English: Mimi & Eunice, “Viral Patent”. Catego...
English: Mimi & Eunice, “Viral Patent”. . (Photo credit: Wikipedia)

Is access to the research paper really the same thing as access to the research results themselves? Are funding agencies creating a false equivalency by confusing the two? And does this confusion favor researchers in some fields over others?

The principle that the results of research that has been publicly funded should be freely accessible in the public domain is a compelling one, and fundamentally unanswerable.

The quote above, from the Finch Report, stands at the heart of the reasoning for many of the public access mandates for research papers from funding agencies. Despite this very clear statement of an unanswerable principle, funding agencies including the RCUK, the Wellcome Trust and a host of US Government agencies continue to develop policies that ignore the fundamental question of what is meant by “results”.

What constitutes “results” is field dependent. For many researchers in the Humanities, the analysis, the “argument,” is what comes out of a research effort. For the comparative literature researcher, the dataset (perhaps the plays of William Shakespeare) is already well known and publicly available. The intellectual property generated by the research is the set of ideas and analyses about the dataset. For these researchers, the paper, or in many cases, the book, is indeed the research result.

But in the sciences, things can be very different. For the diabetes researcher, the result isn’t the paper written about the cure for the disease, the result is the cure itself. And given the intellectual property policies of the RCUK, Wellcome and the US government funding agencies, the researcher is encouraged to lock up access to those results behind a patent paywall. On the surface, this comes off as something of a bait-and-switch: it’s vital that the taxpayer be allowed to read about the cure for diabetes but if he wants to use that cure, well, he’s going to have to pay again.

One immediate effect of these differences is that it has created an unbalanced, unfair Open Access (OA) system for Humanities researchers. Much of the OA movement has been driven by biomedical researchers with a certain narrowness of vision. This is how my field works, thus every field works the same way so my principles are universal. Hence, everyone is well-funded (or funded at all) and can afford article processing charges, and is able to patent any valuable IP that results from research.

There is a small sect of researchers who insist that the definition of “open acess” has been completely settled, and that no debate will be tolerated regarding the one true pathway to progress. The rigid adherence to dogma (or at least an intense pedantry regarding the specific vocabulary used, and we all know how futile attempts to control popular use of language are) seems to dominate over notions of pragmatic flexibility, of finding different paths that offer maximum benefits for very different types of research.

A world where everyone must follow the rules tailored to one group of researchers is an imbalanced world. Under the BOAI definition of OA, Humanities researchers are asked to give up the fruits of their labors, their intellectual property, via the CC-BY licensing requirement, while STM researchers get to keep theirs through the continued acceptance of patents.

Tim McCormick has written eloquently about this imbalance as of late. His very pro-OA blog posts and other writings which fully embrace the concepts of adaptability and openness to new ideas are worth reading for those trying to make sense of OA for the Humanities.  He writes:

It’s not clear that scientific knowledge (factual, non ©) can be equated to humanistic (expressive, interpretive) knowledge. For STEM-based OA, expression (“literature”) is just a carrier for factual information, the possible basis of invention. It’s natural to see the carrier as non-ownable, like information, and want it to be as free and fluid as possible, to maximize opportunities for discovery and invention, which are high-value (and patentable). However, often in HSS, the expression (“literature”) IS the high-value output, from which the research may earn any future gain e.g. by licensing, compilation, book sales.

Beyond the STM/Humanities disparity, the societal value of how we treat researcher intellectual property is vastly more complex than the simple Finch statement that if the taxpayer funded it, the taxpayer should get it for free.

The NIH has just released a report titled, “Profiles of Prosperity, How NIH-Supported Research Is Fueling Private Sector Growth and Innovation”:

NIH-supported research added $69 billion to our GDP and supported 7 million jobs in 2011 alone. But the economic reality is about more than just numbers—stories involving lifesaving breakthroughs and real-world products and companies can be found all over the country.

Part of the mission of many funding agencies, particularly government funding agencies, is to improve national prosperity, often through investments that help drive the national economy. The NIH has had great success in this regard. Similarly, the Bayh-Dole Act, which allows US federally-funded researchers and their institutions to retain the intellectual property from their discoveries has been hailed as one of the most effective and important pieces of legislation of the last half-century, resulting in an estimated 5,000 companies formed since 1980 and an estimated 30% of the NASDAQ stock market’s value.

The question then should be asked, if federally (and otherwise) funded researchers were indeed required to relinquish all rights to their discoveries, would the same benefits have been created for society? Would progress and economic prosperity been improved or worsened by free access to actual research results?

One thing for certain is that universities would lose out on significant levels of financial support. The cash-strapped University of California system made more than $100 million in revenues from technology transfer in 2011. For a funding agency, there seems great benefit in investing in this manner. Rather than funding a short term research project, investing in a project that yields valuable patents can fund research for decades. If you teach a man to fish, and all that.

Academia seems better served by a system that requires private industry to contribute to the continued development of raw materials for exploitation. Patents require that those looking to exploit discoveries pay their fair share of support back into the system. This creates a symbiotic relationship between academics, industry and funding agencies. Each benefits from the other. If instead industry is given free access to research results, then the relationship turns parasitic, with the private sector feeding off of public funds with no need to return revenue back into the system.

These revenue streams are so important to UC that they’ve recently launched a new program to tightly lock up all researcher intellectual property with a small number of private corporations and investors. This lockdown on access seems somewhat paradoxical, given the UC’s recently released OA policy for faculty, which states that, “The Faculty of the University of California is committed to disseminating its research and scholarship as widely as possible.” The UC Regents, like the Finch Report, appear to draw a clear line between access to abstract knowledge and the ability to put that knowledge to practical use. The combination makes for a compelling business strategy though, with wide, free dissemination of the research papers serving as a marketing tool for the expensive licensing rights.

Similarly self-contradictory, Harvard has an OA mandate stating that they, “have an essential responsibility to distribute the fruits of our scholarship as widely as possible”, yet they charged some $14 million in 2011 for access to those fruits. These policies represent a practical recognition of the value that researchers generate, and an implementation that serves to fund that research by sharing the burden of costs with those looking to exploit those results.

To be fair though, while these potential rewards are highly attractive to universities, like lottery tickets, they only pay off for a minority. The majority of schools make no money, or lose money on technology transfer activities. “Only around twenty universities earn more than $20 million (15.5 M euros) in revenue from technology transfer every year, and many lose more money,” says Columbia’s Orin Herskowitz.

This gets to the other great imbalance offered by some routes to OA, namely that the financial burden shifts from the consumer to the producer. We’re already seeing demands that subscription fees be lowered for universities that have an increased financial burden due to paying Gold OA article processing charges. This seems an odd complaint about a system that is doing exactly what it was designed to do.

Estimates show that for the more productive universities, Gold OA charges would end up costing them more than the savings gained from dropping journal subscriptions. If we’re asking productive universities to shoulder the financial burden of access to scholarly papers for non-productive universities, should they also be asked to sacrifice the income technology transfer offers to their operating budget in order to make things cheaper for pharma and other industries? If the public paid for the STM research, then does it own the research or just the set of words and pictures used to describe the research?

In an era of reduced research funding and public support, asking the UC system to voluntarily give up that $100M or Harvard $14M per year is a non-starter. The many researchers who have benefited financially from patents, or who have started spin-off companies would also be equally unhappy. Along with the proposed benefits of OA, it is important to also recognize the necessary sacrifices it asks producers to make for consumers.

And if we’re not willing to ask universities and STM researchers to make those sacrifices, why insist that Humanities researchers do so? Is it better instead to continue to move away from “one size fits all” thinking, to let each field drive progress in its own optimal manner rather than forcing everyone to follow a set of rigid rules put together by a self-selected group with a seeming focus solely on science?

For the ten-year OA veteran, these questions may all seem settled and set in stone. But for many in academia (and funding agencies and governments), these are the early days of translating the OA ideal into a practical implementation, and the rules have yet to be written. Understanding exactly what the taxpayer or funder is entitled to and what each stakeholder is being asked to sacrifice are important first steps.

David Crotty

David Crotty

David Crotty is a Senior Consultant at Clarke & Esposito, a boutique management consulting firm focused on strategic issues related to professional and academic publishing and information services. Previously, David was the Editorial Director, Journals Policy for Oxford University Press. He oversaw journal policy across OUP’s journals program, drove technological innovation, and served as an information officer. David acquired and managed a suite of research society-owned journals with OUP, and before that was the Executive Editor for Cold Spring Harbor Laboratory Press, where he created and edited new science books and journals, along with serving as a journal Editor-in-Chief. He has served on the Board of Directors for the STM Association, the Society for Scholarly Publishing and CHOR, Inc., as well as The AAP-PSP Executive Council. David received his PhD in Genetics from Columbia University and did developmental neuroscience research at Caltech before moving from the bench to publishing.


27 Thoughts on "Is Access to the Research Paper the Same Thing as Access to the Research "Results"?"

The concept of the results of research is vague indeed. I am curious as to the legal principle that gives a funder a right to a document written after the research is over? If I work a summer on an onion farm and later write about what I learned does the onion farmer have a license to my writing? Surely not. How then is research different?

Funding agencies put all kinds of requirements and conditions into their grants–you must promise not to spend this money on hookers and blow, for example. I don’t see why they couldn’t make claims to the resulting papers a condition for accepting the funds.

I can think of several reasons why they either cannot or should not. First as stated the contract would be void for vagueness, just as the Finch claim is. Second the government would be over reaching. Third it might violate the Constitutional bar to uncompensated taking of property. In any case we would first have to see the proposed language. For example does it apply for the researcher’s life? Or to the heirs? Or just for some specified time. This is a very hairy concept to unpack.

Just to elaborate, at present a US Federal research contract has a specific end date. Most journal papers are authored after this date, some long after. Adding an OA provision claiming a license right to all these papers would make the contract open ended, covering events not known and not able to be valued at the time the contract was signed, with no compensation to the paper’s owners. This violates the basic principles of contract law.

The enforcement methods seem to be based around withholding funds, so there’s little a funding agency could do after a grant period has ended (other than withhold potential future grants).

It is not a question of enforcement. The issue is what the Federal funders can and cannot do. There are significant limits on Executive Branch powers. We may see this played out in the rule makings to come. The government cannot unilaterally bind people for the rest of their lives as a condition for receiving research grants. If you have a contrary example I would love to see it.

Then too, patents and copyrights are very different things. Patents are typically for physical processes while copyrights are typically for physical arrays of words, pictures or symbols. While it may be argued that the processes are developed during the research this is surely not the case for documents written after the research is ended. Note too that knowledge and ideas can be neither patented nor copyrighted, yet these are arguably the real research results.

How are research results defined? There was an interesting ding-dong a few years ago in the UK when an author did not want to supply to an interested party the underlying data from which the author had derived his ‘results’. From memory, the interested party tried to get the raw data through a FOI claim, but the university (successfully, I believe) resisted that on the grounds that the data was ‘proprietary’. In an OA context, will, as well as the paper arising from the research, all the details that made possible the publication of the paper, also be made available?

This is an area of immense legal confusion so no one knows at this point. Research results are not defined and that is the challenge. For example the US has no such FOI requirement. In fact all federally funded data not contracted for regulatory purposes is IP of the researcher, or perhaps their institution, because it is valuable. But the OSTP memo suggests that data should be made available, but how and how much is not clear.

There are no rules. This is the fog of revolution.

To add to that, in giving a presentation on OA mandates to a committee of members recently, several wondered how they would handle research funded by corporate money and federal money. If taking even a small percentage of federal money requires OA publication (and presumably the release of data), but the corporate funders refuse to agree to that, what will happen? The guess was that the researchers at universities will no longer get the corporate money. Corporations will go back to doing it all themselves.

There is no Federal mandate for release of all data in the OSTP directive. On the contrary Federal research regulations and contracts go to great lengths to protect data which is claimed as proprietary, which corporate co-sponsorship might well lead to. The mandate is to make non-proprietary data as available as possible. NSF has already done this by requiring a data management plan with every proposal. Other agencies may follow and that may be all there is to the mandate. Or they may create repositories or mandate the use of non-federal repositories. But I see no threat to proprietary data at this time.

And I think that’s going to be a big problem in terms of compliance. If researchers are given the option of declaring data to be proprietary with no negative consequences for doing so, then a significant number will likely do so rather than putting the work into organizing, archiving and making publicly available their data.

Love thy loophole. This is all taking place in the context of the Federal Acquisition Regulations (FAR) and their authorizing statutes, which have a lot to say about data. It is not like the OA mandate (which conversely may not have statutory authority).

This all begs the question, “What is the proper role of government in paving the way for private enterprise?” There is plenty of evidence that the role is one of long-term investment in basic science R&D that universities, private enterprise, and solo entrepreneurs are unwilling or unable to shoulder. From the postal system to railroads to the Internet (and telecommunications) to nanotechnology and biotechnology, the government has had a major long-term funding interest. Only after the government has done its work do venture capitalists and entrepreneurs start filling in the commercial landscape we come to understand. Too often, we think of VCs and private enterprise as the risk-takers, but actually the government is taking a lot of risk in basic science funding. The question here is whether it’s appropriate for universities to benefit from taxpayer-funded risk mitigation. I think it is. Universities benefit from a clear public-private partnership in this regard, and patents are the baton being passed. Without this kind of basic support, economies become tepid or worse.

There’s an interesting wrinkle to patents that people don’t always appreciate — they actually become publications themselves, and probably meet most of the criteria of Green OA. Patents can be found freely online. Filing a patent requires you to tell the world what you’ve invented. Google’s search algorithm, while grant-funded (some funding came from NSF, NASA, and DARPA, so it was government-funded) and published as a paper (http://infolab.stanford.edu/~backrub/google.html), was never patented because they wanted to keep an unknown number of aspects as trade secrets. Should OA advocates encourage patenting as a requirement of government-funded research, because patents explain the findings in practical terms? Are patents publications, even more so than research reports? Are patents actually a form of Green OA?

I don’t know about the medical world, but with natural sciences journals it is fairly common to require the data that supports a paper be available to any qualified researcher. (Not necessarily free, but available.) This creates a bit of a problem for authors from countries which are not so forthcoming with data.

Many (most?) life sciences journals have similar policies. Here’s Cell’s:

One of the terms and conditions of publishing in Cell is that authors be willing to distribute any materials and protocols used in the published experiments to qualified researchers for their own use. Materials include but are not limited to cells, DNA, antibodies, reagents, organisms, and mouse strains, or if necessary the relevant ES cells. These must be made available with minimal restrictions and in a timely manner, but it is acceptable to request reasonable payment to cover the cost of maintenance and transport of materials. If there are restrictions to the availability of any materials, data, or information, these must be disclosed in the cover letter and the Experimental Procedures section of the manuscript at the time of submission.

The problem, of course, is that journals have little power to enforce these rules, and even less incentive (are they really going to refuse papers from a top lab because they won’t send out their cell lines to others?).

I have long been arguing for looking at OA in the humanities differently from how it is viewed in thre sciences. Among other differences, science depends heavily on the “language” of mathematics and hence does not require “translation” into other languages. Humanities research, by contrast, is very dependent on good translation for its dissemination, yet the CC-BY license that the BOAI definition treats as the only “true” open access requires authors to cede control over the quality of translations of their work. Another difference is written right into copyright law itself: the second factor of fair use gives more weight to a use being fair if the work copied is closer to the fact end of the fact/expression spectrum. Indeed, it could be argued that scientific writing of the type typically found in journals is entitled to, at most, very “thin” copyright protection.

To tie Kent and David W’s comments together, I think Google is a really interesting example. If they were compelled to release all of their algorithms because of NSF funding, then there’s no Google, just a slightly improved Yahoo! and Alta Vista.

And David W. is prescient in bringing up the OSTP’s mandate for data archiving and release. That’s a really good idea, and would be helpful for scientific progress, but there doesn’t seem to be any legal way of compelling a researcher to do so. Under Bayh-Dole, the researcher and the institution own the intellectual property. The OSTP memo specifically states that data policy must recognize IP rights and avoid any negative impact on them.

Organizing and archiving data to be made available to others is a time consuming and effort requiring activity. It means serious time taken away from doing new experiments. If researchers are given a Google sort of out, the ability to say, “no, this is my IP and I’m not releasing it because I plan to exploit it,” then that makes any data policy toothless and unenforceable. I suspect many researchers would take that route purely out of efficiency and time saving.

This article is good at pointing out the nuances in scholarship that should be taken into account as general OA policy develops. That said, I do wish there had been more of a mention of data sharing beyond the sort of tech research departments can monetise. Indeed, the most important policies around data sharing surely must come from the biomedical side–the Human Genome Project, for example, or the sequencing of the 2011 Ecoli outbreak). Open data and crowdsourcing in such areas saves lives, so one could possibly argue there is a moral aspect to it. And if we’re talking about cost, what about the inefficency costs and long-term healthcare costs of not sharing such data (as well as negative data)? I’m not unaware of the problems that come with sharing data and making it reusable. But the funder, universities, and publishers together need to work to enable researchers to practice better data managment as well as work to automate as much of this as possible. Some people are already doing this (see, for example, work on the ISA-tab format (http://isacommons.org/) and the journal GigaScience (http://www.gigasciencejournal.com/) and the SOAPdenovo paper). Full disclosure: I manage GigaScience 🙂 But other journals (eg, Scientific Data from Nature) will be doing similar things soon.

Hi Amye,

These are all worthwhile points, perhaps each deserving its own blog posting (there’s only so much one can cover in one post before it becomes a novel). As you note, there are indeed obvious advantages to data sharing, but it does have its costs (particularly in time and effort). It is a situation where a researcher is being asked to sacrifice for the greater good (taking time away from doing new experiments to put together a resource for others to exploit).

There’s also an enormous variety of data generated, some with much more potential for reuse than others. Thinking back to the dark ages of my thesis project, the sequence data I generated was publicly shared and useful for others in their own projects. But the data behind attaining those sequences, the thousands of minipreps run to generate subclones to cover the short stretches of DNA one could cover with the techniques of the time are of no interest to anyone. Sure, if someone finds a problem in my sequence, they could go back to the subclones and restriction maps to double check my work, but it would be vastly faster and cheaper to just re-sequence the region itself. Or as a postdoc, I published the new technique I’d worked out for mouse embryo culture which was widely adopted, but the data behind my specific experiments, where I was manipulating one particular gene in one particular cell type during one particular stage of development would be difficult for others to do much with.

So there has to be some reasoned set of guidelines for what’s worth making the effort for.

I also have strong concerns about much of what is considered “negative data”. For many experiments, it is impossible to tell if the hypothesis was wrong or the execution of the experiment itself was wrong. Is cell type X not involved in cancer, or did I make up my media incorrectly which altered the behavior of the cell? If the latter, then I could be thwarting important medical developments by stopping others from investigating further. Science does require some level of redundancy. I know that if I was going to risk my reputation, time and research funds based on someone else’s data, I would probably do my best to replicate their experiments myself, just to be sure.

Full Disclosure: I’m a big fan of GigaScience and think it’s a really smart and forward looking venture, and used to work with Laurie Goodman and consider her a good friend.

Hi David,

These are all really good points and illustrate well why opening up access to data is far more complicated and nuanced than opening up access to article PDFs. As you mention, what we need is subject specific standards. I also have had some talks with Susanna Sansone recently that have made it clear to me that there’s no perfect, one-size-fits-all solution to these data sharing issues. The solution is not going to be perfect (ie resulting in complete reproducibility) but more along the lines of what’s the minimal amount of effort that puts out the most results, the old pareto principle. I don’t think the system we have now where 99% of data is not made public is a good system. It has a tendency to lead to bad science and complete unaccountability. That said, I’ve also seen (through GigaScience) what it takes, when possible, to reproduce published experiments and it’s expensive and timely. Setting up subject-specific standards and the infrastructure for authors and publishers to make data sharing easy will completely change how we view these issues, in my opinion. Funders are going to push this forward even if publishers don’t, I reckon.

Glad you’re a fan of GigaScience! They’re great to work with.


The basic principle of green OA mandates seems to be that just because you funded my work you have a rght to demand and publish whatever I write about it for the rest of my life. This makes no sense. It is wrong.

Comments are closed.