Friday, October 14, 2011

Academic Publishing

* The following is a longer version of an editorial that is due to appear in the journal Political Geography later in the year.

First, a little background. In May, 2011, The Economist carried an article, somewhat ambiguously titled ‘of goats and headaches’ (the title remains obscure and it’s not worth explaining: it and related comments can be retrieved at The gist of the piece was that some publishers—Elsevier and Springer were named—are doing remarkably well despite the recession and seem to be weathering the digital information revolution more adeptly than their counterparts, notably in the music industry. The commentator pointed out that this occurs despite some serious problems: many universities cannot afford to purchase the expensive package deals that publishers provide for full access to their thousands of titles (the buffet system), nor do they like the alternative, which is to pay download prices for individual articles (the à la carte option). As the unnamed commentator goes on to point out, this is all a little absurd as academics voluntarily give the publishers the copyright to their work but then have to pay later to use the rest of the academic commons.
            This issue clearly resonates with many who resent paying to use digital information. George Monbiot goes further by comparing Elsevier, Springer and Wiley to ‘feudal landlords’ and decrying them as ‘monopoly racketeers’ for charging $30-50 dollars for published items [see]. He concludes by stating that ‘Let’s throw off these parasitic overlords and liberate the research which belongs to us’.
            Like anyone associated with a large organization, I have intermittently grappled with the issue of how my standards diverge from those of my university or my publisher. I have been an editor for Elsevier since 1995, when I took over Cities: the international journal of urban policy and planning and with which I remain connected; I am also developing a new journal, again for Elsevier, titled Current Research on Cities. Prior to that, I was also Review Editor for Political Geography Quarterly—now an Elsevier imprint but at the time published by Robert Maxwell, whom Monbiot describes as ‘one of the biggest crooks ever to have preyed upon the people’.
Over this long run, then, I have heard the company line at various workshops and observed the evolution of Elsevier’s rhetoric. Twenty years ago, I wrote a somewhat acerbic editorial roundly condemning academic publishing and suggesting that all we needed to throw off our shackles was a photocopier (Kirby, 1992).  This current posting is a rather different interpretation of ‘corporate publishing’.  
What is the value added?
It is a common criticism of publishers like Elsevier that they are rapacious and that they do not deserve the largesse that they extract from the transfer of knowledge. Twenty years ago, this seemed to me to be about setting a price point, but today there is a more complex context in which digital information resides. Basically, the erosion of the frictions of distance increases the expectation that information will be free: if you don’t have to deliver the newspaper, why should I pay to see the on-line version?; if I’m not getting the entire CD, why should I pay to download a few tracks?, and so forth. The music industry tried to throttle what we might call ‘open access’, and nearly went out of business; now, the structure of music supply has changed dramatically, and many musicians expect to survive on touring and selling ‘merch’ rather than on their royalties, as that is where the value added most clearly resides for the consumer.
Now, the same logic applied to academic 01 data, as opposed to musical 01 data, would imply the following business model. We create, we make the information available on our Facebook page, and we make our money on the lecture circuit, where we might make a little extra by signing copies of our papers or selling T shirts. And indeed, there are those at the top who do command ridiculous fees for their appearances, but, of course, most of us are not in that kind of demand and so we have our day jobs; aspiring musicians work as graphic designers, we teach. But the basic issue is that we can afford to give away our information because we do not expect to live on that revenue stream. [1]
This is the logic that fuels the very cross in the blogosphere. Here I think is a representative offering, from a mathematician—“It is a truth universally acknowledged that journals fail to add significant value in a way that justifies their high prices (we write, typeset, referee, edit, and they do basically nothing except charge an arm and a leg for it” ( Just to make the point absolutely clear, the author goes on to suggest that commercial publishers actually remove value by wasting authors’ time requiring weird formats, professional quality diagrams and so on. Monbiot concurs, suggesting that long periods ‘in press’ actually delays the dissemination of knowledge.
Twenty years ago I would probably have agreed with this assessment, but no longer. Mathematicians are representative of a previous scientific era, and by that I mean pre-Enlightenment rather than pre-digital. Their notion of science communication is essentially to pass around proofs and to vote upon their elegance. In such a medieval guild (Johnston also uses the metaphor of the village), there is little to be gained from the contemporary publisher; indeed, this type of closed community barely benefits from the printing press (Johnston, 2005).  For them, the vast expense of digitizing the collective wisdom lodged within literally tens of millions of journal articles, borne by publishers over the past decade, amounts to nothing; the same is probably true of electronic submission systems.
The value that is added for the rest of us by the contemporary publisher lies in an entirely different understanding of the scholarly article, the academic journal and information transfer. [2] In this model, the raw material is the scholarly journal article. In a genuflection to the past, these are notionally assembled into issues of the relevant journal, but for the most part no-one sees the print copy, as most readers now use digital downloads. Elsevier handled its first 1 billion downloads between 1999-2006, and numbers continue to increase dramatically.
In a fairly traditional version of scientific relations, the worth of a paper is assessed in terms of the citations that it garners (regardless of their positive or negative tenor).  Since the political attacks on the public sector that began approximately thirty years ago, these citations have become commodified, and are aggregated as a measure of the worth of the author/researcher (and, by extension, of the academic system as whole). Metrics such as the h-score are now routinely employed to provide a measure against which individuals can be assessed against their peers in a discipline, department, laboratory or institute. [3] The value of the citations themselves can be assessed by addressing the prestige of the journals within which the researcher is published, and there are formulas routinely employed to compare journals (e.g. the Impact Factor, measured over one or five years).
These metrics are, to repeat, ways in which to commodify the journal and the individual researcher, and by extension, the discipline itself. Publishers have played a significant role here, especially in terms of organizing the measurement of citations, although this is not a role that they necessarily created for themselves: h-scores, for example, were developed by scientists who clearly still believe in striving for total objectivity in measurement and who have consequently created burdens for all of us in the process (Hirsch, 2005).    
If these metrics were all that we need from publishers, then it would be relatively easy to find other ways to measure productivity [and the transformation of the global university system, with dramatic changes to the role of the researcher/teacher, demand just that]. However, the real value of the journal article—and by extension, the journal article publisher—lies elsewhere. To understand this, we move beyond citations to clickstreams. In this new data universe, users behave as do all users of the internet, that is, they move from site to site as something catches their attention. To do this efficiently requires a platform such as Web of Science (Thomson Reuters) or SciVerse Scopus (Elsevier). There, the user searches by keyword and finds a matching paper; however, that result then points the user towards other matches, in closely linked journals or in entirely different disciplines. Before going further, it is worth reminding ourselves just how vast the scholarly publication landscape actually is: Scopus, for example, is a large bibliographic database, providing search access to over 18,000 journals—but even this is only a small fraction of the 200,000 journals produced globally (Kähler, 2010). As an additional indicator of the vastness of this repository, there are now over 35 million papers and reviews in existence.
Here lies both the challenge, and the unfolding potential, of digital information. While it is merely nice to be offered new music or a new book by Amazon based upon our preferences, it is truly revolutionary to be able to browse efficiently through thousands of journals and millions of articles to find potentially relevant information. Furthermore, the ways in which this browsing happens can, in turn, reveal interesting information on the structures of knowledge. Figure 1, for instance, is an analysis of over one billion clicks in 2006-7, aggregated to reflect the journals visited. [4]

Figure 1: The linkages between disciplines, based upon clickstream data, from Bollen et al. 2009 [see note 4].

What this reveals is interesting as guide to current research behavior and to future funding needs. The authors of the study provide some insights:
“The connections between the journals in the map's rim cross multiple domains. For example, alternative energy (rim, 3PM) connects to pharmaceutical research and chemical engineering, which itself further connects to toxicology studies and biotechnology. Brain research (rim, 6PM) is connected to genetics, biology, animal behavior, and social and personality psychology. Human geography studies connect to geography, plant genetics, and finally agriculture. A number of clusters are well-connected to both the natural science and social science clusters. For example, ecology and biodiversity (5PM) connects the domains of biology (rim, 5PM) and architecture and design (hub, 5PM). Production and manufacturing (12PM) bridge the domains of physics and engineering (rim, 2PM) and economics (hub, 11PM)” Bollen et al. 2009, p.6).

While it not really central to my argument, geographers may be interested to note the bridging role performed by human geography, while noting that the key social science journals that do transcend disciplines are the American Economic Review, American Historical Review, American Anthropologist and Annals of the American Academy of Political and Social Science.  For the record, mathematics journals tend to be connected only to other math journals.
The potential of bibliometrics
The emerging fields of bibliometrics, webometrics, informetrics and scientometrics are part of an empirical effort to rethink how we understand academic knowledge, using the power of current computer systems to trace relations within the vast data collections (e.g. Frenken, 2009). This augments and, in some ways, displaces more traditional ways to think about academic knowledge, which have a rich deductive foundation (Broughton, 2006). While we may recoil from inductive approaches, which can seem like crude trawling exercises, there is enormous potential here to transcend the limits to knowledge acquisition that have arisen during the past 150 years. When Darwin wrote, it was not impossible to strive to know something about everything—an aspiration that has since become impossible to even contemplate. Now the successful academic is one who knows everything about something, however narrow and arcane that something might appear to the outsider. With this level of specialization come barriers that anyone who attempts to transcend a specialized field knows only too well, with the result that subfields tend become ever smaller and more entrenched.
In this environment, some researchers, who do need to transcend these limitations, require help: a health professional, for instance, may need to understand “the neighborhood” in order to design a research project on suburban health. This is where the publishers come in, offering different types of research support. To be sure, some of these are probably not relevant to social scientists. Streaming video of clinical procedures pushed to the smart phone is valuable to a nursing professional but not to us. However, being able to move in an efficient manner from article to citation to article is a valuable resource. We can also go a step further, creating meta-journals that provide the same information in a more structured manner. The new journal Current Research on Cities is, for instance, designed to provide summaries of information about urban studies and urban areas for those working in the field, but also those whose home base may be in geology or public health and who would not think of themselves as urban researchers, but whose work is taking them in exactly that direction (see
Let me summarize my argument in two ways. First, publishing is no longer about passing journal articles back and forth within a small group of colleagues; and in consequence, publishers do more than simply provide legitimacy for the journals they produce in order to extract value from them.
If the transfer of information were indeed a simple business, then it would be equally simple to get rid of publishers entirely. Open access journals with peer review do of course exist—Bollen’s work [cited above] is such an example—but it is interesting that to find the paper, the user has to work through Google Scholar.  In this example, the production of the research is supported by the taxpayer who funds the Los Alamos Laboratory (where the computing was undertaken), and the paper is found by using Google, which is supported by advertising. The authors also had to pay to have the paper published. There is really nothing to stop open access other than the (opportunity) costs involved for the author—and these may be hefty. So long as we stay with a Royal Society model of knowledge, in which the Transactions constitutes a subscription newsletter, by and for a restricted membership, this is acceptable—but it is also noteworthy how few academic researcher/authors have in fact migrated to open access. [5]
            This brings us on to the bigger picture of contemporary publishing. Its contribution lies in its ability to provide access to a literal world of information (that cannot be obtained via simpler search procedures such as Google Scholar: see Jasco, 2005). In this, the academic enterprise is beginning to mirror other commercial ventures that provide a means to manage vast amounts of data about daily life.  What is unusual is the way that this is funded: users do not pay to search for products on Amazon .com or in the way that libraries do. Perhaps there is no philosophical problem in allowing advertising to subsidize scholarly publishing, although I would point to two practical issues. First, there is the ethical question: in many areas of science, peer review is not about supporting one’s chums, it is about endorsing patentable products, some of which are potentially lethal drugs. The dangers of allowing too much cash to pass through publisher pockets are pretty clear (and that is true of the suggestion that we pay reviewers, also). Second, have a look at your G-mail account, especially the advertisements that hover above and alongside the messages and which subsidize the service. Now start typing a message that contains words such as HOA—the ads will change to give you lists of HOA lawyers in your area. That is fine, and sometimes helpful: but what if you are sending a message dealing with research in which the keyword now becomes ‘pedophilia’ or ‘terrorism’? Who knows what kind of outcome this might trigger—instead of advertisements, it is quite possible that some form of surveillance might be triggered [and I do not intend to test my theory]. Cautious scholars would be extremely circumspect about the potential for this kind of intrusion into their research activities, and while this is an extreme example, it does point to the limitations of depending upon advertising to support some of the costs of publication.
            Do publishers deserve to make big profits—and what constitutes ‘big’?  That is not for me to say: the reader is once more directed to what the publishers themselves say on their own behalf (see note 2). I argue only that they are overseeing a much bigger enterprise than most users understand, and without their innovations, things that we can do today—and might do tomorrow—would  be little more than wishlists.

1. I am obviously simplifying the situation here. Most musicians support themselves: most academic researchers are supported, directly or indirectly, by the tax payer, with the result that corporations are in turn being indirectly subsidized.
2. Publishers themselves are keenly aware of the value added debate: see for instance .
3. The h-index assumes that a scholar with ten publications should be cited on ten occasions, and notionally provides a yardstick against which to measure reputation: formally it is expressed as ‘a scientist has index h if h of his/her Np papers have at least h citations each, and the other (Np − h) papers have no more than h citations each (Hirsch, 2005, see also Kähler, 2010).
4. This is a dataset of approximately 1 billion user actions logged in 2006-7 by Web of Science, Scopus, JSTOR, Ingenta, the University of Texas and California State University. Figure 1 covers March 1st 2006 to February 1st 2007, a total of 346,312,045 user actions relating to 97,532 serial publications, mostly but not exclusively scholarly journals. The illustration is a manipulated version of the original which is used, without requiring explicit permission, from an open access site; the complexities of copyright are an issue that I do not have enough space here to discuss.
5. The Open Access journal in question charges authors $1350 to publish papers that are accepted after review; Elsevier offers authors an OA option (in which the latter maintain copyright) for $3000. In passing, it would be interesting to measure the delay in publishing in peer-reviewed OA journals: as the biggest delay in publication is always the problem of completing serious review, I doubt that the production times, complained about by Monbiot, are any shorter.
Bollen J, Van de Sompel H, Hagberg A, Bettencourt L, Chute R, Rodriguez M.A., Balakireva L. (2009) Clickstream Data Yields High-Resolution Maps of Science. PLoS ONE 4(3): e4803. doi:10.1371/journal.pone.0004803.
Broughton, V. 2006 Essential Thesaurus Construction, London, Facet.
Frenken, K. Hardeman, S. and Hoekman, J. (2009) Spatial scientometrics: Towards a cumulative research program Journal of Informetrics 3 222-32.
Hirsch, J. E. (2005) An index to quantify an individual’s scientific research output Proceedings of the National Academy of Sciences 102 (46) 16569–16572.
Jasco, P. (2005) As we may search – Comparison of major features of the Web of Science, Scopus, and Google Scholar citation-based and citation-enhanced databases Current Science, 89(9) 1537-47.
Johnston, R.J (2005) Editorial Environment and Planning A 37 2-8.
Kähler, O (2010) Combining peer review and metrics to assess journals for inclusion in Scopus Learned Publishing, 23:336–346 doi:10.1087/20100411.
Kirby A. (1992) Editorial comment: Publishing deca(ye)de Political Geography Quarterly 11(3) 235-7.

No comments:

Post a Comment