National Institute of Health
National Institute of Health (Photo credit: peroty)

My years of avoiding all things Open Access (OA) are clearly over. We have been hearing about possible OA mandates for all federally funded papers for almost as long as we have been hearing that print “will be dead in 5 years”. Now that we are on the verge of a mandate and I have become our in-house expert on the topic, there is one detail that continues to confuse me.

Under inevitable mandates, what constitutes a “federally funded” paper?

The OSTP public access memo does little to define what it would consider a federally funded paper. It does mention “research that directly arises from Federal funds, as defined in relevant OMB circulars (e.g., A-21 and A-11).” I thought maybe those circulars would be helpful but I was wrong. One defines financial reporting for federal programs at universities and the other appears to be the actual federal budget.

I turned to the NIH as the likely model for others.  Their OA mandate applies to the following manuscripts:

  • Any direct funding from an NIH grant or cooperative agreement
  • Any direct funding from an NIH contract
  • Any direct funding from the NIH Intramural Program
  • Authored by an NIH employee

Now we are getting somewhere. What do they consider direct funding? Here is what the NIH says about that:

direct cost is any cost that can be specifically identified with a particular project, program, or activity or that can be directly assigned to such activities relatively easily and with a high degree of accuracy. Direct costs include, but are not limited to, salaries, travel, equipment, and supplies directly benefiting the grant-supported project or activity. Most organizations also incur costs for common or joint objectives that cannot be readily identified with an individual project, program, or organizational activity. Facilities operation and maintenance costs, depreciation, and administrative expenses are examples of costs that usually are treated as F&A costs. The organization is responsible for presenting costs consistently and must not include costs associated with its F&A rate as direct costs.

So, everything. But I still have questions.

Some stuff is cut and dry–NSF gives you money for a research project and you write a paper about the project and your findings. This seems pretty clean. If you revisit that data in say 10 years and use some of it in a new paper, is that new paper federally funded?

If you used time on DOE equipment, do the results of that work count as federally funded?

If a federally funded fellow assists in some analysis on the data, is that paper now federally funded?

If you use equipment or get data from a state transportation department lab and that lab is funded by the Department of Transportation, does that count?

What if 95% of your money comes from industry and only 5% from the government? What if it’s 98/2?

The U.S. House of Representatives Science committee has a draft bill in discussion called Frontiers in Innovation, Research, Science, and Technology Act of 2013 or FIRST. It is basically the bill that re-authorizes the America COMPETES act. FIRST echoes the OSTP memo with two important differences:

  1. Accepted manuscripts must be made publicly available no later than 24 months after publication with the possibility to extend that period by 6 or 12 months if the agency feels that any stakeholder would be “uniquely harmed” by a shorter embargo, and
  2. The public access policy applies to manuscripts that results from research funded in whole or in majority part by a grant from a Federal science agency.

While this is still a draft bill and is certainly moving at a glacial pace given the lack of any work being done in Congress, if passed, this would raise the question–how do we know what percentage of a paper is federally funded?

We need to ask the authors, of course. More submission questions to come! And then we must take their word for it. With the exception of trying to look up grant numbers on the websites of individual agencies, there is no easy way to validate an author’s claim to federal funding prior to publishing the paper. Even post-publication, we would only know there was a problem if the agencies are actually digging through FundRef to see what their money has produced, and if they report any discrepancies to the publisher.

Lawmakers want to use peer-reviewed manuscripts as a way to show taxpayers where their money goes. Federal agencies are responsible to Congress for reporting productivity and will need a lot of published papers in order to justify more tax dollars. Authors, who want to take advantage of OA that comes without fees will want to maximize listing federally funded research dollars now that publishers are forced to make those papers publicly available.  Might an author feel compelled to claim federal funding of a nominal amount when perhaps they might have left that acknowledgement off in the past?

What about accountability? Where are the checks and balances? What is the rationale for setting funding thresholds? Is it in any way data-driven, or just a randomly chosen number? If publishers are being asked to give away content that today can be sold, shouldn’t they have a voice in these decisions as part of the public-private partnerships envisioned by the OSTP?

Is there even a proposed system to validate that funding claims are real? Should researchers be required to keep careful timesheets and spending records to delineate the specific percentages of salary and funds spent on each individual experiment, and then collate those records with the specific data included in a paper?  Who should have access to these records? Funding agencies? Publishers?

Without clear, verifiable procedures and rules in place, these mandates threaten to become either a burden on researchers who must provide a mountain of paperwork evidence or essentially unenforceable—if OA requirements are to be based on rough guesses, the easiest path for a researcher is to always claim that only 49% of any project was federally funded, and be relieved of all responsibilities.


29 Thoughts on "What Does “Federally Funded” Actually Mean?"

Great inaugural post.

In addition to the poorly considered statements about federal funding, which can be interpreted in numerous ways and which allow even incidental (or unproven) federal funding to trigger OA requirements, the notion that someone can be “uniquely harmed” because they might have to pay for content at any point gave me pause. Since when is paying equated with harm? We have truly gone down the rabbit hole here.

That text really refers to publishers, authors or others being harmed by short embargo periods.

Yes, you’re right, I misread it. I need to not comment so early in the morning!

The NIH/PMC threshold of “any” funding is as low as you can go. It might even include a casual conversation with an NIH grantee by an author. Moreover, my understanding is that PMC accepts entire journals where many of the articles may have no federal funding whatever. Practices like this make impact analysis of federal funding impossible. It amounts to padding the books.

The scientometric issue is how valid the impact analysis of government agencies can be when “government funding” is poorly defined, so that research with very little actual government funding is included? Conversely, what role might the scientometric community play in resolving this issue?

Related but not off topic, authors like to put every possible paper from their CV on their competing and non-competing grant renewals. It makes them look productive.

The problem arises when they are posting papers that aren’t funded by the federal government. The NIH kicks the grant renewal back because the papers on their reference list haven’t been submitted to NIHMS, which sends the assistant or the librarian back to the journal.

I’ve had to express to authors and their assistants that both the NIH and certain drug companies would be mighty surprised to hear that the NIH funded a pharmaceutical drug registration trial.


Don’t worry, whatever is not covered by federal funder mandates will be covered by institutional mandates like Harvard’s, MIT’s etc. All refereed research output, both funded and unfunded, in all fields, is the obvious target for all. No problem for researchers to figure that out: they won’t have to think twice, And publishers — for all their moaning and groaning, and dire warnings of doom and gloom — will, of course, figure out a way to adapt. All the FUD they keep trying to raise at each juncture is so unmistakably just smoke and delay tactics: futile efforts to stave off the obvious, optimal and inevitable outcome for research, researchers, the vast R&D industry, and the tax-paying public in the online era: 100% Open Access to all peer-reviewed research immediately upon acceptance for publication.

The problem though, is that most researchers are going to want to be compliant with the requirements of their funders. This will be necessary to continue to receive funding, pay one’s employees, feed one’s families, keep the lights on, etc. Researchers in this position can’t be expected to deposit their papers in every single repository on earth and can’t be expected to decipher an unclear set of confusing rules for compliance. Depositing your paper in Harvard’s repository may not satisfy the NIH. If these mandates are to succeed, the simpler the process and the clearer the rules, the better.

Yes, Green OA self-archiving mandates need to be simplified and harmonized, and they are indeed evolving in that direction. One deposit (institutional) is enough, and the rest can be done by software (import, export, harvesting). Institutions will monitor and ensure compliance with both their own institutional mandate and funder mandates. The most effective mandate model is probably the Liege/HEFCE model: immediate institutional deposit of the refereed final draft, immediately upon acceptance, whether the journal is subscription-based or gold, and whether access to the deposit is immediately set to OA, or to Closed Access during any embargo. (The repository’s automated request-a-copy Button can then tide over user needs during any allowable embargo with one click from the requestor and one click from the author.) Immediate institutional deposit is also designated as the sole mechanism for submitting publication for performance review, research evaluation, funding applications, or standardized CVs.

If the US government were to adopt this model it would have to impose it on every grantee institution, then regulate those institutions for compliance. That is far from simple. Nor does it address the question raised by this post, which is which articles fall under the Federal mandate?

Institutional repositories need to do a better job of maintaining the version of record if they want to be the holder of the content. Corrections and retractions are not being handled well in repositories.

Indeed! In a study of scientific articles later retracted by journals, just 5% of publicly accessible copies located on non-publisher websites and repositories included retraction notices. As I noted, “the benefits of decentralized access to scientific articles may come with the cost of promoting incorrect, invalid, or untrustworthy science.” see:

Davis, P.M. 2012. The persistence of error: a study of retracted articles on the Internet and in personal libraries. Journal of the Medical Library Association 100.

Publishers prevent or embargo the deposit of the version of record in institutional repositories. But for open access, the author’s final, refereed, revised, corrected, accepted version is just fine. (Retractions are so rare I am not sure why mention them at all. What evidence that they are not being handled well? And compared to what?)

Consider that before Internet the situation was worse: Libraries did not exorcise retracted papers from their print archives. So if the user did not happen to see the issue that contained the retraction, they could unwittingly rely on a retracted article. Internet helped that situation; OA will help it more, by making it much more likely that a user of a retracted paper gets notified. (But as far as the desirability and benefits of OA are concerned, this issue is a minuscule one in any case,)

It is not clear that OA is either desirable or beneficial, but in any case this is all irrelevant as far as the question is what federally funded means?

FWIW the Research Councils in the UK adopted a straightforward and pragmatic definition that relies on the authors common sense. If the author considers the contribution of the funder sufficient to acknowledge it in the article, then it is in scope for the funder’s open access policy. As a result, whether an article is in scope or not is simply answered by inspecting the credits in the article.

This basically makes compliance voluntary, which is not the intent of either the OSTP model or FIRST. NIH tried a voluntary system and it failed. Moreover, what keeps authors from gaming the system by claiming significant federal funding in order to get public access for their articles?

I don’t see why this should be acceptable to publishers. It seems that if publishers are being asked to give something up (subscription revenue, PPV income, eyeballs on the page) with zero compensation for federally funded papers, we should not need to depend on defining federal funding as whatever the author feels like.

Indeed, the twin issues of who, and which actions, are covered is central to regulatory design. This is why regulations typically have an extensive definitions sections. The difference between FIRST and the OSTP memo are huge in this regard.

I guess all I would say to David and Angela is that there is a lot of mistrust of authors being expressed. When the work being described in the article has arisen from a number of funding sources, I can’t think of anyone better than the author to assess the relative contributions of funding sources. Some complex set of rules (and I am assuming they would be complex) seems like an overhead that would still end up relying on good faith on the part of authors in any case.

Also, your question was “what does ‘federally funded’ mean?” I was offering one answer, not suggesting that publishers would like it. OTOH if you don’t like it, perhaps you have a better idea..?

I would like a verification tool and I would like guidelines on how much federal funding qualifies a paper for government mandated OA, when that time comes.

I like FIRST’s “majority of funding” concept, since it focuses on clear cases, with obvious Federal impact. NIH’s “any funding” or the oft used “funded in part” are hopelessly vague.

But “majority of funding” is also too vague. What counts and what doesn’t? How do I tell 49% funding from 51% funding?

The point is that that difference does not matter because either rules out trivial contributions. What counts is a different issue, also important. Direct cost is a well understood concept in grants.

Our head of reference has been very involved in helping investigators comply with the NIH Public Access Policy. While the majority of cases are clearcut, there is still a substantial minority of cases where the connection of a particular paper to a particular grant is more ambiguous. This is often the case where an investigator may receive some funding through a large center grant. Lee has gotten pretty good at sorting through most of these questions but she still sometimes has to ask NIH for clarification in a particular case and the answers are not always as clear as we would wish. The situation may become even more complex if each agency, in implementing its own policy, interprets the ambiguous cases slightly differently. The FundRef initiative will help immensely in tracking this stuff, but it doesn’t provide guidance to the investigator trying to decide whether or not a particular manuscript should be associated with a particular grant in the first place.

Comments are closed.