Academic publishing is broken
I have been meaning to write this blog post for a while, but it has taken me a long time to collect my thoughts into a coherent narrative, and to consider the various aspects and nuances of the topic in as objective a manner as I am able. I’ll also note that this essay is very much my own opinion as opposed to the opinion of Scientists for Labour, though the other committee members have been kind enough to publish my thoughts via the SfL blog. That aside, let us dive in with an overt statement of my opinion of the current state of scientific publishing: it is utterly, completely, not fit for purpose.
Scientific publishing has existed in its current form for around a hundred and fifty years – while the first journals began publishing in the 1600s and the first peer-reviewed journal (Medical Essays and Observations) was published in 1733, today’s powerhouses (such as Nature and Science) were first published in the late 1800s. The scientific journal came about as a necessity; scientists needed a method of communicating their findings to the scientific community and the wider public in a universally accepted format, ideally accompanied with some “seal of approval” provided by the community as a badge of this-is-good-science authenticity associated to the work.
The solution to this problem began with the publication of formal letters, sent back and forth between scientists, in publicly accessible journals. These letters evolved into longer articles, and the journal paper was born. This route to publication solved the accessibility problem, with other scientists and the public now able to access findings. Peer review came later to act as the desired seal of approval, noted generally within science as both the gold standard of scientific rigour and personal a badge of honour. Many of us who work in science recall the publication of our first paper – the trepidation before its submission, the terror as the reviewers’ reports return, the final elation that arrives with a formal acceptance email.
This model has evolved over time, with journals joining together to form publishing houses with greater power to distribute printed papers and full-time staff employed to copy edit and co-ordinate the review process. This model has long been required, as academics cannot feasibly provide the review and article distribution infrastructure that allow the readership to access their material. Publishing houses are fundamentally required to distribute the findings of scientific research, to ensure that new information is reported to scientists and the communities that funded their work.
Or, at least, they were required for that function. Recent advances in information technology have provided new services that, in my opinion, render publishers functions mostly defunct (it’s called ‘the internet’, you may have heard of it). As businesses, scientific publishers have always taken some income, generally from readers purchasing printed copies of research papers. They would print physical copies and distribute them for some charge, typically on a subscription basis. All very reasonable, until quite recently as the vast majority of research is now read in a digital format. University libraries still often maintain subscriptions to physical journals, but I (and most of my colleagues) have never read a single physical paper; all the science we now consume comes in a digital format. Of course, we should note the substantial overhead to the digital infrastructure required to maintain a journal website and the setup of such a service is a significant task, but the internet giants have shown that services can be provided free at the point of use, with Wikipedia being a particular shining example of such a service that operates without advertising on a not-for-profit basis.
With digital document sharing, published science is now essentially available in abundance. We can access almost every piece of work ever written, in seconds, for negligible cost. Or at least, we could, if it weren’t for the stranglehold that the scientific publishing industry maintains over academic publishing. Today’s journals have evolved into ever bigger and broader beasts – “megajournals” such as PLoS ONE or Scientific Reports now accept tens or even hundreds of papers per day, often operating with journal scopes that cover all of science and medicine. Indeed, with the paper-by-paper consumption that is generally the norm in today’s scientific communities, these megajournals appear to no longer see the necessity for specific journals that focus on some small subsection of science.
For papers published in any journal, two funding models prevail. This first is an ‘author pays’ model, where the author pays some fixed fee on publication (often a few thousand pounds). That fee then allows free-at-the-point-of-use open access to the paper for its full lifetime, regardless of the number of downloads, reads or citations it receives. The second is a ‘reader pays’ model, where a reader pays some smaller fee per view, often a few tens of pounds. Typically, in both instances, authors and readers do not literally pay for publications – more commonly a funder or the authors’ host institution pays, often using public money in the case of universities and research institutions. The former model is arguably the most popular, from the authors’ perspective because it allows the easiest access by the most people and from the journals perspective because every article earns a significant flat amount of revenue. This means a few popular articles comparatively provide less profit for the journal, but the vast majority of less popular articles comparatively provide more profit. The journals still require the popular articles to maintain the standing of the journal, but there’s good money to make from the articles that never get seen. More journals are moving to the open access model by default, and it’s generally given a flashy name; ‘gold open access’ being the most common.
Both models, however, beg the question of what any of us are actually paying the journals for. In many cases, the public is double funding the research – once to pay for scientists’ salaries and bills, and again, to read the outcomes of their work. In the pre-internet world, this model was reasonable given the service that publishers provided in acting as the middleperson between author and reader. But now, scientists can publish what they want online, whenever they like (for example via ResearchGate, or a personal website). In scientific self-publication, when the goal is generally for dissemination as opposed to financial profit, what use is there to funnelling thousands of pounds per paper of public money into private for-profit corporations, that now do very little more than manage a website? And while these websites are substantial efforts in themselves, could a more socialist approach be taken towards some alternative digital solution?
Let us recap exactly what happens in scientific publishing. A group of scientists conduct an experiment and gather some results. Their morals, the conditions of their funding and/or their job prospects require them to publish those findings, so they choose to publish a research paper. They submit that paper to a journal, and it is received by an editor, who, in all but the most prestigious journals, is an unpaid volunteer. That editor sends the document for peer review, where it is reviewed by experts in the field, who are invariably also volunteers. There is an ongoing discussion between author, editors and reviewers until the article is deemed acceptable, and it is then passed for the first time (directly) to an employee of the journal who generally performs some light copy editing and formatting, and then uploads the document to a website. The article may then be made immediately available or may await publication in the next journal “issue”, which may or may not include physical printing, and is then considered “published”. For posterity, we should note that the journal also maintains their websites that handle submissions and facilitates the publication process generally. In my eyes, the publishers’ role was a significant task when conducted on printed paper and by written letter, but in today’s ultra-connected, always-on world, to me they feel utterly defunct. As I mentioned earlier, establishing a digital alternative will be a significant task, but I believe there is a way forward that provides research outputs free at the point of use on a not-for-profit basis. The peer-review process is also generally closed, in that very few journals publish review comments, or name reviewers, or use double blind peer review. Additionally, published articles must be in a format appropriate for print media, particularly they must be static written documents. It is an old format that science continues to deem necessary and ultimately puts up with, but do we really need to?
So why does it continue? In my opinion, this process is generally followed because publishing in a journal comes with a soft guarantee of readership – journal x is likely to provide my paper with y reads and z citations, as suggested by some metric employed by the journal to assess their performance (evaluation of such, oft-rubbished metrics being the potential fodder of some future blog post). All very well, until we consider that much of the established publication process may not be needed or could be automated using modern technology.
I posit the following: without article authors that need to publish, review and edit to keep their jobs, there would be no journals.
We now have other solutions to the original reason behind the creation of journals (i.e. dissemination of results that are established as fact by peer-review). As scientists, what prevents us from self-publishing our work online? Plenty of free-at-the-point-of-use solutions exist, from personal websites to scientific social media sites such as ResearchGate, to alternative research presentation formats like LinkedIn, YouTube or Facebook pages. Services like GitHub present a highly impactful alternative method of disseminating the outcomes of research, with version branching and cloning that provides code as a living document. Given that the UK Government generally judges us as scientists by our so-called ‘impact’, are we really assuming that we have a higher likelihood of a journal article being read and cited than a well-produced YouTube video being watched and cited in a non-standard publication? To take the argument to its extreme, what is better for the UK plc – writing a paper that is ultimately cited by one author and forgotten, or creating a YouTube video that inspires a bedroom-based inventor to create something that changes the world? Our outputs are outdated, our metrics make no sense, and no one knows how to judge if we are effective or not. With self-publishing we can share our own research – journals generally rely on scientists self-sharing their work anyway, as very few scientists keep up with new editions of journals anymore. I find new publications on other scientists’ social media feeds, not on a dusty journal website.
While self-publishing seems like an obvious way to cut out for-profit journals, an elephant obviously remains in the room, however: if I self-publish, how can I keep the gold standard of peer review? The current system generally requires a paper to be formally reviewed by two (or more), generally anonymous, experts. This was an excellent system when it was invented in 1733, but by today’s standards, in my opinion it leaves a lot to be desired. Sometimes experts are biased, sometimes the wrong expert gets called to review and sometimes poor research filters through to publication. Compared to what we could provide with an online solution, I think that format is wildly defunct – imagine if papers were living documents instead of static snapshots, constantly updated with new research as it arrives, commented on by experts in a transparent and open manner, collaboratively improved as experts from around the globe add to the knowledge presented in any one format. Why have a pdf when you can have a wiki, with live animations of data analyses, tours of laboratory setups, live footage of experiments, supporting media files presenting the work as it happens?
There are some things about this new publication mode I haven’t solved and I think require some more thinking. Firstly, I think it is likely we will need some people to adjudicate on disputes and disagreements between experts from the same field (‘editors’ in the traditional sense). While I think this role should be filled by the scientific community as a whole and that if there are disagreements then the disagreements themselves should be prominent, keeping everything free from corruption will be difficult, though I think no more or less so than under the current system. Secondly, and probably more importantly, we will almost certainly need a (likely large) team of people who will have to aggregate, clarify and organise information. In the best case, I think this team will look something like Wikipedia’s own wealth of volunteer content moderators, but this could also be used as a means for international job creation. Again, I am not sure exactly what form this team should take, and more thinking is required.
I also appreciate that some journals have taken steps to improve transparency, by facilitating the creation of data repositories, allowing for supporting files to be attached to journal papers, implementing double-blind review etc. However, publishing houses ultimately remain for-profit entities, that exist to make money from the work of others by maintaining and perpetuating a defunct publication model. Our employers, in subscription to this outdated model, still adhere to “publish or perish” employment practices, which make it very difficult to try new methods of sharing our work and create impact – on a fixed-term research contract, every second in employment matters.
If we are to change the current practices, there is an incredible uphill struggle ahead – we are battling against cultural norms that, while they may no longer make sense, have been established for hundreds of years.
That said, there are smaller steps we can take towards democratising scientific research. We could all look to lead in the creation of new dissemination methods – creating video abstracts for our papers, sharing them online and learning how these formats perform compared to traditional papers, in terms of impact. We could attend internal meetings, sit on university boards and make these arguments. We could join other members of our research communities to new format organise conferences and events with a focus on new methods of dissemination and one of the small positives of the COVID-19 pandemic is that we have now been provided with new tools to do so. We could build co-operative publication houses, owned and operated by scientists, for scientists, that exist not-for-profit and outside of the traditional stranglehold of the big publishers.
In all honesty, I don’t know if any of these ideas will take hold and we may fail at the first hurdle, should we try to advocate for them. But we now have so many options available to us that were not available when scientific publishing was first invented, and I think we should give them a go.
If you want to talk about starting a collective of academics who want to change the way we publish and communicate science, hopefully for the better, please get in touch with me at firstname.lastname@example.org – I would love to talk.