Thursday, February 11, 2016

Are ‘predatory’ journals completely negative, or also a sign of something positive? *

* This post was first published on the SciELO in Perspective blog, on February 2nd, 2016 (and translations in Portuguese and Spanish are available on SciELO in Perspective as well).

It’s not nice for parents to find that their toddler child is telling lies – or at least trying to tell lies. But is it so bad? It seems that the ability to tell lies is a sign of a well-developing theory of mind. And a well-developed theory of mind is very helpful in social interaction and things like having empathy would be very difficult without it. So if your child is beginning to tell lies, it might be cause for celebration rather than regret! [1] Of course, it doesn’t mean that lying should be unconditionally encouraged, but a well-developed theory of mind may actually help in developing proper morals as well when the child grows up.

The emergence of so-called ‘predatory journals’ (see Beall’s list of predatory publishers) could be seen in a similar light. Predatory journals are not desirable, it goes without saying, but the fact that they come about is a sign of a developing market, and a true market in scientific publishing services is a good thing, in my view. A true market offers choice to the party who pays. In the classical subscription model of scholarly journal literature, the party who pays is typically the institutional library, and they have little or no choice as to which journal to subscribe to and which not, if the journals are considered to be relevant to the research being carried out at their institution. And if they are forced to make a choice, for instance because of lack of funds, it is an agonising decision which choice to make.

In some countries – in Latin America, but also The Netherlands, for instance – the government is playing a direct role in negotiations with publishers to provide access to scientific literature. In principle, this results in a single buyer, countering the effect of a publisher monopoly by being able to bring much more negotiating clout to the table, and as a result quite possibly wider access (as is indeed being reported from Latin America [2]). But it doesn’t amount to a functioning market. The absence of a real choice is not remedied. Even a government cannot just choose to provide access to one publisher’s content at the exclusion of others, simply because no single publisher provides all the content the research community needs. And a monopoly (single provider)/monopsony (single buyer) system has drawbacks as well. It has a tendency to lock in a certain balance of powers (or an imbalance; Stockholm Syndrome has been mentioned to me in that regard), and in doing so, presents an impediment to progress in the quest for increasing efficiency at decreasing cost. It removes, or at least diminishes, the incentives of individual or small players to decide to buy or sell at a range of prices and service levels. For authors, for instance, it removes any economic motivation to decide to publish in low-APC journals or to communicate their research results via preprint servers. And it makes it very difficult for new publishing initiatives to gain a foothold.

A true functioning market is not just ‘peaches and cream’, though, since any true market also attracts rogues. And predatory journals certainly are. (By the way, I remain to be convinced that all publishers and journals on Beall’s list are indeed genuinely predatory; it is my impression that some of his criteria are rather shaky.) The existence of rogues is an inevitability of human nature, I’m afraid, but to authors, the attraction, insofar as there is one, of potentially predatory journals comes in part from their usually low charges and quick publication.

Yet, even with the drawback of being polluted by predatory journals, a functioning market is preferable to a quasi-market, completely dominated by monopolies or monopoly-like players. A system of subscriptions, in which the party who pays – the institutional library – has practically no meaningful choice of what to buy, differs from one of article processing charges (APCs, which make open access possible), in that the party who pays ­– the author – is the party who does have a meaningful choice of where to submit and publish. So ‘flipping’ the system from subscriptions to APCs does deliver something much more akin to a functioning market, and ‘caveat emptor’, ‘buyer beware’, applies to all markets.

It is not strictly true, of course, that in an APC system the author pays. It is most often the funder who does, by including the means to pay in the authors’ grants. This leads me to think that the subscription system could also be made to work as a quasi-market, if authors (the ones having the real choice) were made aware of the cost of the choices they make. This would be the case if authors had to find or request money in their grants for articles published in subscription journals as well, and not just for articles published in APC-supported open access ones. Imagine what will happen if authors were presented with a bill – of, say, $5000, a reasonable estimate of the collective cost per article of subscription journals; more for so-called ‘prestige’ journals – by their institution for every article they publish in a subscription journal. The likelihood is that they will more often choose APC-supported open venues, especially when it is slowly dawning upon the scientific community that openness in itself is an essential part of the quality of a published article [3].

Making researchers responsible for the financial aspects of their publishing decisions also fits into the general logic of public policies, which put the responsibility of how grants are used on the researchers, after all. Besides, responsibilities that focus on the per-article publishing costs incurred are naturally more scalable with the growth of research output.

This may even help solving the underlying problem with the current journal publishing system: the conflation of scientific communication and career-advancing reputation management [4]. This goes beyond what publishers do: in many cases, the same institutions in charge of purchase decisions of journals are also in charge of research promotion and evaluation, which is mostly – if unjustifiably – based on journal prestige, which may well constitute a conflict of interest. The consequence of separating the two may be that more articles are posted as so-called ‘preprints’ (for communication) and that not all of those are subsequently published in journals (for career purposes), given the cost of the latter. Global preprint repositories such as arΧiv, bioRχiv, and others, would acquire critical importance in that scenario. I am not dismissing the need for reputation building and career progression, but should that really stand in the way of communication for the sake of scientific progress? Is it right that the results of research – particularly of research carried out supported by public funds – should be used primarily for the purpose of authors’ reputation stratification (establishing a ‘pecking order’) via journals, many of which are not openly accessible?

The publishers of journals will say that establishing a pecking order is not their only task, or even their most important one. Most of them maintain that their gate-keeping function, via the peer review they commission, is their true raison d’être. This is a strange argument, as they are only in control of the gate to their tiny little patch of the vast forest. Virtually every article can find a journal in which to be published and so be added to the scholarly literature. What is presented as gate-keeping is in effect just sorting and ranking articles according to vague criteria such as ‘quality’ or ‘relevance’ and therefore not distinguishable from establishing a pecking order. It may help some scientists, but it doesn’t help science. On the contrary: scientific communication is being held hostage by pecking order concerns. And what’s more, these efforts are costing the scientific community a great deal. In time, money, and impediments to sharing knowledge.

It bears repeating: scientific communication and reputation management should not be combined in the same system.

Apart from the obvious costs mentioned, this kind of ‘gate-keeping’ is abusing peer review. Not only does it cause ‘cascading’, many articles being rejected, resubmitted elsewhere, being rejected again, et cetera, until they are eventually accepted. At each stage needing peer review again, and so being a tremendous burden on peer reviewers. Peer review should be aimed at helping authors to improve their articles, by questioning assertions, methods, and the like, and only as a last resort rejecting for publication (in the way of some open access journals such as PLOS-One). In this light, the emergence of ‘easy’ journals, even ‘predatory’ ones, is not miraculous. It takes some getting used to, I presume, but I see great potential in an approach like this: author-arranged open peer review [5].

Jan Velterop

References
1. PAN DING, X; et al. Theory-of-Mind Training Causes Honest Young Children to Lie. Psychological Science. 2015, vol. 26 nº 11, pp. 1812-1821. DOI: 10.1177/0956797615604628
2. Personal communication.
3. VELTEROP, J. Openness and quality of a published article. SciELO in Perspective. [viewed 27 January 2016]. Available from: http://blog.scielo.org/en/2015/12/16/openness-and-quality-of-a-published-article/
4. VELTEROP, J. Science (which needs communication) first, careers (which need selectivity) later. SciELO in Perspective. [viewed 27 January 2016]. Available from: http://blog.scielo.org/en/2015/10/29/science-which-needs-communication-first-careers-which-need-selectivity-later/
5. Peer Review by Endorsement (PRE). ScienceOpen. Available from: http://about.scienceopen.com/peer-review-by-endorsement-pre/

Sunday, August 16, 2015

Peer Review By Endorsement

Below is the original proposal that eventually led ScienceOpen to give it a go and make 'Peer Review by Endorsement' a legitimate, efficient, and affordable alternative to the generally (very) expensive publisher-mediated peer review mechanism we all know.

It was originally called JONAS, for Journal Of Nature And Science, as a way to indicate the wide potential for this approach in terms of disciplines covered, and yes, a gentle play on the titles of prestige journals, also known sometimes as 'glam' journals :-).

(The idea of JONAS has something in common, of course, with the idea of overlay journals that have at various times over the last decade and a half been suggested for manuscripts deposited in arXiv, whereby peer review reports were added to them and so giving the article some 'official' publication status.)

This is the original proposal:
JONAS (Journals Of Nature And Science – working title) is a new approach to open access publication of peer-reviewed scientific literature.
Establishing a publishing system that addresses:
  • Open access, 
  • Fair and efficient peer review,  
  • Cost of publishing
  • Speed of publishing
  • Publication of negative/null results
Open Access. The JONAS publishing system focuses on the superb technical publication, in various formats/versions, of peer-reviewed articles for optimal machine and human readability and re-use.
Fair and efficient peer review. Anonymous peer review has problems around issues of transparency, fairness, thoroughness, speed, publisher-bias, specious requests for further experiments or data, and possibly more. JONAS is a system using signed, pre-publication peer review, arranged by the author(s) (many publishers ask authors who to invite to review their papers anyway), and just verified by the publisher (peer review by endorsement). Reviews will be open, published with the article that’s endorsed, non-anonymous, and the rules are that peer-endorsers must be active researchers, and not be, or for at least five years have been, at the same institution as, or a co-author of, any of the authors. Such a peer-review-by-endorsement system is likely to be at least as good as, and quite probably better than, the currently widespread ‘black box’ of anonymous peer review. As reviews/endorsements would be signed and non-anonymous, there is very little danger of sub-standard articles being published, as endorsers/reviewers would not want to put their reputations at risk. The review process between authors and endorsers is likely to be iterative, resulting in improvements on the original manuscripts. “Author-arranged” may include peer review being done by services specifically set up for that purpose, as long as the reviewers are not anonymous and conform to the JONAS rules.  
Cost of publishing. A system like this can be very cost-effective for authors. The technical costs of publishing are but a fraction of the cost usually quoted for organising and arranging peer review. First indication is that an amount in the order of a few hundred pounds sterling (£) per article can be sustainable, given sufficient uptake. Tiered charges could be considered depending on the state of the manuscript when submitted. If the manuscript needs very little work to bring it up to proper publishing standards, or if the author doesn’t want or need those services, the cost can be very low indeed. 
Speed of publishing. Since the peer-review-by-endorsement process has already taken place before the article arrives at the publisher, publication can ensue within days, even hours, depending on the state of the manuscript.
Requirements for manuscripts: ORCIDs for authors and reviewers/endorsers; inclusion of (permanent links to) datasets used, underlying data for graphs, a section “details for replicability and reproducibility” with clear and unambiguous identification of materials used, including reagents, software and other non-standard tools and equipment.  Input: Properly endorsed articles to be accepted in the form of Word, Pages, (LA)TEX, XML, HTML, Markdown, and Excel or CSV for data, and high-resolution image files (where possible scalable vector graphics) via an upload site or attached to emails. 
Output: Articles would be published as XML, HTML, PDF, ODF and ePub formats, as much as possible semantically enriched and aesthetically formatted, plus Excel/CSV for data (tables extractable and rendered in Excel from PDFs with software freely supplied). 
Commenting and post-publication review: to be enabled and encouraged for all articles, links to comments to be provided with each article. Comments may be made on different sites, and will be linked to, if that is the case. Anonymous comments will be ignored. 
Access Licences: CC-BY or CC0 — DOIs for the articles would be assigned/arranged by JONAS. 
The core of the JONAS system is effectively to have OA journals with a low-cost structure, with superb and highly optimised technical quality of the published articles (machine-readability and re-useability!). The principal difference with other OA journals would be the pre-arranged open peer review ("peer-review-by-endorsement"), organised by the authors themselves, according to a set of rules that ensures a reasonable level of assurance against reviewer bias (because of its openness and non-anonymity, actually more assurance than is provided in the usual anonymous peer review as widely practiced). Since arranging peer review is one of the major costs of any publisher (mostly staff costs), leaving that part of the publishing process in the hands of researchers and the academic community can make a great difference to the cost of publication. So far, efforts to reduce the cost of publishing have been concentrated on technical issues. Changing the mechanism (emphatically not the principle) of peer review offers much greater scope for cost reduction. 
What JONAS' job would be is to take such peer-endorsed articles and make them into professionally published and complete (including data and metadata) documents, adhering to all the technical, presentational and unique identifier standards, in a number of formats, linked and linkable to databases and other relevant information, human- and machine-readable and suitable for widespread usage, for text- and data-mining, for structured analysis (incl. semantic analysis) and further knowledge discovery, and, crucially, for long-term preservation in repositories and archives of any kind. 
Any manuscript submitted in advance of peer-endorsement having been procured, would be placed, ‘as is’, on JonasPrePubs, a ‘preprint’ server, at no cost. 
The JONAS publishing system is also superbly suited to scientific societies and other groupings that wish to have their own journal. Such a journal can be fully integrated in the JONAS system, provided the manuscripts are submitted fully peer-endorsed or peer-reviewed (whether or not arranged by the author(s) or the scientific society in question). The charges per manuscript for individual authors and for societies wishing to publish their journals in the JONAS system would be the same. 
The JONAS methodology could be implemented on various publishing platforms.

Jan Velterop

Monday, July 20, 2015

Levelling the Open Access – Paywall Playing Field for Authors

Early career researchers are often reported to express the view that they face a dilemma. Submit to – and hopefully publish in – an open access journal, with possibly a relatively low impact factor, or in a traditional, pay-walled journal with a relatively high impact factor.

Given the large number of traditional pay-walled journals with low, or no, impact factors, I find this not the most credible argument. And even for ‘glam’ journals there are now good open access alternatives.

And yet, there are moments when I understand researchers when they are having to decide where to submit their papers. Do they choose an older subscription-supported journal, or a younger APC-supported open access journal? In the latter case, they’ll have to find the funds to pay the Article Processing Charge; in the former, they don’t, since subscriptions are paid out of the library budget.  It does make a difference to a researcher's perception. Even though in many cases it is the funder who provides the money for the APCs, the researchers are aware of the cost and part of the decisions they take are financial/economic ones, even if sometimes subconsciously. They are not confronted with financial/economic decisions if they submit to a paywalled journal. Convenience may set in, perhaps in the form of a certain laziness, and a decision to stick with the old hassle-free subscription journals is easily taken.

It may happen here and there, but what I have not seen is attempts by the library community to confront researchers with the cost of paywalled journals. I'm not talking about the subscription price, but about the cost to the system of a single paper published in such a journal. It is a significant cost. For subscription journals published by the major publishers, this is on average in excess of $5000 (there are differences depending on the publisher), and for the ‘glam’ journals presumably more, much more (Phil Campbell, editor-in-chief of Nature, estimated costs of $30,000–40,000 per paper in 2013. That’s costs to the publisher; costs to the system will be higher, as they include profits.)

Now imagine that universities, perhaps via their libraries, take care of any payment to publishers, be they subscription charges or APCs, and then reclaim a per-article fee from their grants whenever researchers publish their articles. The amounts for APCs identical to the amount charged by the open access journal in question, of course; the amounts for articles in subscription journals on the basis of the average per-article revenue of the publisher of those journals. (These amounts may be reasonable estimates, I imagine, as they will seldom be known in detail.) The amounts thus reclaimed for articles in subscription journals could then be used for the journal acquisitions budget. 

I have no illusion that this would solve all the problems of the cost of scientific publication, but it will increase general awareness of the true cost of publishing in subscription journals, and may help to level the playing field, to use an old cliché, between open access and pay-walled literature in the mind of scientists at the point when they decide where to publish their papers. 

Jan Velterop

Tuesday, July 07, 2015

Dealing with information overload*

Are we overwhelmed by the amount of scientific information that is being published? PubMed adds an average of more than two abstracts a minute to its database, and that is just in the life and medical sciences. If that doesn’t amount to information overload, what does?

[More...]

*This post was first published May 18th 2015 on the SciELO blog. The 'more' link redirects to that blog.

Thursday, May 07, 2015

Keeping the Minutes of Science – Looking back 20 years

I attended the Royal Society discussions on the Future of Scholarly Scientific Communication, held in two sessions of two days, the last one ending yesterday. Most interesting meetings. (Tweets from the meetings under the hash tag #FSSC).

The discussions brought back memories of an article I wrote 20 years ago, on "Keeping the Minutes of Science". I re-read it this morning, and it is interesting (to me at any rate) how my thoughts have evolved, from just before the introduction of what later became known as the 'Big Deal', when this article was written, via becoming an Open Access advocate a few years later (before it was even called open access yet) to proposing to change the way peer review is done.

I was at Academic Press at the time (well before it was taken over by Elsevier), which may explain some of the thoughts I then had. I am posting the full article below. (The bibliographic reference to the original is J.J.M. Velterop, “Keeping the Minutes of Science.” In: Electronic library and visual information research (Elvira 2). Edited by Mel Collier and Kathryn Arnold. Proceedings of the second ELVIRA conference, held in May 1995 at De Montfort University, Milton Keynes. Aslib. 1995, pp 11-17.)

The article:
KEEPING THE MINUTES OF SCIENCE
INTRODUCTION
Scientific journal publishing differs markedly from most other kinds of publishing. Born out of the exchange of letters on scientific topics and results, it has evolved into as much a service to scientists who need to publish the results of their work, as a service to those who need to be kept abreast of scientific developments elsewhere. Unlike most other publishing, scientific journal publishing has as much to do with the proper recording of scientific activity as with the conveyance of information. As far as the latter is concerned, scientists do not rely on scientific journals alone any more, in order to keep informed. Journals are only one of the variety of ways in which scientists gather scientific information. However, scientific information that comes to a scientist via a scientific journal still carries the highest degree of authority for information, as it has been peer reviewed and gone through a certification and validation process before reaching the reader.
For the author a scientific journal is essential. There are currently no other ways for a scientist to get his work certified and validated than publishing in a journal of good reputation. This certification and validation process is of immense value to science. It is to a large degree the result of a long self-organising process that has grown into a highly sophisticated structure (including a “pecking order” of journals) in which scientific results are placed in a hierarchical context, are ‘taxonomised’, standardised, formalised, made accessible through a uniform and globally accepted reference system. It forms the backbone of a scientific archive.
Scientists, libraries and publishers share a responsibility to protect and safeguard this elegant system of “keeping the minutes of science”. Collapse of this system is something science can ill afford, even though in the future it is quite possible that new systems will arise and eventually be standardised and become globally accepted. A hiatus of a number of years, however, will do great damage to the continuity of scientific research and if there is to be a transition into a new situation, it has to be carefully managed.
This paper does not deal with the issue of print journals versus electronic journals. Journal publishing is discussed as a concept, independent of the medium. It is obviously recognised that the introduction of digital technology can change the way journals are being used and this technology can help achieve efficiencies that were hitherto impossible. But the concept of a journal remains essentially unchanged.

PROPOSITIONS
1. If not properly published, scientific results are fairly useless — because usually indistinguishable from speculation.In most democracies, anybody is free to publish anything that is not obscene or libellous, and there are quite a few countries where one can get away with that as well. Whether it is true, whether the argument is not flawed, whether it is ambiguous, whether it is informative, whether it is new, whether it is original, etcetera, is quite another matter. The level of reliability of unaccredited publications is not known. It may be high or dismally low. Although the unaccredited material may be intrinsically reliable, the fact that it is not reasonably certain that it is, makes it unreliable. In order to substantially increase the chance that all or most of the above can be relied upon as representing the current state of knowledge, peer reviewed scientific journals evolved.
In the beginning, at the time the first scientific journals were established (1655, LE JOURNAL DES SÇAVANS, France, quickly followed later in the same year by PHYLOSOPHICAL TRANSACTIONS of the Royal Society of London), peer review did not yet exist. But it became clear fairly soon that not everything submitted was fit for publication and criteria for admission were developed, although journal articles were initially not regarded as definitive publications, as the mature research results would still finally be written up in monographs1.
It would be a mistake to believe that peer review ensures quality and integrity of the material at all times. The American of Research Integrity (ORI) can testify to that. However, the chance of finding total nonsense in established peer reviewed journals is slight. Peer review does not only ensure a good measure of QIVAS (Quality, Integrity, Veracity, Accuracy, and Security)2, but also a hierarchy, not just of importance, but of relevance to the central issues of the discipline concerned as well. On top of this, publication in a journal provides priority and authority (in the original sense of established authorship). Results cannot easily be ‘hijacked’ any more once they are received by, and published in, a reputable journal and (undetected and uncontested) plagiarism becomes pretty difficult, too. Published results have, by virtue of having been published in a peer reviewed journal, become part of the accepted and acknowledged body of scientific knowledge or theory.

2. The growth in numbers of scientific papers is outside the influence or control of publishers.The laws of supply and demand do not apply in a straightforward way to scientific information. As Bernard Naylor has described it: “If there is an excess of supply over demand in the journals industry, and there seems no prospect of an increase in demand, the obvious alternative is that supply ought to fall. However, whichever way you look at it, supply is tending to increase. The normal self-readjusting tension between supply and demand fails to operate”3.
But we should not forget that it is the scientists’ very job to uncover and add to the body of knowledge!
Currently, the number of scientific articles published per year seems to be increasing fairly constantly within the order of 4 to 6 percent. This poses great problems to the global scientific community. ‘Twigging’ of disciplines and further specialisation is still the most common response. As long as the need and desire for insight into natural phenomena grows, scientific activity grows, and with scientific activity, scientific publications. Scientific papers are, however, not just used to record and convey results. They are also used to advance careers and boost egos. Indeed, in many situations they are the single most important measure of a scientist’s performance. And not just scientists are being measured by the number of their papers, but entire departments and whole research institutions as well.Growth of scientific material seems inevitable. It is becoming more and more difficult, of course, to be sure that one’s paper will actually be seen by the majority of the intended audience. So apart from ‘slicing’ their results into many papers which one would do for reasons of career advancement, some researchers resort to this technique in order to increase their chance to be ‘found’. Further growth of the amount of published material will only make this a more attractive (some would say necessary) option, in order to increase the ‘signal to noise ratio’, so to speak.
Publishers do not increase the amount of scientific articles; scientists do, driven by ‘publish-or-perish’ and performance criteria, but essentially just doing what they are expected to do and what they are paid for. Publishers only respond to the phenomenon and as often as not, also experience it as a problem. The irony is that the very same governments that insist on proving maximum performances with published papers are the first to cut the budgets that, by way of paying for subscriptions to the journals, support the mechanisms that make publication of those papers possible.

3. Scientific articles should be published only for their scientific merit, not for their commercial merit.One of the attractions of the current model in which articles are published in the context of journals is that no commercial judgement is passed. The only reason why an article is accepted or rejected is its quality or relevance to the particular forum that the journal constitutes. This is done exclusively by the active scientists who act as editors-in-chief, members of editorial boards, and reviewers of the journals, with no interference from the publisher. Particularly not if the publisher is independent of a particular scientific constituency and is unencumbered by any possible scientific (or even political) controversy or secret agenda.
Independent journal publishing differs markedly from book publishing, where the market potential also has a large influence on the decisions, made by the publishers, on whether or not to publish. Scientifically worthy books are, as a rule, not published if insufficient prospective buyers can be found.
The fact that journal articles are published on their scientific merits only is worth preserving. Market forces may (and will) introduce an improper bias. Just as it is improper to sell academic degrees, it would be improper to give undue preference to scientific results coming from wealthier institutes or companies. This would be likely to happen, though, if commercial criteria were introduced for the publication of primary research results. Wider distribution than the normal journal circulation is already being sponsored for certain articles. This is a welcome source of income to publishers and likely to influence decisions if the current system of publishing purely on scientific merit is compromised.
This is a realistic risk, though, as publishing on exclusively scientific criteria is being threatened by the advent of P3, or Pay-Per-Paper, a.k.a. Pay-per-View (in electronic environments), document delivery (DocDel) or individual article supply (IAS). After all, except in extraordinary circumstances, no publisher, independent or associated with a scientific society, will want to commit resources to the publishing of an article which offers no, or very limited, potential for recouping the investment through sales of the article. Such an IAS system is likely to favour publication of articles with clear commercial potential. In practice, this is likely to mean articles written by well-known scientists from English speaking countries, especially the United States (for that is where a major market will be found), and from universities or research centres with resounding names, or companies with deep pockets which are willing to sponsor the publication. It is not difficult to imagine claims of cultural and commercial imperialism in such a scenario.
 
4. Information is not a commodity.
The Oxford dictionary defines ‘commodity’ as “...a thing of use or value, spec. a thing that is an object of trade, esp. a raw material or agricultural crop”. The only faithfully recorded instant when commodities behaved like information occurred almost 2000 years ago. It concerned loaves and fish, and the emphasis should be on ‘faithfully’, not on ‘recorded’. Since then, it has not happened again that one could give away a commodity like a loaf of bread and still have it. It can be done with information, though!
It follows that information cannot truly be expected to have the same economic properties as commodities. Sharing information does not mean the same as sharing a bushel of wheat, unless it takes the form of keeping the first five pages of an article to oneself and giving the other five to someone else. Hence the virtually worldwide establishment of legal constructions like copyright, which are awkward and imperfect, but the only means currently available to make sure that the necessary resources remain to be committed to the recording and dissemination of information. It would be attractive for purchasers of information to treat it as a commodity, especially given its ‘loaf-and-fish’ properties. What really is copying of information and document delivery is known as ‘resource sharing’, or sometimes goes under the euphemism of ‘interlibrary loan’. The potential is enormous: take out one subscription per country and share it with all the other libraries. The law permits it! It doesn’t take a rocket scientist to see the consequences, though.

5. The real product of scientific journal publishers is not paper, not distribution, not content, but the service of providing a structured forum for scientific discourse. But alas, information is not a commodity and it is not even the publishers’ product. Stevan Harnad, editor of the electronic journal PSYCOLOQUY, describes in his paper “Implementing Peer Review on the Net: Scientific Quality Control in Scholarly Electronic Journals”4 how his eyes were opened in a conversation with a television network executive, who told him that TV‘s product is not the programmes that are broadcast, but the viewers’ eyeballs which are sold to the advertisers (although in many countries outside the USA this would be either not or only partially true). In a similar vein, scientific journal publishers’ product is not the content (that is the authors’), not the printed paper (that is the printers’), not the distribution (that is the mail’s), but the provision of a forum for the conduct of scientific discourse, which facilitates the proper keeping of the minutes of science. A journal is a concept, not necessarily a physical entity. The publisher provides a structured (‘pecking order’) and controlled (‘quality label’) forum, complete with gatekeepers (editor and reviewers), organisers (indexers for secondary literature and databases), and ‘translators’ (although the accepted lingua franca of science seems to have become English, the ‘real’ language of science is more than this: it is a closely knit framework of standardisations, rules and conventions, in the interest of precision and the avoidance of ambiguities, which amounts to a ‘grammar’ and ‘idiom’ that few scientists fully adhere to; hence the need for ‘translators’, better known as subeditors or technical editors, who often also provide conventional translation services for those authors whose native tongue is not English). Almost as an afterthought, publishers also arrange for composition (typesetting for print; SGML-coding for electronic dissemination), printing or mounting on servers, and subsequent dissemination.The service of providing a structured and controlled forum is as important for authors as it is for readers of journals. It is not for nothing that many scientific societies are charging the authors for publication of their articles. It is inherently fair that both authors and readers contribute to the maintenance of the forum. Societies are finding it difficult to keep up the system of page charges, as most independent publishers are not levying them. The American Physical Society is currently examining its page charge levies, after an experimental period of suspending page charges for manuscripts submitted electronically to PHYSICAL REVIEW C. One of the arguments used for continuing the system of page-charges is “...that it is reasonable to expect research grants to bear some of the publication costs, since publication could be considered an important aspect of research”5. Most independent publishers have concentrated on subscriptions as the sole source of financial support for the journals. Should this lead to journals accepting only papers from researchers at institutions or‘ companies underwriting and supporting the journal via a subscription? It would be fair, but hardly feasible and introduce a similar potential bias as discussed under Proposition 3.

6. Economic models for journal publishing exclusively based on subscriptions are becoming less viable; the ones based on individual article sales (document delivery) have never been, nor will ever be viable.The flaw in the previous proposition is, of course, that publishers in reality derive their income basically from treating content, printing and dissemination as their product. Content is first converted into a pseudo-commodity, with the aid of copyright, and then sold on a just-in-case basis by the subscription, and, reluctantly, on a just-in-time basis (or just-too-late, as Bernard Naylor aptly describes it!) by the individual article. The transactional document delivery model is particularly badly suited to the forum concept of a journal. It reduces the intricate fabric of written scientific discourse to the one-way street of information provision, devoid of much contextual and meta information that makes the package of a journal so valuable (even aside from concessions document delivery makes to browseability and serendipity). Also, each individual article would have to be purchased hundreds of times at the prices that currently seem usual, or carry a price tag that is substantially higher, in order to make it possible that investments in the publication of the articles could be recouped. But a more serious danger lies in the fact that document delivery leads to a model in which articles are no longer published on their scientific merit, but on their commercial merit, as already discussed under Proposition 3. Subscriptions are not satisfactory any longer either. Library budgets have decreased in real terms over the last decade, pretty much worldwide. Resource sharing and inter- library loan are a result, and subscriptions, in combination with fair-use and library privileges provided by law, are ideally suited to that. Or so it seems. The natural course of events is now that numbers of subscriptions will go down, interlibrary lending up, subscription prices also up as a consequence, with the result that users’ access to material becomes more difficult or cumbersome; libraries spend their budgets on administrative chores connected with inter-library loan instead of building collections in order to optimally serve their constituencies; authors see the potential of chance encounters of their articles with readers dwindle; scientific societies cease to exist or are forced to minimise the service to their members; and publishers go out of business. In short, everyone loses, except, of course, the suppliers of photocopiers and the paper they churn out.
7. Academic Press is convinced that there are viable alternatives that do much more justice to the needs of authors, libraries, users, and publishers, without, on the whole and in total, costing more.Fortunately, there may be other scenarios. Academic Press clearly sees the plight of the libraries, which is, by extension, the plight of the entire scientific community. Supply is not decreasing but increasing, demand is not increasing but decreasing (economic demand, by the entities that pay; not the end-user demand or the demand from authors for a vehicle to publish their papers in), so the only alternative left is to reduce costs. Academic Press is working on paradigms in which subscriptions are essentially replaced with licences, giving subsequent free electronic access to every user affiliated with the library taking the licence, thereby essentially reacting to increasing end-user demand, while at the same time accommodating the dire financial situation the libraries find themselves in. Options to grant major incentives to consortia of libraries for taking licences that span a whole range of journals are currently being explored. We are even including the possibility of offering such licences to loose consortia of all libraries in a given province, state, or even country, which are then free to share all the material amongst all members of the consortium in whatever form (electronically or on paper) and however frequently as is desired. This scheme is called APPEAL, for Academic Press Print and Electronic Access Licence.
In the view of Academic Press, the paradigm of an APPEAL licence would potentially offer promising benefits to all parties concerned:
the authors would have the assurance that their papers are directly available to a much larger potential audience than is the case now;
the libraries would be able to offer the research, teaching and student community much wider instant access to much more material;
the barrier to turning to and browsing through many more sources would be removed for researchers, students and other library users;
the publishers would be able to make the necessary investments in improving the sophistication of the ‘forum’ and the cost per unit-of-information ratio.
All this for substantially the same amount of money that is being spent on scientific literature by libraries now. On top of that, the libraries would be able to make appreciable savings on costly inter-library loan and cut down, or even eliminate, expenses for commercial or semicommercial (BLDSC) document delivery, at least regarding current journal material.
CONCLUSION
A form of formal publishing, whether in print or electronic, remains essential for the structuring and preservation of the body of scientific knowledge, however many problems the unrestricted growth of scientific knowledge poses to the global scientific community. It is imperative that sufficient resources continue to he made available for this. However, the resources currently used for ‘keeping the minutes of science’ can be used far more efficiently, doing much more justice to the scientific efforts carried out and catering much better to the identified needs of the scientific community.

REFERENCES
1 Manten, A.A. (1980), Development of European Scientific Journal Publishing before 1850, in: Meadows, A.J. (Ed) Development in Science Publishing in Europe, Amsterdam, Elsevier.
2 Lindsay, John (1995), Academics Can Do It For Themselves. Presented at the UK Serials Group Meeting, UMIST, Manchester, 1994. STM Newsletter; no. 96.
3 Naylor, Bernard (1994), The Future of the Scholarly Journal. Presented at the General Meeting of LIBER, July 1994.
4 Harnad, Stevan (1995), Implementing Peer Review on the Net: Scientific Quality Control in Scholarly Electronic Journals. In: Peek, R. & Newby, G. (Eds) Electronic Publishing Confronts Academia: The Agenda for the Year 2000, Cambridge MA, MIT Press.
5 APS News (American Physical Society), Member Input Sought on Page Charges, 9 March 1995 (11158). 


Jan Velterop

Wednesday, April 29, 2015

How to take peer review out of the clutches of publishers

You may remember that some time ago I wrote about JONAS – Journals Of Nature And Science – the essence of which was to take peer review out of the clutches of publishers and make it a purely academic responsibility again, what it should be in reality anyway. The result will not only be a cheaper system (by an order of magnitude compared to the current one), but also likely a better, fairer, more expert and possibly faster one. And because publishers' main focus will necessarily be on the technical issues of producing correctly XML-coded, archivable, preservable, findable, machine-readable as well as human readable, text- and data-mineable articles in a variety of formats for different purposes (e.g. XML, HTML, PDF, ePub), the currently often sloppy production may be greatly improved (you'd be surprised at the number of errors in the material published by even the publishers priding themselves most on quality – mixing up β with ß or + with ±, for instance!).

ScienceOpen has decided to follow up on the JONAS idea and recently announced that they will give authors the option of publisher-free peer review. Needless to say that I'm very pleased with that. Several scientists have remarked that "this is an important experiment" and expressed their hope that it will take off.

ScienceOpen will probably face some substantial hurdles, as is generally the case with new ideas in the area of scholarly publishing. I'm reminded of the early days of open access, in that regard. However, tenacity and persistence will do a lot to overcome those hurdles, as does help from those in the academic community who would like to progress peer review reform and open access.

Such help doesn't have to be onerous. Simply talking to colleagues and peers about it, retweeting relevant tweets, mentioning it in blog posts, et cetera, will be of tremendous value. The fact is that ScienceOpen, as a small new outfit, doesn't have a big marketing budget, and therefore relies on word-of-mouth. Moreover, even if they had a larger budget, they would rather refrain from email spamming and the like. In my view, they should be rewarded for that attitude with whatever help those who are sympathetic to new approaches in scholarly publishing can offer.

And, of course, if you could consider trying out this approach by publishing a paper with the Peer Review by Endorsement method, that would be super.

Propagating this blog post would be highly appreciated, too, obviously. Many thanks in advance.

Jan Velterop

Wednesday, October 08, 2014

Journals of Nature and Science

Joe Esposito's recent post on the Scholarly Kitchen prompted me to post the following proposal, which I have discussed with various people, but which has no takers yet. But who knows what the future holds...

I called the proposed system JONAS (for 'Journals Of Nature And Science' – working title, obviously). It is, I think, a new approach to open access publication of peer-reviewed scientific literature. If it isn't I've missed something (entirely possible).

JONAS is about establishing a publishing system that addresses:
  • Open access,
  • Fair and efficient peer review, 
  • Cost of publishing
  • Speed of publishing
  • Publication of negative/null results

Open Access — The JONAS publishing system focuses on the superb technical publication, in various formats/versions, of peer-reviewed articles for optimal machine and human readability and re-use.

Fair and efficient peer review — Anonymous peer review has problems around issues of transparency, fairness, thoroughness, speed, publisher-bias, specious requests for further experiments or data, and possibly more. JONAS is a system using signed, pre-publication peer review, arranged by the author(s) (many publishers ask authors who to invite to review their papers anyway), and merely verified by the publisher (peer review by endorsement). Reviews would be open, published with the article that’s endorsed, non-anonymous, under the rules that peer-endorsers must be active researchers, and not be, or for at least five years have been, at the same institution as, or a co-author of, any of the authors. Such a peer-review-by-endorsement system is likely to be at least as good as, and quite probably better than, the currently widespread ‘black box’ of anonymous peer review. As reviews/endorsements would be signed and non-anonymous, there is very little danger of sub-standard articles being published (not worse than is currently the case anyway), as endorsers/reviewers would not want to put their reputations at risk. The review process between authors and endorsers is likely to be iterative, resulting in improvements on the original manuscripts. “Author-arranged” may perhaps include peer review being arranged on behalf of the authors by services specifically set up for that purpose, as long as the reviewers are not anonymous and conform to the JONAS rules. The LIBRE service is one example (currently in prototype).

Cost of publishing — A system like this can be very cost-effective for authors. The technical costs of proper publishing are but a fraction of the cost usually quoted for organizing and arranging peer review. First indication is that an amount in the order of £100-150 per article can be sustainable, given sufficient uptake. Tiered charges should be considered depending on the state of the manuscript when submitted. If the manuscript needs very little work to bring it up to proper publishing standards, or if the author doesn’t want or need those services, the cost could be very low indeed.

Speed of publishing — Since the peer-review-by-endorsement process has already taken place before the article arrives at the publisher, publication can ensue within days, even hours, depending on the state of the manuscript.

Requirements for manuscripts: ORCIDs for authors and reviewers/endorsers; inclusion of (permanent links to) datasets used, underlying data for graphs, a section “details for replicability and reproducibility” with clear and unambiguous identification of materials used, including reagents, software and other non-standard tools and equipment.
 
Input: Properly endorsed articles to be accepted in the form of Word, Pages, (LA)TEX, XML, HTML, Markdown, and Excel or CSV for data, and high-resolution image files (where possible scalable vector graphics) attached to emails or via a simple upload site.

Output: Articles would be published as XML, HTML, PDF, ODF and ePub formats, as much as possible semantically enriched and aesthetically formatted, plus Excel/CSV for data (tables extractable and rendered in Excel from PDFs with the software to do that, Utopia Documents, freely supplied).

Commenting and post-publication review (signed comments and reviews only) would be encouraged for all articles, links to comments to be provided with each article. Comments may be made on different sites, and would be linked to, if that is the case. Anonymous comments would be ignored.

Access Licences: CC-BY or CC0 — DOIs for the articles, and where appropriate for individual elements within articles, would be assigned/arranged by JONAS.

The core of the JONAS system would effectively be to have OA journals with a low-cost structure, with superb and highly optimized technical quality of the published articles. The principal difference with other OA journals would be the pre-arranged open peer review ("peer-review-by-endorsement"), organised by the authors themselves, according to a set of rules that ensures a reasonable level of assurance against reviewer bias (because of its openness and non-anonymity, actually more assurance than is provided in the usual anonymous peer review as widely practiced). Since arranging peer review is one of the major costs of any publisher (mostly staff costs), leaving that part of the publishing process in the hands of researchers and the academic community can make a great difference to the cost of publication. So far, efforts to reduce the cost of publishing have been concentrated on technical issues. Changing the mechanism (emphatically not the principle) of peer review offers much greater scope for cost reduction.

What JONAS' job would be is to take such peer-endorsed articles and make them into professionally published and complete (including data and metadata) documents, adhering to all the technical, presentational and unique identifier standards, in a number of formats, linked and linkable to databases and other relevant information, human- and machine-readable and suitable for widespread usage, for text- and data-mining, for structured analysis (incl. semantic analysis) and further knowledge discovery, and, crucially, for long-term preservation in repositories and archives of any kind.

An added service could be that manuscripts submitted in advance of peer-endorsement having been procured, would be placed, ‘as is’, on JonasPrePubs, a ‘preprint’ server, at no cost. This could help to secure priority (as a kind of 'prophylactic' against high-jacking of ideas – which would never happen in science, of course, but better to be safe than sorry, right?).

The JONAS publishing system would also be superbly suited to scientific societies and other groupings that wish to have their own journal. Such a journal could be fully integrated in the JONAS system, provided the manuscripts are submitted fully peer-endorsed or peer-reviewed (whether or not arranged by the author(s) or the scientific society in question). The charges per manuscript for individual authors and for societies wishing to publish their journals in the JONAS system would be the same, I imagine.

The JONAS methodology could, of course, be implemented on various publishing platforms.

Jan Velterop