Like many of us who work in the general area of scholarly communication, I was very interested to see the August 25 update to the 2013 White House Office of Science & Technology Policy (OSTP) Memorandum on Increasing Access to the Results of Federally Funded Research (a.k.a. the Holdren Memo). Issued by Dr. Alondra Nelson, currently the acting director of the OSTP, the new memo (which I’ll call the Nelson Memo for convenience) advances several significant and substantive changes to the terms of the Holdren Memo – while, at the same time, putting those changes forward with less directive language.

All Agencies, not Just Those Making $100m+ in Grants

One of the most significant changes in guidance is the fact that whereas the directives in the Holdren Memo applied only to agencies that make more than $100 million in research grants annually, the Nelson Memo is directed at all federal grantmaking agencies. Those making more than $100 million in grants each year are asked to submit to OSTP their plans for new policy implementation within 180 days, while those making $100 million or less in annual grants are given 360 days. These plans should be completed and published by the end of 2024, and fully implemented by the end of 2025.

No More Embargoes Allowed

Another change that has really caught the scholarly communication world’s attention is the elimination of embargoes from the OSTP guidance. The Holdren Memo enjoined “each agency” to “use a twelve-month post-publication embargo period as a guideline for making research papers publicly available” (while, at the same time, giving each agency leeway to “tailor its plan as necessary to address the objectives articulated in this memorandum, as well as the challenges and public interests that are unique to each field and mission combination”). It also required agencies to establish “a mechanism for stakeholders to petition for changing the embargo period for a specific field by presenting evidence demonstrating that the plan would be inconsistent with the objectives articulated in this memorandum.”

Neither embargoes – of any length – nor the idea of “leeway to tailor” plans in light of “challenges and public interests that are unique to each field and mission combination,” nor a possibility of petitioning for exceptions is contemplated in the Nelson Memo. Instead, “federal agencies should update or develop new public access plans” to ensure that “all peer-reviewed scholarly publications authored or co-authored by individuals or institutions resulting from federally funded research are made freely available and publicly accessible by default in agency-designated repositories without any embargo or delay after publication.” Full stop.

The White House in Washington DC on a sunny day

No More Talk about “Public-Private Collaboration”

The Holdren Memo specified that “each agency plan shall… encourage public-private collaboration to maximize the potential for interoperability between public and private platforms and creative reuse to enhance value to all stakeholders, avoid unnecessary duplication of existing mechanisms, maximize the impact of the Federal research investment, and otherwise assist with the Agency plan.” Furthermore, it required that each agency’s plan include a strategy for leveraging existing archives, where appropriate, and fostering public-private partnerships with scientific journals relevant to the agency’s research.” No such language exists in the Nelson Memo (though it is mentioned that a Subcommittee on Open Science will “coordinate engagement with” various stakeholders, including publishers and libraries, “on federal agency public access efforts”; what exactly this means is quite unclear).

What about Research Data?

The Holdren Memo discussed, at length, the importance of providing public access not only to research publications, but also to underlying research data sets. Section 4 of that memo laid out detailed requirements regarding the management, archiving, distribution, and discoverability of data sets produced in the course of federally funded research, and (again) required each plan to “encourage cooperation with the private sector” in doing so. The Nelson Memo, by contrast, makes no mention of the private sector beyond preserving the Holdren Memo’s injunction to ensure that data access plans protect “business confidential information.” More importantly, whereas the Holdren Memo explicitly allowed for limitations on public data provision based on “feasibility,” the Nelson Memo does not acknowledge any reasonable limitations on data provision based on practical considerations, nor the possibility that there may be “challenges and public interests that are unique to each field and mission combination.” And it goes further, asking funding agencies to “develop approaches and timelines for sharing other federally funded research data that are not associated with peer-reviewed scholarly publication” (emphasis mine). This significantly expands the scope of the data-access requirement, while at the same time eliminating the one-year grace period allowed under the Holdren Memo — and, of course, applying to all grants made by the federal government, not just those from the biggest grantmakers. This will create a significant burden for publishers – something the Holdren Memo took some care to avoid, but which seems to be considered of no consequence in the Nelson Memo.

Still “Public Access, Not “Open Access”

One important element of the Holdren Memo that has not changed in the Nelson Memo is the use of the term “public access” rather than “open access.” While there is still no universally accepted definition of “open access,” the distinction between the two terms is generally understood to be the difference between “free to read and download” (public access) and “free to read, download, and reuse without functional restriction” (open access).

Subtle (But Possibly Significant) Differences in Language

One of the intriguing differences between the Holdren and Nelson Memos is that of language. From the beginning, the Holdren Memo is explicitly directive: “(OSTP) hereby directs each federal agency…” it begins following an introductory section, and the Memo proceeds to lay out its specific directives using verbs like “shall” (“Each agency plan shall…”, “each agency shall… provide a mechanism…”, etc.) and “must” (“Each agency plan must be consistent with the objectives set out in this memorandum,” “Each agency plan… must contain the following elements,” etc. – all preceding emphases mine). Interestingly, no such language exists in the Nelson Memo. From its introductory paragraph, it is framed in language that indicates its provisions are optional rather than mandatory: “In accordance with this memorandum,” the document begins, “OSTP recommends that federal agencies, to the extent consistent with applicable law…” (emphasis mine). Subsequent language in the Nelson Memo continues to give the impression that its provisions are suggestions rather than prescriptions: “Federal agencies should develop new, or update existing public access plans as soon as possible…”; “Plans should describe…”; “Scientific data… should be made freely available and publicly accessible by default at the time of publication…”; “Federal agencies should report to OSTP, when requested, on the status of their public access plans…”; etc. (emphases mine). The words “must” and “shall” appear only in reference to abstract overarching principles (“A federal public access policy… must allow for broad and expeditious sharing…”; “Nothing in this memorandum shall be construed to impair or otherwise affect…”; etc.). At no point are the terms in the Nelson Memo referred to as “requirements,” whereas those of the Holdren Memo are explicitly characterized in that way.

Many Questions

Public discussion of the Nelson Memo has already begun on the listservs, and various stakeholder groups including the Association of Research Libraries, SPARC, and the Association of American Publishers have issued public responses. But many questions remain to be answered. Among them:

  • Where the Holdren Memo laid out requirements, the Nelson Memo seems only to advance suggestions. Is this difference in language meaningful in practice?
  • What is the nature of the OSTP’s statutory authority over federal agencies? If an agency opts not to follow the terms laid out in the Nelson Memo, will there be any consequences? If so, what would they be?
  • Does the Nelson Memo wholly supersede the Holdren Memo, or are terms of the latter not directly altered by those of the Nelson Memo – such as, for example, the requirement that agencies incorporate in their plans a “strategy… for fostering public-private partnerships with scientific journals relevant to the agency’s research” – still in force?

Karin Wulf and I have invited Dr. Nelson to be interviewed for a future Scholarly Kitchen posting, and if she accepts we’ll be sure to ask her questions such as these. We await her reply.

Rick Anderson

Rick Anderson

Rick Anderson is University Librarian at Brigham Young University. He has worked previously as a bibliographer for YBP, Inc., as Head Acquisitions Librarian for the University of North Carolina, Greensboro, as Director of Resource Acquisition at the University of Nevada, Reno, and as Associate Dean for Collections & Scholarly Communication at the University of Utah.

Discussion

35 Thoughts on "A New OSTP Memo: Some Initial Observations and Questions"

This phrase leads me to another question: “all peer-reviewed scholarly publications authored or co-authored by individuals or institutions resulting from federally funded research are made freely available and publicly accessible by default in agency-designated repositories without any embargo or delay after publication.”

I’d like to know more about what they mean by “by default.” Does that imply an automated system/procedure to submit published papers to the appropriate repositories? How do they envision that working for small publishers who may not have the technological or financial resources to build such a function? Or, will author self-archiving be permitted, even though it would not be done “by default”?

Maybe the reference to ‘by default’ suggests that other repositories might be acceptable where more appropriate? A lot down to the detail in the guidance from the agencies though.

I read “by default” as simply it should be made public in agencies’ repositories with no embargo, unless there is a good reason otherwise. Nothing to do with automated or not automated. This memo isn’t directed to publishers – it is directed to Federal agencies and their funded researchers.

Hi Rick. Thanks for this excellent post. I can clarify one of your final questions. The memo says “Any provisions of the 2013 Memorandum that are not updated or superseded by this new policy guidance are maintained.”

Thanks, David. If Karin and I succeed at getting an interview with Dr. Nelson, it will be interesting to see what her views are on the extent to which the new memorandum “update(s) or supersede(s)” the terms of the previous one.

Thanks for queuing up these questions. A few related ones come to mind.

What about copyright? Does that stay with the publishers? Some copyright transfer agreements I’ve seen transfer copyright from the authors to the publisher in consideration for the publisher incurring the costs of publication, indexing and long-term preservation, etc.. Then the publisher licenses back the right of authors to post accepted manuscript (AM) versions on repositories. I didn’t see anything in the Nelson memo about AM vs publisher versions of record.

What about costs? Are societies and publishers expected to eat them? It’s certainly within the purview of agencies funding research to require that those accepting funding or employment will ensure that the resulting scholarly research articles be published open access. But the researchers will howl that that will mean less money to do the research. Publishers are loath to share their costs, but I’d guess they run between $1000 to >$4000/article for most reputable journals. (That guess is from OA fees or adding up paid staff fractions at a couple society journals I’m familiar with, guessing on living wages, and dividing it by articles published per year). Depending on the proportion of federally funded articles at a journal, that could be a big bite if journals Will agencies think it’s already paid for, and journals should thus give it away for free?
And what is federally funded research anyway? >50% of direct funding? Any indirect fraction? A gift that keeps on giving?

Interesting decree. Or recommendations that might or might not become interpreted as decrees.

I don’t see anything in the new memorandum that addresses the issue of copyright transfer, which should mean that it’s business as usual in that regard. Of course, what “business as usual” means is that funded authors can no longer transfer all copyright privileges to the publisher — but that’s been the case since the Holdren Memo.

As for costs — yes, societies and publishers are supposed to eat them, except to the degree that they can be recouped through article processing charges. But according to the White House study that seems to have largely informed the Nelson Memo, now that we’ve moved to a largely digital publishing environment, these costs consist mainly of those “primarily accosted with staffing and services related to moving the article from submission to publication, such as editing or proofreading.” So it seems publishers have nothing serious to worry about in that regard. [insert shrug emoji here]

Could you please give more details about this comment: “Of course, what “business as usual” means is that funded authors can no longer transfer all copyright privileges to the publisher — but that’s been the case since the Holdren Memo”?

I work with researchers often funded by a smaller government agency and actually under contract to this agency, part of which is to publically display the published work. With the way most copyright forms are currently written, they cannot check “US govt work” on the form because they are not employees of the govt. Also, even if they were, I think all authors have to be employees of the govt for that to work too, not just one or some of the authors. So they often end up signing over copyright, because there is no other appropriate box to check on the copyright form, and then asking for a final version to post given the circumstances, with mixed results. The articles often end up paywalled. If one tries to discuss with the publishers before signing over copyright, they often say the only other option is to pay for open access.

Clearly things would change with the Nelson memo, but in the meantime I’m wondering how exactly govt-agency-funded authors (again, not employees; and again, sometimes some are agency funded and some aren’t) avoid signing over copyright to the publishers. Although I’m asking from a smaller agency perspective, the same would apply to NIH-funded authors too I would think unless they were (all) NIH employees as well.

I have not been encountering the copyright forms that C.M. describes, but either way the task is to post the final version of agency-funded published manuscript not the accepted version (often quite different with edits/clarifications and/or corrections made before finalization). And ideally have it free access on the publisher site too, not paywalled.

Thank you.

Copyright is the exclusive right to decide who may publish, copy, distribute, reuse, create derivatives from, etc. (beyond fair use or fair dealing) a piece of original intellectual work. If you accept research funding from a government agency on the condition that you will only transfer those rights on a nonexclusive basis, then you are agreeing not to transfer all copyright privileges to a publisher. You can still technically assign copyright to the publisher if required, but it won’t really mean much because you can’t give the publisher the right to say how and where the work will be copied or distributed — which, since the Statute of Anne, has been the core privilege of copyright ownership. The whole point of the public access provision is to stop people from having to pay for access, which means taking away the copyright holder’s right to control access.

To be more specific, authors can transfer copyright, but in this case accepting a grant and working for the Federal government comes with a global nonexclusive, and irrevocable worldwide license to the US Federal government for the publications from that work.

Government rights are a little more complicated than this. Where a work is created by a federal government employee on the clock, copyright doesn’t even apply. For a work created for the federal government under contract (such as a funding agreement) other than the rare “work made for hire”, it depends on the terms of the contract. There is no default other than general copyright law if the contract is silent. This leaves us with some problems.
1. What if there are both employee authors and non-employee authors, and their contributions cannot be distinguished?
2. What if there was a written agreement and it cannot be found? In the Department of the Interior, contracts are routinely destroyed three years after the acceptance of the last deliverable, with no consideration of copyright as a “deliverable”.

There is a section of the Federal Acquisition Regulation that may apply. 48 CFR 52.227-14(c)(iii) says “… the Contractor grants to the Government, and others acting on its behalf, a paid-up, nonexclusive, irrevocable, worldwide license in such copyrighted data to reproduce, prepare derivative works, distribute copies to the public, and perform publicly and display publicly by or on behalf of the Government”–where “data” includes pretty much any recorded information. But this only applies if the Federal Acquisition Regulation applies, which it doesn’t for grants or permits, and if the contract doesn’t specify otherwise.

I appreciate the replies. It is complicated, indeed. The authors do add an Acknowledgment with some standard wording from the Federal Acquisition Regulation that E.M.M. mentions, and it’s good that that does seem to cover the work then being posted elsewhere. But there are always different scenarios depending on the authors involved and so on. Will certainly be staying tuned on this important development. Thank you for this informative discussion.

At the risk of oversimplification, it appears to me that so long as an accepted manuscript is deposited by author or publisher into the funding agency’s accessible repository, the obligation to provide public access to that manuscript is met. Publishers and societies are not obligated to link or a road map to said repository. Everything else is as business as usual (leaving off any discussion of data). What am I missing?

Bob, is your question directed to me or to a previous commenter? I’m not sure what you’re referring to with “What am I missing?”.

Rick, it was just a general inquiry/observation, and perhaps premature without the impending guidance from the agencies.

That’s a very thoughtful and very impressive explainer you folks banged out in a very short amount of time. Kudos. It’s worth a more prominent discussion on implications for societies. Hopefully SK will be hosting more of these discussions as these new policies get worked out.

I’m glad you mentioned that the cost of managing and hosting research data has been ignored and the complications given short shrift. That is a major cost to do well, and data management is its own skill set that I think few scientists get training on.

12 months is probably the short answer. At present, these repository releases aren’t on day zero.

I have not seen any definition of data which distinguishes between quantitative and qualitative data.
We have been asking questions of ECRs about making data open [ciber-research.com/harbingers-2] and it is clear that ECRs whose data is qualitative (in much of medicine and some social sciences) cannot see a way to enable openness. Usually the contractual arrangements with patients and interviewees prevents this. I would appreciate understanding of where the new mandates leave researchers in this category

Do government agencies conduct any kind of peer review? In funding research, you’d think the government woukd want some assurance that what it funds is quality research. Do agencies just accept anything the author submits, no questions asked, or is there some kind of internal review process going on, at least by agency staff? And, if so, does THAT count as peer review?

I don’t see anything in the memo to suggest that the government is proposing to take over the publication function itself. It’s just mandating (or at least “mandating”) that articles and data sets be made freely available immediately upon publication. Funded researchers will still be free to submit their work to regular publishers — but they’ll only be able to publish in places that will allow them to make the publication freely available without embargo.

The memo doesn’t imply publication or peer-review by the government. It also doesn’t mandate APC charges unless the publishers demand that all things submitted by government-funded researchers must be only “Open Access”.
My reading is that some agencies will just continue as is, obtaining the authors final peer-reviewed manuscript and making it available, but without the 12 month embargo.

Some require peer review and policy review for products authored by their employees. This may be in the form of pre-submission requirements, meaning that work gets pre-vetted before it gets to the journal process, which conceptually should be a good thing. The policy review part can get touchy. For example, I know from co-authoring that NOAA, EPA, and USGS have such policies. Grant recipients who are not federal employees have generally not been required to submit to agency clearance to publish in my experience, although YMMV. (Disclosure, I’m a federal scientist, although I chime in here in my personal capacity and my official duties are unrelated to setting public access policy).

They generally don’t require a agency clearance but they are required to send a copy of their work (sometimes final peer-reviewed author’s manuscript is required, or sometimes a published version is acceptable)

In Section 3, when describing Peer Reviewed Scholarly Publications, the memo explains what parts of their plans Federal agencies should update allowing for public access. An interesting statement is found in point ii that may help answer the Accepted Manuscript question. “How to maximize equitable reach of public access to peer-reviewed scholarly publications, including by providing free online access to peer-reviewed scholarly publications in formats that allow for machine-readability…” Granted that many publishers have multiple ways of doing this, but one ramification of this (if one follows a strict reading) is that full-text XML should be available at publication, which in turn means that neither header XML (abstracts and metadata) or PDF formats will meet this standard. So the AM version of the manuscript as it is currently created in many workflows as a service to authors by publishers will not meet the memo’s guidance. If this reading about the call for machine readability is correct, then full-text XML encoded in something like JATS will be required, which means the easiest path is some version of Gold Open Access or free access that opens all published formats. Being easiest of course does not mean cheapest or most viable to a publisher’s business or to an unfunded author’s ability to publish research.

Question: What is the definition of “federally funded” as used in this memorandum (and for that matter, in the Holden memorandum)? Is funding “in kind”, such as the use of federal facilities or other federal resources, included?

Comment: The existing policy (from the Holden memorandum) and parallel requirements in research permits have been difficult to enforce. Similarly, enforcing public access to “government works” is a problem, especially when there is mixed private-government authorship. I do not think this is due to malice on the part of staff, researchers, or publishers. In the rush of publication, people unfamiliar with legal requirements simply forget. If an open access publication fee was not included in the original budget for the research, it is tempting to publish in a paywalled journal so as not to have to pay it from already stretched budgets or worse, out of pocket. We may never know if anything resulting from funded or permitted research has been published at all unless the permit holder tells us or we stumble across it while searching for something else. Few agencies have the staff to follow every research permit holder and nag them about this. Requirements like these, however meritorious, require implementation and enforcement–neither of which is free.

I don’t have an answer to your question, though it seems to me that both the Holdren and Nelson Memos concern themselves specifically with research that receives direct, financial funding as distinct from in-kind support. I could be wrong about that, of course.

In response to your comment: as discussed in my piece, the whole question of how “mandatory” this mandate is remains unanswered. Apart from the difficulty of enforcing it — which, as you point out, is a real problem — there’s the higher-level question of whether the Nelson Memo lays out literal requirements. Certainly the language in which it’s written strongly suggests otherwise, which means there are two possibilities here, both of them intriguing. Either:

a) the memo’s language (“OSTP recommends that…”; “Federal agencies should report to OSTP…”) was constructed purposefully and means exactly what it says, in which case its provisions are not requirements but rather recommendations and requests, or

b) the memo was intended to lay out binding requirements for federal agencies, but was so sloppily written that it conveys a very different message.

c) That regardless of language (should vs must, etc.) the degree of obligation derives from the context and not the content of the memo. E.g., is the agencies funding going to be negatively impacted by not having said plans submitted? If so, then whether the memo says should or must, the consequences likely mean that the agencies will.

Angela Cochran wrote a post about that particular confusion back in 2013:
https://scholarlykitchen.sspnet.org/2013/12/09/what-does-federally-funded-actually-mean/

To me, the answer is straightforward — if the author lists federal funding in the paper, or submits the paper in progress or renewal reports to the funder, then it’s federally funded. In the past, authors have been generous with giving these credits — they want funders to see them as productive and to see their funds leading to results. If authors in the future have to pay an APC on any papers where they credit federal funding (and not pay one where they don’t), perhaps they may become more conservative with such credit.

Ah, an oldie but a goodie. That was my first TSK post and I can’t believe we are back there again. Having just taken a cruise through some of our data, there are lots of federal funding acknowledgements listed for Review Articles, Editorials, and Letters to the Editor. There are also a lot of collaborations with pharma. Will the NIH policy mandate that if there is an industry or cancer center collaborator and the only option is OA by APC that the money not come from the grant?

“We may never know if anything resulting from funded or permitted research has been published at all unless the permit holder tells us or we stumble across it while searching for something else.”

Many (most?) Federal grants have wording to place federal funding in their acknowledgements section. And most publishers pick up that funding that goes in the CrossRef record, not that it is always done, or sometimes is mis-typed. CrossRef however doesn’t generally have Federal funding listed in the sense of intramural (Federal employee-performed) research. Web of Science and some other indexing platforms do pick up funder information from the performer fields.

As to wording in the acknowledgements section, I suspect it depends on agency rules. For example, I find no such requirement in Department of the Interior grants and I’ve seen grant-funded research published in paywalled journals. For an example of how another agency does it, check out Treesearch (https://www.fs.usda.gov/treesearch/). The Forest Service already requires employee authors to post works to this repository as soon as possible.

Requiring funding information in the acknowledgements section is totally separate from a requirement to be open or behind a paywall. Publishers place the funding information, often with the award numbers, from the Acknowledgements in CrossRef, and one can search CrossRef for all items listed with X funder, or X and their sub-organizations. CrossRef data is also used in the CHORUS dashboards.

Separately, since the 2013 memo, all the large Federal agencies have required intramural authors (authors who are Federal employees) to place publications (at least authors final peer-reviewed manuscripts) in a specific agency repository (many of them use PubMed Central as that repository) and also require the same for extramural authors.

“If an open access publication fee was not included in the original budget for the research, it is tempting to publish in a paywalled journal so as not to have to pay it from already stretched budgets or worse, out of pocket.” Also, a common practical snag is expiring agreements. Federal funding agreements commonly have time limits, and by the time the OA APC invoices would come around, the agreements are expired and the funds had to have been spent down. Thus the comparative elegance of the Holdren compromise and CHORUS. C’est la vie.

OSTP’s announcement did not include the word ‘journal’. Does this leave the door open to authors and their employers to find alternative ways of publishing federally-funded research findings in line with OSTP’s stipulations on things like metadata? I wonder how many will choose to walk through that door to find cheaper ways of publishing (and perhaps in formats free from the strictures of journal article norms). After all, there’s a ton of research being published outside journals (and books) today – including from researchers in university research centres. Maybe OSTP’s announcement will push more to think ‘why do I need to publish in a journal at all?’

Comments are closed.