Tuesday, November 06, 2012

Interview with the Scholarly Kitchen’s Kent Anderson


From the moment it was conceived, PubMed Central was controversial, and it has remained controversial ever since. The brainchild of Harold Varmus — the then director of the US National Institutes of Health (NIH) — the idea for PubMed Central was first mooted in 1999, but originally called E-BIOMED.
Kent Anderson

When Varmus published his initial proposal, publishers quickly concluded that it posed a serious threat to their livelihoods. Specifically, they were convinced that, if E-BIOMED went ahead, the US government would become a publisher, and they would be disintermediated as a result. So they launched a firestorm of protest.

Their protest delivered results. When the service was launched eight months later it had been re-branded as PubMed Central, and was a pale shadow of the revolutionary new “electronic publications” system that Varmus had envisioned. Significantly, Varmus had had to concede that publishers would have final say on whether the papers they published were put into PubMed Central — and most publishers chose not to participate.

Varmus later conceded that he had been naïve not to have anticipated the furore. “I must have known that I was not going to be at NIH for much longer,” he joked to New Scientist in 2003, “because this caused a tremendous political argument: what the hell was I trying to do to destroy the publication industry.”

Nevertheless, it was soon apparent that NIH did not intend to give up on its dream of having a large free full-text archive of biomedical and life science papers, along the lines of the physics preprint server arXiv. This resolve was only strengthened when, two years later, the Open Access (OA) movement came into being. The tide of history, it seemed, was flowing in NIH’s direction.

Socialized science


In 2004, therefore, Varmus’ successor at the NIH Elias Zerhouni published a draft policy entitled “Enhanced Public Access to NIH Research Information. The aim was to persuade researchers to post their papers in PubMed Central.

One again, publishers objected. In a 2004 editorial penned in the Chemical & Engineering News, for instance, C&N Editor-in-Chief Rudy Baum complained, “Zerhouni’s action is the opening salvo in the open-access movement's unstated, but clearly evident, goal of placing responsibility for the entire scientific enterprise in the federal government's hand. Open access, in fact, equates with socialized science.”

For publishers the nightmare scenario was that research funders would gradually squeeze them out of the process of disseminating research. After all, the papers published in scholarly journals are written by researchers, and the peer review process is conducted by researchers — at no charge to publishers. In the age of the Internet, some were beginning to conclude, the need for publishers was beginning to look moot. At the very least, they reasoned, the role that publishers play could be reduced in an online world. This would help ease the burden on the public purse, which many believed was being gouged by publishers charging excessive journal prices.

But this time publisher opposition did not succeed. In May 2005 the NIH introduced its Public Access Policy. While this was initially only a request that researchers post NIH-funded papers in PubMed Central, it was later upgraded to a demand, and today researchers are required “to submit all final peer-reviewed journal manuscripts that arise from NIH funds to the digital archive PubMed Central upon acceptance for publication.”

Publishers continued to mutter about the Public Access Policy, but they had to learn to live with it. And faced with growing calls for research papers to be made freely available, many also began to experiment with OA.

But last year their fears of being disintermediated were reignited, when three large research funders — the Wellcome Trust, the Howard Hughes Medical Institute (HHMI) and the Max-Planck Society — announced plans to launch their own OA journal, eLife.

Since the funders indicated that they would not initially be charging a publication fee, publishers complained that it was anti-competitive. Writing on the Nature newsblog on the day of the announcement, for instance, Nature’s Declan Butler commented, “[F]rom what we know so far from today’s press conference, this new journal appears to offer few tangible novel innovations and may indeed disrupt the thriving open access environment. Its decision not to charge author fees, at least in the journal’s short and medium term, in fact could risk setting back the cause of open-access publishing by undermining — through what might be considered unfair competition — economically successful open access publishers”.

And when at the end of October, eLife announced that it had published its first few articles, critics were angered to see that the papers had been hosted not on the journal’s own website, but on PubMed Central.


Active critic


The most active critic of these latest developments is Kent Anderson, who has responded by penning a series of hard-hitting posts on the Scholarly Kitchen blog. Anderson is a former executive director of The New England Journal of Medicine, and currently CEO/Publisher of The Journal of Bone & Joint Surgery. He is also Editor-in-Chief of The Scholarly Kitchen blog and, as he puts it, “one of the more vocal sceptics of open access”.

In his several posts Anderson has argued that — contrary to its own claims — in hosting eLife papers PubMed Central is now clearly acting as a primary publisher, and inappropriately competing with private publishing technology companies as a result. Moreover, he added, by favouring eLife, PubMed Central has drawn attention to the fact that it routinely gives OA publishers preferential treatment when listing journals in the all-important PubMed citation database, and is thus discriminating against subscription publishers.

Anderson has also suggested that the preferential treatment given to eLife is evidence of “cronyism” — a charge he makes partly on the grounds that a Wellcome Trust employee serves on the PubMed Central National Advisory Committee (PMC NAC).

As I discovered when I contacted him, eLife’s Managing Executive Editor Mark Patterson rejects the last charge. He also dismisses Anderson’s claim that PubMed Central is acting as a primary publisher. And Wellcome’s Head of Digital Services Robert Kiley insists that neither he (when he served on the PMC NAC) nor the Trust’s solicitor Chris Bird (who currently serves on it) have ever asked for any kind of favourable treatment for Wellcome and/or eLife whilst serving on the Committee.

Nevertheless, Anderson maintains, questions still remain, and he intends “to continue to poke around” in pursuit of answers.

Meanwhile David Lipman, the Director of the National Center for Biotechnology Information (which runs PubMed Central), declined to comment on the matter. “At this time, we don’t think that engaging Mr. Anderson in a public rebuttal is productive,” he emailed me.

Below I publish an interview with Anderson that was conducted by email over the last few days.

The interview begins …


RP: In a couple of recent posts on the Scholarly Kitchen blog (here and here) you expressed concern about the way in which you believe that eLife — the new Open Access journal being launched and funded by the Wellcome Trust, HHMI and the Max-Planck Society  — is being given preferential treatment by PubMed Central. Specifically, it is your belief that the normal selection process that journals are supposed to go through before they are listed in PubMed Central has been waived for eLife. If you are right, then why should the research and publishing communities be concerned about this, and what would you like to see PubMed Central do to correct the situation? Do you think it will?

KA: PubMed Central is a government initiative, part of the US National Library of Medicine [NLM] and the National Institutes of Health. These agencies are supposed to act fairly and uniformly according to rules they make obvious to everyone. In the case of eLife, the normal rules were not followed — eLife was accepted into PubMed Central before it had demonstrated any capacity to publish independently.

This is against the stated policies of PubMed Central, as expressed both on their sites and in email communications with other publishers before and during my initial examination of the situation. All of this is documented in my first blog post.

For publishers, this is frustrating because they abide by rules they feel are uniformly applied, so to see a new and unproven journal cut the line with the clear assistance of PubMed Central only adds to their worst fears — that something in the leadership at PubMed Central is bending the rules to suit a larger agenda, that there isn’t a level playing field, and that there’s a hidden agenda.

I see no signs that PubMed Central plans to correct the situation. They have just published the second round of eLife articles, again as the primary publisher. They have not explained themselves publicly beyond responding to my direct questions.

RP: For non-scientists, can you explain why being listed in PubMed is so important?

KA: Most journals aren’t viewed as legitimate until they are indexed in PubMed. It’s that simple. There’s an assumption that the NLM imposes rigorous standards — demonstrated editorial and publishing capabilities, for example — before a journal can be listed in PubMed. This is no longer true. It is true for being listed in MEDLINE, which gets us to the fundamental confusion in the market between PubMed and MEDLINE: for many authors, editors, and readers, MEDLINE and PubMed are viewed as one and the same.

Initially, they were. PubMed, however, has become something different. It is not manually curated, and you can shortcut into it by putting free content into PubMed Central.

RP:  Just to clarify: PubMed Central is the free full-text archive of biomedical and life science papers, and PubMed is a citation database. Is that correct?

KA: Correct.

When PubMed Central gives quick access to PubMed listing, they are putting their thumb on the scale for journals that provide free access — not because of editorial quality. This adds to the impression that they have an open access bias. Their behaviour would be different if they had a quality bias, for instance.

In the case of eLife, they didn’t even wait to see if eLife could independently publish articles before facilitating their entry into PubMed. eLife received this while also receiving primary publishing services from the US government.

Primary publisher


RP: Right, as you point out, eLife’s first few papers have been published directly on PubMed Central. This you believe underlines the fact that PubMed Central is now acting as a primary publisher. If that is correct, why in your view is it significant and to be deprecated? What would you like to see PubMed Central do to correct the situation? Do you think it will?


KA: PubMed Central has always stated, and still does so in multiple places in its online descriptions of itself, that it is not a publisher. This has never been true — it has always been a secondary publisher of primary content. Because they were a secondary publisher, most of us just let it slide — not all, but most. There have been editorials complaining about PubMed Central as a publisher in some high-profile journals, even in its secondary role.

A recent analysis showed that PubMed Central is stealing traffic from participating journals, which has direct financial consequences for them, from advertising impressions to institutional usage reports, both of which depend on traffic. So I think controversy over PubMed Central’s role as a secondary publisher is not dead, by a long shot.

However, PubMed Central has more recently become the primary publisher of journals, including the Journal of the Medical Library Association. And, with eLife, they helped launch a new journal, a major crossing of many lines in the journals ecosystem — government subsidization of a business launch, for example.

To correct the situation, PubMed Central should take the eLife articles down, issue an apology, and begin enforcing its stated policies uniformly.

Another option would be for them to change their policies to allow journals to use them as a hosting platform, and state expressly who can do that, what services they provide, and so forth. But to have it happen in the shadows is improper.

However, I don’t expect them to change their behaviour in the short-term. They seem to feel they are immune to criticism.

RP: I understand eLife has made it clear that this is only a temporary solution, until its own publishing platform is up and running. If that is correct, does it lessen your concern in any way?

KA: What strikes me about this question is that you, like eLife, are putting eLife in the driver’s seat. They said this, too — that they chose to publish on PubMed Central, as if the US government was providing a free publishing platform, ala WordPress, which they chose to avail themselves of. That’s disingenuous, and not how anybody else feels the NCBI (which runs PubMed Central within the NLM) should operate.

So, no, my concern is not lessened because this is “temporary.” In fact, that’s part of my concern. If this was temporary, and because eLife has a contract with a proven publishing platform vendor to launch in 2013, why did PubMed Central feel the need to deviate from its policies to provide this exceptional bridge to a journal start-up?

What was different about it? And was that difference driven by the long-standing relationships between PubMed Central and Wellcome Trust? There are many signs this is the case, including an exceptional meeting in June 2012 in which eLife was the featured project — something that has never happened in the minuted history of PubMed Central — during which the head of the NCBI gave eLife business advice. Clearly, boundaries aren’t clear to those meant to enforce them at NLM.

RP: I emailed Mark Patterson, eLife’s Managing Executive Editor for his views on this matter. In fact, he denied that PubMed Central is acting as a primary publisher for eLife, saying, “Publishing encompasses the entire process from submission and editorial assessment, to content processing and online dissemination of content.” And he added, “Nor is PMC the only venue for current eLife content; as soon as open-access content is available at one site, it can be made available at many more. For example, eLife content is also available at EuropePMC, github, Mendeley and Scribd, and will shortly be available at the eLife journal website hosted by Highwire. 

KA: If you accept Patterson’s argument that you have to fulfil all the requirements of his definition to be a publisher, then nobody is publishing eLife — because eLife is not doing the content processing and online dissemination. So his argument elides the facts, and doesn’t make intrinsic sense.

The links from the eLife site go directly to PMC, which is where they are published. The attempt to say that because the articles are OA means they have no primary source and are freely circulating is weak and nonsensical. A lot of subscription-based articles are circulating freely, as well — in fact, probably more are. That doesn’t mean they don’t have a home or source.

There are degrees of publishing, and I’ve tried to make it clear that these are in effect. PMC has been a secondary publisher of content for most of its existence. Now, with at least three journals, including eLife, it is the primary publisher of that content. It disseminates the upstream, authoritative version.

Cronyism?


RP: You believe that eLife has received preferential treatment due to what you call “cronyism”. I think this is more a suspicion than a demonstrated fact, but what is your reason for suggesting that PubMed Central has succumbed to cronyism and, if you are correct, what in your view are the implications?


KA: Well, I’ll let the facts speak for themselves, and we’ll let your readers determine whether this amounts to cronyism. First, cronyism is defined as, “partiality to long-standing friends, especially by appointing them to positions of authority, regardless of their qualifications.”

In the case of eLife, the friendships are there — Robert Kiley, who promotes eLife in publishing venues on behalf of Wellcome Trust and is an employee there, served on the PubMed Central National Advisory Committee for many years; one of Wellcome Trust’s senior attorneys is currently on the same committee; and Mark Patterson, the managing executive editor of eLife, was an invited guest at the June 2012 meeting of the same committee, a committee that has never had an unlisted outside guest address the meeting about a new initiative.

This is where the partiality begins, in my opinion. Also, as I mentioned above, the head of NCBI, David Lipman, proceeded to give eLife business advice in that same meeting, clearly a loss of perspective and another sign that the two entities are acting as cronies, beyond their proper roles and relationship.

Then, within four months of that meeting, PubMed Central becomes the only place where eLife content is published, despite policies at PubMed Central that would seem to make this impossible — requirements to have published at least 15 articles independently before any evaluation can begin, for instance. I think all signs point to cronyism.

The implications of this are irksome and problematic.

First, it means that publishers who have looked to the NLM to provide a level playing field and an objective process for registering journals and applying various NIH mandates know that in fact the NCBI and NLM waive their own rules when it suits them. This makes their pronouncements about their process ring hollow, and undercuts their authority.

Second, it indicates there are deep management issues at the NLM. For instance, when I asked to know the date that eLife applied for inclusion in PubMed Central, I was told that the NLM doesn’t track these processes, and didn’t know the date. That’s ludicrous, when they impose rules on publishers, who abide by them, but then don’t track their own processes. Again, their credibility as an authority in the field is at risk.

Finally, it means some publishers are “in” and some are not — that there is an inside track based on favouritism and relationships, which suggests at least some personal corruption of the process — that management has become too invested in a certain outcome or vision of the future to act as a fair and disinterested arbiter.

RP:  I also contacted David Lipman and Robert Kiley by email and asked them to respond to your concerns. Lipman replied, “At this time, we don’t think that engaging Mr. Anderson in a public rebuttal is productive.” Kiley responded, “With regard to whether eLife was given any special favours to get its content into PMC, I really don’t know. My sense is that eLife had some good scientific content it wanted to make available, but doesn’t yet have its Highwire site fully up and running. (I understand that this will go live before Christmas).” Kiley also confirmed that neither he nor Chris Bird have ever asked for any kind of favourable treatment for Wellcome and/or eLife whilst serving on the PubMed Central National Advisory Committee.

And this is how Patterson responded to the suggestion that eLife has received preferential treatment, “I can only speak from my experience at eLife and at PLOS. In both cases, PMC has been responsive and effective. Based on the experience of other publishers, such as Marty Frank’s, it seems PMC takes into account the reputation of the organizations behind new open-access projects when assessing their inclusion.”

KA: Well, I think this sounds like equivocation and evasion. Conflicts of interest are a lot about appearances, and we’ve all seen denials that break down later as more evidence comes out.

I’ve heard through the grapevine that there’s more than meets the eye here, and from some well-placed sources. I’m going to continue to poke around as time allows. We’ll see at the end of the day how this all plays out.

The facts remain that eLife received unprecedented access to the PMC National Advisory Committee, that it used PMC as its primary publisher a few months later, and that Wellcome has a strong presence on their board. If it looks like a duck, walks like a duck, and quacks like a duck, it’s fair to conclude it’s a duck.

I find it interesting that instead of engaging in a public discussion, the civil servant David Lipman chooses instead to stonewall. Where is the investigation? Why isn’t he providing some facts and specific information in response to specific questions?

I can’t get a straight answer from anyone in the organizations involved about who approved this, who was involved in the decision making, and when it was approved.

So far, we have David Lipman denying he knew anything about it, big names at Wellcome denying they had anything to do with it, and Mark Patterson saying it was just PMC being “responsive,” but in a manner that blew past all their stated requirements and processes, which they are still imposing on other publishers, OA or not.

Somebody was in the soup for this. Who was it? Who made the decisions? Who told people at eLife to send content to PMC? Who told people at PMC to prepare a place for eLife content? When did they do it? The lack of specifics, answers, and accountability makes me wonder all the more.

Preferential treatment


RP: You have also expressed concern about the way in which you believe PubMed Central is now competing with private technology providers. Why do you believe this to be the case, and for those not versed in the US cultural and legal environment, can you explain why you feel this to be an important issue?


KA: PubMed Central currently serves as the primary publisher of three journals. This means that these three journals don’t have to pay for what everyone else has to pay for — hosting services, XML conversion, and so forth. They are getting a free ride courtesy of the US government.

In the US, and probably in many other countries, this doesn’t comport with the role of government, which is to provide a framework for business but not to compete in businesses that are already in sufficiently competitive markets.

In this case, the contract eLife has with HighWire Press could be in jeopardy if eLife finds the PubMed Central solution to be adequate over the next few months.

In the case of the other two journals, the US government is giving them a free ride. That’s wrong. There are many competitive providers of these services in the publishing market, some of which are free or very affordable.

The US government shouldn’t be a competitor to these companies. Publishers should have to pay to play. Nobody should get a free ride on the taxpayer’s dollar. Ironically, that’s one of the fundamental complaints that catalyzed the open access movement.

RP: eLife aside, you believe that PubMed Central is giving preferential treatment to OA journals in general. However as Patterson points out, a comment on your first post by the Executive Director of the American Physiological Society Martin Frank suggests that this may not be quite as you suggest. Could it be that while eLife may indeed have been given preferential treatment, the incident is not necessarily evidence of a bias towards OA? In other words, while it may demonstrate favouritism, it may not be evidence of a predisposition towards OA (and presumably a prejudice against subscription publishers) at PubMed Central.

KA: I think you misread the comment from Marty Frank. He noted that until he mentioned that a new journal was being published as an OA journal, he couldn’t get PubMed Central’s attention. Once they realized this, they greased the skids. That showed clear OA bias.

However, you’re correct that it’s not uniform. It does exist, and there are signs of it, and it’s the only bias they seem to show occasionally. How they created the fast-track into being listed in PubMed is another sign of bias — it’s based on being OA, not on being a great journal.

What’s more worrisome is that their management practices are so hard to pin down. Do they require what they say they do? Are they consistent in how they treat journals? Their inconsistency is the bigger problem, but it leans OA, because when they make exceptions, it’s always for OA journals.

I think these most recent events should have all publishers concerned, whatever their business model. Clearly, there’s a lack of rigorous process at NLM when it comes to how journals are indexed, measured, and accepted. And while eLife was getting a big assist, other OA journals from other publishers were being given the straight arm. That’s not efficient or effective management.

Intertwined issues


RP: I think there are several intertwined issues here, including issues of transparency and consistency, the role of government agencies, and indeed of the role of government itself. I doubt anyone would disagree that it is important for government agencies to behave in a transparent and consistent way, and to operate a level playing field. So if PubMed Central is not being transparent, or if it is tilting the playing field in some way, I am sure people would want to see you challenge the organisation. But can you clarify your views on PubMed Central’s proper role? Is it your contention that it is inappropriate for PubMed Central to act as a publisher at all, or simply that if it is doing so then it needs to acknowledge as much, and to offer its publishing services to everyone, and on a paid-for, not subsidized, basis?

KA: I think it’s becoming clearer with each passing year that PubMed Central is doing more than it should. It’s interesting, because most of the publishers I knew sniffed this out early — that PubMed Central would be competitive. Now, evidence is emerging that it is taking away customers, page views, and traffic. PubMed Central has been diligent about executing this strategy, as well, putting their links and their links only on the main search results, for instance.

They are driving traffic to their version preferentially through their proprietary interface. That’s way over the line. They are being competitive, and they are manipulating their interfaces to their advantage.

I personally think that PubMed Central could shift into a dark archive, and keep the same deposit requirements and toll-free access requirements in place. This would rectify the problem of competing with traffic, accomplish the same access goals, leave the publishers in a better situation from a traffic, branding, and audience standpoint, save the government money, and preserve the literature.

I also think PubMed Central should not tie indexing in PubMed to inclusion in PubMed Central in any way. PubMed is an index that should be based on passing some quality threshold, not an index partially based on publishing policies.

As for providing publishing services, I don’t think the US government has any place scaling up a commercial publishing platform. But let’s not forget, “free” is a price. Right now, they are offering a very limited number of customers, for unknown reasons, free hosting. That is a commercial venture, just with the price of “free.” 

There is no need in this market for the government to be doing this. Other comparable journals fund their own existence. Why do these get government assistance?

RP: In terms of the proper role of government. If, as OA advocates maintain, private industry is operating in such a way as to restrict access to publicly-funded information, and if it is resisting becoming more open, then is not the duty of a government to intervene — in the way, for instance, that the UK government has done. As UK Minister of State for Universities and Science David Willetts put it to the Publishers Association in May, “Our starting point is very simple. The Coalition is committed to the principle of public access to publicly-funded research results. That is where both technology and contemporary culture are taking us. It is how we can maximise the value and impact generated by our excellent research base. As taxpayers put their money towards intellectual enquiry, they cannot be barred from then accessing it. They should not be kept outside with their noses pressed to the window — whilst, inside, the academic community produces research in an exclusive space. The Government believes that published research material which has been publicly financed should be publicly accessible — and that principle goes well beyond the academic community.” The upshot is that in future all UK-funded research will need to be made OA. Do you not feel that the UK government has intervened in the market appropriately here — with the aim of ensuring that UK-funded research is made publicly available? And is not the NIH’s Public Access Policy equally appropriate?

KA: Taxpayers, at least in the US, have paid for the research reports required by grants the government gives. These reports can be posted on government Web sites, and nobody is objecting to that. In fact, I encourage it, and have publicly said I would heartily join in a movement to get the government to force 100% of its researchers to comply with these requirements. So far, only a fraction do, well below 50%.

The government should focus on enforcing its own mandates with regard to taxpayer-funded reports and on doing a good job of posting these taxpayer-funded reports for discoverability.

What taxpayers haven’t paid for is the peer-reviewed literature. This is often portrayed as a minor difference, but it’s actually a huge and expensive difference. One of the biggest costs is the cost of rejection. This can run into hundreds of thousands of dollars for a good journal.

What does rejection do? It helps papers find their proper homes — the audience, the editorial brand, and the venue that they’ll do best in. A recent study confirmed that this works well for authors, boosting their impact. But it’s an invisible cost.

There’s also a lot of disingenuity in these arguments. For instance, while governments and universities will argue that information should be refined, filtered, and published at no or low cost, and made into a public good, what about the other IP emanating from taxpayer-funded research?

Well, here in the US, other IP coming from taxpayer-funded research is not made freely available to the public. It is patented. These results are kept from the public, exploited for millions of dollars by higher education institutions and researchers, and it’s all based on taxpayer-funded research.

Why is it defensible, even encouraged, for higher education to benefit in this way from taxpayer-funded research, but discouraged and indefensible for publishers to be paid for vetting, publishing, archiving, and disseminating high-quality information, which is distinct from the research reports the taxpayers paid for, and actually not taxpayer-funded?

There is also the question of whether the public really benefits from all these papers being freely available. I have an advanced university degree in business and an undergraduate in English. I once earned my living, for many years, editing medical texts. I’ve been published in some of the top journals. I’m one of the best-educated people in my family, and, like probably everyone reading this, in the top tier of the world when it comes to educational attainment. Can I read 1% of the scientific literature, make sense of it, and put it into practice? No. I’ll readily admit it. I don’t even want to try. I want the experts with more specific training, more authorization, and more experience to get the best information possible. I don’t need it.

There’s plenty of good information for educated lay people like us. Most people — and if you haven’t seen a regular person wrestle with a relatively simple scientific paper, you’ve lost touch — can’t derive much value from the vast majority of scientific papers. Even Wikipedia has been found to be written at too high a level, including its Simple English version.

Scientists in other, closely related fields often don’t find value in reports from one another’s fields. Practitioners in some fields feel their journals have gotten beyond them. Relevance, appropriateness, and timeliness are really important. Access only matters when those three aspects are fulfilled.

The scientific literature isn’t relevant or appropriate for the lay public, and it’s often too far ahead of what’s available in the real world, making it also not important from a timeliness perspective. A new breakthrough in medicine, engineered materials, microbiology, or physics — beyond the news aspect, which is not what we’re talking about — is unlikely to find its way into daily life very soon after publication, usually years later.

I think some foundational ideas have become conflated, and they need to be separated. OA wants to continue to conflate them because the conflation serves their purpose, but there are some real questions that need to be asked still, despite the level of government pronouncement you produce.

Antipathy towards Open Access?


RP: You are the CEO/Publisher of
The Journal of Bone & Joint Surgery. I believe the journal is subscription-based, and I think it does not offer a hybrid OA option.  You are also the founder of the Scholarly Kitchen blog and President-Elect of the Society for Scholarly Publishing (SSP), both of which have been critical of Open Access. How would you respond to those who might argue that your interest in this matter is driven by an antipathy towards OA?

KA: The Society for Scholarly Publishing has no dog in this fight. It’s been completely neutral when it comes to publishing business models and new technology, mainly serving as one of the best and most supportive venues for new ideas to emerge.  

However, because SSP cultivates discussion and networking, there are sceptics and critics as well as adherents and supporters at its meetings and within its membership — and this pertains to anything, from publishing to mobile devices to dealing with international markets to business models to dealing with vendors. It’s a vibrant organization that welcomes all opinions from publishers, librarians, editors, and consultants.

Some of the bloggers on the Scholarly Kitchen have been sceptical of some of the claims made by open access adherents, and we have featured a lot of research and analysis in this regard. But the bloggers on the Scholarly Kitchen pass no “open access litmus test” and have a variety of opinions on all sorts of topics, and very different backgrounds. Some run open access journals, and quite proudly, or work for publishers with many open access titles.

I am one of the more vocal sceptics of open access, because I think it is structurally flawed — I have always thought that more virtues accrue to publishers who cater more to the users of the literature than will accrue to those who cater more to its authors.

I also think that by catering to the readers, you cater to authors sufficiently and properly, but don’t lose perspective. That’s why the journals I run are paid for based on readership and usage. There are too many incentives that can go awry if you forget about your readers and their needs. That’s where my scepticism comes from.  

RP: I hear what you say about the Scholarly Kitchen, and I have seen the disclaimer on the home page distancing the SSP from the views expressed on the blog. But I have yet to meet an OA advocate who is not convinced that the hidden agenda of the Scholarly Kitchen is to talk up the subscription publishing model and to talk down OA. And in so far as the SSP created the Scholarly Kitchen, and presumably funds it, critics assume that it too is biased against OA. You are saying that this is not correct. So why then do you think that there is such a widespread belief that the blog is anti-OA?

KA: I think the Scholarly Kitchen is one venue that has been willing to question many of the assumptions and practices of OA publishers, so it has stood out a bit because of that.

I also think that one of the tactics OA advocates routinely wield is to discredit anyone who questions OA as perhaps imperfect or improvable, and to use a broad brush to dismiss critics. They want to paint us as illegitimate and unsympathetic and ill-informed, when I think most people feel we have a critical eye on many things, are pretty balanced, and are just asking tough questions.

The attacks we’ve received when we’ve talked about OA have been surprisingly vitriolic and immature, even when we’ve said some things that were intended to point out issues the OA community might want to think about, in a helpful way. Some people really have a hair-trigger about anything short of complete OA cheerleading.

When it comes to OA, we’ve had posts on the Scholarly Kitchen about how to make OA work, what authors want from OA, proving that OA increases downloads, and so forth.

But we aren’t patsies, and that’s what OA advocates seem to want us to be. We question the CC licenses, the assertions about citation advantages, the utility of scientific information to lay audiences, and so forth. I think it’s important these ideas are tested and not just accepted as doctrine.

And I’m surprised that scientists or scientific-types would object to having their ideas and assertions scrutinized.

RP: Can you expand on your personal position vis-à-vis OA, both gold and green OA?

KA: I think green OA is untenable financially, redundant in the networked world, and confusing to users.

I think gold OA is structurally flawed because it creates incentives to publish more but has not yet yielded any increment in quality except for those rare cases in which a gold OA journal acts more like subscription journals — with high rejection rates, for instance.

I think gold OA probably has a place as a temporary catalyst mode for nascent fields, in which the authors and readers are truly one and the same. But to get beyond this, a reader-oriented model will work better.

I also think there is no reason to believe that gold OA will continue to be cheaper than subscription publishing — the market is beginning to differentiate, and costs are increasing in desirable venues. There are early signs that if the subscription model ever moves to a minority model, gold OA will show some of its structural flaws — taking money from research budgets, favouring rich institutions and countries, and making it harder for good research from unexpected places to reach its intended audience in a timely fashion.

More information is not better information, and the subscription model aligns a lot of interests who are very effective at getting information to people at the right time, in the right place, and with branding they understand.

What is to be done?

RP: In your second blog post you cited the minutes of the June 2012 PubMed Central National Advisory Committee meeting you mentioned earlier. In doing so you drew attention to the fact that eLife may be thinking of scaling back its initial goal of competing with the likes of Science, Nature and Cell, in favour of a PLoS ONE-style model that would see high acceptance rates at the journal. You have in the past been critical of PLoS ONE and its “lite” peer review model. Were eLife to go down the same road would it in your view be a retrograde step?

KA: Yes, that was an interesting part of the minutes, not only because eLife seems to be abandoning its goals of becoming a premier life sciences journal, but also because the head of the NCBI provided some of this business advice, which was completely inappropriate.

I think PLoS ONE is a repository calling itself a journal, and clearly has lower standards — as stated in its own words — than other journals. It doesn’t care about novelty or importance, but readers certainly do. It only cares about studies being “methodologically sound,” but it’s not clear it’s set up to accomplish this.

It has an acceptance rate that is about as high as the acceptance rate for scientific papers in general — meaning, in reality, it’s accepting about 100% of the non-garbage papers it receives. It does not filter its contents well, which is what readers want — relevance and quality are not on its radar, only volume and a low bar for acceptability.

This is what happens when the open access model is allowed to go too far, in my opinion — incentives tilt toward serving authors and their short-term needs. The irony is that authors are, in the bulk of their working lives, readers. And most of them want relevant, timely, high-quality content, which PLoS ONE can only sporadically provide. If eLife goes down the same road, what will we have gained? And what will the users of scientific information have gained? More noise, or more signal?

The interesting thing about eLife is that it represents a new breed — a funder-sponsored journal with funder-paid and funder-funded editors. There is no broken line between funding and publication. Journals are supposed to provide objective and disinterested review of content.

The principals at eLife have stated that their goal is to increase the “productivity” of their funding through publication. That signals they want more published. The eLife model is actually, in my mind, worse than PLoS ONE because at eLife, it’s like the players (the funders) have hired the referees.

Combine this with their goal to score more, and suddenly, there are fewer penalties called. So, if eLife adopts the bulk-publishing model and the lite editorial review approach of PLoS ONE, I think eLife will be in a slightly less tenable position than PLoS ONE when it comes to wondering whose interests they are serving. At PLoS ONE, they are serving the authors’ interests. At eLife, they will be serving the funders’ interests. Or, at least, that’s how it appears. And in the world of conflicts of interest, appearances matter.

The situation with PubMed Central isn’t reassuring because it shows that eLife is willing to pull strings and use its influence to cut corners. If they can’t even execute their launch without cronyism and conflicts of interest, what can we expect in the future?

RP: Based on the email I have now received from Patterson, it seems that the high-acceptance journal discussed at the June 12th Meeting may be a planned spinoff journal from eLife. That is, any high-acceptance journal would be in addition to, not instead of, eLife. This seems more like the model adopted by PLoS: have a high-quality journal (or several such journals) but also offer a high-acceptance journal as well. I guess you would still view that as a retrograde development?

KA: I think high-acceptance-rate journals have the deck stacked against them in the journals world through sheer positioning. They are emphasizing quantity over quality.

Because of this, they usually diminish at least a few of the functions of journals — dissemination (they are more passive); validation (they use expert peer-reviewers but not statisticians or methodology reviewers); designation (they don’t care as much about novelty or importance); and filtration (they don’t make sure the information gets to the right audience, which is related to the more passive dissemination).

Those other functions are expensive to do right, and high-volume journals tend to focus on quantity, not quality. That stacks the deck against their long-term primacy and vitality, I believe. They risk becoming essentially well-documented and lightly curated repositories.

RP: You said that green OA is untenable financially. Many OA advocates argue that subscription publishing is untenable financially — that, after all, is why there has been a long-standing serials crisis. You are probably right to say that gold OA will not prove any cheaper than subscription publishing. But if you are right, does this not suggest that scholarly publishing will remain untenable financially in the long run, whatever publishing model is used. If so, what is to be done?

KA: There has been a serials crisis for nearly 100 years. It’s like the old joke that the second book that came off the printing press was entitled, “The End of the Printing Press: What’s Next?”

Our industry seems to mete out some standard anxieties, and we always produce more capacity than content — at least, we have since the 1700s. There are more journals on the market now than ever, there are more researchers publishing, and scholarly publishing has continued to find ways to grow and innovate.

One concern I have is that we’re eroding the foundations, which is the primary research record. We can’t build fancy new synthesized content services and “big data” if the primary literature is built on sand. Lower standards around statistical review, methodological review, disclosure, conflicts of interest, and so forth don’t help in this regard. A lack of funding to purchase good information doesn’t help the primary literature thrive.

The subscription model has many virtues, including spreading the costs to the larger population of users and readers. OA advocates may find that because authors produce at their most prolific one paper per year but read 100, the asymmetries will put major price demands on 1% of the current environment — the authors and their institutions. And that will reveal how current OA prices were heavily reliant on the subscription model’s healthy environment.

I think funding needs a hard look — what do we want to be buying, more or better? Who should pay? What happened to all the department subscription budgets when site licenses came? Why have libraries been receiving a lower and lower percentage of university budget for 30 years? There are implications to these choices and trends. I think we need to sort out what we’re aiming for — quality or quantity — and have these discussions and arguments, but with a mind toward finding a better way, not just having battles.

I think OA has a place in the world, but it’s not nearly as big a place as its advocates seem to want to dictate. There has never been a single business model in publishing. Let OA compete, find its place. But to have the government as a tacit player in the market seems wrong, and that’s how PubMed Central is behaving. That needs to stop.

RP: Ok, let’s end on that. Thank you for agreeing to do the interview.

1 comment:

Mike Taylor said...

"But last year their fears of being disintermediated were reignited, when three large research funders — the Wellcome Trust, the Howard Hughes Medical Institute (HHMI) and the Max-Planck Society — announced plans to launch their own OA journal, eLife."

There fears of being outcompeted is a more accurate assessment here. eLife is still a conventional journal -- it doesn't in any sense make an end-run around the traditional publication process. Any fear that existing publishers have of it is based on their sense that it will do a better job than they do, at lower cost.