Open Conference Service, ELPUB2008

Font Size: 
Global annual volume of peer reviewed scholarly articles and the share available via different Open Access options
Bo-Christer Björk, Annikki Roos, Mari Lauri

Last modified: 2008-06-13


OBJECTIVES: In this study we will examine one of the key parameters in the academic peer reviewed journal system, namely the number of scholarly, peer reviewed articles published annually. Several estimates, ranging from 1,2 million to 2,5 million (e.g. Harnad et al., 2004; Houghton & Vickery, 2005; Ware, 2006) have earlier been presented in the literature. We use a repeatable procedure and describe the used methods in detail.

Among the studies which have clearly declared the bases of their estimates probably the most cited is published by Tenopir and King (Carol Tenopir & King, 1997; C. Tenopir & King, 2000). However Tenopir and King base their calculations on the material published by US based publishers only. We claim that the journals indexed by ISI on average tend to publish far more articles per issue than the often more recently established journals not covered by the ISI, and that this should explicitly be taken into account in a calculation of the total number of articles. This aspect was not taken into consideration in the above mentioned, creditable series of research work made by Tenopir and King.

We will present an estimation of the number of openly available (OA) articles published in primary open access journals in 2006. The proportion of the articles published in subscription-based journals for which the articles have been deposited in an OA archive or are directly available openly on the web will also be evaluated. These figures are extremely important in any discussions about Open Access and our method lends itself to being repeated at regular intervals in order to track significant trends.

METHODS: Our study is based on the bibliographic data retrieved from Thomson Scientific's Web of Science databases (WoS), Ullrich's Periodicals Directory and data openly available on the internet. We also use national data available in Finland.

The total number of scholarly, peer reviewed, active journals published in 2006 was calculated using Ullrich's Periodical Directory, setting as parameters active, scholarly and peer reviewed. Secondly, the number of titles indexed in the WoS in 2006 was estimated. After that, we made a simple subtraction and got the number of journal titles not indexed in the Wos (non ISI titles). The extraction of the number of articles indexed in the WoS databases in 2006 followed. Based on that we could calculate the average number of articles that the journals indexed by the ISI published in 2006.

In order to arrive at a total number of articles published we needed to find a way to estimate how many articles the non ISI titles publish on average per year. This was done by making a sample (n=250) of journals from Ulrich's database. For practical reasons we included only online titles. We went through the sample and picked up 104 non ISI titles. From these titles, we counted the number of articles published in 2006. After having the average number of articles on both journal categories we were able to estimate rather accurately the total number of scholarly articles published in 2006.

For calculating the share of the scholarly OA articles we counted separately first the number of articles published in 2006 by four major OA publishers (PLoS, BMC, Hindawi and Internet Scientific Publications). To be able to find out the amount of the rest of the directly open articles ("gold" OA), we searched first from Ullrich's Periodical Directory the collection of the OA journals. This was done by setting as parameters: online, active, scholarly, peer reviewed and open access, and subtracting four major publishers. From that journal list we made a randomized sample of journals and counted the number of research articles on those. The average figure was then multiplied by the number of this collection of OA journals.

Counting the share of the "green" (see e.g. Guedon, 2004) OA is more complicated. To do it we again made two randomized samples of journals, consisting together 300 titles. The first sample consisted of journals indexed by ISI (n=200) and the other non-ISI journals (n=100). The size of these groups was equivalent to the share of yearly published articles on each, journals indexed in ISI representing 70 % of the yearly output. We decided to go through the first issue in 2006 of each title and check the full text open availability of the second research article. Google was used as a general search engine for this purpose and the search was for the sake of simplicity conducted outside the university intranet.

This work is still under way, but as a result we will get the total number of OA articles published in 2006 and separately the share of the "golden" and "green" articles and besides an estimate for the number of self archived articles.

RESULTS: We found out that the total number of scholarly articles published in 2006 was approximately 1 350 000 and the average number of articles published per journal was about four times as high for the journals that ISI tracks (105) compared with journals not covered by the ISI (26). The percentage of the OA articles totally as well as the share of different OA options will be reported in the final conference paper.

CONCLUSIONS AND DISCUSSION: We believe that the estimates represented in this study are more accurate than those presented earlier in different contexts. We define our method accurately and the estimates can easily be replicated and/or adjusted by other researchers in later years.

Different elements in our calculation differ in terms of accuracy. The total number of articles included in the indexes of the ISI should be very accurate, provided that we have searched the database in a correct way. Also the total number of journals tracked by the ISI in a given year is information which can be verified.

The total number of the peer reviewed scholarly journals is much more difficult to estimate accurately. Ullrich's database is the best tool available for this purpose, but its coverage is not 100 % and some journals it includes might on closer scrutiny not turn out to be peer reviewed.

One considerable fact is also that we have not paid attention to differences between subject areas. It has been stated in literature earlier (see e.g. Carol Tenopir & King, 2004) that journal characteristics vary significantly between distinctive domains. This factor needs to be checked in the forthcoming research.


Guedon, J.-C. (2004). The "Green" and "Gold" Roads to Open Access: The Case for Mixing and Matching. Serials Review, 30(4), 315-328.
Harnad, S., Brody, T., Valli?res, F., Carr, L., Hitchcock, S., Gingras, Y., et al. (2004). The Access/Impact Problem and the Green and Gold Roads to Open Access. Serials Review, 30(4), 310-314.
Houghton, J., & Vickery, G. (2005). Digital Broadband Content : Scientific Publishing (No. DSTI/ICCP/IE(2004)11/FINAL): OECD : Directorate for Science, Technology and Industry : Committee for Information, Computer and Communications Policy.
Tenopir, C., & King, D.W. (1997). Trends in scientific scholarly publishing in the United States. Journal of Scholarly Publishing, 28(3), 135-170.
Tenopir, C., & King, D.W. (2000). Towards Electronic Journals: Realities for Scientists, Librarians, and Publishers. Washington D.C.: Special Libraries Association.
Tenopir, C., & King, D.W. (2004). Communication patterns of engineers. Hoboken, NJ: John Wiley.
Ware, M.C. (2006). Scientific publishing in transition: an overview of current developments. Retrieved October 2007, 2007, from