Open Access Archivangelism

Dear Falk,

First, I apologize for my school-masterish tone!

On a planet which still has far too few OA mandates and much too little OA, it cannot be repeated often enough that every single mandate is a step forward, and welcomed by all (in the research community!)

But I hope you will agree that optimizing these first pioneering mandates is very important too, to provide a tested, successful model for others to follow.

This is the reason for focusing here -- in this side-discussion that has arisen from the underlying discussion of the strengths and weaknesses of the US's NIH mandate -- on the detailed breakdown of Austria's FWF mandate's compliance rate and the formal and implementational details that have been generating it.

You have kindly provided some very important benchmark data on the compliance rate for the FWF mandate in your previous postings:

The compliance rate for the FWF mandate in its current form is 65%.

Of this, half (32.5%) is Green OA self-archiving and half (32.5%) is from publishing in fee-based Gold OA journals, one third of them exclusively Gold OA (11%) and two thirds (22%) hybrid Gold (meaning the journal is still charging institutions for subscriptions, but authors can pay an additional fee to make their own individual article OA -- and FWF pays that extra fee for fundees).

The global baseline for the annual un-mandated Green OA self-archiving rate worldwide is about 20% (although Yassine Gargouri will soon be reporting some new results suggesting that this worldwide un-mandated rate may have increased in the last few years).

So it has to be admitted that the FWF mandate's net gain from raising the baseline un-mandated rate of 20% to 32.5% for Green OA self-archiving is not very great (and much smaller than the Green OA rate generated by other current mandates -- such as Belgium's FNRS mandate -- that may be formulated and implemented in ways that could help Austria's FWF mandate increase its Green OA compliance rate too).

The bulk of the OA generated by the FWF mandate (32.5%) comes from articles published in Gold OA journals or in subscription journals offering optional ("hybrid") Gold OA publishing for an extra fee; FWF hence increases by 22.5% the global Gold OA baseline rate, which is under 10%.

But this Gold OA increase, as you will agree, has been bought at a price -- up to 3000 euros per paper. I ask you to keep that figure in mind in some of the replies I make, below, to the points you have raised.

It is also very important to know the discipline breakdown for Green and Gold compliance rates, because, as Andrew Adams has noted, PMC/UKPMC are just for biomedical research, and, as we all know, physicists have been self-archiving in Arxiv for over two decades at very high rates, un-mandated. Hence their contribution is not a result of the FWF mandate.

On 2012-05-21, Falk Reckling [FR] of Austria's FWF wrote on GOAL:

FR:
In 2006 no Austrian institution had a mandate or an IR. Today only some Austrian institutions have an IR [Institutional Repository] but no one has an OA mandate.

Austria's IR tally in 2006 and today, as well as Austria's mandate tally in 2006 and today are roughly comparable with those of other countries: Increase in the number of institutions with IRs (but with those IRs remaining near-empty) and still extremely few mandates (either from funders or from institutions) -- although Austria does seem to be unusually low in its number of IRs relative to its number of universities.

So the question is: What can be done to generate more IRs and more mandates in Austria?

This is precisely where the FWF can help, in three ways:

1. Require Green OA self-archiving of every FWF-funded article, whether or not it is published as Gold OA.

2. Designate the author's IR as the locus of the deposit (with OpenDepot as the interim alternative locus, for fundees whose institution has not yet created an IR).

3. Only offer to pay for Gold OA if conditions 1 and 2 have been fulfilled.

This will ensure that FWF fundees self-archive. 

It will encourage fundees' institutions to create an IR. 

It will engage fundees' institutions in monitoring and ensuring compliance with the FWF OA mandate.

It will motivate and facilitate the adoption by fundees' institution of a Green OA mandate of their own, for all of their research output, in all disciplines, not just FWF-funded research.

It will ensure that most of the compliance with the FWF mandate is not just Gold OA paid for by FWF at 3000 euros per paper.

FR:
By organizing a nationwide network we now try to tackle these problems.

Before organizing a nationwide network in place of upgrading the FWF mandate with conditions that induce institutions to create their own IRs, it would be well to look at the experience of France's HAL, which is a nation-wide repository just as empty as individual IRs that have not mandated deposit.

Years more can be lost travelling down that garden path...

FR:
At the same time, we noted that a lot of Austrian scholars were/are voluntarily willing to deposit their articles in central disciplinary repositories like arxiv, Repec, SSRN, Citeseer or PMC.

I have already replied about the profound denominator error that you are making here. The un-mandated deposit rate for central disciplinary repositories is just as low as the un-mandated deposit rate for institutional repositories. The crucial factor is not the repository but the mandate. 

And convergent, collaborative mandates (from both institutions and funders) designating the author's IR as the locus of deposit will generate far more institutional mandates than divergent, conflicting ones, for the many reasons I've already described.

One also has to be careful how one counts one's central repositories: Arxiv, as noted, is one of the prominent exceptions to the global un-mandated Green OA self-archiving rate. Physicists self-archive in Arxiv, un-mandated, at a much higher rate than other disciplines self-archive (anywhere) un-mandated. 

In over two decades, however, the only other discipline that seems to have followed the example of physicists un-mandated self-archiving) is mathematics. It would seem to be a strategic mistake to wait yet another two decades hoping that un-mandated self-archiving will generalize to other disciplines, rather than just to go ahead and mandate it.

Moreover, there is a third prominent exception to the global un-mandated self-archiving rate (20% overall, but much closer to 100% for physics and maths, in Arxiv) and that is computer science, which has been doing high rates of Green OA self-archiving without needing to be mandated to do so -- but they have not been depositing" in Citeseer (which is not a repository at all, but a harvester):

Computer scientists have been self-archiving on their own institutional websites (since long before IRs were invented). But their admirable un-mandated practice has not generalized in over two decades either.

Citeseer does provide a good case in point, though, for the power and efficacy of central harvesting, navigation and search across distributed local deposits.

Google and Google Scholar are further examples of the power and functionality of central harvesting across webwide distributed contents: one does not deposit centrally in google. 

There are more examples of central harvesting and navigation/search over distributed content.

But there is no point in further developing the potential of metadata harvesting and functionality while OA content is still so sparce.

That, again, is what OA mandates are for -- and why it is so important to optimize them, so they maximize compliance and OA.

FR:
With BMC, PLoS and others the need for covering APC arose. And we found it useful to support an alternative business model, as other renown institutions did [see COPE] :

[On COPE, see these critiques of paying pre-emptively for Gold before or instead of [effectively] mandating Green OA.]

Paying the additional costs of Gold OA article processing charges (APCs) is fine, if one has already done what is needed to maximize all OA generated by one's OA mandate (and one has the spare cash). 

But (it seems to me), it is very far from fine to spend all that extra money without first having done what is needed to maximize all OA generated by the mandate.

FR:
I do not agree with your position that Gold OA is costly and Green OA is nearly for free. In practice, Green OA costs a lot of time and money for creating repositories, establishing mandates, having well-informed supporting staff, interpreting publishers policies, advising researchers, depositing papers, e.g.

I would be very grateful to see what actual Green OA costs you have in mind, Falk. Like the denominator fallacy, it is crucial here to compare like with like. Gold OA costs are from 500-3000 euros per paper.

IR software casts nothing, server space costs next to nothing, IR one-time set-up time is a few days of sysad work, and annual IR maintenance is a few more days of sysad work, per year. IRs are set up for a variety of reasons, not just OA, but let us pretend as if the IR costs are just OA costs: How much do you think that adds up to, per paper deposited? (And bear in mind that adopting a mandate costs nothing, and greatly increases the number of papers deposited, hence decreases the cost per paper.)

Yes, extra money can be spent, and is being spent, on "having well-informed supporting staff, interpreting publishers policies, advising researchers, depositing papers [in place of authors]".

But the very same thing can be said about these additional expenditures as what was just said about expenditures on Gold OA fees: It's fine to spend this extra money if you have the extra cash -- but not if you have not adopted a mandate that will maximize self-archiving. Most IRs are spending all this money without a mandate (since most IRs don't have a mandate, let alone an optimized one).

So we are again speaking apples and oranges, if we try to rationalize spending scarce cash on Gold OA instead of optimizing our OA mandate in the direction of institutional Green OA self-archiving on the grounds that IRs are costly:

If the costs of Green OA and Gold OA are compared on a per-paper basis (as they need to be, to make sense), there is no contest: Green OA is incomparably cheaper, and Green OA mandates generate incomparably more OA.

FR:
That might give you some reasons why we find, for example, that UKPMC offers a very good option to solve some these problems 

UKPMC, a central repository for UK biomedical research, populated mostly by funder mandates, does not even address the matter at hand here, which is about ways to optimize those funder mandates so that they will generate more OA.

The UK too, like Austria (and the US) would benefit from much greater funder mandate compliance and would also generate many more complementary institutional mandates were it to:

1. Require Green OA self-archiving of every FWF-funded article, whether or not it is published as Gold OA.

2. Designate the author's IR as the locus of the deposit (with OpenDepot as the interim alternative locus, for fundees whose institution has not yet created an IR.

3. Only offer to pay for Gold OA if those two conditions have been fulfilled.

This will ensure that fundees self-archive. 

It will encourage fundees' institutions to create an IR. 

It will engage fundees' institutions in monitoring and ensuring compliance with funder OA mandate.

It will motivate and facilitate fundees' institutions to adopt a Green OA mandate of their own, for all of their research output, in all disciplines, not just funded research.

It will ensure that most of the compliance with the funder mandate is not just Gold OA paid for at 3000 euros per paper.

FR:
Finally, we see no contradiction to support both Green and Gold the same time, but we think in the end a change of the business model should be envisaged.

It is question of priorities, contingencies and timing.

Green OA self-archiving has to be made universal first, by both funder and institutional mandates, both designating institutions as the locus of deposit. That will generate 100% (Green) OA. That, in turn, will eventually make subscriptions unsustainable, reduce costs, and induce a conversion to Gold OA, while also freeing institutional subscription funds to pay for it.

All the best,

Stevan Harnad

"FWF requires all project leaders and workers to make their publications freely available through open access media on the Internet."

Message to Falk Reckling (FWF):

Dear Falk,

I fervently hope that you are communicating with us on GOAL open-mindedly, with a view to gaining information you perhaps did not have, and with a readiness to revise policy if a valid case can be made for the fact that it would help.

Because all too often, I have alas found, those who come to OA policy-making tend to make some initial judgments and decisions, implement them, and then when either practical evidence itself, or those who have more and longer experience in OA and OA policy, call into question those initial judgments and decisions, the response is: "My mind's made up, don't bother me with facts!" and the initial policy simply becomes more and more firmly entrenched, regardless of the consequences.

It is too early for such rigidity, Falk. And Andrew Adams and I (and many others) are trying to explain to you what is amiss with both the FWF policy and the rationales that you are voicing here.

On Sun, May 20, 2012 Falk Reckling wrote on the Global Open Access List (GOAL):

FR:
[1] an [Insitutional Repository] has to exist and that is often not the case.
[2] And if an IR exists it is often not used,
[3] mandate or not.
[4] Here some central disciplinary respositories are much more successfull as PMC/UKPMC.

1. Many, many institutions have repositories, and those that do not yet have one are merely a free piece of software and a server sector away form having one.

2. Yes, almost all existing repositories are unused (at least 80% of annual institutional refereed research output is not deposited). But that is the point! That's precisely why deposit mandates are needed.

3. It is an enormous factual error, however, to say that institutional repositories are unused whether or not they have a mandate. Again, that is the whole point. There is abundant evidence that institutions that mandate deposit are not near 80% empty but near 80% full! (And especially when they have adopted the optimal ID/OA mandate of U. Liege.)

4. It is also an enormous factual error to state that central repositories like PMC/UKPMC are exceptions to the 20/80 rule (i.e., that only 20% of total research output is deposited un-mandated). The total research output of an institution is all the refereed journal articles, in all disciplines, that its authors publish each year. The total research output of a central discipline-based repository is all the refereed journal articles published each year by all authors in that discipline, in all institutions worldwide. To imagine otherwise is to fall into the denominator fallacy:

The annual percentage use of a repository is the annual ratio of deposited articles to all target articles within its ambit. For an institution, the denominator is obvious, and easily estimated. For an entire discipline, it is far from obvious, but it too can be estimated. And I can assure you that the un-mandated Bio-Medical Research content of PMC/UKPMC is no higher than the global 20% baseline for all other disciplines. What gives the illusion that it is otherwise is two things, one trivial, one nontrivial:

The trivial reason for this profound error and misconception is the simple fact that disciplines are much bigger than institutions. So the absolute number of articles in a disciplinary repository is much bigger than those in any institutional repository, even though their un-mandated content is just 20% in both cases.

The nontrivial reason for this profound error is the fact that much of PMC/UKPMC content is mandated (by NIH, MRC, Wellcome Trust), and for that subset the percentage deposit is of course much higher -- exactly as it is with institutional mandated content.

So the overall error is to conflate central repository content and mandated content, and incorrectly (and misleadingly) deduce that central repositories are doing better than institutional repositories because they are bigger and have more deposits.

Reflection will show that it is mandates that generate deposits, not centrality or disciplinarity (irrespective of whether the mandates are institutional mandates or funder mandates). 

(The Physics Arxiv is the sole exception, where un-mandated deposits are close to 100%, and have been for two decades: But two decades is far too long to keep waiting in the hope that the physicists' spontaneous, un-mandated self-archiving practices would generalize to other disciplines: they have not. That's why the OA movement has moved toward supporting mandates.)

And as several of us have now stated, the functionality of a central repository for navigation and search (which is certainly incomparably better than the functionality of any single institutional repository, where no one would ever dream of doing navigation and search) is fully preserved if the central repository harvests the metadata and links to the full-text from institutional repositories. 

The point being made here about the importance of ensuring that both institutional and funder mandates collaborate and converge on institutional deposit instead of diverging and competing is that it makes a huge practical difference -- both to the burden on authors and to the probability of persuading institutions (who are the universal providers of all refereed research, funded and funded, in all disciplines) to adopt deposit mandates of their own -- whether funders mandate institutional deposit or institution-external deposit.

But Falk, you do not seem to be hearing this in these exchanges so far: you seem instead to return over and over to the funding of Gold OA fees rather than the mandating of Green OA. Is there any hope of drawing your attention to this much more fundamental and urgent question, on which the prospects of OA growth in upcoming years hinges?

FR:
b) We believe that OA is better supported by the Gold road and therefore a change of the business model is needed. That means costs should be covered by APCs or institutions or mixed models.

It is a great pity if you are rigidly committed to this belief, which is not only erroneous (for the many reasons we have been describing) but costly, because of the premature, pre-emptive focus on getting OA by paying Gold OA fees instead of by mandating Green OA -- and designating institutional repositories as the locus for direct deposit.

If funders mandate institutional deposit, institutions (the universal providers) will mandate Green OA too, and we will have 100% OA (Green). That will already solve the research accessibility problem, completely.

Moreover, that is also the fastest and surest way to eventually convert journals to Gold OA (and liberate the subscription money to pay for it.)

Solving the research access problem does not immediately solve the journal affordability problem too -- but does make it into a far less urgent, life/death matter (since with 100% Green OA, all users have access, whether or not journals are afforded or cancelled.)

I profoundly hope you will set a good example for other policy-makers, by showing some open-mindedness, flexibility and reflection on these crucial questions.

Best wishes,

Stevan Harnad

This is a comment on
"Horizon 2020:
A €80 Billion Battlefield for Open Access"
an article in AAAS's ScienceInsider which notes that:

"Elsevier's embargoes for green open access currently range from 12 to 48 months"

First, it has to be clearly understood that the existing EU mandate (i.e., requirement) that the EU is now proposing to extend to all of the EU's €80 Billion's worth of funded research -- while something similar is being proposed for adoption in the US by the Federal Research Public Access Act (FRPAA) as well as by the currently ongoing petition to the White House (rapidly nearing the threshold of 25,000 signatories) -- is a mandate for the Green OA self-archiving, by researchers, of the final drafts of articles published in any journal. (What is to be mandated is not Gold OA publishing: You cannot require researchers to publish other than in their journals of choice, nor require them to pay to publish.)

And although some publishers do over-charge, and do lobby against Green OA mandates, the majority of journals (including almost all the top journals) have already formally confirmed that their authors retain the right to self-archive their final drafts immediately upon publication -- not just after an embargo period has elapsed.

Moreover, that majority of journals that have formally endorsed immediate, un-embargoed Green OA include all the journals published by the two biggest journal publishers, Springer and Elsevier! The only difference is that unlike Springer, Elsevier has just recently tried to hedge its author rights-retention policy with a clause to the effect that authors retain the right to self-archive if they wish but not if they must (i.e., not if it is mandated!). (See Elsevier's "Authors' Rights & Responsibilities")

(Curious "right," that one may exercise if one wishes, but not if one must! Imagine if citizens had the right to free speech but not if it is required (e.g., in a court of law). But strange things can be said in contracts...)

Elsevier is having an increasingly severe public image problem: It is already widely resented for its extortionately high prices, hedged with "Big Deals" that sweeten the price package by adding journals you don't want, at no extra charge. Elsevier is also itself the subject of an ongoing boycott petition (with over 10,000 signatories) because of its pricing policy.

So Elsevier cannot afford more mud on its face. Elsevier has accordingly taken a public, formal stance alongside Springer, on the side of the angels regarding Elsevier authors' retained right to do un-embargoed Green OA self-archiving of their final drafts on their institutional websites -- but Elsevier alone has tried to hedge its progressive-looking stance with the clause defining the authors' right to exercise their retained right only if they are not required to exercise it..

Elsevier's cynical attempt to hedge its green rights-retention policy against OA mandates will no doubt be quietly jettisoned once it is publicly exposed for the cynical double-talk it is. Meanwhile, Elsevier authors can continue to exercise their immediate, un-embargoed self-archiving rights, enshrined in Elsevier's current rights agreement, with hand on heart, declaring that they are self-archiving because they wish, not because they must, even if it is mandated. A mandate, after all, can either be complied with or not complied with; both choices are exercises of free will, yet another basic right…

Elsevier has formally acknowledged its authors' right to self-archive their final drafts free for all online since 2004.

Under "What rights do I retain as a journal author?", Elsevier's "Authors' Rights & Responsibilities" document formally states that Elsevier authors retain the right to make their final, peer-reviewed drafts Open Access immediately upon publication (no embargo) by posting them on their institutional website (Green Gratis OA):

"[As an Elsevier author you retain] the right to post a revised personal version of the text of the final journal article (to reflect changes made in the peer review process) on your personal or institutional website or server for scholarly purposes"

More recently, however, an extra clause has been slipped into this statement of this retained right to self-archive:

"but not in institutional repositories with mandates for systematic postings."

The distinction between an institutional website and an institutional repository is bogus.

The distinction between nonmandatory posting (allowed) and mandatory posting (not allowed) is arbitrary nonsense. ("You retain the right to post if you wish but not if you must!")

The "systematic" criterion is also nonsense. (Systematic posting would be the institutional posting of all the articles in the journal. But any single institution only contributes a tiny, arbitrary fraction of the articles in any journal, just as any single author does. So the mandating institution would not be a 3rd-party "free-rider" on the journal's content: Its researchers would simply be making their own articles OA, by posting them on their institutional website, exactly as described.)

This "systematic" clause is hence pure FUD, designed to scare or bully or confuse institutions into not mandating posting, and authors into not complying with their institutional mandates. (There are also rumours that in confidential licensing negotiations with institutions, Elsevier has been trying to link bigger and better pricing deals to the institution's agreeing either to allow OA to be embargoed for a year or longer or not to adopt a Green OA mandate at all.)

Along with the majority of publishers today, Elsevier is a Green publisher: Elsevier has endorsed immediate (unembargoed) institutional Green OA posting by its authors ever since 27 May 2004.

Elsevier's public image is so bad today that rescinding its Green light to self-archive after almost a decade of mounting demand for OA is hardly a very attractive or viable option:

And double-talk, smoke-screens and FUD are even less attractive

It will be very helpful -- in making it easier for researchers to provide (and for their institutions and funders to mandate) Open Access -- if Elsevier drops its "you may if you wish but not if you must" clause, which is not only incoherent, but intimidates authors. (This would also help counteract some of the rather bad press Elsevier has been getting lately...)

Stevan Harnad