Home | Search | Demo | News | Feedback | Members Only
Michael Mabe, Elsevier Science Ltd, Oxford, UK
What do authors and readers really want from journals? Do these desires differ when the journals are electronic? Do they really use the features they say they want? How do readers actually use the journals they say they read? Surprisingly, many of these questions, so important to publishers and librarians, remain unanswered or unclear, despite the efforts of many to determine them. Electronic journals offer a unique and relatively non-invasive way to find out. The SuperJournal Project of the Electronic Journals division of the Electronic Libraries programme (eLib) does exactly this. It is a unique collaboration between publishers, universities and libraries which has developed a multiple electronic journal application covering both the sciences and the humanities on which the behaviour of users at 13 UK universities can be assessed. Four clusters of journals in political science, cultural and communications studies, molecular genetics and proteins, and materials chemistry are now operational and results are already coming in from the first ones launched. In addition to showing how users actually use electronic journals (and the barriers that exist to use), the data also reveals the different patterns of scholarly activity in the humanities and the sciences.
It is a truth generally acknowledged that Oxford Cocktail Parties usually have at least one philosopher present. If pushed to define "the wise man" the philosopher will generally fall back on Plato's description of Socrates:
Socrates is a wise man because he knows what he does not know
If this description is true, then journal publishers must be very wise indeed, for what we do not know about the way users use journals is very considerable. There can be few products with the venerability of the journal (333 years young last month (1)) where so little is actually known about how (and how often) it is used by its consumers.
A non-exhaustive list of questions we might ask (and not know the answer) will contain at least some of the following:
The answers to these questions will have tremendous impact on the future of journal publishing. Establishing new economic models requires detailed research into usage. It is this concern that lies behind the great interest shown by 17 European journal publishers in being involved with the SuperJournal Project (2).
SuperJournal is a project in the electronic journals division of the Electronic Libraries Programme (eLib). The Joint Information Systems Committee (JISC) of the Higher Education Funding Councils and DENI established eLib as a direct result of the Follett report on the future of libraries. SuperJournal is a project researching the factors that can make electronic journals successful and of value to the academic community. Many studies have explored useability: the barriers that need to be minimized so that readers will use an electronic journal. SuperJournal focuses on usefulness: the key features that make electronic journals valuable to researchers and regularly used.
The SuperJournal project is a collaboration between four different parties:
The project has been funded by the eLib programme of JISC and supported by the SuperJournal Consortium of publishers.
The SuperJournal Project has a number of objectives. Its principal purpose is to learn what features and functionalities of electronic journals are of real value to the academic community, and to explore this value from the perspective of all the stakeholders (authors, readers, publishers and libraries).
We can identify four objectives each aimed at one of the stakeholders:
The main objective of the research is to answer one simple question: what do authors and readers really want from electronic journals (4). The main problem with this question is that it cannot be asked directly, or if it is, readers usually cannot answer it in a meaningful way. Terence Conran (5) described this problem quite neatly in an interview in the Independent last year: the trouble with market research was "[it] only tells you history...If you only show someone a red dress, they will say they want a red dress." Research into electronic journals has been similarly plagued.
Hands on experience of using electronic journals is an essential pre-condition for answering the question of what readers really want. It is only when such a context has been provided for views and opinions to form that meaningful answers can be obtained. Bearing this consideration in mind, SuperJournal has therefore adopted the following methodology:
The SuperJournal application now consists of four subject clusters of journals in the following areas: communication and cultural studies (CCS); molecular genetics and proteins (MGP); political science (PS); and materials chemistry (MC). Each cluster has been designed to provide as great a critical mass of quality material as possible within the constraints of the coverage of the participating publishers and the time available. Each cluster contains roughly ten or more journals from four or more of the publishers.
The purpose of the SuperJournal application is to deliver features and functionality so that readers can tell us those they value most. The application is therefore more a "test-bed" rather than a true electronic journal. A list of value added features was developed for the application to test. These features were in the following areas:
A key consideration in designing the application was timing the resources. As the project is three years in length the application needed to be up and running and available to all user sites within the year. A collection of off- the-shelf software was used by Manchester Computing to provide the functionality, assembled in such a way as to be seamless from the users' point of view. The first release of the application, available with the first journal cluster had the basic functionality. As each successive cluster was launched the application was upgraded with new features. By phasing these features with the launch of each new cluster, response to these features could be monitored as part of the evaluation study.
Thirteen UK universities are participating in the research and are partners in the SuperJournal project. Three or four universities participated in the evaluation research being conducted by the HUSAT Research Institute for each journal cluster, but all thirteen have access to the application and all of its content.
The purpose of the baseline research is to generate a picture of how users currently use existing (electronic) journals and to assess the expectations and enthusiasm for electronic journals. At the time of writing, the baseline research has been completed on universities participating in the first three clusters. Potential users at the participating universities are asked to fill in a questionnaire available on the web. Representative users are invited to focus groups to explore their perceptions and expectations in greater depth. A typical focus group includes 4-10 users, commonly researchers and academic staff but also some postgraduate students. While similar topics are covered, identical questions are not posed to each focus group. The agenda and the level of detail of the discussions are largely in the hands of each particular group. Each session might last about 1-2 hours and is held under the guidance of a researcher from HUSAT.
The attached tables (6) summarize the results of the focus groups (see Table 1, 2 and 3). It should be noted that even if a box is not ticked, it may not mean it is unimportant; the topic simply did not occur in the particular discussion. The tables are divided by cluster (MGP, CCS, PS) and by target user site (C-Cambridge, O- Oxford, U-UCL, N- NIMR, B-Birmingham, Ul-Ulster, W-Warwick, S-Sussex, D-De Montfort, L-LSE).
There is general agreement about many issues facing readers and authors. Interestingly, there is some divergence between scientists and humanities users. The results can be discussed in three main areas: problems using printed journals; a wish list for electronic journals; problems with electronic journals.
Discussion in the focus groups centred on four main areas:
Almost all the focus groups were concerned about the large numbers of missing items in their libraries. These items were missing through theft of the journal issue, vandalism or issues being bound up into volumes at the binders. A lower level concern, among the humanities readers, was the local availability of journal titles.
All participants felt they had difficulty in keeping up-to-date with the literature and were concerned with time constraints generally. A leading concern, in respect of performance issues, was having to go elsewhere (i.e. other libraries or using I.L.L.) to retrieve the information they needed. A large majority had difficulty in finding information in their library and with the slowness of the process. All participants complained about long queues for the photocopying of articles. A secondary concern from scientists was the difficulty of reproducing photos or colour images accurately on a photocopier.
Discussions in the focus groups centred on the following key areas: content and coverage; availability and access; timeliness; searching; linking; multi-media; customization; interactivity and education.
All participants wanted coverage to be as comprehensive as possible. The availability of back issues was especially important to humanities researchers. Almost all participants wanted desk-top access to information. Scientists, in particular, were concerned that access should be fast, instant and guaranteed. Where more rapid publications times were the norm for their subjects, participants expressed the hope that electronic journals were there to speed-up publication. The vast majority wanted to be able to research items over a wide range of journals. They also wanted better browsing and searching facilities in general. Many respondents wanted searching to be a seamless process with a search linking them to abstract and article, in many cases the searches being conducted on full text. Linking was felt to be important by the majority, especially from references to articles and to supplementary data. Humanities users wanted direct links to source materials while scientists wanted direct access to deposited data such as structures, x-ray diffraction data etc.
Screen issues, technical or performance issues and author-related matters were foreseen as the most common problems for electronic journals. A majority of focus groups expressed the view that reading on screen was difficult and unattractive and that "flipping" through articles is desirable but easier in print. The availability of good printers was a key technical issue. For scientists, long down-load times were also critical. There was a generalised, if low level, concern about multi-media matters for authors. Primary among these were the copyright position on multi-media elements and the dangers of other researchers taking and using your deposited data for their own work.
So far an enormous amount of log file usage data has accumulated. Its collation and analysis into meaningful datasets is proving to be more labour intensive than first expected. Although analysis of the key usage data is still on-going, some preliminary results can be reported. I am indebted to Ken Eason, Sue Pomfrett and colleagues at HUSAT for making this currently unpublished material available.
For all the clusters usage has built up slowly over time. The growth of usage has been uneven between the various clusters. There has been considerable growth of non-target site cluster usage. That is, users at universities chosen as the selected sites for a specific cluster have also been using other clusters in the application.
The number of users of the molecular genetics and proteins cluster (MGP) have grown faster than the communications and cultural studies cluster (CCS). In addition there are three times as many users in MGP as CCS. When considering repeat usage of the application a similar picture emerges. Nearly twice as many users of MGP are repeat users in comparison with CCS. The category of users who are repeat users divides into 34% researchers, 25% academics and 17% librarians. However, for MGP most of the repeat users fall into the research category.
Analysis of the actual use made of functionality and features is complex. It really needs long term repeat usage for a confident assessment of actual behaviour. However, some trends are already being discerned which may, or may not, be confirmed at the end of the project. Despite provision of a number of types of search engines in the SuperJournal Application, the actual usage of any of these engines is low. In contrast there are high levels of browsing among all users. Browsing tends to lead to yet more browsing, a point emphasized by the low usage of abstracts compared to articles for both clusters.
Despite the differential in the numbers of users of the MGP and CCS clusters (145 repeat users MGP, 36 CCS) there is an unexpected lack of differential in terms of actual usage by the two groups. MGP users accessed 153 articles per month while CCS users used 98. One possible explanation for this behaviour may be current awareness browsing. MGP users are scanning the journals regularly to check that there are no articles relevant to their research area rather than to find some specifically to use.
Humanities users made specific comments in the focus groups about the need for a large backfile to make electronic journal use more attractive to them. Despite this, the present usage data suggests the reverse is true: humanities users are accessing current issues more often than back issues and using the currents issues more often than the science users. Whether this result is significant or whether it is indicative of much greater back issue reading for all users than anticipated must wait on further results.
With the exception of the capacity to browse from one article to another via hypertext links, users have made little use of other additional functionalities. It may well be that the value of these functionalities will be appreciated when repeat users become more familiar with the application, but there are already indications that the key functionalities for all users are the simple ones: content, browsing, printing.
Most recently (7), it has been possible to analyse the usage data obtained for one of the clusters in an attempt to begin to answer some of the questions posed at the beginning of this article. For the MGP cluster, Figure 1 shows the absolute number of items accessed per journal in the cluster for the period ending February 1998. Figure 2 recasts this information as percentage of items accessed per title by normalizing the usage levels of Figure 1 by the actual number of viewable items per journal for that period made available on the SuperJournal Application. While the final usage levels will naturally change as the journal articles remain available to users of the Application over the coming months, the data already shows that three of the eleven titles (Journals Nos. 2, 12 and 15) have had over 30% of their articles accessed and (it is presumed) read, and in the case of one of these usage has exceeded 45%.
Figures 3 and 4 examine the nature of this access in more detail on three of the journals in the set. Figure 3 shows the absolute number of articles that were accessed once only, twice, thrice etc. Figure 4 normalizes this result by the total number of items that it was possible to view for each title during the period the cluster has been available. The similarity between the curves for differing journals is intriguing. While as expected the highest proportion of the articles are accessed once only (18-26%), small but significant accesses are occuring between three and ten times. While it is not yet possible to say if all these accesses are by different users (authors admiring their own work cannot be ruled out!), there are glimmerings of reproducible behaviour patterns among readers which may have profound consequences for publishers and librarians as they consider new models for the sale and distribution of the academic journal.
So what have we learned?
On the basis of results achieved so far, it appears that the clustering of journals is of higher relevance to science users than to humanities researchers. In the focus groups, the lack of a large archive was said to be more of a problem by humanities scholars; at the time of writing the usage data does not support this.
For all users, content, browsing and printing are the most important functionalities, although it must be emphasized that this conclusion may change as further data is analysed.
It has become clear that obtaining articles containing true multi-media from authors for any of the journals in SuperJournal is going to be impossible within the project time scale. Most authors are reluctant to submit multi-media unless the creation of that multi-media element is directly part of their research work. All authors appreciate the advantage of including hypertext links and the use of colour. More complex multi-media elements such as video, audio, etc. are deemed less important. The benefits of inclusion are outweighed by the costs in terms of difficulty and time available to a researcher.
From a technical perspective, building the application from existing software was relatively easy, and certainly easier than expected. However, getting consistent data formats from the 17 publishing houses involved (and by the extension the technical suppliers and typesetting companies engaged by them) was problematic. There are no universal de facto standards on-line. Even PDF, the closest approach to such a standard, during the duration of the project has evolved from ADOBE Acrobat 2.0 to ADOBE Acrobat 3.0, the files prepared by the latter not being readable by the former.
Many unanswered questions remain. Despite the first elements of detailed usage analysis reported in the last section, we still have to answer most of the academic usage questions posed at the beginning of this paper. We have had tantalising glimpses of the attractive and repulsive forces acting on readers and authors in their approach to electronic journals and a glimmering of the way readers actually use journals. One unexpected outcome has been the difficulty of making qualitative judgements from quantitative data: usage levels are numbers, what exactly constitutes "high" or "low" usage? Benchmarks need to be established.
Getting satisfactory answers to these questions was the starting point for the SuperJournal Project. Over its final year, answers are beginning to emerge that will illuminate some of the most critical issues facing the world of scholarly communication.
The contributions of the staff of all the project partners, the Superjournal Consortium (2) members, HUSAT Research Institute at Loughborough University, Manchester Computing at the University of Manchester, the thirteen participating UK university libraries (3), is gratefully acknowledged. Special thanks is due to Christine Baldwin, SuperJournal Project Manager, Ken Eason and Sue Pomfrett of HUSAT, and David Pullinger, SuperJournal Project Director, for their help and advice in preparing this paper. The support and funding of the project provided by the SuperJournal Consortium of publishers and Higher Education Funding Councils JISC eLib programme is also gratefully acknowledged.
Further information about the SuperJournal Project is available at http://www.superjournal.ac.uk/sj/
This web site is maintained by email@example.com
Last modified: June 21, 1998