Although there are many definitions of “Open Access” (OA) journals, a useful guide is that offered by the “Budapest Open Access Initiative,” (BOAI) which speaks of “The literature that should be freely accessible online […] which scholars give to the world without expectation of payment” [2]. This description fits best with Willinsky’s definition of an “unqualified open access journal” [3] which allows its articles to be available “immediately, completely, and exclusively free-to-read” [3].

OA journal publishing has strong international support amongst researchers and librarians [4-6], and even amongst Research Councils and Institutes [7; 8]. At the time of writing this article, some 5,000 individuals and more than 460 organisations had added their signature supporting the BOAI [2].

There is a growing body of evidence to suggest that OA journals are being cited as much as, and even more than, non-open access (NOA) journals [9-11]. These and other issues, such as the principle of free access, faster publishing times, wider readership, and greater number of citings, are strong motivators to publishing in OA journals [12; 13].

In OA journals, the publishing costs (or a portion of them) are sometimes carried by the authors, or their institutions, or by journal advertising [14; 15]. The author-pay models are not popular amongst researchers [16], and there is some discussion about the possible disadvantages that this model has for researchers in the developing world [17].

The path of OA publishing has not been smooth, and the debates around OA publishing continue. Researchers who do not publish in OA journals have cited unfamiliarity with OA, low prestige and impact, author payment issues, and low readership [12; 13]. The arguing points include copyright and other legal issues, peer-review, the value of OA articles, business models, publishing costs, technology infrastructure, business models, indexing services, standards, rewards for researchers and marketing [4; 5; 9; 18-20].

In the midst of this are the researchers who want access to the information in NOA journals. The term “researchers,” of course, has a very broad application. The Public Library of Sciences (PLoS), now a publisher of the OA journal PLoS Biology, aims at a “much wider audience, including millions of students, teachers, physicians, scientists, and other potential readers, who do not have access to a research library that can afford to pay for journal subscriptions” [21]. While the term “digital divide” is usually used to apply to the divide between those who have access to the Internet and those who don’t, there also appears to be a digital divide between those who have access to NOA journals and those who do not.

To those who do not have access to NOA journals, the information is, as it were, a closed book.

Unless they take matters into their own hands.

The environment that facilitated the growth of OA publishing has been the Internet. Simultaneously, the Internet has allowed for the dissemination and sharing of information on an unprecedented scale. The music industry was a prime example of this sharing, when the sharing (frequently illegally) of digitised music exploded. Given that so many journal articles, both OA and NOA, are now available electronically, it is not difficult to imagine a scenario in which researchers share NOA articles with others, without the permission of the authors, publishers or copyright holders. These researchers would be people who have access (usually through institutions) to journals, and they use that access to share the articles with other researchers who do not have access to those journals.

Indeed, web sites that facilitate this exchange do exist. Question that need to be answered include: How many articles are affected? How many journals are affected? Which journals are most commonly affected? What is the estimated cost to these journals? This paper examines a site that is used for such a purpose, to begin answering these questions.

The web site to be discussed (referred to hereafter as “the site”) is aimed specifically at medical professionals and students. At the time of this study (January 2009), the site had a total of 127,626 registers users, and nearly 300,000 postings in the electronic discussion forums. Although aimed at the medical field, the site did not prevent non-medical persons from joining.

The site had a wide range of electronic discussion forums that one might expect to find on a medical site, including “Medical Education Forums”, “Allied Health & Nursing Forums”, “Medical Student Forums” and “Physician and Resident Forums.” In addition, however, there was a set of forums collectively labelled “Electronic Library.” “Electronic Library” contained forums and sub-forums aimed specifically at the sharing of information that one might typically find in a library. This paper focuses on a sub-forum named “Databases & Journals – Requests and Enquiries.”

In “Databases & Journals – Requests and Enquiries”, users who do not have access to a particular NOA journal or conference proceedings post a request for a desired article. Other users who have access to those journals then download the requested articles from those journals, and make them available to the requester by posting the articles either into the forum or to a publicly-accessible web site. Although many participants on this site are students, other users have identified themselves as practicing professionals or academic staff.

Strict rules govern the requests. For instance, only three articles may be requested by each user each day, and the hypertext links to the articles must be submitted in the request. There are, however, no rules concerning the type of journals, licences, etc, of the articles requested or posted. Naturally, not only does the requester have access to the downloaded article: anyone reading the forum also has access to the article, and, in most cases, the site tracks the total number of times an article has been accessed or ‘viewed.’

Until January 2009, the site was publicly available. Until then, although all the forums could be read by guests, full participation required a free registration, requiring merely a user name and an email address. (The site and forum messages were routinely indexed by search engines such as Google). Since January 2009, the site has been available to a small group only, and has also frequently been offline. This study deals with data that were publicly available during 2008.

The aim of this study was to investigate the interactions in the “Databases & Journals – Requests and Enquiries” forum and to report on the number of requests and postings, the journals from which articles are being accessed, and to comment on the broader implications of the findings.

Within the six-month period, the total number of unique articles requested in the sub-forum Databases & Journals – Requests and Enquiries” was 6,587. Of the 6,587 articles, 6,463 (98.1%) were in English, 68 in German, 42 in French, and 14 in other languages. The total number of unique articles returned (i.e. found by other users) was 5,464, giving a success rate of 82.95%. The oldest article was from Science, dated 1884 [23].

The success rate of 83% would probably have been higher if the requesters had more frequently followed the basic rules of the site, which included requesting a maximum of three articles per day, supplying direct links to the articles, and ensuring that the text was available electronically. Although the rule of three articles per day was usually strictly enforced, there were occasions when people requested and obtained a greater number of articles. This included requests for entire issues.

Of the 5,464 articles returned, the number of ‘views’ could be determined on 5,251 articles (96.1%). (Others had been placed in formats in which the number of views could not be determined). From these 5,251 articles, the total number of views was 23,461, the minimum number of views was 0 (for 69 articles), and the maximum was 177 [24]. The mean number of views was 4.47, the median 3, and the mode 2. These last three descriptive statistics, however, do not show the skewness of the views – this is shown in Figure 1 below:

Figure 1

Figure 1: Number of views per article. N=5,251.

It is also obvious that, although many of the articles are viewed by participants who did not originally request the article, generally, the article is viewed by the requester, and only a few others.

A question to consider, then, is, how much money was saved by the people who viewed the articles? (Conversely, how much was lost by the publishers?) The prices of individual articles ranges. Some are available for as little as $7.00, and others for as much as $50.00 for academics, and $100.00 for non-academics (e.g. [25]). At an approximate value of $30.00 per article, the 23,461 views of the articles amount to $703,830.00 over the six-month period. Extrapolated across a year, this translates into a little more than $1.4M.

Articles from a total of 2,867 journals and conferences were requested. The 22 journals with the highest number of requests are summarised in Table 1 below:

Figure 2

Table 1: Table showing the 22 journals with the highest number of requests, ranked according to number of unique articles requested.

At the top of the list are two general scientific journals, Nature and Science. Not surprisingly, some of the major medical journals are also found in this list. Other journals that one might expect to find (e.g. BMJ), allow free access to their articles after 6 or 12 months, and so would not necessarily be placed in this list. (As a matter of interest, 11 articles had been requested from BMJ, of which 9 were successfully delivered).

The author is a signatory of the BOAI.