There have been some heated criticisms online about scientific publishing as an enterprise. While some of the arguments are valid, a lot of them are rather reductive and misleading — as often is the case for most public discourse on the internet. I want to help sort through which of these criticisms are more legitimate than others.
The crux of these arguments are usually along the lines of: Scientists do all the work in science, from research to penning the paper to peer review, therefore journals are parasitic middlemen who profit from the hard work of others and adds nothing of value to science. Therefore, it’s better if we abolish journals.
Here’s the tl;dr for this article: journals play an indispensable role in the progression of modern science, but their current business model is problematic.
What do peer review journals do?
Peer-reviewed journals are where researchers publish the results of their research.
Scientists are not paid to just do research. They are also paid to share the results. Publishing a peer review paper proves to whoever funding the research that the scientists are delivering their promises. Without publication, there is no tangible product for the funders. A scientist cannot just take the money, do the research, and not tell anyone.
Peer-reviewed journals are also a main avenue where researchers communicate their results to each other, which is how science progresses as a collective enterprise.
Because scientific papers need experts from very specific research areas to review them, no individual can be a full time reviewer, because (1) there probably aren’t enough papers in this person’s specific expertise to keep them busy year-round, and (2) if this person has a broader range of expertise, their knowledge for any specific field is probably not enough to be a reviewer.
That’s why we have journal editors.
The necessary labor of journal editors
Because peer-reviewed journals play an indispensable role in scientific research. Their workers need to be compensated for their labor.
So, what kind of work do journal editors do?
Mainly three things: I call them filtering, editing, and publishing.
First is filtering. At Science, we reject about 80–90% of all submissions before sending them out for peer review. Without journal editors to read and sieve through the bulk of the submissions, researchers may have to review 5–10 times the amount of papers they are currently reviewing, and reviewer fatigue is already a problem. Without a broad expert to preliminarily gatekeep, the whole peer-reviewing process will be much more inefficient. It will be like calling for a jury for every single court case.
Then there is editing. Professors are not necessarily good writers for the same reason that they are not necessarily good teachers. They are simply not hired based on their merits as writers or teachers, even if writing and teaching are core elements of their job. From my experience as an editor, depending on the quality of the first draft I receive, I usually have to do detailed editing work for 20–70% of the manuscript. I do often feel like I’m a ghost writer without a byline.
Writing and communication skills are one of the main job requirement for journal editors, in addition to research experiences, which usually requires a PhD. Editors are at the frontend of the assembly line that communicates scientific information to other scientists, and also to journalists and policymakers, including those who approve grants, and the public.
The editing process also plays an important role in leveling the playing field for researchers whose first language is not English — the lingua franca of the scientific world as a legacy of European colonialism. If a piece of research is important but may be disadvantaged by the writing skill of the researchers, the editor can help close that gap.
Finally, there is publishing. This include things like proofreading, typesetting, licensing or creating artwork or illustrations, printing, distribution, etc., which all incur a cost.
It is disingenuous to say that because researchers has already done the research and written the paper, that they can essentially also moonlight as journal editors. This is similar to saying: why should we have cooks and restaurants if the farmers are the one who are producing all the ingredients? The cooks don’t actually “make” anything! Why should the farmers lose out on the extra money when the cooks do not add any “physical” value to the ingredients except for their labor?
An effective division and specialization of labor is the key for any collective enterprise. It is far more efficient to have a small number of journal editors to do what they do, then to have all researchers to split up that necessary labor.
So, how should we pay for this labor?
Paywalls are bad
I agree that the current subscription model is problematic, but this grievance needs not to be justified by invalidating the labor of those who works for a scientific journal.
You can criticize the tipping culture in the US without invalidating the labor of restaurant workers.
It is not right that researchers have to pay to read their own research. Although this is rarely true, because at least in the US, institutions such as universities and national labs usually pay for the journal subscription fees. One justification for this system is this: Although the researchers are credited with their research, the research actually belongs to the institutions that fund the research, and so the researchers do not necessarily have the assessing rights to their own research outside of the umbrella of the funding institutions. Researchers do not own their research, just as miners do not own the coal they excavate.
Under this arrangement, the institutions would provide access to the journals for the researchers as part of the funding — because having access to journals is required for researchers to do the job. But outside of this contract, the institutions are not obliged to give the researchers access to their own research. Just like a CIA agent wouldn’t have access to his/her own files upon termination or retirement.
The retired CIA agent analogy is also a good counterpoint for the argument that taxpayer funded research should be available to the public. There are all kinds of taxpayer funded programs that contain information not disclosed to the public, for all kinds of reasons. There are information that are classified for intended national security reasons. And there are scientific research papers that are behind paywalls for unintended business reasons.
The business reasons are unfortunate, and they are a bug, not a feature of the scientific enterprise. As with most unfortunate arrangements, this has proliferated because the community has so far found the system tolerable. But as information technology continues to outpace societal changes, journals and researchers are exploring new avenues of this area, sometimes with competing interests.
From my perspective, there are currently two main ways to publish research without a paywall.
The first are non-peer reviewed online archives. Depending on the culture of the research field, it is sometimes a common practice to upload one’s research paper to an open archive (e.g. arxiv.org, biorxiv.org, etc.) before or after the paper has been peer reviewed, or been edited by a journal editor, or been accepted by a journal. The main disadvantage of this is the lack of quality control, and most importantly, the absence of the peer-reviewing process, because the model does not provide resources to facilitate the functions provided by journals, as discussed in the previous section.
The other option is to publish your paper in an open access journal, where journals instead of charging readers to pay the editors, charge the authors instead. The downsides of this are more subtle. One that comes to mind is that this shifts the burden from the institutions that usually pay for the subscriptions, to the individual research groups, which now have to include the cost of publication in their grant applications.
The upside for both alternatives is obvious: everyone will get to read your research for free.
The dose makes the poison
But how many people will read it? Unlike newspapers and other publications, scientific journals do not have the same volume of readership to spread out their revenue stream. As a result, the bill is usually too pricey for ordinary individuals to read, or to pay for open access publication.
While the profit margin for for-profit academic publishers is reported to be as high as 40%, is grotesque, I do not think this contributes greatly to the price tag. I do think that this is a problem, but for different reasons of which I will explain later in this section.
More about this 40% number. Now, I am not sure if this is the case, but if this number also includes textbook sales, then I’d reckon that should be the bigger fish to fry — in terms of the volume of sales and the vulnerability of those who are directly affected by this. Undergraduate and graduate students who have to pay out of pocket for an overpriced textbook are more worthy of our attention than the battle between university libraries and publishers about journal subscription packages. Because the burden to pay for overprice textbooks rests more directly on individual students, who have much less bargaining power than institutions, this should be the focus — if we are to criticize academic publishers.
Ultimately, I do agree that the profit-driven model of scientific publishing is exploitative, but that’s because all profit is exploitative. Science — the journal where I work at, is one of the few non-profit publishers. If you compare our revenue (~$100mil/yr) to the giants of the industry, e.g. Nature-Springer (~$2bil/yr), or Elsevier (~$10bil/yr) — it’s not even close! And they probably have a higher year-over-year growth as well, because capitalism naturally favors for-profit businesses. Do I think if it will be good if we force all publishers to become non-profits? I don’t know. But that’s no longer a critique on scientific publication, but a critique on profit itself, and you can just read Marx instead of my blog.
Epilogue: the publish or perish grind
This is the same as the “campaign or lose” problem for politicians. Researchers have to prove to their funders that their dollars aren’t being wasted, just as politicians have to prove to their constituents that their votes aren’t being wasted.
So, how do you communicate that information to the funders? The funders do not have the specialized expertise to assess the value of your research. So, how do you convey the message that “my research is valuable according to experts in this field”?
That’s a practical problem presented by how research funding is structured, that journals have become the de facto kingmaker. Just like how politicians are elected based on their popularity and not necessarily their ability to govern, scientists are funded based on their ability to publish and not necessarily their ability to solve problems. I do not have a solution for closing this gap. But I do think the publish or perish culture is a greater source of stress and distraction for researchers (taking up maybe 10% of their time), more so than the money they have to pay to read and publish papers (costing maybe 1% of their funding).
