velterop looked), he asks, “How many UK adults have ever gone into a scientific reference library?”
I find his next two comments a little discouraging, frankly:
(a) Many of the libraries who receive a copy of BioMed Central journals that have a subscription component, such as Genome Biology, have asked us not to send the print, as they actively find print a problem.
(b) We offer any library the opportunity to receive, at cost, a print archival copy of all or a portion of the research that we publish. Not one library has so far taken us up on the offer. Print seems to be of rapidly decreasing importance to libraries. [Emphasis added]
I know: It’s STM journal literature. I’ve gone on record in the May American Libraries saying it’s probably a good thing that 90% to 95% of STM journals are likely to become electronic-only. Still…shouldn’t libraries take some interest in the only proven method for long-term retention of this information?
Velterop goes to town on the longer version of Elsevier’s attack on OA quality and significance. He notes that authors choose journals based on reputation, so every journal has an incentive not to damage its reputation; that many respected journals already have page charges (at least for color figures); and—again—that the temptation to accept more papers is precisely the same for subscription journals that charge more as they get bigger as it is for OA journals: More articles means more revenue.
Then he catches Elsevier in a flat-out mistruth: The claim that ISI only provides impact measures for two OA journals. “BioMed Central alone has 6 journals that currently have impact factors… ISI explicitly tracks 22 BioMed Central journals and several more of these will get impact factors in June 2004. And citations of the other 80+ BioMed Central journals are already captured and tracked in ISI’s cited reference database, so although ISI does not yet produce journal impact factors for these journals, if one wants to find out how many times an article has been cited, one can do so.”
Adam Hodgkin described the hearing in a LibLicense post: “By my estimate, in addition to the Committee members and Expert Witnesses, there were 80+ citizens and interested parties in the Committee meeting room. Deep green and gilt wallpaper—worthy of a Lord Chancellor—four large chandeliers, leather back chairs…and three enormous full-length portraits of 18th c. parliamentarians on the walls.” The first witnesses were from Blackwells, John Wiley, and Nature Publishing Group; two Elsevier representatives appeared later. Hodgkin draws particular note to the point at which Wiley’s Dr. Jarvis might have lost the sympathy of his audience: “when he appeared to be arguing that it was a good thing that the general public cannot get access to specialist scientific journals.”
The uncorrected transcript of the hearing was made available a few days after the hearing itself. I’m obliged to note that neither witnesses nor Members have had the opportunity to correct that record, and that the transcript is not yet an approved formal record of those proceedings.
In the uncorrected transcript, Jarvis makes it clear that Wiley doesn’t provide delayed open access—“we make quite a lot of sales of back-issue information.” Robert Campbell of Blackwell weighs in early claiming the dangers of OA: “We think there is a danger that an author-paid model could lead to lower standards.” He also claimed OA was not “popular amongst authors.” Nature’s Dr. Charkin repeats the bizarre claim that they’d have to charge “£10,000 to £30,000 per article” ($18,100 to $54,300) “to replace our revenues.” At least he recognizes that it’s revenues (not costs) that are being used to arrive at these estimates.
Campbell calls Britain a “net exporter of knowledge” based on the claim that it produces 5% of journal articles and only pays 3.3% of total journal subscriptions. I, for one, would assert that, if British researchers read 20 times as many articles as they produce, Britain must surely be a net importer of knowledge by any rational measure—and, based on the 5% to 3.3% discrepancy, that importation is being subsidized by the rest of the world.
The chair asked whether there was a demand for open access publishing. Seems like a simple question. Here’s the answer:
Dr. Charkin: We are just running a survey through all the authors to Nature to find out. We ran an open access debate about a year ago within Nature and there really was not overwhelming support. Clearly, there is some sort of a groundswell, but it certainly was not overwhelming, and early indications from procedures at the National Academy of Sciences in America and such like have not really supported the contention that it is huge.
Of course, the chair didn’t ask whether there was “overwhelming support” for OA; he asked whether there was a demand, which “a groundswell” would seem to affirm.
The chair dug into Nature’s absurd per-article costs. The response: “Very crudely, £30 million of sales: we get income of £30 million and we publish 1,000 papers a year.” That doesn’t speak to the advertising revenues of Nature or to the considerable portion of its content that is something other than refereed scientific papers. In essence, this representative is asserting that an author-paid model would mean abandoning all other forms of revenue and supporting all existing subscriptions and costs through the small pipeline from submitted papers—even though existing OA journals offer value-added features at a price.
Sure enough, Dr. Jarvis speaks out against access to STM articles on principle:
One of the things that intrigues me is that there is some evidence that some of the support for open access is coming from outside the research community. There are some very high-profile stories of members of the public who want to read this kind of information. Without being pejorative or elitist, I think that is an issue that we should think about very, very carefully, because there are very few members of the public, and very few people in this room, who would want to read some of this scientific information, and in fact draw wrong conclusions from it…. I will say again: let us be careful because this rather enticing statement that everybody should be able to see everything could lead to chaos. Speak to people in the medical profession and they will say the last thing they want are people who may have illnesses reading this information, marching into surgeries and asking things.
I omitted the middle section in which Jarvis wholly undermines his incredibly elitist argument: “I think the mechanisms are in place for anybody in this room to go into their public library, and for nothing, through inter-library loan, get access to any article they want.” So, on the one hand, we shouldn’t have open access because it’s too dangerous and, you know, doctors don’t want patients to be asking questions—and, on the other hand, we don’t need it because any member of the great unwashed can get anything they want anyway. All clear now?
One of the committee members didn’t buy it:
That is not what Dr. Virginia Barbour is saying, the molecular medicine editor at The Lancet [an Elsevier journal]. She feels that patients should be able to access papers about their medical conditions. What are you doing to ensure that patients who are not scientists have access to quality medical journals that could help them have a better understanding of their own illnesses?
Jarvis then says, sure, they can get access at no cost, but not immediately on their desktop screen at home. “Again I would take issue with that view. This is something that sounds like a very good idea, but there is a lot of information in the world which most of us need help with and to be talked through. You could get yourself in a terrible mess if you go and read this kind of information, which is pretty archaic, much of it.” (I’m guessing he either said or meant “arcane,” unless Wiley has unusually long publishing delays.) This stuff is too dangerous for lay people to read—and, to be sure, they can read it if they really want to. Better they should just use Google and believe whatever they find there, right?
Later discussions included pushing at price inflation and an attempt by at least one publisher to lay the blame for library budget problems at those damn librarians (albeit not in so many words) thanks to good old Andrew Odlyzko, who apparently still believes that the only worthwhile function of academic libraries is to move STM articles from one researcher to another. To wit, the problem with library budgets is “the library overhead.” “If you look at the whole system, two-thirds of the cost of journals is the library, not the publisher”—a true statement if and only if the only function of a library is to provide access to journal articles.
One interesting interchange came when a committee member noted that he was “old enough to remember when there were very few commercially published journals around, and when scientists and people in humanities published in the journals of their learned societies… The evidence shows that not-for-profit journals—and a lot of those are published by learned societies—are more highly cited than your journals are; but they are a damn sight cheaper. How can you justify it?”
Wiley’s Jarvis danced around the question in a remarkably convoluted statement—and, when the committee member pushed on the question again, asserted that scholarly societies “subvent their costs” through member fees. Amazing: Now the members of scholarly societies are subsidizing the journals, not the other way around!
Here’s another price point: For Blackwell, with a 15% profit margin, the total revenue per article came out to £1,250 ($2,262), a bit more than $1,500 but a lot less than $3,750—and Nature’s Charkin admitted that this figure was more or less accurate for Nature Publishing Group’s academic journals, albeit not for Nature itself.
Much of the rest of the hearing involved copyright. Here’s Dr. Jarvis’ definition of a copyright system: “the unimpeachable right of an author to publish their work wherever they want for no cost.” Hmm. I’d really like to publish Cites & Insights as an insert in TV Guide, since I’d reach a much larger audience; does copyright give me that right? (Bad example. How about AARP Magazine, with the nation’s largest circulation?) For that matter, I wasn’t aware that the author of a third-rate piece of pseudoscientific claptrap had the “right” to publish that work in, say, Science—but maybe I don’t understand copyright all that well.
Sure enough, later on we have the claim that publishers need copyright assignment to protect the authors from plagiarism and infringement. “If your author’s work is then stolen or changed, what publishers can do because of their scale and their reach is to do something about that. Individual authors would find it very difficult if their article was used and changed.”
The second session had as witnesses Crispin Davis and Arie Jongejan, both from Elsevier. Elsevier’s position paper includes many of their key arguments, but a few items may be worth noting. Elsevier feels put upon by this whole discussion of price increases: After all, the company is making much more of their backfiles available online and downloads keep increasing, so libraries are really getting bargains thanks to Elsevier’s beneficence. And, Elsevier says, every customer has a wide range of options: the Big Deal is just one of many choices.
According to Davis, every one of those 1.2 million articles “is in a respected journal, distributed to 250 countries round the world, reaching some 12 million scientists.” He’s making that claim on behalf of every single refereed STM journal in the world, an estimated 16,000. Every one respected, every one reaching 12 million scientists, every one distributed to 250 countries. Even those with three-digit circulations.
Naturally, Davis raises the specter that third-world institutions and authors wouldn’t be able to afford OA publishing; you’ll never see Elsevier recognizing that waivers exist. Davis also repeats the odd metric that, because Elsevier claims that UK institutions pay 3.3% of total subscription fees (which he rounds down to 3%), this means “we consume three percent of the world’s research.”
One questioner notes the charge that pharmaceutical companies are paying scientists to claim authorship for research articles they didn’t write. After Davis says that accepting such articles would be against their policy, Jongejan can’t help himself:
That is exactly the reason why we are concerned on the open access model.
He may not be saying in so many words that OA invites fraud and malpractice, but I can’t think of any other way to read that sentence.
Elsevier also claims to be a moderating influence on pricing in STM publishing—which might even be true, since its “modest” 6% to 7.5% annual increases may pose a problem for other STM publishers that want to emulate Elsevier’s 34% profit margin but haven’t yet raised prices high enough. One questioner notes that when Davis worked in consumer goods he would have given his eyeteeth for 6-7% annual price increases. Davis says that libraries look at increased usage and realize they’re getting “fantastic value for money.” That’s why no institutions are questioning the Big Deal: It’s such a bargain.
Elsevier’s position paper claims lots of evidence that it costs $3,750 to publish STM articles with proper peer review. They cite the Open Society Institute as a basis for that claim. On March 3, Melissa Hagemann of that society’s Open Access Project sent out a note to appropriate lists, saying in part:
Unfortunately, the [OSI] Guide has been misquoted to the effect that the authors estimate the cost of a published journal article at $3,750. Such a claim is incorrect. As the Guide text makes abundantly clear, the table containing this number serves only to illustrate a simple method by which such fees may be determined, and all the figures used in the illustration are identified as hypothetical.
Citing such a heuristic example will only be perceived as uncritical. As all the numbers in the Guide’s illustration are contrived and clearly identified to be so we obviously adduced no evidence to substantiate them. None of the numbers in the illustration are represented to be industry averages, nor can they reasonably be mistaken as such.
We ask that all those who have been misquoting the OSI Guide desist from doing so in the future.
Sally Morris of the Association of Learned and Professional Society Publishers (ALPSP) was unwilling to accept that statement, saying (in part):
I cannot imagine that the authors plucked a figure out of the air believing it to be misleading. Surely OSI/SPARC aren’t backing off this figure simply because publishers agree with them?
Until this discussion started, they had gone up considerably in many publishers’ eyes for having taken a much more rational approach to costs than had some other OA enthusiasts; it would be a pity to undermine this perception now.
To which David Prosser, director of SPARC Europe, responded (in part):
What I do know is that the guide authors did not want to mislead readers and that is why they described the figure as a ‘hypothetical example,’ a ‘sample author fee projection’ and a ‘simple illustration.’
They did not describe the figure as an ‘estimate of the average cost of publishing a paper across all journals’ or even as a ‘figure for a single journal.’ Unfortunately it is being quoted as such and that is why there was felt to be a need to issue a clarification.
Nobody at OSI, SPARC or SPARC Europe is ‘backing off’ from the figure in the business guide for the simple reason that none of us ever put it forward as an example of the real costs of publishing a paper!
Morris still wasn’t satisfied:
Illustrations illustrate something. The authors must have thought the illustrative figure was in the right ball-park, surely?
The table illustrates how to do the calculation—that’s all. It really is that simple.
Since Blackwell turns a profit with total revenue per article around $2,250—including all the costs of print subscriptions and licensing negotiations—it certainly seems reasonable to believe that $3,750 as a cost per article for purely electronic publishing was pulled out of thin air. Particularly since those responsible for that number say so.
I didn’t read the transcript of the second day of hearings, but would note a few items from Richard Poynder’s April 1, 2004 report at Information Today. Vitek Tracz of BioMed Central: “The role of publishers in the process of publishing scientific papers is wildly, incredibly exaggerated and overblown. We publishers are facilitators. It is the scientists who do the research, who publish, who referee, who decide.”
Harold Varmus (PLoS) challenged the suggestion that OA would reduce access. Maybe developing countries don’t have computers at every desk, but “every institution has a desktop computer, and you can download the appropriate articles.” Varmus would like to see public research funding made contingent on OA publishing for the results.
Nonprofit publishers—society publishers, by and large—are nervous. “They fear that any action by government to curb commercial publishers’ excessive profits could inflict damage on them.” Julia King (Institute of Physics) doesn’t believe OA as currently defined is a sustainable business model. The Royal Society issued a press release claiming that OA would require an extra $3.5 million annual funding.
Varmus accepted the possibility that OA was a potential threat to societies, which might have to adjust their business plans—that is, accept that library subscriptions should not be used to subsidize non-publishing activities. Varmus went a bit further: “Maybe there are too many societies.”
Some time after the March 1st hearing, BioMed Central issued a 12-page document stating 11 “myths” about Open Access. After the first-page summary, each page includes a one-sentence myth, a direct quotation arguing that myth, and a response from BioMed Central. Without repeating BioMed Central’s response to Elsevier’s position paper, a few of the myth responses are worth noting.
Ø Myth 4: Patients would be confused… (See the Jarvis comments above). BioMed Central: “This position is extremely elitist. It also defies logic. There is already a vast amount of material on medical topics available on the Internet, much of which is junk. Can it really be beneficial for society as a whole that patients should have access to all the dubious medical information on the web, but should be denied access to the scientifically sound, peer-reviewed research articles?... Patients suffering from diseases are understandably motivated to put in the effort to learn more about their conditions, as the success of patient advocacy groups in the USA has shown. Patients absolutely should have the right to see the results of the medical research that their taxes have paid for.”
Ø Myth 5: It is not fair that industry will benefit from Open Access. (Jarvis’ claim that corporate subscribers would be big winners with OA because they don’t produce many research articles.) BioMed Central: “To say that they do not contribute significantly in terms of publishing research is inaccurate. Industry publishes a significant amount of research itself, and also funds much research within the academic community that then goes on to be published.”
Ø Myth 7: Poor countries already have free access to the biomedical literature. BioMed Central: “The list of eligible countries has many notable omissions [such as] India, Pakistan and Indonesia… Countries such as Brazil and China…ar also excluded from the eligibility list, even for discounts. There is an obvious explanation for these omissions. These larger countries have significant research programs, so publishers can generate substantial income by selling subscriptions to them. It appears that traditional publishers will only offer Open Access to the developing world when they can be sure it won’t affect their profits… Many low-income countries have already started their own Open Access journals. Meanwhile, BioMed Central currently offers a full waiver of the article processing charge to authors in low and low-middle income countries.”
Ø Myth 9: A high-quality journal such as Nature would need to charge authors £10,000-£30,000 in order to move to an Open Access model. BioMed Central: This only applies to Nature itself, and even there it’s wildly off the mark. In practice, a significant fraction of Nature’s revenue is spent to commission and produce the rest of its content—News & Views, book reviews, commentaries, etc. Even if the research articles were freely available online, this other content would drive healthy subscription revenue. This “front matter” is far more widely read than the research articles; it’s not clear whether making those articles OA would have any negative impact on subscriptions. Nor would it be likely to eliminate or substantially diminish Nature’s impressive ad revenue (included in the per-article calculations!), yet those are assumed to fall to zero in Nature’s calculation. And, for that matter, the supporting claim that Nature rejects 9 out of 10 research articles is misleading: Rejected papers can be passed along to other Nature Publishing Group journals (e.g., Nature Medicine) with the referee work already done, sometimes allowing immediate acceptance. BioMed Central does the same with its top-tier Journal of Biology.
Ø Myth 11: Publishers need to take copyright to protect the integrity of scientific articles. BioMed Central notes that it’s “exceptional[ly] rare for a scientific publisher to use copyright law to defend the integrity of a scientific paper on behalf of an author. In fact BioMed Central knows of no situation where this has happened.” BioMed Central asserts, with some evidence, that the insistence on copyright transfer is to protect publisher profits by controlling access.
The first session had library people as witnesses—Lynne Brindley (British Library), Peter Fox (Cambridge), Frederick Friend (Joint Information Systems Committee, JISC) and Di Martin (University of Hertfordshire). Note the same caveat as above: These are uncorrected transcripts. I’m only covering roughly the first half of the first session.
The chair noted that librarians say there is a crisis in the provision of scientific publications, but publishers deny it. “Tell me the truth. Who is right? They cannot both be right.”
Friend: “There is certainly a crisis, in that libraries are not able to buy all the content that they need to supply their users, and the reason for that is that the periodical side of our budgets is rising much more rapidly than the cost of other information. That is the key factor.” (He did not note that STM journals increase in cost much more rapidly than journals in the humanities or other periodicals.)
Fox offered a specific example: Ten years ago, scientific journals took 25% of Cambridge’s materials budget; that has gone up to 33% “and rising,” which means “taking about half a million pounds a year out of the resources available for purchasing books and journals outside the scientific area…”
The chair, in a lively mood, continued:
Publishers tell us that the problem lies with libraries and their failure to promtoe themselves to university authorities. They are saying you are bunch of wimps really, I guess. Is that true?
Brindley began, “That is an unacceptable comment, lacking in any evidence, frankly.” She noted the efficiency gains in university libraries and BL’s need to cope with 43% inflation in journals over five years.
The chair again: “it is rumoured that a lot of people say that with the digital age you do not need those vast ranges of buildings that you have now, and the huge acreage they cover…” Martin notes contrary evidence: “We have seen an exponential rise in use of digital information, but we have seen no reduction in usage statistics of our buildings, or indeed in our book loan figures.”
How about “overheads”—which, again, seem to be taken as “everything except the journal subscriptions.” Trained reference librarians? Overhead. Selectors? Overhead. Buildings? Overhead. Martin notes that they’ve had a reduction in staffing even as they needed to add a new post to deal with licensing; she believes that overhead has been reduced. When asked if they could be more efficient, she says they’ve been being more efficient.
Brindley responded to a question suggesting all-digital publishing. “The evidence…suggested that we would be living, at least until 2020, with a very hybrid system of both digital and print publications… The evidence…is very much that people do still come in to consult material.” She also noted the problem of secure, long-term preservation of access to digital material.
One of the committee members asked why bundling (the big deal) is so unpopular with libraries. Fox: “Bundling requires us to buy journals that we do not necessarily want in order to acquire things that we do want, and is pushing more and more of our budget into the pockets of a smaller and smaller number of publishers… [It] is reducing the amount of money available for the output of the publishers that do not bundle.”
When asked about best and worst practices—and given the statements of publishers that they offer lots of flexibility—Martin noted that her experience does not show flexibility to be the case: “We find that publishers tend to approach us in terms of selling us a fixed product, and we have to negotiate very hard to get any flexibility within those products.” “The starting point is ‘take it or leave it.”
Later, continuing the question of which companies are best and which are worst, Friend noted that ALPSP members are easier to deal with. Two examples of publishers that “have been very difficult to deal with”:
One would be Elsevier, where last year we [JISC] spent about six months doing national negotiations, and we are still spending another four months in sorting out the details at proposal level. You agree [on] a national price of, say, 5 percent [above] what you paid last year; but then, when the detail gets down to local level, you find that the reality is very different. That negotiation has been extremely time-consuming, and is still not resolved for many universities. Another example I can give you is the American Chemical Society, where we have had great difficulty on long-term access.
Are the universities finding that learned society publishing is being squeezed out or forced into bundling by the commercial houses?
Friend: “The short answer is ‘yes.’”
Friend believes bundling is on the way out, “and I see it being replaced by open access” in a gradual change.
Then there’s the supposed competitive marketplace and what it means in reality.
Fox: “The problem is that we are in a monopolistic situation. If an academic needs an article froma particular journal, an article from a different journal will not do; and therefore they have to subscribe to that journal.” He went on to note that required copyright transfer maintained that monopoly. Then there’s the other problem: “The people that are paying for the journals, i.e., the libraries, are not the people that are [making] the decision whether or not they are purchasers.”
Discussion continued with problems of long-term access on electronic-only subscriptions (when you cancel a print journal, you get to keep everything you already paid for; not so with most ejournal subscriptions), licensing issues, limits on access, and UK tax issues.
Moving to open access, Friend has no doubts: “I am certain open access is viable in the long term.” JISC has a membership with BioMed Central to cover publication payment for authors in UK universities, and is also putting money into institutional repositories.
This group of witnesses included four professors and the chief executive of the Authors’ Licensing & Collecting Society. It became clear that at least some of the professors were editors for commercial journals and, from appearances, seemed to have been well coached for the session.
Professor Williams (tissue engineering and editor of an Elsevier journal), who does not believe he is conservative: “I do not see that there is any significant problem in S&T publishing at the present time. I think it is a very robust situation.” He noted nothing about library costs, just that he finds online access very good.
On the other hand, Prof. Crabbe (biology) claimed to speak for “colleagues not only in my own university but others” in being “totally supportive” of OA, and “at the forefront of open access for scientific community.” When asked about a split, Williams admitted there was—and, in stating his opposition, came down hard on OA: “Right now, in the way it is going—and I compare some of the journals which I see in my own area with that which I edit myself—I see a very big difference in quality. It is the quality of the science that is being published and the quality of the publication media that is of greatest interest to me.” Prof. Hitchin (math) chimed in, saying “up-front payments in particular are a big issue. They create large problems for certain disciplines in one of the open access models”—specifically, problems for independent researchers not supported by grants.
Prof. Fry (microbial ecology) saw “tremendous problems with the proposed models for open access”—both claiming that printed journals don’t really cost much more than internet-only journals (printing and distribution is “a very small part of the overall cost”) and that OA could hurt learned societies. He said that learned societies support the majority of conferences and congresses, “largely from their profits from publications.” In other words, the libraries are paying for the congresses indirectly—and, for Fry, this was the proper way of things.
As to bundling, Williams saw the need for commercial publishers to bundle—and Fry said, “Bundling has been extremely valuable for the users of journals because it has increased their access to journals enormously.” (Fry is publication manager for a society that publishes through Elsevier, and gains enormously from online income as a result of Elsevier’s bundling.)
There was more, most of which I’m not discussing here. The scholars didn’t seem to think turning over copyright was an issue at all (the licensing person wasn’t so sure). Institutional preprint archives were discussed, with some jabs from “the current system works just fine” people about “any old scientific garbage” being on such sites. One committee member suggested that OA might result in “pressure, direct, indirect, perceived or otherwise, on the journal to publish, with less stress on the quality, and, secondly, to speed up, possibly to the detriment of quality, the process of review and publication…” Prof. Crabbe said flatly that, if that happened, no one would publish in such a journal. “It only takes one journal, one paper, one bad paper in a journal for that journal to get a very bad reputation.”
Prof. Williams also owned up to saying there was no reason that all colleges and universities needed the same access to scientific publications—those not working at the cutting edge may not need access to the highest quality publications. And, for that matter, he believed that “[the vast majority of institutions] could not, in fact, understand what we publish, and I think one has to be very careful in determining policy on the basis that everybody should have free access to what we publish.” This variant on Jarvis’ earlier suggestion that it’s dangerous for the public to have access to arcane medical research—this time saying that even most academic institutions ought not to have such access—was very nearly the ending point.
But not quite. Mrs. Carr (from the licensing society) responded:
Well, words failed me there, for a moment. I think if somebody does not understand what they are reading then they do not understand it, but not to have access to it, if it is the author’s wish that they should, or indeed if the community needs it, must be a cause for concern. I am speaking there personally, in a sense.
And the chair closed with the kind of statement I just don’t think you get in Congressional hearings in the U.S. (more’s the pity):
I think you had better take him for a drink, Jane, and beat him over the head.
After which, of course, he thanked them all.
According to a February 24, 2004 Open Access News post, Rudy Baum offered an attack on OA in Chemical and Engineering News, entitled “The open-access myth.” In the piece (which I haven’t read directly, but I know that Peter Suber’s reporting is trustworthy), Baum states the “myth” as this: “STM publishers add little value to the research the publish and therefore should not charge institutions for subscriptions to the electronic versions of their journals, or, at the very least, they should provide open access to the public a short time after publication.” While it’s true that some open access advocates diminish the value added by STM publishers, that’s certainly not a general stance of the movement; instead, OA advocates believe that up-front payments are a better way to pay for the value that publishers add.
Baum isn’t clear “what advantage is conferred by shifting the cost of publishing from libraries to researchers.” Suber responds that open access itself is the benefit. From my perspective, an equally important benefit is that such a shift makes the costs evident to the researches, which might yield savings through shifts to more cost-effective methods, which in turn might free up library funds—which might then enable libraries to carry out their whole range of missions (only one of which is transmitting STM articles) effectively.
Baum again: “The open-access movement’s demand that an entirely new and unproven model for STM publishing be adopted is not in the best interests of science.” Suber notes that it’s certainly not entirely unproven, since OA journals have been around for considerably more than a decade—but another answer might be that most OA advocates do not demand that all STM publishing immediately convert to article-payment models.
The Stanford Report had two Vantage Points in its February 26, 2004 issue, both from very high-profile Stanford professors. Donald Kennedy, president emeritus, Bing Professor of Environmental Science, Emeritus, and editor-in-chief of Science, says that “subscription journals are here to stay.” Early on, he misstates the general economics of OA publishers:
The “open access” movement means that neither individuals nor institutions, like libraries, will pay to receive the journal through subscriptions to the print journal or site licenses for the online version.
But many, perhaps most, OA journals do charge for print subscriptions, a charge to cover the cost of printing and postage. Online access is free—and site licenses are irrelevant, thus eliminating one significant cost to the publisher.
Kennedy says he thinks “it is a good thing that we will now have both models in play.” He also says that the “author-pays” model is plausible in biomedicine—but “in less populated and well-supported fields, such support is far less readily available.” He goes on to assert that a rising tide of OA submissions will make the author-pays model more difficult to sustain. “That’s because it costs almost as much to reject a paper responsibly as it does to accept one. The higher the rejection rate, the larger becomes the expense budget that must be met from the fixed revenue from author fees.” He goes on to mention additional costs for journals that add news and perspective pieces to refereed articles—which, to be sure, are part of the priced value-added extras in some OA models. The problem with the assertion is that it assumes that a rising tide of submissions means lower quality submissions in general, thus a higher rate of rejection. If there are more submissions and they’re good quality, there’s no problem: e-journals don’t have inherent page limits.
Kennedy hopes “that Science will continue to serve, as it has for many years, the world’s largest general scientific society.” He also says he hopes PLoS succeeds—and finishes by saying “I know of no normative standard by which their [model] or ours can lay special claim to the moral high ground.” Personally, I see no plausible scenario in which Science will lose its stature or its subscription status; like Nature, it’s not a key part of the problem.
Patrick Brown, professor of biochemistry and cofounder of PLoS, offered a piece entitled “Free online scientific journals make sense.” He notes that the U.S. government spends more than $50 billion a year on nonclassified research and explicitly raises the issue of lay access to the results of that research:
But if your mother learns she has breast cancer and desperately wants to find what researchers have discovered about her disease, or when your daughter in high school reads a story in the New York Times about the latest research on climate change and wants to see it with her own eyes, they face a perverse and unnecessary obstacle. They, and countless others around the world who would benefit from timely access to scientific and medical knowledge, cannot freely access the published results of research financed by their own tax dollars.
He goes on to complain that, although Stanford has access to most of the scientific literature, they can’t “Google” the millions of scientific articles. He goes on to call the traditional STM business model “a vestige of an era when printing articles in paper journals and transporting them in trucks and boats was the most efficient way to disseminate new scientific discoveries and ideas.” Now, he says, research articles are “delivered much more efficiently and conveniently via the Internet” and concludes that charging for access is “therefore no longer economically necessary, rational, or fair.”
Brown throws in one argument that makes me cringe, since I believe its implications are clear:
An “open access” system for scientific publishing will not entail new expenses, nor should it place a financial burden on the authors. The governmental and private institutions that finance the research already pay most of the costs of scientific publishing indirectly—through the funds they provide to research libraries. These same institutions would accomplish far more with the same money by phasing out subscription payments to restricted-access journals and, instead, paying for open-access publication of the research they support.
There it is: Don’t add OA fees to research grants; take the money away from the libraries.
This March 2 issue begins with an “objection-reply” on the theme, “Whether the upfront payment model corrupts peer review at open-access journals.” It’s an excellent, detailed discussion. If you don’t normally read SOAN, I’d suggest you get this issue (www.earlham.edu/~peters/fos/newsletter/03-02-04.htm) and read the first three pages. Much as I love to poke at weak points and overstatements in OA advocacy, I find no fault with anything in Suber’s multipart refutation of the “corruption” suggestion, one of the most common attacks on OA.
The next article offers “Top 10 priorities for the OAI community.” I don’t talk much about OAI, partly because of personal issues and the incredibly cavalier attitude of some leading OAI proponents toward libraries, but it’s a good list to consider.
This viewpoint, which appeared in EMBO reports 5:3, is a useful overview of some of the issues regarding OA and OAI. It’s worth reading and should be readily available online. I think I detect a slant, but maybe I’m wrong, so I’ll just note that Grivell considers viewpoints on several sides of the issues and writes clearly.
This Nature “web focus” began April 2 with an introduction from Declan Butler, European correspondent for Nature. You’ll find the whole thing at www.nature.com/nature/focus/accessdebate/, and I suspect new contributions will be added to the discussion after this issue appears. I reviewed the introduction and the first twenty documents in the debate, but only mention a few of those documents here. (Two of them were the Kennedy and Brown Vantage Points from Stanford Reports, discussed earlier in this section.)
Butler’s introduction frames the issues fairly and includes two paragraphs I find particularly telling:
One jarring aspect of proposals to reform scholarly publishing is that, all too often, they implicitly consider ‘journals’ as a single homogeneous entity, to which one universal publishing model can be applied. On the contrary, diversity is everywhere. In any discipline, journals range from high quality ‘must reads’ with high rejection rates—which in turn result in higher costs per published paper—to publications which add little value to the articles as submitted, and are read by few apart from the authors themselves.
Journals are also published by a range of patrons, from individuals, and commercial publishers, to learned societies who use publication revenues to support their community in other ways. Likewise, a journal might be run largely by scientists working for free, or by professional editors. Some are electronic only, some have print editions. The list goes on. Any discussion of publishing models must surely take into account this heterogeneity. There is no one-size-fits-all solution.
One caveat, one that I would apply to the whole discussion about much higher costs for journals with high rejection rates. For most refereed journals (STM or otherwise), referees work for free. For many refereed journals, so do the editors. With electronic submission and routing, it’s reasonable to believe that the actual costs to the publisher for rejected articles should be very low. Most of the costs associated with published articles are for copy editing, preparation of graphics, preparation of metadata, and conversion of the article into a form appropriate for publishing (and, for print journals, printing and postage). But the cost per published article for all of that work is the same, regardless of whether the journal publishes four of every five articles submitted or only one of ten. If $1,500 is a plausible publication cost (and that seems high), wouldn’t $20 be a sufficient submission cost to make an efficient all-electronic refereeing system work?
Here’s the next-to-last paragraph—and you probably already know my answer to the questions: Yes and Yes.
Could these costs be paid for in other ways than the traditional reader-pays subscription model, and under what circumstances? Or will the outcome be a mix, with open access prevailing for certain types of publication, and elsewhere, market demand for greater access to the literature driving imaginative deals between publishers and libraries to make such access more affordable.
ALPSP is clearly in a conflicted position on this whole issue, as this piece shows. It begins: “It is no accident that much experimentation with the Open Access journals model, where costs are covered by payments made on behalf of the author rather than on behalf of the reader, is being carried out by learned societies and other not-for-profit publishers.”
She goes on to express cautions, then turns to the critical issue—and the one where I’m least sympathetic to ALPSP and learned societies in general. She provides an argument in the second paragraph that I find remarkable and, at least for the U.S., wholly disingenuous.
Where [societies] do make a surplus—and not all do—it is typically invested in such activities as promoting public education, subsiding conference fees or membership subscriptions, and providing research grants and bursaries. If, as seems likely, the author-end cost-recovery model were further to reduce surpluses—which are already modest compared with some commercial publishers’ profits—these other services would inevitably suffer, and it is arguable that both science and society would be the poorer.
Some people argue that it is not right that library budgets should pay for societies’ other activities. But it is perhaps fair to ask where those library budgets come from: ultimately, they come from the taxpayer, meaning, primarily, business. If, on the other hand, these society services were no longer subsidized, who would have to pay? In many cases, it would be the individual scientist—paying more for society membership, more on conference fees and travel. The alternative of more direct subsidy from taxpayers’ money, whether to the societies themselves or to the individual scientists, might have considerable drawbacks in terms of independence and academic freedom.
Say what? Now, ALPSP works in the UK, and maybe the primary source of all library budgets in the UK is business taxation. I don’t believe that to be true in the U.S., and especially not for private universities. In any case, it’s a shell game: The thesis that it’s acceptable for societies to subsidize their own operations on the backs of libraries, while not rallying to improve library budgets. I argue that it is not right that library budgets should pay for societies’ other activities. Period.
She calls the essay “Open Access: yes, no maybe.” But what you’ll read is her assertion that even PLoS is “charging substantially below the actual cost of publication” with its $1,500 fee. Thus, for Elsevier to “experiment,” they would have to charge the “real cost of publication”—that is, the entire amount that Elsevier makes from print and electronic publications, including profit—and “we would be endorsing a model that at the moment is unsustainable.”
“We think that the so-called lack of access is a red herring.” After all, there’s always ILL—but not to the electronic version, given most licensing restrictions. She raises the usual hobgoblin of editorial independence, those who lack the funds for publication payments, and the disruption of a model that “has evolved over centuries” (but has become primarily commercial over a very few decades). It’s a classic Elsevier piece, albeit slightly less heavy-handed than some.
After several other commentaries (including the two republished from Stanford Report), Worlock weighs in with “Open access and learned societies,” an unusually long piece for this debate. It has some gems: A survey found that 80% of scientists belong to at least one learned society. That means 20% of scientists find no learned society worth joining: A startling figure. There’s another estimate of how many refereed journals are out there: this time it’s 21,000. And one poll found 87.5% of responding learned societies getting a surplus from their publishing activities. If that’s true, Morris’ “and not all do” is true, but just barely.
Then things get weird. There’s an argument that one reason for membership is that scientists are “paying for exclusivity and receiving it” through their membership subscriptions to journals, and that membership would be less valuable if everyone has access. There’s a new counter-OA argument: Scientists want restricted access to articles!
Another numbers paragraph notes 765 journals in the Directory of Open Access Journals as compared to 21,000 journals in Ulrich’s—but weakens that by noting that the 21,000 includes monographic series. (DOAJ now lists more than 1,000 OA journals, so it’s at least 4% of the journal world at this point.)
Unfortunately, Worlock goes on to quote at some length the Open Society Institute “$3,750” cost. She then goes on to find that ALPSP benchmarks show a median cost of £450, which, at $815 or so, is considerably below PLoS’ $1,500 charge.
After that, there’s the usual anti-OA confusion: Societies will wish to continue print (and can, since there’s no bar to charging cost recovery for print subscriptions). The majority of journal authors aren’t dissatisfied with the present system (but libraries are going broke). Humanities research is cheap, making publication charges difficult to cover (but humanities journals aren’t the problem, as they’re relatively inexpensive in any case). “The move to open access is still very much an untested concept,” after more than a decade of journal publishing.
“Learned societies exist to foster and disseminate knowledge, and undertaking an open access experiment which could threaten the economic survival of the society would be unlikely to be seen as an acceptable risk.” Doesn’t that reason for existence argue forcefully for experimenting with OA?
His piece is “The myth of ‘unsustainable’ Open Access journals,” and he argues forcefully that the current STM publishing system is “unsustainable.” He quotes one of UC’s academic senates to that effect. He also notes that specific publishing-related services can be unbundled, each step charged for separately. This raises interesting possibilities: The print version of a journal might not come from the original publisher at all, for example. (The Creative Commons BY license required by some OA publishers makes this straightforward: Any other agency is legally free to reprint those articles for a price.)
Walker discusses hybrid publications, “open access by the article.” These are journals that charge for subscriptions, but if an author pays a publication fee, the online version of that article will be free for all use. For the Entomological Society of America, which began using this model four years ago, 62% of published articles were open-access last year; for another society, the figure reached 66% by 2003.
The publishing charges are very reasonable. The American Society of Limnology and Oceanography charges $350 for a 10-page article; ESA charges $124 for an 8-page article. More recent attempts at the hybrid models involve charges ranging from $995 to $2,160, “but I would argue…that these fees are unnecessarily high.” He believes the hybrid model offers many benefits, including increased loyalty from members—that is, scientists who want broad access rather than exclusivity.
“Is free affordable?” To JCI, yes. The journal’s been around since 1924; when it went online in 1996, it immediately provided free and unrestricted online access. They’ve found that the policy is one “that the journal can afford.” It’s a detailed paper, including what’s involved in publication, the percentage of rejections (almost 90% for JCI), and other aspects.
I’ve omitted quite a few for various reasons. We have the proud open access publisher saying, well, no, it’s not really sustainable and we’re pulling back. We have Thomson ISI saying that, as far as they can determine, there’s no correlation between the publication model and the journal’s impact: OA neither helps nor hurts measurable ISI-style impact—but it’s awfully early to say.
All in all, a fascinating ongoing debate.
All original material in this work is licensed under the Creative Commons Attribution-NonCommercial License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc/1.0 or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.