We can shift academic culture through publishing choices

Dear Björn,
Thank you very much for taking the time to read and comment on my article. I really appreciate the time you invested in helping me make the piece better. Please see my responses (marked by >>>) to your comments below.

Many thanks,
Corina

1) In this manuscript, the author proposes a more ethical publishing system compared to the one we have today. The author starts by explaining, in simple and broadly understandable terms the current parasitic relationship between corporate publishers and academia today. She correctly notes that the main driver for these developments on the academic side was "pursuing prestige".
    However, from reading the article, the reader is forced to conclude that the author believes that today, scientists do not strive to pursue prestige any more, as essentially every single suggestion the author puts forth asks scientists to do the opposite of pursuing prestige, by asking them to "publish ethically", regardless of the consequences for their salary, funding or other career aspects.
    Such a text constitutes a laudable appeal to the selflessness of scholars, echoing many similar appeals that have been formulated over the last 20+ years. Clearly, those are lofty ideals, but I have strong reservations as to the mass applicability of such a plan. After all, her predecessors have asked their colleagues exactly the same thing without much tangible effect for the last 20+ years. I doubt that more of the same will drastically alter anything.
    In contrast to what the author (and this reviewer) might hope for, it is highly unlikely that the prestige factor historically and currently dominating publication practices everywhere will disappear tomorrow. Thus, at the very least, this article needs to deal with how scholars either a) may be convinced more effectively to adjust their publication practices against their own self-interest (and if one needs to refer to "changing the incentive structure", please explain how this could be done in realistic steps without brainwashing of about 7 million 'full-time equivalent' researchers) or b) remove the current source of prestige differential: journal rank. Without such an explanation, I see no value in adding this article to the already bulging literature on this topic.

>>> One confusing aspect of version 1 was that I phrased the aim of my paper as changing the incentive structure; however, as Chris Hartgerink pointed out, this wasn’t actually the aim of the paper. I agree because I don’t focus on incentives or prestige in this article. The focus of my paper is to explain how the current publishing landscape works because this is what researchers always find surprising when I give this paper as a talk. After my talks, researchers often approach me saying how they want to change something about their publishing choices. So it seems that researchers don’t generally know much about how publishing works (it’s no wonder, we are so removed from it when we submit papers. And this kind of understanding isn’t taught or mentored unless their mentor is an expert in this topic). For me, if the information in this paper is able to change the publishing choices of even a few researchers, I will consider it a success (it has already influenced at least one researcher: https://twitter.com/van_tho/status/855317905759547392). Indeed, it is only because a colleague questioned me about why I publish where I publish that I started learning how publishing works in the first place, which caused me to change my publishing choices. There is some evidence that researchers at universities with increased awareness of OA policies also have more experience publishing OA (Zhy 2017). This suggests that education about OA has the potential to change behavior. I changed the aim of the paper to reflect its actual aim: My aim here is to explain how the current publishing landscape works and to highlight ethical and exploitative aspects that are not always obvious. I argue that it is in the best interest of researchers, academia, the public, and research rigor to adopt ethical publishing choices. Adopting such choices will instigate a cultural shift in academia.
    Addressing prestige and incentive structures is beyond the scope of this article; however, to elaborate on how individual publishing choices can result in a culture change, I added: academia is made up of individuals, and the values of these individuals create academic culture. If researchers align their publishing choices with ethical publishing practices, then academic culture changes. We can all easily change our values and actions right now.


2) First paragraph and Figure 1:
It makes little sense to compare the subscription costs of a subset of libraries of a single country with the online archiving costs of some file on the internet. There is no relation at all between such two completely arbitrary numbers.

>>> Good point. For a cost per article comparison, I added: 
The global average cost of publishing a paywalled article is $5000 ( van Noorden, 2013). 


3) We do know what the annual cost of publishing scholarly articles is: several sources mention converging ballpark figures just under US$10b. With the number of articles per year at about 2m, we arrive at a current consensus figure of ~US$5k per article the taxpayer is currently paying.  We also know that a whole slew of publishers operate on per-article costs of just under ~US$100 up to ~US$500, which constitute the lower bound of actual per article costs. In other words, anything above ~US$500 requires an explanation (in some cases even costs above US$100). In the case of current subscription publishers, this difference includes (but is not limited to) profit and paywalls. In the case of gold publishers, it is not at all clear where the difference to 100-500 goes. Hence, in Fig. 1A there is a lot missing and in Fig. 1B, it is not at all clear why charging $2900 should not be similarly exploitative as in Fig. 1A.
    I suggest to drop all the numbers in Fig. 1 and just show profit and paywalls in A as excluding scholarship, while whatever costs accrue in B are investments and not lost.

>>> Regarding the difference between the actual cost of publishing an article and the wide range of APCs on the ethical route in Figure 1B, I added: Given that the actual cost of publishing an article is $1.30–318 ( Bogich et al., 2016), it is important to consider where the additional money goes when paying APCs. Some journals charge higher APCs to cover their additional costs, which might involve paying staff for editorial services, promoting the journal, writing news stories, or developing new publishing technology (e.g., see eLife’s cost breakdown at: https://elifesciences.org/elife-news/inside-elife-what-it-costs-publish). For some journals, their higher APCs also provide income for the publisher’s shareholders (see the Exploitative route to publication). If you choose to pay a higher APC, make sure the activities the public’s money will be invested in are aligned with the three ethical principles above.
    I want to keep the numbers in Figure 1 because I think it is important to directly connect researchers with the costs of publishing. If readers can see price tags for the different routes to publishing, it will help connect them with the reality of their choices.


4) Exploitative route:
The author writes: "Publishers obtain the product (the journal article) for free, as well as many of the services involved in the peer review of the product (e.g., volunteer editor and peer reviewer time)."
This wording invites misunderstandings: scholars don't work for free, many if not most of them earn a (in some cases more than decent) wage. They provide their services mostly for the authors, which coincidentally means at no cost for the publishers. This is not to be confused with "free" - it is actually a coincidental subsidy of publishers inasmuch as the scholars' salary is paid out of the public purse.

>>> Thanks for pointing this out. I clarified the sentence to: Publishers pay nothing for the product (the journal article) or the services involved in the peer review of the product 
    In the next sentence, I replaced the word “free” with “publicly-funded” products and services.


5) ”When the paper is published, only individuals at institutions that can afford journal subscriptions can read the research." In principle, this is correct. However, this statement is complicated by, e.g., the fact that some institutions may be able to afford subscriptions, but choose not to subscribe to certain journals and that most publishers offer reduced or even waived subscription fees to developing countries on the IUGG or UNDP lists.

>>> Regarding waived subscription fees to some developing countries, I added: What’s more, publishers breach these agreements by denying previously-promised access (Perez Koehlmoos & Smith 2011).


6) The author also cited an "impact factor" without reference. In the case of Clarivate Analytics' Impact Factor, the author cannot cite the IF as if it were computed rather than, at least in part, negotiated, without clarifying citations.

>>> I added a reference and changed the sentence to: If a generalist journal in the sciences accepts papers from less cited fields, their journal’s Thomson Reuters impact factor would decrease (PLoS Medicine Editors 2006).


7) In this section, the author also neglects the standard acquisition rules in academia (and indeed in the entire public sector!) that acquisitions need to follow a bidding process. Subscriptions these days, especially the "Big Deal" bought by large public institutions, are negotiated behind closed doors, commonly with professional publisher negotiators completely outmaneuvering their hapless librarian counterparts. Any mention of costs should reasonably also mention the way academia pays for them: by breaking or at least bending commonplace rules.

>>> Great point. I added: Additionally, universities breach their standard practice of choosing the most competitive bid: publishers do not compete with each other to obtain university subscriptions on the premise that each publisher’s goods are unique (Eve 2016).


8) Ethical route to publication:
Already in the first paragraph, the author paints a misleading picture, contradicting her own text until this point. Above, the author stated: " researchers have played a key role because they pursue prestige " Indeed: researchers pursue prestige. Even if all journals were OA provided by NP organizations, they would still pursue prestige, all else being equal. Hence, the authors would *not* choose a journal with an "APC they can afford", but with a *prestige* they can afford. This, of course, makes all the difference in the world: if a lab can afford, say, 50k for a Nature article, of course they will pay for it. If a lab cannot, then the authors will have to pay out of their pocket what is required to secure a permanent position. Hence, without eliminating prestige, the injustice and discrimination so rightfully called out by the author above, will simply be transferred from reading to publishing: today, the scholarly poor can't read. In a gold-OA world as described in the article so far, the scholarly poor can't publish (at least not where they get noticed). Given sci-hub et al., the gold-OA route described so far seems even less ethical than the exploitative publishing system where the rich subsidize an obscenely expensive anachronism, such that at least the poorest countries can read and publish for free.
    There remains much work for the author to convince anyone that just because there are no profits and no paywalls, the proposed system will be any fairer.

>>> I agree that incentive structures need to change to address the massive prestige issue. The Smaldino and McElreath (2017) paper is particularly useful for providing the incentive to change incentives because their model indicates that selective journals actually select for bad science. Since prestigious journals are selective, they are more likely to have selected for bad science than a journal that selects papers based on scientific validity and ignores subjectively determined potential impact. A nuanced treatment of the prestige issue is beyond the scope of this opinion piece, but I added: Academia is made up of individuals, and the values of these individuals create academic culture. If researchers align their publishing choices with ethical publishing practices, then academic culture changes. We can all easily change our values and actions right now.
    Regarding discriminating against who can pay to publish, I added: There is a further argument to be made that no money should be exchanged when publishing research products, neither via journal subscriptions nor APCs, because the public has already paid for the research. Any costs that are charged in addition to the initial funding creates inequalities in who can pay to publish or read (Fuchs & Sandoval 2013), and violates ethical principle 1.


9) Table 1:
Likewise, there is little to convince at least this reviewer that all the journals listed here are really that much more ethical than the current corporate parasites. Certainly, the RoySoc journal looks perfect, but the reader doesn't know where the money is coming from and has to trust the name of the publisher in terms of functionalities, such as, e.g. digital long-term preservation, TDM, data and code requirements, and many more. PeerJ (which I support) are a business where we have to trust their founders that they really use our money wisely. eLife is published by the MPG and only publishes a small fraction of submitted articles at a cost prohibitive for most scholarly poor. In terms of reproducibility, we do not have any data, yet, but if eLife can be lumped in with the GlamMagz in this regard, the statistics tell us that eLife will be part of the problem, rather than the solution - and who wants the public to have access to unreliable research? CCBR publishes with a very restrictive license, which can hardly be called "OA" (e.g., no TDM allowed!), PLoS APCs are also much higher than they need to be in case of P1 due to this journal subsidizing their community journals and for the community journals due to their selectivity, which increases unreliability (statistically, on average). Neither ScienceOpen nor Biology Open (nor any of the other journals!) offer competitors to take over their services in case users are not pleased with what they get.
    Thus, in brief, the list in Table 1 looks like a half-hearted attempt at saving a 20th century industry from obsolescence and badly mangling product functionality and market effectiveness as unintended consequences.

>>> It is my intent with this paper to list the most ethical journals that are available right now so researchers can go out tomorrow and change their publishing choices without having to wait for the infrastructure to change. This doesn’t mean that all of the journals in the list are ideal models for how publishing should work, nor does it mean that they will remain ethical (according to the criteria in the paper) in the future if they are, for example, sold to Elsevier. PeerJ is pretty open about where they spend their money (e.g., https://peerj.com/blog/post/115284878682/new-publication-prices-at-peerj/), and eLife provides APC waivers (https://elifesciences.org/articles/21230). I contacted CCBR, bringing it to their attention that they should use a CC-BY licence instead - thanks for pointing that out!
    I agree with you that selective journals are part of the problem (see Smaldino & McElreath 2016), which is why I make it clear which journals select based on subjective impact in Table 1. Since I do not address the prestige issue in this paper, I’m not going to go into the details about the difference between selective and non-selective journals. I lay out the options for people so they can make their own choices based on information about these options.


10) Availability to read by everyone leads to additional benefits:
Actually, PDF is probably among the worst formats for TDM. What would be required is a scholarly mark-up language that can be easily converted into any format the user desires. Just flipping our existing journals to ethical publishers and hoping that the "invisible hand" of the market will then automagically create such scholarly standards will likely not be sufficient.

>>> I agree with you and have been learning about reproducible manuscripts (see Hartgerink 2017 http://onsnetwork.org/chartgerink/2017/03/30/reproducible-manuscripts-are-the-future/). Depending on the publishing platform a publisher uses, it could be very easy for a journal to change from publishing PDFs to reproducible manuscripts. Indeed, eLife is considering doing so in response to a conversation a few of us researchers had with one of their staff members (https://elifesciences.org/labs/cad57bcf/composing-reproducible-manuscripts-using-r-markdown). In contrast, it would likely be difficult for a publisher like the Royal Society to switch because they use ScholarOne, which is inflexible and expensive. If journals are aware that their researchers want to publish reproducible manuscripts, then they can adapt accordingly, but we need to speak directly to the journals to make this happen.


11) Researchers can change the incentive structure by changing publishing choices:
While the author is merely simplistic and/or naive in her approach thus far, this last paragraph borders on wishful thinking. For more than 20 years we have had the possibility to make our work OA at point of publication with just a few clicks and haven't done so: as long as hypercompetition demands that we publish in certain venues, just making people pay won't change a thing. If I'm an early-career researcher and Nature has accepted my manuscript, I will publish there, as long as it carries the prospect of getting a job. In that case (and this is how it still is), this researcher will publish there if it is TA, hybrid or OA, (almost) regardless of cost. In the US (and increasingly in the UK and other countries as well), people go into debt for the prestige of a degree from certain universities. Surely they will go a little more into debt for the prestige of a certain journal? No, scholars are not free to choose where they publish and just making it expensive for them won't change that - other than making the procedure more hateful than it already is.

>>> I removed references to incentives (see my response to comment 1). I have found that by educating researchers about how publishing works and how we exploit ourselves and discriminate against others in the process, many are interested in making changes. I know researchers who only publish OA and in non-glamorous journals and they have gotten excellent jobs (including myself). Additionally, there is growing evidence that publishing OA gives researchers an advantage in their careers (McKiernan et al 2016 https://elifesciences.org/articles/16800).


12) In conclusion, I'm far from convinced that the world the author describes will be neither more ethical nor fairer than today. In fact, from most relevant aspects, it seems it will make things even worse than what we have today, as bad as it currently is. Other than from a historical perspective, the author completely fails to account for the main driver of publication practices: prestige. For her suggestion to actually improve anything, she needs to explain why the prestige factor should completely disintegrate overnight, which seems highly implausible. I hence cannot see anything that this article could possibly contribute to the debate on this topic that hasn't already been said elsewhere, with more competence and persuasion.

>>> Please see my responses to comments 1 and 8.


References
Zhu, Y. (2017). Who support open access publishing? Gender, discipline, seniority and other factors associated with academics’ OA practice. Scientometrics, 111(2), 557-579.