This paper was refereed by the Journal of Electronic Publishing’s peer reviewers.

Introduction

An open access article is freely available to all interested readers and is preserved for posterity (Suber 2003). At this time, open access to scholarly research in the biosciences remains partially realized. According to research published by Mark Ware in 2012, approximately 30% of STM content is open access in some form, whether immediately or after a delay (Ware 2012).

This mixed approach remains in effect today. Publishers such as PLoS and BioMed Central make their articles immediately available to everyone upon publication. In the case of publishers like Elsevier and NPG, most of their articles are provided immediately to subscribers or licensees, and eventually to everyone else after an embargo period. In this article, “immediate open access” designates an article that is open access upon publication.

It is feasible to distribute new articles online at little to no marginal cost, which supports the notion that open access is economically feasible. This contention has driven the acrimony between open access advocates and publishers, with advocates claiming that research content should be free and publishers responding that they continue to deliver worthwhile value in the scholarly publishing chain (Alliance for Taxpayer Access 2015; Anderson 2014). Both perspectives exhibit ends-based reasoning, and for that reason are overly simplistic. I support open access publishing, and also believe that publishers would have an essential role to play in a fully open-access world.

Although the rancor between open access advocates and publishers is usually expressed in financial terms, economics is not the most important factor in this discussion. If online scholarly publication were simply a matter of more efficient distribution, then immediate open access would now be a reality (Clarke 2010). Michael Clarke made this observation in 2010, in the course of arguing that much of scholarly publishing practice supports the cultural values of the academy. These values are currently impervious to financial considerations.

I will focus on Clarke’s piece in the next section of this paper. For now, it suffices to say that the imprimatur of journal publication remains essential to success as a researcher in the biosciences. As yet no new means of credentialing have supplanted publication in prestigious journals. As long as this is true, established subscription-based journals will remain an essential component of scholarly publishing in the biosciences.

Given this truth, economics become more important on the demand rather than the supply side. Journals that serve essential credentialing functions in their fields carry handsome costs, which is why increases in subscription or licensing fees have long exceeded the rate of inflation (Association of Research Libraries 2012). Even though researchers are always free to publish their research in an open access venue, in some cases it remains harmful to their careers to do so.

This is no doubt depressing for the many advocates who have worked tirelessly for more than a decade to bring about immediate open access. One argument advocates cite for open access is that greater access to a paper will lead to increased citations later. This is known as the “citation advantage” (OpCit Project 2013). This advantage is disputed and depends upon which assumptions are used to draw this conclusion (Davis 2011). But there is no doubt that text and data mining efforts can be much more robust with open access to a paper as well as all of its supplementary materials (BioMed Central 2015). This is one reason why government agencies and private funders throughout the world are prioritizing open access distribution of research results, albeit generally with an embargo period. One such example is the US National Institutes of Health Public Access Policy (National Institutes of Health 2014).

Complete open access in the biosciences remains highly likely in the long term, which is an eventuality I will return to in the conclusion of this paper. In the meantime, we should not be surprised that the transition to open access is still incomplete, because nothing changes easily. To cite one analogous example, it took decades for monograph publishing to standardize after the invention of the printing press (Wikipedia 2015).

Bearing this truth in mind, this paper will detail three proposals for providing immediate open access to the scholarly literature in the biosciences. After describing each proposal I will argue that modifying the copyright status of scholarly articles is essential to providing immediate open access, and that the aegis for doing so rests with the authors of scholarly papers. The copyright status of scholarly papers is at the crux of this discussion.

Although the open access discussion remains vital, scholarly papers no longer comprise the complete spectrum of bioscience literature. I will conclude by describing why librarians and publishers should work together to enhance access to the many forms of scholarly expression.

The Cultural Functions of Journals in the Biosciences

In order to analyze why open access in the biosciences remains incomplete, it is useful to understand the cultural functions of journals within the academy. To do this I often turn to Michael Clarke’s aforementioned piece, “Why Hasn’t Scholarly Publishing Been Disrupted Already?” Clarke’s essential insight is that open access journals do not yet respond to all of these cultural drivers as effectively as established titles.

Clarke delineates five functions that scholarly journals have come to assume: dissemination of articles, registration of the ideas developed by an author or authors, validation that the new ideas are sound and should become part of the literature, which usually occurs with peer review, filtration of the vast quantity of scholarly literature into manageable amounts of material to read, and designation of an article as being of permanent value (generally denoted by subsequent citations and a high impact factor).

He divides these five functions into two categories: technological and cultural. The Internet has radically simplified dissemination and registration, the two technological functions. This truth underlies much of the advocacy for open access.

As Clarke notes—and I reinforced above—if these two functions comprised the full role of scholarly journals, then open access would have occurred long ago. The other three cultural functions—validation, filtration, and designation—are the key to understanding why this has not yet occurred. Publishing a paper in the “right” journal remains critical to academic respect and advancement. An author may be intellectually supportive of open access and sympathetic to librarian concerns about the budgetary implications of maintaining access to key journals. However, if their mentors and advisors recommend publication in a subscription-based title as the key to achieving tenure, almost all authors will follow this path.

Implicit in these actions is the assumption that our established norms for scholarly publication are the optimal means of achieving validation, filtration, and designation for the scholarly literature. Many observers disagree, and have offered trenchant criticisms of how peer review works (Smith 2006) and of the weaknesses of the impact factor (PLoS Medicine Editors 2006). For this reason there are now several alternate means of presenting and filtering the scholarly literature that do not rely upon the established cues of peer review or impact factor (Priem, Taraborelli, Groth, & Neylon 2010).

To date, these efforts—for which I am appreciative and sympathetic, and wish to see continue—have not shaken the everyday academic’s continued embrace of peer review or of the impact factor. Academics, like everyone else, are more primed to resist losing what is familiar than they are to develop new approaches to how they work. Daniel Kahneman and Amos Tversky documented this tendency toward “loss aversion” several decades ago, and it is very much in play here (Kahneman & Tversky 1984). This is why it took so long for monograph publishing to become standardized after the invention of the printing press, and it is why the transition to immediate open access publishing is not yet complete.

Three Proposals for Immediate Open Access

The three proposals in this section are arranged in decreasing order of plausibility.

Proposal 1: Reallocate funds that currently support research to underwrite immediate open access to articles published in all bioscience journals

At the present time, open access in bioscience journals occurs via payments to the journal, either by the author directly or by their funding agency. Although many open access journals charge no fees, the most prestigious generally do. According to research published in 2012 the average author fee to publish in an open access journal is $906 (Solomon & Björk 2012). Fees are significantly higher for flagship open access journals such as those published by PLoS (PLoS 2015).

These funds offset the revenues lost since the article cannot be sold via subscription or licensed. The “author-pays” model is inherently unsatisfying because, as Stuart Shieber notes, it cannot escape the unfair taint of being a form of vanity publishing (Shieber 2009). In addition, it locks out authors who are unable or unwilling to pay the author fees (Peterson, Emmett, & Greenberg 2013).

Institutionalizing support for immediate open access publishing is better than pushing this task onto authors. This was Stuart Shieber’s essential insight in another 2009 paper, “Equity for Open-Access Journal Publishing” (Shieber 2009). In proportion to the entire amount of resources spent to fund bioscience research, the amount necessary to provide immediate open access to resulting articles is trivial. For example, the Finch report estimates additional costs of £50–60 million per year to support complete open access in the UK (Finch 2012), which is only .01% of the approximately £4.7 billion that UK Research Councils spends to directly fund research (Department for Business Innovation & Skills 2014).

In a recent paper, longtime US publishing observer Andrew Odlyzko estimates that $800 million annually, which is the cumulative amount that members of the Association of Research Libraries spend on their journal purchases, is enough to support a fully open access publication system (Odlyzko 2014). The 2014 budget of the National Institutes of Health is $30 billion (National Institutes of Health 2014), and Odlyzko’s estimate is .25% of this.

Even if the Finch and Odlyzko estimates are significantly low, the point holds that the cost of providing complete open access remains a small fraction of the amount of the overall funds dedicated to funding bioscience research.

One idea, therefore, would be for universities that publish research to set aside 1% of the overhead from funds that support research to make the resulting articles immediately available open access. This set aside would support author payment charges to publishers up to a generous cap, beginning at $3,000 per paper. Although some publishers claim their cost of publishing each paper is significantly higher than this (Van Noorden 2013), the $3,000 figure is designed to be well above the average author fee as a way to incentivize a transition to immediate open access.

That likely publisher resistance points to the many hurdles facing such a proposal. The mere existence of ample funds to support immediate open access publishing is insufficient to make it happen. Even though there is enough money overall to sustain immediate open access in the biosciences, this money is spread across many entities each with their own agendas. Publishers that could command a higher than average author fee, or who would prefer to maintain a subscription model, would object that such a plan interferes with their business prerogatives. Meanwhile university administrators would resist a carve-out of 1% of overhead, as this currently funds other activities within the institution.

Finally, some open access advocates would likely argue that this approach entrenches the current system of journal publishing. My response to that objection is that this system is likely to remain entrenched anyway, and at least in this scenario all research articles are immediately open access.

The heart of publisher resistance would be the fact that they generally control the copyrights over the articles they publish. This makes the articles a form of intellectual property that can be sold on the open market. As Shavell has observed, reform to copyright laws would be necessary to achieve the aims of open access advocates (Shavell 2010). Such a reform would vest ownership of the articles in another entity besides publishers—perhaps with universities or with the funding agencies that support the research, who would then grant publishers a license to make articles immediately available open access. Of course, publishers would fiercely contest any such change as it fundamentally alters their business model. Lawsuits and Congressional lobbying would be certain.

Despite these significant hurdles, in terms of Clarke’s analysis there is one clear advantage of this approach to the other two I will present. This effort does not require changing any of the cultural values of the current publication system or ask authors to publish in different venues than they currently do. Although the financial landscape surrounding publishing would change dramatically, all behaviors, incentives, and rewards within the publishing system would remain the same.

Proposal 2: Develop the burgeoning publishing infrastructure within academic libraries to publish new open access titles

One alternate approach is for librarians to publish new open access journals themselves, and then continue the practice of collecting and preserving scholarly content. This is the aim of the recently formed Library Publishing Coalition (Library Publishing Coalition 2013), which would essentially circumvent the role of traditional publishers.

An advantage of this approach is that library directors could reallocate their budgets to support publishing more readily than the larger university could reorient itself to support open access. In addition, it aligns library operations with long-standing values in support of open access.

There are challenges. For starters, it requires library staff to develop the skill sets of publishers. This is not insurmountable, but it does take time. And—thinking of Clarke’s insights—the resulting titles would have to compete with established titles that already allow readers to filter the literature. There is also nothing to prevent publishers from offering to create and preserve their own digital content, in an end-to-end solution that cuts out librarians. What’s good for the goose is good for the gander.

In my view, library publishing efforts will serve as a valuable test bed for developing innovative publishing approaches and will also result in new titles that become key in various fields. While these efforts will place some competitive pressure on established journals, they are unlikely to yield a full-fledged alternative to them.

Proposal 3: Enhance author-initiated efforts to release the final version of their paper in an open access manner, even if that paper is formally published in a subscription-based journal

It is possible for authors to “self-archive” their papers, a step long advocated by cognitive scientist Stevan Harnad (Poynder 2013). Even if a paper is formally published in a subscription-based journal it is generally permissible for authors to provide the final submitted version free to all readers. The process for uploading materials to a self-archive is generally trivial and can often be automated.

Although self-archiving readily achieves the dissemination function of journals with little effort, relatively few authors complete these steps even when such actions are mandated as a condition of grant awards. Many funders require authors to authorize release of their articles in an open access manner after an embargo period, but not many actually adhere to these strictures (Davis 2012).

Archival and dissemination functions are not in the typical author’s workflow, which is why publishing companies and libraries exist. For this reason, reliance upon authors to spearhead the logistics of providing open access to their content is a fool’s errand. Even if self-archiving became a widespread activity, this would likely lead to papers scattered across the web rather than in a handful of easily accessible and well-managed places. Self-archiving is best thought of as a metaphor for authors maintaining control over the distribution of their work, rather than as an actual practice to be encouraged.

Authors asserting how to access their work is vitally important. Copyright generally rests with publishers because authors release their copyrights as a condition of publication. If authors maintained their copyright through systems developed and curated by publishers and librarians, immediate open access could become much more prevalent in the biosciences. This would not require changing any of the credentialing functions of established journals.

Opportunities Ahead

Authors, and to a lesser extent funders, are in the strongest position to effectuate immediate open access. Authors set the terms about how scholarly discourse occurs in their disciplines and are best able to dictate how to access their work. Funders can influence how publication occurs as a condition of their grant awards, which can influence discussion about changing the copyright status of journal articles.

In a testament to the power of funders, the Gates Foundation has become the first major biomedical funder to require immediate open access articles reporting on the research it supports (Bill & Melinda Gates Foundation 2014). Although this policy does not come into full effect until 2017, it is a strong move in support of immediate open access.

Librarians and publishers are effectively agents for authors and would need to respond in different and collaborative ways to any widespread increase in immediate open access.

Both librarians and publishers should get ready for this day. However it happens—whether in the fashion I’ve described above or in a different way entirely—immediate open access to bioscience articles will eventually become a reality. Librarians, researchers, and even some publishers have all predicted this (Lewis 2012; Shieber, 2012; Jha 2012).

As a longtime supporter of open access, I hope this comes to pass. In such an event there are many ways that publishers could maintain viable and profitable businesses. There will always be a need for tools that analyze and organize the corpus of bioscience literature, even if the underlying articles themselves are free. Business models could shift from licensing access to licensing such tools.

In addition to these new business opportunities, there is an ever expanding body of work that comprises the scholarly record. This includes blog posts, data visualizations, and sometimes even tweets. OCLC recently published a useful paper about strategies and considerations for stewarding this “evolving scholarly record” (Lavoie, Childress, Erway, Faniel, Malpas, Schaffner, & van der Werf 2014). In addition, the Council on Library and Information Resources has offered guidance on the preservation of scientific research data (Halbert, ed. 2013).

The burgeoning world beyond the PDF offers rich terrain for librarian and publisher partnership. Finding the right approach to providing immediate open access will be politically thorny and frequently charged, but the reward will be increased energy to devote to the preservation of born-digital scholarship.

Marcus Banks earned his MLIS in 2002, and has been deeply interested in questions of access and scholarly communication reform almost since the day he graduated. Banks has worked at the National Library of Medicine, the Georgetown University Medical Center, the NYU School of Medicine, UC San Francisco, and Samuel Merritt University.

References