Philosophical Transactions has the distinction of being the first periodical scientific publication, as we currently understand the concept — dating back to 1665. It was initially a private affair, organized under the auspices of the Royal Society of London, a collection of prominent researchers who gathered in meetings which are precursors to our modern conventions. Transactions was curated and financed by Henry Oldenburg, who also pocketed the proceeds, meagre though they were. This pioneering venture constituted the framework on which our modern academic publishing system is built. Unlike those early efforts, however, journal publication has grown into a lucrative industry, with all the advantages and pitfalls that this entails.
I want to wander a bit through this system, because I believe it to be a fundamental driver of our current scientific process, for better or for worse.
What do I mean by that? First off, let's have a look at the evolution of publishing houses since Oldenburg's day. As discussed in more detail here, the scientific journal as we know it came about in the 19th century, when well-known titles like Science and Nature were established. These were private ventures, and thus required profitable business models; consequently, the cost of printing, editing, and shipping journals had to be offset by revenue from subscription fees.
Interestingly, these early for-profit publishers did not commonly utilize a peer review process, in which other outstanding researchers in the field are invited to provide their opinion of an article manuscript with respect to quality, accuracy, and consistency (discussed in detail here). While it was not unheard of for an editor to seek an external review, typically a manuscript was assessed by the editor alone. According to the linked article, even Einstein, who published roughly 300 articles in his prolific career, was only subjected to peer review on one occasion. Today, of course, peer review is a de facto standard of the scientific process, and one around which proponents of traditional publishing practices most often rally. It did not begin to attain this status, however, until the end of the Second World War.
Another prominent aspect of the modern scientific article is its perceived impact on future scholarly activity. For various reasons — including the vast scope of scientific effort, ranging from the highly specific to the very general, the microscopic to the macroscopic, and everything in between — scientific impact is a concept that is notoriously difficult to quantify, although not through lack of effort. At present, there is a plethora of proposed metrics which target journals or individual articles. The simplest and most utilized metric, however, remains the impact factor, which constitutes a sort of base currency for scientific research. A journal's impact factor is, quite simply, the number of articles it publishes, divided by the number of times those articles are cited — resulting in a metric which reflects the average citation per article.
I use the term currency intentionally, because it provides a segway into the topic I really want to address: the business of scientific publishing. The traditional business model of a scientific publishing firm went something as follows. For a given journal, overhead expenses consisted of: printing costs (extra costs for colour plates); salaries for editorial, type-setting, and administrative staff; marketing costs; and delivery costs. Profits were derived through subscription fees and advertisements. This made complete sense in a world where all media was physical media, and it still made sense when I was a M.Sc. student wandering an immense and mystifying library to find and photocopy the articles I was interested in (for the record, I'm not ancient — I did have access to Medline, but this was all still new on the scene).
It does not make sense today. I cannot recall the last time I've had a physical periodical in my hands, or even for that matter visited a physical library. If I want an article, I access it online and download and/or print it out. This has indeed become the standard of scientific communication in our modern era, but while the major publishers have adapted in the sense that they without exception provide web-based platforms for disseminating articles, their business models appear to be stuck in the 1960s. For the modern publisher, overhead no longer includes printing or delivery costs, but rather the substantially smaller (and fixed) expense of maintaining a web server. Moreover, a large part of the effort that goes into the peer review process — that of managing or reviewing articles — is work done by highly qualified but unpaid academics. As far as I can see, only the administrative, type-setting, and high-level editing work constitutes additional overhead. And marketing, of course.
Despite this, the subscription fees typically applied to journals do not seem to have been adjusted accordingly. Many journals (without naming names) still charge enormous fees for colour prints, as if authors had any interest in physically printing their articles. The library of my alma mater, Memorial University of Newfoundland, recently announced its decision to cancel subscriptions for 4000 journals, for which it was paying $1.4 million CAD. Indeed, the Canadian Association of Research Libraries recently reported that journal costs have actually increased 25% over the previous four years, estimating that Canadian universities collectively spent $250M in 2015 to access online articles. To quote that site:
Recent research shows that journal prices are much higher than the true cost of publishing. The top five publishers, who control over 50% of the market and above 70% in some disciplines have profit margins in the order of 28-38.9% in their companies' scholarly journal divisions, bringing them in close proximity to pharmaceuticals industry leader Pfizer (42%), and vastly outpacing corporate giants such as Disney (14%) or Toyota (7%). Meanwhile, researchers are finding that open access publishing costs are far lower costs for publishing research. It is becoming increasingly clear that the subscription-based publishing system is simply unsustainable.
Wow. Scientific publishing is not only an industry, but it is a very lucrative one. Not bad for an enterprise which does little to actually produce their product, beyond sending emails and paying someone to typeset (and even here, using the proper Latex template, one can pretty much do this on their own). And marketing, of course.
How, you may reasonably ask, have they been allowed to continue milking this cash cow? A fine question. I think it lies mostly in the way academia is currently organized. Namely, universities geared to produce an overabundance of advanced degrees, creating a whole lot of junior researchers trying to find their path in science while at the same time distinguish themselves from their numerous peers. Increasingly, the means by which one distinguishes oneself (i.e., to a funding or hiring committee) is through the publication of articles in high-ranking journals. The impact of one's articles actually takes second fiddle to the impact of the journal in which they are published. Here again, we come back to the impact factor, which basically defines how highly a journal ranks. Journals with high impact factors (typically traditional journals such as Science, Nature, Cell, New England Journal of Medicine, etc.) are clearly in high demand, and as such, their publishers have the luxury of setting high fees. In addition, publishers (emulating cable providers) tend to bundle their high-demand titles with lots of lower-demand ones that few serious scientists have an interest in, jacking the prices for these as well.
This business arrangement also has an impact on the nature of the scientific process itself. The types of research lines that are typically high impact are also of a more general nature, since they appeal to a wider audience of researchers. However, in the field of science, general observations are almost always built on a body of evidence derived from narrower, specific research lines; as well as articles with a more technical or methodological focus. Since such approaches are lower impact, they are less attractive to young scientists trying to distinguish themselves. This leads to the tendency for research articles to focus on general interest topics (e.g., the relationship of functional connectivity to some cognitive or disease state), without rigorously testing the assumptions that are necessary to interpret such results. The focus on impact factor thus risks sacrificing detail for popularity. While excellent research will happen regardless, it is important to note the ways in which our reward systems can influence our priorities.
One alternative that is becoming increasingly popular is the open access model. Here, publishers eliminate subscription fees and make articles available to everyone, replacing these instead with submission fees, to be paid by the researcher. Open access fees are typically in the $thousands, and still constitute a substantial cost for research institutions. As the fee varies drastically depending on the journal, there is clearly still a high profit margin to be gained from this model. Another annoyance with the open source model is a new breed of start-up journals with dubious names, editorial boards, and review processes. Because this form of publishing has little overhead, it has become a new form of spam; few days go by when I do not encounter a handful of emails soliciting papers or offering me editorial positions.
As a closing note, I'd like to point out that not all hope is lost. On the contrary, in many countries academia is beginning to reassert itself, most conspicuously by means of a growing chorus of voices demanding open access publishing models and open data obligations. Dutch universities, for instance, have recently reached a deal with Elsevier, ensuring that 10% of output from Dutch universities will be published open access without additional cost to the researcher, a percentage which will grow by 10% annually. In the U.S., the NIH has stipulated that studies which it funds must be made open access via PubMed Central. The trend in many other countries is also going in this direction. New journals with modern philosophies, such as eLife and PeerJ, are cropping up, and promise to be a major headache for tradition publishing practices in the years to come. I know I plan to move in their direction as much as possible.
