Canadian Journal of Communication Vol 45 (2020)157–171 ©2020 Canadian Journal of Communication Corporation http://doi.org/10.22230/cjc.2020v45n1a3765
Rowland Lorimer, Simon Fraser University
Rowland Lorimer is an Emeritus Professor, SFU, Project Director of the Canadian Association of Learned Journals Readership Analytics Project, Director of the Canadian Institute for Studies in Publishing—Journal Services, and serves on the Board of Directors of the Canadian Journal of Communication. Email: firstname.lastname@example.org.
Background The historical context for this article is the embrace by the Canadian Journal of Communication of emerging technology and, most recently, article analytics.
Analysis The article focuses on four years of usage data for articles published since the 1974 founding of the CJC. It is intended to assist the journal, authors, and readers in understanding the potential for usage of articles published in CJC. The article details overall usage; explores journal penetration of global, national, and special markets; depicts the range of subject matter published; outlines the performance of the journal website and two secondary aggregators; summarizes usage levels throughout the journal’s collection; and reports on the year of publication and age of article data.
Conclusions and implications The details of the findings provide insight into the nature of usage of articles in the field. In general, but of less interest than the detailed findings, one can surmise that usage is influenced by subject matter, presentation style, editorial vision, an initial quick rise to peak usage, access dynamics, but not authorship in and of itself.
Keywords Scholarly journal publishing; Communication; Scholarly communication; Online publishing; Usage analytics
Contexte Le contexte historique pour cet article est l’adoption par le Canadian Journal of Communicationde technologies émergentes et, tout récemment, d’analyses de données sur les articles.
Analyse Cet article porte sur des données d’utilisation recueillies sur quatre ans pour tous les articles publiés dans le CJC depuis son lancement en 1974. L’objectif de l’article est d’aider les revues savantes, leurs auteurs et leurs lecteurs, à mieux comprendre le potentiel pour l’utilisation d’articles parus dans le CJC. Cet article recense l’utilisation globale; explore la pénétration de revues dans les marchés mondiaux, nationaux et spéciaux; indique l’éventail des sujets traités; expose les grandes lignes de la performance du site de la revue et de deux aggrégateurs secondaires; résume les niveaux d’utilisation pour l’ensemble des numéros de la revue; et présente des données sur l’année de publication des articles.
Conclusion et implications Les données recueillies illuminent le type d’utilisation d’articles dans le domaine. Bien qu’il soit préférable pour le lecteur d’examiner les données en détail, on peut tout de même les résumer en suggérant que, parmi les influences sur l’utilisation, il y a le sujet de l’article, son mode de présentation, la vision éditoriale, un laps de temps avant l’utilisation maximale et une certaine dynamique d’accès. Quant à l’identité de l’auteur, celle-ci ne semble pas avoir autant d’influence.
Mots clés Édition de revues scientifiques; Communication; Communication savante; Édition en ligne; Analyse d’utilisation
In 1994, reflecting enthusiasm for the internet’s capacity to make information available around the world at minimal cost, the Canadian Journal of Communication (CJC) became one of the first established journals in the world to upload its back issues to the internet and make them freely accessible. It also developed a subscriptions management module so the journal could publish online and maintain the necessary revenue to staff its operations. Upon completion, CJC shared its subscription module with what became Open Journal Systems (OJS); CJC then began using OJS’s manuscript management software, thereby bringing its entire operations online while maintaining a print version.
The Canadian Journal of Communication reported on these developments to the Canadian Association of Learned Journals with regularity and encouraged other journals to take advantage of emerging technology. In 1997, CJC was involved in co-convening “Scholarly Communication in the Next Millennium,” a conference that discussed not only emerging technology-based administrative and dissemination opportunities but also explored what librarians were calling the “scholarly communication crisis.” This “crisis” was brought about in the late 1970s and early 1980s by library budget increases that fell short of providing their clients with access to the quickly expanding body of scientific research combined with the increased costs of maintaining subscriptions brought about by post-war price-, profit-, and title-maximizing practices of first some, and then many, international science, technical, engineering, and mathematics (STEM) journal publishers.
While CJC initiatives were demonstrating how technology could ease administration and vastly expand dissemination, the librarians were giving notice that, as consumers, they expected technology-facilitated price drops. Indeed, in the late 1990s and 2000s both librarians and publicly minded scholars questioned the need for publishers and even journals themselves, a question that was taken up by some authors who felt that the peer review process was a too stringent barrier to sharing research results. This questioning became a dominant theme in the open access movement, and it is reflected in the success of predatory journal publishers.
These developments laid the foundation for a fundamental transformation of the realities of scholarly communication over the past two-and-a-half decades, which has unfolded in two streams. One stream, led by journal innovators, emphasized technological facilitation. A second stream, led by libraries, claimed partial if not total rights to the benefits of technology on behalf of their institutions.
Setting the streams aside, consider a brief history of journal publishing. In the print world, articles have been disseminated in three ways: personal subscriptions, mailing off-prints, and library subscriptions. Conferences also played a role, as they do today. In the years following World War II, a war that was won, at least partially, by the application of theoretical physics, demand for access to such knowledge soared. Every university wanted access to expanding research activities in physics and chemistry, education was seen as a valued investment, and the participating countries had to cope with returning soldiers and the ensuing baby boom. This quick and dramatic expansion of the research literature devalued personal journal collections because they were increasingly inadequate. The circulation of research results via off-prints was clumsy, reflective of small-club dynamics, and often required out-of-pocket expenses by researchers who rarely saw any benefit sharing off-prints with people they had not and never expected to meet. And while, for researchers, libraries were increasingly essential in staying current with one’s sub-discipline, there was little central coordination of singular collections via interlibrary loans. Most importantly, the exchange that did exist was far too slow and not necessarily reliable. In short, the circulation of print-based research reports was constrained: only the richest universities could fully serve their researchers and students.
A case in point: At its print zenith, CJC had only approximately 600 subscriptions. Many went into libraries and claims for the “non-receipt” of issues, often a library worker’s euphemism for lost or stolen, together with the establishment of hour-based lending “reserves,” suggested that those copies were used frequently. Duplicate subscriptions by departments also suggested that use was extensive. High-quality photocopying, then called Xeroxing, which was introduced to universities around 1967, further increased article usage. But still, the system was both slow and clumsy.
Usage of print-based publishing content was also opaque. Both authors and publishers were quite blind to the readership levels. In this environment the founder of the Institute for Scientific Information, Eugene Garfield, developed the Impact Factor. It measures the sum of the citations the articles in a particular journal receive in a year, a crude aggregative measure of readership to be sure. For social sciences and humanities (SSH) journals there was no parallel metric.
Print-based journal publishing may have resulted in hundreds of readers for an average article and perhaps thousands of annual readers for an average journal. No doubt some prestigious titles reached the tens of thousands. And in certain disciplines other mechanisms ensured the flow of necessary knowledge. But, for the most part, researchers were blind to the size and nature of the readership of any of their published articles.
In 2020, established journals are seeing hundreds of thousands of article views. Internationally prestigious journals have reached the millions. In short, the transition from print to internet- and digital technology-facilitated publishing has created a fundamentally different information universe. Production management has been eased substantially; usage far beyond levels in the print era is now entrenched, and opacity has been replaced with an emerging transparency. We can see how often an article is seen, where, for how long, and, in some instances, by whom. Usage data can be crunched with relative ease to gain “big data” insights into the nature of knowledge production and acquisition. As a result, there has been an explosion of meta-publishing initiatives—some legal, others semi-legal, and still others clearly illegal.
The most obvious example is social network sites such as Academia and ResearchGate. They have sprung up to capitalize on informing authors about usage of their articles. As they build their databases, these social networks provide additional community-building information. However, their metrics are based on workarounds to gain access to articles and recirculate them. They measure the usage of their copy version of an original article by those who sign up as members. Next come “preprint” servers, usually not-for-profit academic-run operations that attempt to work around cost of access and delays in publishing by accepting non-peer reviewed work. Then there is Digital Science, an organization owned by an international publishing conglomerate that has developed a rich meta research ecosystem operation, the tentacles of which are ever evolving. Joining the mix are university- and field-based repositories often run by university libraries. And joining them are other organizations such as Crossref, an organization that administers Digital Object Identifiers (DOIs). This list could continue almost indefinitely. It would include Google Scholar and its analytics services, a slew of indexing and abstracting services—including Elsevier’s Scopus, the Publisher Analytics Reports from the Web of Science Group—and the pirate site Sci-Hub. Indeed, the field is becoming so populated with various initiatives that its complexity is threatening the exclusion of small-scale operations and even major efforts in less developed economies.
This explosion is a dramatic demonstration of the unanticipated value of the digitization of the generation, funding, carrying out, reporting, disseminating, and derivative exploitation of research. It calls for thoughtful, inclusive design and strategic participation. Ironically, digital was supposed to be cheaper.
The Canadian Journal of Communication believes there is an opportunity, perhaps even a necessity, for independent scholar- and society-run journals to extend their efforts to include collecting and analyzing metrics that reflect the usage of the actual site of publication. This would parallel the analytical efforts that are de rigueur with commercial journal publishers. It would also help keep the benefits of publishing research in the research community by providing an alternative to a shift from financially exploitative publishing to financially exploitative meta-publishing. The challenge is to create a joint analytical initiative by independent scholar- and society-run journals that is of sufficient size to match private-sector initiatives. Thus, in keeping with its embrace of technology, CJC has entered the world of usage analytics.
The analytics reported in this report are a beginning. It is intended that the first beneficiaries of these analytics be authors. Thus, this report is cast with that perspective in mind. We also believe that, more broadly, journal workers, other researchers, and students can benefit from gaining greater insight into article usage and the digital dynamics of scholarly communication. Finally, the following analytics are intended to spur the interest of other journals. The next step, already in motion, is to assemble a small group of journals and provide a group report.
This article provides descriptive data drawn from three sources: the CJC website and two secondary aggregators with which CJC has contractual arrangements: EBSCO and ProQuest. It does not include data from other sources that carry CJC content, including course management systems, ResearchGate, and Academia.
In overview, it appears that a number of major factors influence article usage. These factors are suggested as heuristics to assist in understanding the data presented:
In 2018, the following performance measures (see Table 1) describe the market penetration and reach of the collection of articles published by CJC.
Of all the article views on the journal website, 86 percent were HTML views and 14 percent were PDF views. On the ProQuest site, 78 percent were HTML views and 22 percent were PDF views. There is no differentiation between HTML and PDF views on the EBSCO site. As a percentage of the total views, PDF views have been declining over the past four years (2015 to 2018). Articles are available in both formats on the CJC website.
The reach of CJC articles in 2018, facilitated by indexing and abstracting services, was 216 countries. The top ten countries according to usage values are listed in Table 2.
Figure 1 depicts the data from the highest-use countries over four years, from 2015 to 2018. Note the 2016 data for Canada brings it into first place in usage.
Articles in CJC are used across Canada. As Table 3 shows, setting aside Québec, generally speaking, the greater the provincial population, the greater the per-capita usage of CJC. The presence of post-secondary communication programs within provinces also appear to stimulate usage.
Interestingly, but not shown in the table, the ratio of HTML to PDF usage varies slightly from province to province. University library acquisition policies appear to account for some of this variation.
An examination of the 20 most frequently accessed articles in 2018 on the journal website (which, in general, accounts for 80 percent of usage) gives a sense of both the content and the number of times each of these high-usage articles was accessed. There are several elements to note in Table 4.
Not yet shown in any of the data reported are the different levels of access to CJC articles gained in various countries through secondary aggregators. Generally, as Table 6 indicates, the aggregators used by CJC are active in large English-speaking countries and less so in non-Anglophone markets. The high use by readers of aggregators within Canada is, from a business perspective, disappointing. The revenue the journal receives from secondary aggregators per article view is woefully inadequate to pay its employees and other operational costs.
A focus on the top 20 articles neglects the performance of the totality of articles published by CJC. Such a focus provides certain insights that would be lost if an attempt was made to include even 100 of the 1,000 or so of the articles CJC has published since its beginning in 1974. But, by no means does a focus on the most frequently accessed articles tell the whole story. A sense of that story can be gained by Table 7, which details the usage levels of the top ten, then the following ten, and the following three tens, up to the top fifty articles, as well as the remainder of the total published collection of the journal. While the top ten do stand out, all the articles that go mildly viral are within the top ten, the usage levels of the top 11 to 50 drops dramatically. Most dramatic is the usage of the remainder of the collection. Those “remaining” 900 or so articles account for over 50 percent of website usage of CJC content and over 90 percent of the usage on ProQuest’s presentation of the articles. This finding was unexpected, and its significance is that, in general, the entire collection is being accessed. This appears to reflect curiosity-driven, subject-oriented behaviour rather than a narrow focus on “outstanding” articles, which is a real plus from an educational perspective.
One measure of market reach is the number of countries from which readers accessed content. It would be reasonable to expect that there might be a close correspondence between usage levels and the number of countries reached by any particular article (the more countries, the more views). As Table 8 shows, there is some relation but not a close correlation. This suggests that, at least for some articles, the audience may be very small but widespread. There is also room for follow-through research. An exploration of which countries were reached by which articles might show interesting patterns that could suggest expansion opportunities for journals.
Moving back to the bigger picture, and furthering an effort to frame the usage of the whole collection, the following table indicates that, in the years 2015 to 2018, there have been just over three million article content views—that is full-text views plus titles and abstracts. Table 9 provides the data, and Figures 2 and 3 present them in two different ways. Table 9 also indicates a steady annual increase in all views of the growing collection of articles and the presence of a substantial viral article. This three million-plus level of use is not only reassuring for CJC authors but also represents a considerable effort by journal staff and a substantial contribution to knowledge in Canada and around the world.
Figure 2 presents article and journal performance by highlighting the frequency of use over four years of articles published between 1994 (when, 20 years after it was founded, CJC began its online backlist open access operation) and 2018. Notably, there are two spikes, one for articles published in 2009 and one for articles published in 2006. The highest spike (over 450,000 in 2016)and unmatched in the history of the journal) is attributable to a sudden interest in 2016 in an article that was published in 2009 (Basen, 2009). It appeared in a lengthy issue entitled “Rethinking Public Relations” and the article dealt with political framing (a.k.a. branding in a commercial context). The article explored how the Conservative Party of Canada framed the ultimately unsuccessful Liberal candidate, Stéphane Dion, in the lead-up to the 2008 Canadian election. Nearly two years before the election, February 4, 2007, the Conservatives aired a Superbowl attack ad portraying him as a weak leader. As the election grew closer, the Conservatives, inspired by American politics, attacked Dion not for his weaknesses but calling into question his strengths.
The second spike is attributable to a number of strong articles published in 2006 that generated considerable usage in the ensuing years. Three appear in various top 20 listings. Particularly notable was a special issue (31.1) (Canadian Journal of Communication, 2006), which reported a number of studies resulting from a 2004 conference focused on the social impact of culture. The conference was funded by Canada’s Department of Canadian Heritage, along with others, and its title was “Examining the Building Blocks of Cultural Citizenship.” The conference concluded that more research on the social effects of culture was needed in six main areas to accompany economic analyses and to provide the foundation to warrant policy development. The issue contained some of the resulting studies.
While the four-year snapshot of usage depicted in Figure 2 provides some interesting data that suggest the long-term value of published articles and some outstanding performances in certain years, Figure 3 portrays growth, delay, and the duration of usage.
Figure 3 presents usage data by the age of article. For example, one-year-old articles, five-year-old articles, and seven-year-old articles. Thus, the usage in 2015 of all articles published in 2010 is combined with all 2013-published articles in 2018 and so forth. In contrast to Figure 2, Figure 3 depicts an immediate rise of usage within a year of publication that is sustained for three years. Demand appears to decline slightly for four to eight-year-old articles (excluding the demand spike of the 2009 “framing” article in its seventh year). Nine to 12-year-old articles then see some resurgence in demand, but 13-year-old articles begin to see a slow decline in demand through to year 20, where demand rests at 64,000. Within those trends are various ups and downs that are probably less important than knowing that the average level of annual usage of all articles from their year of publication to year 12 is 142,314, and from year 13 to year 20, the average usage is 79,969. The latter is far from nothing.
The last column on the chart underlines this “far from nothing” element. Articles more than 20 years old published in CJC from 1974 up to 1998 attracted over 582,000 views between 2015 and 2018.
As noted in the abstract, in general, usage is influenced by subject matter, presentation style, editorial vision, an initial quick rise to peak usage, access dynamics, but not authorship of multiple articles. However, the particulars, which themselves are overviews, are far more interesting. Each of the findings opens possible tentative insights, and more research on more journals is required to begin to draw conclusions about usage. However, CJC authors can conclude that they are in good company in terms of single article usage level, the length of time usage can be anticipated, and a reach of over 200 countries, including all Canadian provinces. This applies not just to a few, frequently used articles but to many. Authors can be confident that their articles live alongside complementary articles spanning a substantial breadth of topics in Canadian communication; that their articles can reach subsets of readers in substantial numbers; and that usage is widespread among many published articles—readers are not artificially pulled to high-use articles, celebrity authors, or a narrow set of topics. Readers can also be assured that they are tapping into a well and widely used body of research. Those who contribute in other ways to the CJC can take pride in these various elements, alongside CJC’s financial management and its record of innovation. This history of innovation started with CJC being the first established Canadian journal to make back-issue content available online without charge and extends to making its performance public in this article.
Canadian Journal of Communication, https://cjc-online.ca/index.php/journal
Digital Science, https://www.digital-science.com/
Google Scholar, https://scholar.google.com/
Open Journal Systems, https://pkp.sfu.ca/ojs/
Sci-Hub, Sci-Hub is a pirate site with a changing URL
Canadian Journal of Communication. (n.d.). Advertising information/indexing. URL: https://cjc-online.ca/index.php/journal/about/editorialPolicies#custom-6 [January 16, 2020].
Basen, Ira. (2009). A schlemiel is the elephant in the room: The framing of Stéphane Dion. Canadian Journal of Communication 34(2), 297–305.
Canadian Journal of Communication 31. (2006). Volume 1, table of contents. URL: https://cjc-online.ca/index.php/journal/issue/view/114/showToc [January 16, 2020].