The view that the results of publicly financed research should also be publicly accessible enjoys broad support in the academic community. Where their own articles are concerned, however, many authors hesitate to circulate them openly, for example by publishing them in Open Access journals or placing them in their institution's repository. They ask themselves whether that will not be at odds with the copyright rules and whether they will gain – or perhaps even lose – prestige. For their part, institutional managers wonder whether switching to Open Access will not make things more expensive than sticking with the traditional system of publication.
This article analyses the current situation regarding these three issues. The only possible conclusion is that the academic community finds itself in the course of a transition – from paper to digital – as regards the dissemination of knowledge, a transition that urgently requires an active and directive approach on the part of universities and research institutions. This conclusion is in line with a recent recommendation  by the European University Association, with the primary conclusion being that 'Universities should develop institutional policies and strategies that foster the availability of their quality controlled research results for the broadest possible range of users, maximizing their visibility, accessibility and scientific impact.'
In academia, it has traditionally been the author of a publication who has been viewed as the copyright holder. But that is not entirely a matter of course. Businesses and the US and Australian governments, for example, assume that it is the employer who holds the copyright under the so called work for hire  clause. Discussions also regularly arise as to the copyright in teaching materials. This article will restrict itself to scientific/scholarly publications. Where the identity of the copyright holder is concerned, it will reflect current academic practice.
It is the copyright holder who decides on the conditions for distribution and reuse of their publication. Laws do, however, provide for fair use  exceptions. Libraries also have certain rights that are mainly associated with their task of preserving material.
When everything was still published on paper, authors of certain types of academic/scholarly publications (articles, books, some conference proceedings) often transferred their whole copyright exclusively and irrevocably to the publisher of the work concerned. The publisher required them to do so in exchange for the layout and the distribution of the publication and for organising the review process. It should be noted that the work would be reviewed by colleagues – "peers" – as part of their research duties. The publisher would then capitalise on the rights it had thus received through subscriptions (in the case of articles in journals) and sales (in the case of books). Significant restrictions would apply to the distribution and reuse of the publication; in certain cases these could be bought off, for example by paying reproduction fees for reuse in educational materials.
This publication model became controversial because of the continual and significant increases in the cost of subscribing, an average of some 11% to 12% annually. It was hoped that digitisation – with distribution via the Internet and word processors for the layout – would cause prices to fall. That hope has evaporated, however, and many people would say that this is because of a lack of competition between publishers due to their copyright monopoly.
Early in the present century, the concept of CSR  found its way into academia under the name 'Open Access'. Authors, universities, and bodies funding research wished to improve their scientific/scholarly and social profile by offering Open Access to the results of publicly financed research. That, however, requires the consent of the copyright holder. If this is the publisher, consent is hardly ever granted. Only a few of the smaller publishers allow a published article to be included in the repository of the institution concerned and released after an embargo period of between six months and two years. More frequently, consent is given for an author's version to be distributed, also often after an embargo period. Authors hesitate to make use of this option, however, because of the confusion that can arise with their officially published version, for example as regards citations. The SHERPA-RoMEO Web site maintained by the University of Nottingham (UK) gives an up-to-date overview  of the copyright policies of virtually all publishers. These policies are extremely varied and are very hard for authors to follow. Actually, there are more policies than publishers as some publishers even apply a different policy to the different journals they publish. The extent to which an author must transfer his copyright to the publisher is also becoming increasingly a matter for negotiation. To assist authors in this regard, SURF has developed a legally sound publication licence . By issuing the licence, the author grants the publisher the right to publish his article but retains all the other rights, including the right to make the published article freely accessible – after a maximum wait of six months – via his institution's repository.
A new generation of publishers – Open Access publishers – has now come into existence who no longer require that the copyright be assigned to them in return for publishing an article but who require an immediate publication fee to be paid. Articles are published immediately after acceptance in a freely accessible digital journal, and thus become available throughout the world. The author also retains the full copyright in his article and can therefore do whatever he wants with it, for example upload it to his own Web site, publish it again in a different language, or include it in his institution's repository. In the latter case, the author can also specify what the repository is allowed to do with the article. SURF has also produced a standard repository licence . Some traditional publishers, e.g. Springer, have associated themselves with this model in that for some or all of their journals they allow the author to choose between transferring his copyright and paying a publication fee. Lund University (Sweden) maintains a list of all peer-reviewed Open Access journals  (currently 3600).
Transferring the copyright in a publication has become a relic of the past; nowadays a "licence to publish" is sufficient. The author retains the copyrights. Institutions should make the use of such a licence part of their institutional policy.
Lust for Prestige
The principal gauge for the prestige of scientific/scholarly authors is the impact factor of the journals in which they publish. That yardstick is not an undisputed one. Michael Mabe , the then director of Public Relations for Elsevier, was of the following opinion: 'Extending the use of the journal impact factor from the journal to the authors of papers in the journal is highly suspect; ... [impact factors] are not a direct measure of quality and must be used with considerable care.'
The impact factor of a journal is a fraction calculated by counting all the citations of all items in that journal that can be found in a selected group of indexed journals, during a specific period, and dividing it by the number of 'real' scientific/scholarly articles in the journal. The quality of an article is only one of the many factors that can influence the number of times it is cited. Other known factors are the age of the article (the older it is, the more citations), the number of authors (the more authors, the more citations); the type of article (a wide-ranging subject review article will be cited more regularly than a highly specialised article); and the discipline (researchers in the 'hard' sciences give more citations than those in the humanities). Articles that have become notorious in some way – perhaps because of fraud – are also cited very frequently. 'Decorative' citations are also a well-known phenomenon: it looks sophisticated, for example, to cite a Nobel Prize winner even if you have not actually read his article. There are also 'citation syndicates', in other words groups of authors who like to cite one another.
These are a few of the factors that can determine the numerator in the fraction. But the impact factor can also be influenced via the denominator, i.e. the number of "real" scientific scholarly articles. After all, the smaller the number of items in the journal that are considered to be "real" articles, the larger the result when one uses it to divide by. The measurement period (usually two years) and the selected group of indexed journals are also constantly the subject of discussion.
But the most fundamental criticism of the use of the impact factor as a yardstick for an author's prestige is that it represents an average number of citations for all the articles in a journal. It is as if one were judging a school pupil according to the average performance of his or her classmates. Managers and bodies that finance research do realise this, but they often justify using this method because they lack of anything better. Rankings for research institutes are in fact based on these impact factors. A manager whose institute is high up in the rankings will probably tone down any criticisms that he has.
Until 2004, the ISI company had a global monopoly on determining the impact factor for some 7500 journals (out of a total of about 25,000). Since then, Elsevier has provided impact factors (via Scopus) for roughly double that number. In competition with one another, the two organisations have since expanded the scope of their services, including providing the number of citations per article; this has not yet led, however, to any price reductions. Also in 2004, Google brought out Google Scholar (GS). GS basically covers all scientific/scholarly publications for which the metadata are available on the Internet, giving all the citations for each publication. This approach responded to the criticism levelled at the use of the classic impact factor as a yardstick for the prestige of individual authors. One disadvantage, however, is that the ranking of publications in GS is based on a secret algorithm. Moreover, the fact that GS is only available in the form of a beta version does not contribute much to its status as a measuring tool. There are now a number of other citation databases available, for example CiteSeer (computer and information science), Repec (economics), Scirus and getCited (both cover all disciplines). Like GS, all these databases are freely accessible.
Now that the number of citations per article is becoming clearer, it is possible to base an author's prestige directly on that statistic rather than on averages for journals. In this regard – and despite criticism – the h-index  is steadily gaining in authority. If an author has published n articles, he has an index of h if h articles are cited more than h times and the other n - h articles are cited a maximum of h times. The h-index reflects, as it were, both the breadth and depth of the author's oeuvre.
Analysis of citations has shown, for example in a study by Eysenbach , that Open Access publication greatly increases the number of citations. Partly because of this, a number of Open Access journals have been able to achieve a good impact factor within a relatively short time. Nevertheless, it is still, in general, the established traditional subscription journals that are ranked highest in their respective disciplines. One would seem to be dealing here with the 'Matthew effect' .
In addition to citations, the number of times an article is downloaded can also serve as a yardstick for the extent to which it is used, for example in teaching. It is obvious that the download score for Open Access articles will be higher than that for pay-for-use articles. The Los Alamos National Laboratory in the United States has announced the release of MESUR  in the autumn of 2008, an application that will enable one to monitor the number of downloads of articles. That number can then be used, in addition to citations, when determining impact measurements at article level and the prestige of authors.
The classic impact factor for a journal is not a good yardstick for the prestige of an author. Modern digital technology makes it possible to tailor the measurement system to the author. Institutions should, when assessing scientists and scholars, switch to this type of measurement and should also promote its further development.
The question that remains to be answered is whether the new Open Access model is cheaper than the old subscription model. Expectations are that Open Access will turn out to be cheaper for the academic community as a whole because the significant costs for contracts and shielding will be eliminated. Moreover, the Open Access model is based on the operation of market forces, while the traditional model is based on copyright monopolies. The advantages to society of Open Access to knowledge would also seem to be considerable, although it is no easy matter to arrive at reliable estimates of those advantages. In their study John Houghton et al.  make a significant contribution towards helping us answer this question for the Australian case.
At institutional level, it is relatively simple to carry out a cost comparison. In the Open Access model, the institution pays a publication fee – which differs from publisher to publisher – for all its articles. An up-to-date price list  for these fees can be found on the BioMedCentral Web site. The calculation can take account of the fact that the bodies that finance research are increasingly viewing publication as the culmination of the study concerned and therefore accept the publication costs as forming part of the research budget. The SHERPA-JULIET Web site gives an overview  of the policies of the major research financiers.
In the traditional model, an institution not only pays subscription and/or licensing fees but also reproduction fees, charges for individual articles that are requested (via interlibrary loan or directly from the publisher), and the contract costs already referred to. The latter are difficult to calculate, but they are nevertheless considerable. The library must first of all determine its purchasing policy; this often involves time-consuming consultations with scientists/scholars. The accessions then need to be registered in a complex system of subscription records or laid down in licences; these are legal documents, specifically regulating access rights, that are by no means trivial and usually only drawn up after lengthy negotiations. The institution must then implement the agreed access restrictions by means of shielding constructions such as IP addresses, passwords and proxy servers, or special software such as SFX. Finally, these restrictions mean that filling the institution's repository is a laborious matter and therefore needlessly expensive.
The calculations will show that the cost advantages of the Open Access model are not distributed equally between institutions. In the age of Open Access, genuine research universities may even find themselves having to pay more than in the age of the traditional model. Such institutions also need to spend more (than for example institutions focusing on teaching) on other research facilities such as laboratories, supercomputers and grids, and on attracting top researchers. Indeed, noblesse oblige. Estimates of the costs involved in publishing research results using the Open Access method range from 1.5% to 2% of the research budget.
Recently two Higher Education institutes in the Netherlands have done such a calculation, the results of which have been summarised in appendices to this article. As expected, open access is very beneficial to the Institute of Social Studies with its focus on an educational programme. The outcome for Utrecht University, a top research institution, is surprisingly more nuanced. If research funders are prepared to pay for the publication fees for articles that result from their projects, open access is also financially advantageous to this university.
One pressing problem has become the phase transfer from A to B. An institution that decides to use Open Access does not immediately cease having to pay its subscription and/or licensing fees. In fact, it pays not only for publishing its own articles but also – through the subscription or licensing fees – for publication by institutions that have not yet made the switch. This effect can be prevented if large groups switch together, although that requires co-ordination. One example is CERN's SCOAP3  Project. This involves CERN – acting on behalf of the high-energy physics community – defining the conditions for quality control and open distribution of articles and requesting publishers to submit quotations. This discipline is consequently switching to Open Access all in one go. Other examples are the Dutch university library consortium UKB and the Max Planck Gesellschaft which in experiments  with Springer could make the switch to Open Access publishing for all their authors in Springer journals. The economist Ted Bergstrom has already referred, as far back as 2001, in a now famous parable , to the need for such co-ordination.
The traditional subscription model for circulating publications is needlessly complex and expensive. Switching to Open Access, however, requires co-ordination that goes beyond the level of individual institutions. Supra-institutional organisations, for example the European University Association, should take the necessary initiative.
Cost Comparison: Open Access v. Subscriptions at Utrecht University
Established in 1636, today Utrecht University is an internationally renowned top research university with almost 30, 000 students and over 8, 500 staff. Annually UU publishes 5000 articles in peer-reviewed journals; of these 1500 result from externally funded research (figures 2005). Utrecht University is a signatory of the Berlin Declaration on Open Access.
|Publication fees UU1||
|Subscriptions and licences||
||Document supply 3 + copyright clearance fees + collection management 4||
|Publication fees research funders 2||
|Remaining costs UU||
1. 5000 peer-reviewed articles @1300
2. 1500 peer-reviewed articles @ 1300
3. Includes out of pocket costs to third parties (other libraries, publishers) and in-house handling costs.
4.Includes defining the collection, acquisitions and administration, shielding access and copyright issues.
University Librarian of Utrecht University
Cost Comparison: Open Access v. Subscriptions at the Institute of Social Studies
Established in 1952, the Institute of Social Studies in The Hague is an international graduate school with typically 400 students per year. Its research programme results in books, reports and about 60 articles in peer-reviewed journals annually. An intangible open access advantage is the free access that ISS alumni, who are often based in developing countries, will have to these articles.
|Publication fees ISS 1||
|Subscriptions and licences||
||Document supply 3 + copyright clearance fees||
|Platform or aggregation costs 2||
|Collection management 4||
1. 60 peer-reviewed articles @ 2000. The publication fee is an estimate based on the pricelist published by BioMedCentral. The list refers mainly to STM journals . For social sciences the fees may be lower.
2. ISS assumes that even in a full open access world still some aggregation or platform fees will be needed.
3. Includes out of pocket costs to third parties (other libraries, publishers) and in-house handling costs.
4.This is a rough estimate. Includes defining the collection, acquisitions and administration, shielding access and copyright issues.
Head of the Office of Library and Information Technology Services
- The European University Association (EUA): EUA urges universities to develop clear strategies to advance open access, 3 April 2008
- Wikipedia: Work for hire, 26 October 2008 http://en.wikipedia.org/wiki/Work_for_hire
- Wikipedia: Fair use, 26 October 2008 http://en.wikipedia.org/wiki/Fair_use
- Corporate Social Responsibility (CSR): the view that socially responsible behaviour is rewarding for an enterprise.
- SHERPA/RoMEO, Publisher copyright policies & self-archiving http://www.sherpa.ac.uk/romeo/
- JISC/SURF, Copyright toolbox http://copyrighttoolbox.surf.nl/copyrighttoolbox/authors/licence/
- SURF, Copyright in higher education http://www.surffoundation.nl/smartsite.dws?fs=&bw=&ch=AHO&id=13660
- DOAJ, Directory of Open Access Journals http://www.doaj.org/
- Amin, M., Mabe, M., "Impact factors: Use and Abuse", Perpectives in Publishing 1, Elsevier, October 2000
- Wikipedia: H-index, 26 October 2008 http://en.wikipedia.org/wiki/H-index
- Gunther Eysenbach, "The open access advantage", Journal of Medical Internet Research 8, no 2 2006
- Wikipedia: Matthew effect, 26 October 2008 http://en.wikipedia.org/wiki/Matthew_effect
- MESUR: MEtrics from Scholarly Usage of Resources http://www.mesur.org/MESUR.html
- John Houghton, Colin Steele & Peter Sheehan, "Research communication costs in Australia: Emerging opportunities and benefits", DEST Research Communications Cost Report, 29 September 2006 http://firgoa.usc.es/drupal/node/31598/print
- Comparison of BioMed Central's Article Processing Charges with those of other publishers
- SHERPA/JULIET, "Research funders' open access policies" http://www.sherpa.ac.uk/juliet/index.php
- SCOAP3 : Sponsoring Consortium for Open Access Publishing in Particle Physics: Towards Open Access publishing in High Energy Physics http://scoap3.org/
- "Together Again: Springer, Max Planck Agree To New "Experimental" Deal", Library Journal, Reed Business Information, 5 February 2008
- Ted Bergstrom, "The Parable of the Anarchists Annual Meeting", Excerpted from "Free Labor for Costly Journals?", Journal of Economic Perspectives, Fall 2001
- BioMed Central | for authors | Article processing charge comparison