Lluís M. Anglada
Based on statistics of use of electronic journals licensed by the Consortium of University Libraries of Catalonia (CBUC) during the period 2000-2003, this article presents the evolution of journal use over time and the degree of dispersal as a framework for calculating and assessing the consortial gain, which is considered to be the percentage of use of titles not previously subscribed to in paper format that have been licensed thanks to the economies of scale of consortial purchasing. The aim of the CBUC study was to analyse the period of change from the paper to the digital format in the collections of the CBUC member libraries. During the period covered by the study, users maintained access in paper format to retrospective collections and some current subscriptions, and the comparison with previous collections on paper was made with an acceptable degree of chronological proximity.
1. Introduction and objectives
The Consortium of University Libraries of Catalonia (CBUC)1 is composed of all the public universities of Catalonia and the National Library of Catalonia, as members, while CBUC allows all types of libraries to participate in its programmes as users. Since 1999 it has followed a clear line of action with regard to negotiating licences of electronic resources through consortial purchasing of journals, books and databases, which are offered under the "trademark" of the Digital Library of Catalonia together with electronic resources produced by the member and user institutions of the Consortium (Anglada 1999).
Work began for the collective licensing of electronic resources in 1998, with the aim of providing a common core of electronic information for all users of the libraries of the CBUC. Most of the subscriptions are paid for by the member libraries, though the Government also provides complementary aid. The expenditure is distributed according to an internal formula that takes into account, among other parameters, the previous expenditure of each member with the publisher in question and the dimension of the university within the consortium. In addition each member is assessed an equal part of the expenditure. Cooperation among members has proven to be fairly positive, because in addition to the considerable increase in the number of titles that all the libraries have obtained, the larger universities have realized a saving on what they would have paid to maintain the journals they were already subscribed to. On the other hand, the universities with few previous subscriptions increased their expenditure on subscriptions but gained access to an amount of information that they would have been unable to obtain through individual subscription (Anglada, Comellas, Roig, Ros and Tort 2003).
After a period of almost three years in which the Digital Library of Catalonia has been in full operation, currently the CBUC is engaging in several lines of study that are especially necessary with a view to renegotiating licences when the first ones begin to expire. For most packages of journals the licences were acquired under the "Big Deal" model intended especially for consortia, which includes cross access, new additional titles or the whole catalogue of the publisher at very attractive prices. This model has been the subject of debate in the professional literature in recent years (Frazier 2001; "To the editor..." 2001; Rowse 2003), but the "Big Deal" model has undoubtedly offered a prime opportunity for achieving a spectacular increase in the accessibility of scientific information, thus putting a halt to the previous trend of continual cuts in library collections. This form of licensing has become so popular among experts that it has even been called the "Good Deal" model. However, throughout this work we will use the more neutral term "Big Deal" to refer to this model because it has been adopted in the literature as the most widely accepted name.
The consortia used the first period of change from paper to digital formats in scientific publishing to expand their collections in line with the interests of the publishers to generate new demand by offering in electronic format, at a marginal cost, journals that were not previously acquired for the collections on paper. However, in the words of Rowse (2003), this model is very likely to be subject to revision by both publishers and libraries at the time of the first or second renewal of licences. Leaving aside the specific licensing models that may arise in the near future, it is clear that the economies of scale obtained by consortia are considerable, as is evidenced by the continual growth of these institutions. It is therefore highly foreseeable that they will continue to play a key role in the acquisition of scientific information in the second stage of negotiation of licences. The benefits observed are pointed out by King and Xu (2002) in a very detailed study on the role of these bodies: "Library consortia have grown in number and size because they are shown to be highly beneficial to their member libraries since they can perform the services better (i.e., with higher quality, faster, etc.) and/or at less cost due to economies of scale."
This is the context in which we have placed the analysis of statistics on electronic resources and the assessment of the return on the investment and the satisfaction of users. The data obtained at the CBUC make a highly interesting empirical contribution to a major international discussion on the models for acquisition of scientific journals in the university libraries of the future. The CBUC study falls within the framework of a now fairly extensive literature on the use and users of electronic resources, which was reviewed and consolidated by Tenopir (2003) and by the works contained in the proceedings of the seminar that the Ingenta Institute organised on the subject of consortial purchasing and its results (The Consortium site license... 2002). However, only a limited number of detailed studies of the consortial gain in the acquisition of journals and the dispersal of use in large collections obtained through "Big Deals" have been made, and we will use them as a reference to make comparisons in our study (Sanville 2001; Xenidou-Dervou 2001; Keller 2002; Piguet 2002; Nicholas, Huntington and Watkinson 2003).
Data were gathered from the third quarter of 2000 to the second quarter of 2003, but covered a different period for each of the four publishers considered: Academic Press, Kluwer, MCB Emerald and Wiley. The Academic Press package was introduced in the Digital Library of Catalonia in January 2000, but because it was the first package introduced by the CBUC and reliable figures were not available by titles, its use during the first six months was not taken into account. The period for gathering data was extended to the second half of 2003 to cover the package of Wiley for a period of at least a year, because though it was the most recent package, it was the one with the most intense use. This decision complicated matters, because it forced us to work without detailed data of the first six months of 2003 for Academic Press, which was affected by the acquisition by Elsevier and the integration of its titles in the ScienceDirect portal as of January 2003.
The libraries considered in the study for all four packages were the Library of Catalonia and the eight founder universities of the CBUC: UAB (Autonomous University of Barcelona), UB (University of Barcelona), UdG (University of Girona), UdL (University of Lleida), UOC (Open University of Catalonia ), UPC (Technical University of Catalonia), UPF (Pompeu Fabra University) and URV (Rovira i Virgili University). The participation in certain packages of the following user libraries was also considered: Jaume I University (Kluwer, MCB and Wiley), Ramon Llull University (Kluwer) and the University of Vic (Academic Press and MCB).
The identification of the titles and the attribution of data of use to the titles previously subscribed to on paper was quite controversial, because the publishers do not state in the statistical tables the specific titles included in the licence of the CBUC, nor do they offer data to relate several titles that are linked by mergers, splits or changes. It was therefore decided not to aggregate data under the heading "subscription", but to disaggregate the subscriptions both on paper and in electronic format, into as many publications for statistical purposes as the number of different titles linked to them. This method of work explains the increase in the number of titles that figure in the documentation of the Consortium on the licences negotiated.
The figures on the use of the journals of the different packages were obtained from two sources of data provided by the publishers: 1) chronological files of general use of all the publications available in the portals of the publishers, and 2) detailed files by titles, which were used to calculate the exact use that could be attributed to the package licensed by the CBUC from each publisher. Except in the case of the quarterly evolution of use (Table 2), those that were not subscribed to within the specific licence of consortial purchase were not counted. One must take into account that the use of sample articles and of titles subscribed to independently by some of the consortial institutions may modify the percentage of participation of each institution in the total use of each package.
Finally, one should also take into account that it was not possible to correct the distortion generated in the data obtained by the use of proxy cache servers in the protocols of web client configuration of some of the universities studied. Another possible cause of distortion that could not be taken into account is the lack of a uniform definition of what is considered to be an effective download of an article: though the new standards of COUNTER (2002) already specify that the repeated demand for the same article by a user within an interval of 10 seconds for html, and 30 seconds for PDF, will be considered as a single article downloaded, we do not have a clear idea of the degree of consistency applied by all publishers in the past with regard to this correction factor.
Evolution in use
Over the three years analysed, the consortial purchase of these four packageswithout considering other, more recent, packages and full text bibliographical databases that also make up the offer of the CBUCled to a great increase in the access to electronic scientific journals, which between 2000 and 2003 rose from 195 to 1,495 titles (Table 1), and from 5,409 to 93,367 articles downloaded per quarter (Table 2). It should be noted that for most libraries the new titles obtained with these licences represent percentages higher than 94%, whereas for UAB and UB they represent 83.24% and 73.97% respectively (Table 4). There is thus a great deal of elasticity in the use of information due to the use of titles not previously subscribed to (Table 3) and the progressive absorption of new packages without a stagnation in the total number of articles downloaded (Figs. 1 and 2).
Table 1. General use of each package
* PP titles (CBUSC): Titles previously in paper format in the CBUC collections. Single titles are counted as a result of the aggregation in the CBUC of all the different titles subscribed to by each member library.
The use of two parallel sources of datathe files by titles and the chronological files by institutionmade it possible to adjust the recorded use to the specific characteristics of the licensed package. In Table 1 one can see that, in the case of two publishers (MCB and Wiley), the difference in articles downloaded between the CBUC package and the whole portal of the publisher is not very significant, whereas for the other two publishers the difference is greater. In the case of Academic Press, this is due to a large number of individual subscriptions by several libraries to titles not licensed within the CBUC package, whereas in the case of Kluwer, it is due to the use of a series of sample titles which were accessed provisionally as part of a promotion by the publisher.
The greatest rate of use corresponds to the Wiley journals, at 39.05 articles downloaded per title per month, and the smallest was that of the MCB package, at 6.52 articles per title per month. One must take into account, however, that the time period and the volume of the retrospective collections may affect the rate of use, so until the use of each title can be considered in relation to the number of articles that it offers, the conclusions on this point will not be definitive. Nevertheless, the rates of use are fairly representative if they are assessed package by package, due to the relative stability observed in the quarterly evolution of the number of articles downloaded in each package (Table 2).
The evolution of use was analysed using the data obtained from the monthly summaries for each institution, and some viewed articles are therefore counted that do not belong to the package of the CBUC. However, they represent an insignificant percentage and affect all quarters of each package equally. As can be seen, though the growth in the use of each package is moderate, it must be considered from the viewpoint of the total quarterly growth and the total accumulated consumption: the progressive incorporation of newly offered packages generates a new demand, while it moderates the growth in the use of packages already available. Furthermore, in the first few months of use of each package, two opposite phenomena hinder the assessment of the evolution of use: firstly, there are a significant number of "exploratory" uses by users and librarians, and secondly, the new electronic documents may remain "invisible" during the first few months because many users still receive the issues in their libraries on paper and do not feel the need to find the electronic version.
Table 2. Quarterly evolution of the articles downloaded in each package
* Provisional and approximate figures estimated downward due to the integration of AP in Elsevier's Science Direct.
Use of new titles previously not available on paper
The proportion of use of articles of titles not previously subscribed to on paper (NPP titles) to the total is very high, though in many libraries the striking increase in new titles makes it hard to make a comparison with the previous situation. The main interpretation of the figures must therefore be made for the whole CBUC using a package-by-package analysis (Table 3) though the data are also disaggregated by universities participating in all four packages (Table 4).
Table 3. Percentage of use of new titles not previously subscribed to on paper
Table 4. Distribution of use by universities
University libraries that figure as users of all four packages were considered individually. The Libraries under the heading "Rest" are: UVic, UJI, URL, BC, Proxy CESCA.
The analysis of the consortial gain, calculated as the percentage of articles downloaded from titles not previously available on paper (% use NPP titles), is shown in Table 3 in relation to both the previous collections of each library (xBIB) and the sum of the single titles that the CBUC had through the collections of the participants in the consortial purchasing. In both cases one can see a significant use of the new titles. However, in the absence of precise data on the previous demand for interlibrary lending, the figure that shows the clearest comparison with the previous situation is that which is calculated by relating the articles downloaded by each library to the titles that this library had on paper before the consortial purchase, that is, 61.49%.
Analysis of the degree of dispersal of use
Table 5 and Fig. 3 show the dispersal of use of articles for all titles licensed in each package. Though the periods of analysis do not coincide for all four publishers, they were sufficiently long to consider the comparison of the curves of dispersal as valid. It is thus observed that with an average of 35% of the titles one achieves just over 80% of the downloads of articles. Please note that the greatest dispersal of use corresponded to the package with the lowest use per title per month, MCB at 6.52, whereas Wiley, the package with the least dispersal, showed the greatest intensity of use at 39.05 articles per title per month.
Table 5. Accumulation of articles downloaded by groups of titles
Table 6. Presence of new titles in the most used groups of titles
* The totals for the CBUC for 80% of the articles downloaded are the sum of the values of each package rather than the dispersal calculated by individually aggregating the data of the 1,498 titles.
The analysis of the dispersal of use provides a fair measure for evaluating the consortial gain. Indeed, as is shown in Table 6, though most of the titles that are used little or not at all are those that were not previously subscribed to on paper, their participation in the group of 80% of the articles downloaded is as high as 43.22% for the whole CBUC. There was also a considerable concentration of use in certain titles: 80% of the articles downloaded in the whole CBUC were obtained with 516 titles of a total of 1,498, which means that the core consumption was satisfied by 34.45% of the titles, whereas only four libraries used more than 75% of the titles at least once in the three years (Table 4).
The main points that stand out from a general consideration of the results are the magnitude of the consortial gain and the dispersal of the use of all the titles, especially if the whole CBUC is considered as a valid unit of assessment, because it was the unit considered in the negotiation of the licences. A library by library assessmentwhich is not the aim of this articlewould certainly qualify these conclusions according to the university analysed. A weak point, or one that remained undefined due to the limitations with which we worked, was the moderate growth in the use of each package considered individually.
Tables 3 and 4 clearly show the magnitude of the consortial gain expressed by the high level of use of titles that were not previously included in the paper subscriptions of each library. The level of use of the titles not previously subscribed to on paper was very high, whether one considers the articles downloaded by each library in relation to their previous collections on paper, or the total number of single titles on paper that are attributed to the whole CBUC: 32.22% and 61.49% respectively.
Furthermore, the absence of reliable historical data on the use of the paper subscriptions used to establish the starting point for calculating the consortial gain makes it impossible to detect whether, in addition to the increase in the use of titles not previously subscribed to, there has been a significant increase in the use of titles previously subscribed to on paper. This is highly likely, because the conditions for consulting documents are more favourable to use in electronic than in paper format: simultaneous location of all the titles in all the libraries of a university (thus saving the cost in time and money of the circulation of photocopies on the campus), accessibility from the offices of teaching and administrative staff, and remote access for some users from consultation points off the campus. This point should be the subject of a future survey of users to gather their practices and opinions, and is therefore the area in which we feel we should focus further work along the lines started with this article.
A fuller analysis of the use of titles not previously subscribed to would require statistics on interlibrary lending (ILL), which is beyond the scope of the present article. Because consortial purchasing supplies titles that were previously covered by ILL, the cost of ILL should be compared with the marginal cost with which titles subject to marginal use are obtained in the new model. Though this has not been studied specifically for the CBUC, it seems highly likely that there is a margin for assuming a consortial gain in this area. Everything points to the existence of benefits in making this comparison because the marginal cost of many titles can be clearly written off as of a certain threshold of requests for ILL or pay-per-view acquisition. Though the CBUC has no studies of the cost of ILL for new titles resulting from the licences, an ARL study in 1996 stated that an ILL-type transaction in research libraries cost $28 (Jackson 1997). Both the time spent waiting for the article and the attribution of costs to the final user of the ILL have always had a dissuasive effect on demand, whereas instant access to the document in these large electronic packages stimulates the dispersal of use.
The evidence observed thus supports the thesis of a substantial consortial gain, which would be even clearer if one made a comparison with the real figures (if available) of use of the paper collections in previous years and with the demand for titles through ILL. However, these positive data on performance must be considered with caution, especially taking into account the rate of growth of use and the degree of dispersal of use of the total number of titles.
With regard to the evolution of use over time, a certain stagnation in growth is observed if each package is considered independently (Table 2 and Fig. 1). This should be studied in more depth in the light of the continual growth in new electronic resources offered both by the CBUC and by the different libraries independently. Indeed, in addition to these four packages of journals, in the last two years the CBUC has also licensed other electronic products (full text bibliographic databases and packages of journals from more specialised publishers) that were not analysed in this study. Another element that has not been fully assessed is the influence of maintenance of retrospective or current collections on paper in some of the libraries analysed, especially with regard to the use of titles that are also available in electronic format. Indeed, for certain titles some libraries considered in the study maintain the "electronic+ paper" model, according to which one can continue receiving the paper version for a small surcharge. At present it is very difficult to estimate to what extent the presence of certain titles on paper in the libraries reduces, stimulates or has no effect on the final figures for electronic use.
With regard to dispersal, the consumption observed corresponds to the 35/80 model rather than the traditional, theoretical 20/80 model of paper consumption, thus coinciding closely in this aspect with previous results observed in the literature consulted (Sanville 2001). This reinforces the idea of a freer, more exploratory and often more accidental consumption of the collections (Nicholas, Huntington, Watkinson 2003), in which the participation of students can lead to a considerable dispersal of use. In this sense, a combined reading of Tables 4 and 6 and Figure 5 indicates that there would be much room for manoeuvre in establishing priorities for licensing titles in the future. Indeed, though in the first stage of the change of model the consortial gains are striking, on closer inspection it can be seen that the proportional gains in articles are less than those in titles, and that the figure 35/80 represents a limit that is difficult to exceed in dispersal of use, though it is greater than that traditionally observed for paper documents.
The results of this study provide a basis for reflection on future acquisition policies based on objective figures of use. Though consortial purchasing is leading to a considerable increase in access to large volumes of information without making a previous analysis of the individual demand for each title, this does not obviate the need for a joint assessment of the performance of the packages, which would help university libraries to report to their respective academic authorities in this transitional stage in the acquisition model.
Note Consortium of University Libraries of Catalonia (CBUC), <http://www.cbuc.es>.
Anglada, L. (1999). "Working together, learning together: the Consortium of Academic Libraries of Catalonia." Information technology and libraries. Vol. 18, no. 3, pp. 139-144.
Anglada, L.; Comellas, N; Roig, J; Ros, R.; Tort, M. (2003). "Licensing, organizing and accessing e-journals in the Catalan university libraries." Serials. Vol. 16, no. 3, pp. 299-306.
The Consortium site licence: is it a sustainable model? (2002). Edited proceedings of a meeting held on 24th September 2002, London, organised by the Ingenta Institute. Oxford: Ingenta Institute. 246 pp.
Jackson, Mary E. (1997). "Measuring the performance of interlibrary loan and document delivery services." ARL: a bimonthly newsletter of research library issues an actions. Issue 195 (Dec.). <http://www.arl.org/newsltr/195/illdds.html> [consulted on 19 February, 2004].
Keller, Alice (2002). "Electronische Zeitschriften: was sagen Nutzungsstatistiken aus?" B.I.T. online. Vol. 5, no. 3, pp. 213-232.
King, D.W.; Xu, Hong. (2002). "The role of library consortia in electronic journal services." In: The consortium site licence: is it a sustainable model? Edited proceedings of a meeting held on 24 September, 2002, London, organised by the Ingenta Institute. Oxford: Ingenta Institute. pp. 9-76.
Nicholas, D.; Huntington, P.; Watkinson, A. (2003). "Digital journals, Big Deals and online searching behaviour: a pilot study." Aslib proceedings: new information perspectives. Vol. 55, no. 1/2, pp. 84-109.
Piguet, Arlette (2002). User needs and usage as seen through a consortium. A presentation made at the Seminar on Digital Journals, organised by the Faculty of Librarianship and Documentation of the University of Barcelona and the CBUC, 21-22 November, 2002. <http://lib.consortium.ch/external_files/barcelona.ppt> [consulted on 19 February, 2004].
Rowse, M. (2003). "The consortium site license: a sustainable model?" Libri. Vol. 53, no. 1, pp. 1-10.
Sanville, T.J. (2001). "A method out of the madness: OhioLink's collaborative response to the serial crisis: four years later progress report." Serials. Vol. 14, no. 2, pp. 163-177.
Tenopir, C. (2003). Use and users of electronic library resources: an overview and analysis of recent research studies. Washington, D.C.: Council on Library and Information Resources.
"To the Editor (2001): [letters in response to the opinion piece, The Librarians' Dilemma: Contemplating the Costs of the "Big Deal", by Kenneth Frazier in the March 2001 issue of D-Lib Magazine]." D-Lib Magazine. Vol. 7, no 4. <doi:10.1045/april2001-letters> [consulted on 19 February, 2004].
Xenidou-Dervou, C. (2001). "Consortial journal licensing: experiences of Greek academic libraries." Interlending & document supply. Vol. 29, no. 3, pp. 120-125.
Copyright © 2004 Cristóbal Urbano, Lluís M. Anglada, Angel Borrego, Carme Cantos, Antonio Coscullela, and Núria Comellas