To preface this post, I have no problem with genuine evidence-driven experiments in publishing or with ideas I don’t necessarily agree are the best route forward. However, I think it is a requirement on those of us who advocate for change to do so responsibly and to also raise concerns when others do not. I have given talks in support of the PRC model but increasingly believe that it is being forced into a negative direction and risks damaging the wider preprint movement. My issue is two fold; firstly that the proponents of the model are forcing a personal agenda on the wider space and secondly that those proponents are ignoring evidence and engaging in extremely questionable behaviours that range from exclusion to deliberately misleading claims.
What is the PRC model?
The PRC model (Publish-Review-Curate) is one of a number of different publishing models being proposed. It suggests breaking publishing down into three parts:
Post articles as preprints or in open-access repositories.
Peer review occurs on the preprint or article in an open-access repository. Ideally, the review is linked to the article.
Articles are “curated” which could be by journals, peer review services, individuals or other means.
The model is driven by a number of key actors in the open science space, namely; HHMI, eLife and preprint peer review services who have formed a “PRC Alliance”. A range of funders are also supporting this model, particularly those who prefer to mandate change rather than work with researchers more broadly.
Issues with how to PRC model is portrayed
Proponents incorrectly represent the model as being more widespread than it is
This is one of the most widely used mis-representations of the PRC model in the communications that I’ve seen so far.
This is a mis-representation because what proponents are usually referring to is the adoption of preprints, not the adoption of the PRC model. Indeed, the model itself is not at all widespread.
Our own recent report shows that preprint peer review remains extremely limited (approximately 2% of preprints and preprints are themselves only ~10% of the literature to begin with). The PRC model is not “just” preprints and so it’s incredibly misleading to take credit for the adoption of preprinting so far, which has occurred very successfully and independently of the PRC model.
Proponents mis-represent the models support for preprints
PRC proponents regularly represent the model as being founded on preprints. However, the definition of the model itself states that “publishing” is not only achieved via preprints but also open-access repositories.
A reliance or encouragement of open access repositories, regularly restricted to grant recipients or those working at an institution, would lead to significantly more inequity than a model genuinely based on preprint servers that are open and accessible to anyone. This is perhaps one of the worst aspects that the open science space could advocate for. Even the traditional journal system is more equitable than institutional or funder repositories as that relies on simply having enough funding to be able to pay to play.
Proponents mis-represent the model as being about widening trust signals
The model is very much based on peer review and peer review in the human reviewer sense, not AI mediated feedback. There is no real effort to widen trust signals (at least not beyond peer review) or to consider a different or more future-proofed approach.
There are some example articles from PRC advocates that claim to discuss trust signals yet rest entirely on suggesting that preprints are only trustworthy because they are peer reviewed. This goes against all of the current evidence looking at preprint “quality”. Quite ironically, an expert would not have allowed such an article through a peer review process.
Proponents mis-represent the model as improving research assessment
A system that still rests on curation or where a preprint is reviewed does not improve research assessment. It simply shifts the goalposts. We replace the word “journal” with “curator” or “review platform”. That’s not a real change, it’s just dressing up a wolf in sheep clothing. Much like Diamond OA is never going to shift us away from proxies such as journal names or brands. It is a fact of human behaviour that a reliance and increased focus on where a preprint is reviewed will simply shift journal to preprint review service as the name proxy.
The model does aim to increase the recognition of peer review as a scholarly activity. I fully support this aspect but I do think it’s telling that this is the part most advocated for (PRC conflicts of interest coming up) and done so without also advocating for recognition of things like teaching or other important academic activities.
Proponents consistently fail to discuss the lack of evidence around the model
One of the more unforgivable failures by proponents of the model is their active ignorance towards evidence. Proponents regularly fail to discuss the lack of evidence in support of the model or mis-represent evidence that does exist in support of preprints as being in support of the entire PRC model. There are a range of preprint peer review services that have existed for long enough to collect a decent evidence base. However, this has still not yet materialised, suggesting that either the proponents are not interested in making evidence-driven decisions or that the evidence does not support the model. Some key questions that are easily addressed include:
- How often are peer review reports actually read and how are they used?
- Are preprint peer review services producing materially different reviews to journals?
- What perspectives do researchers hold towards preprint (or transparent) peer reviews? Do these differ from traditional journal-organised peer reviews?
- What perspectives do researchers have around a model that changes preprints into tools of formal assessment
Proponents do not offer detailed explanations as to how the model improves peer review
I’m a strong advocate for transparency, which is the one thing PRC would achieve. However, transparency alone is not sufficient and should not come at the cost of author choice, experiments or real improvements. There is some very limited evidence that transparent peer review is associated with more trustworthy/higher quality papers. However, this is based on authors opting in to a transparent process.
The model does not tackle or discuss the use of AI, particularly as a tool for author feedback or its use in peer review. This is partly what makes the PRC model feel outdated. Furthermore, the PRC model offers no evidence or explanation for how it would solve the “crisis” within peer review where peer reviewers are increasingly difficult to find or are themselves shipping the review work out to AI tools.
Whilst there is nothing inherently wrong with not having a clear plan or answers while experimenting, there are problems with making unsubstantiated claims and promises. There are also issues when the model is being pushed as the next step and dominating discussions/funding that could be used for genuine experiments.
Proponents fail to disclose clear conflicts of interest
The largest supporters of this model are preprint peer review services. In a recent update to the model proposal, Stern & O’Shea outlined that authors should pay preprint peer review services for the reviews, introducing new revenue streams for these services.
Indeed, this represents a clear conflict of interest. These services are reliant on grant funding (including from HHMI) which is not a long-term solution. Therefore, the call to have separate for peer review and curation represents a potentially lucrative business model for these services.
Is this funding worthwhile? Not many services are transparent with their funding, with PREreview being a noticeable example of a service that actually is very transparent in this regard. However, most preprint peer reviews are only accepted by a limited number of journals and we are still yet to see a truly centralised and portable pee review service. Another cost is that of reviewer time. These services all continue to rely on free volunteer labour from academics. In the background of increasing calls to pay peer reviewers, can these preprint review services justify these practices and what happens if they were to pass the cost of paying reviewers onto authors?
Open access publishers are enticed by a model that would allow separate charges for peer review and curation. If publishers bought into this model on a bigger scale and introduced these separate charges this could very quickly become the worst thing to happen to open science since the unregulated introduction of APCs.
A further conflict of interest that is rarely accurately disclosed is that many of the advocates sit on overlapping boards across the PRC space. This may be the prime reason that there has been such cohesion towards this model to begin with and is exceptionally limiting for the wider space as there are few fresh faces or ideas. Moreover, this people use their positions of power to gate keep against ideas and more meaningful change.
There’s a need for a deeper investigation into the money and people involved in this model to more accurately expose conflicts and who stands to personally gain.
Proponents mis-represent the costs and benefits involved
Coming back to costs, one of the arguments from proponents for moving away from journals is the associated costs and profiteering of current publishing business models.
However, the proposal that authors pay for the preprint peer review(s) and then funders or institutions pay for curation leads to double dipping and very likely an overall higher cost than a single APC. To use an analogy here, consider TV and subscription services. It used to be that you would be a single cable charge and receive a large number of TV channels. As streaming services became more abundant that single charge for any one service is significantly less but to get the same overall experience one must pay for multiple services. These smaller but numerous costs add up to significantly more than the single cable charge used to be. This is also likely to result in greater inequity within the system as those who can afford multiple different peer reviews (or from services that may be considered more prestigious) will be the already well funded researchers.
Another key cost is the ability to monitor and regulate. Greater dispersion results in a system that is much more difficult to regulate, monitor or hold to account. Together, this increases vulnerability to bad actors. Just as we witnessed with the OA movement which paved the way for paper mills, predatory journals and more, the PRC model is exceptionally vulnerable to abuse or cooption by bad actors.
As an aside, the financial status of a publisher (for-profit vs non-profit) is another mis-direct. The values of the publisher are much more important. Take AAAS (Science) as an example. AAAS have been one of the largest blockers of positive changes throughout the history of scientific publishing, yet they are a non-profit organisation. They blocked the initial efforts to establish preprints within the Life Sciences in the 1960s.
Proponents mis-represent the model as each step being genuinely independent
The biggest supporters of the PRC model are journals (or curators). The model is very regularly described as separating the functions of journals and yet the largest examples are failing to achieve this (see two examples below). Very few examples of the model in practice actually separate out the peer review and curation functions. This significantly limits what the model is actually achieving; effectively replicating the traditional model of publishing but with greater transparency and potentially more costs. This is the most likely, realistic, implementation of the model; one in which a single service performs the two chargeable functions (review-curate). However we also see some platforms, such as ORE, perform all three steps.
Not only is this not communicated sufficiently when discussing the model, potentially misleading people, but the current evidence clearly demonstrates that the primary implementations of the PRC model are built around journals. Not only that but the model maintains a non-discussed (inconvenient) curation step before the the publishing/review step in that the journal editors/preprint review editors are still selecting which preprints they will review.
Example implementations of the PRC model
Open Research Europe
Open Research Europe (ORE) is the publishing platform for recipients of Horizon Europe grants. It presents itself as a PRC platform.
ORE does not use editors to make binary accept/reject decisions. Instead, once your manuscript passes basic checks it is accepted and then cannot be removed from the ORE system. At the point of submission you are agreeing to publish in ORE and cannot then submit your work to other journals.
Once the submission checks are complete, the work is posted as a preprint “awaiting peer review”. Outside of ORE, preprints do not need to be peer reviewed, making the use of the term here slightly misleading. Effectively, this is early posting of work that is already under peer review – similar to the “In review” label on Research Square. Whilst this does help share work earlier, it’s much more restrictive than posting a preprint. Indeed, in comparison to the current publishing system, ORE is actually more restrictive on authors. It is incumbent upon the open science community to ask if transparency is worth sacrificing author choice.
This clearly demonstrates that PRC does not require separate organisations and can be an even more restrictive environment than current publishing practices. This also shows that the PRC model can easily skirt around preprints; potentially impacting the adoption of preprinting.
eLife
In many ways, eLife is held up as the poster child of the PRC model. The switch from a traditional “glam” journal to a journal that only reviews preprints has not been particularly smooth. In the process, eLife lost its impact factor* and a number of staff have been made redundant in the last couple of years. Despite this, it should be recognised that eLife took a bold step in pursuit of changing a system and leading the charge is never easy or straightforward.
However, in the context of PRC, eLife still effectively functions as any other journal in that it peer reviews and curates as a single step. One key difference is that once a preprint is accepted for peer review, eLife commits to publishing that preprint with the transparent peer review report.
This again highlights that the implementation of the PRC model is more likely to be extremely similar to existing journals. If the PRC advocates achieve their business model goals then in the future these single “RC” services may begin charging for peer review and then charging for curation.
*eLife have long argued that the impact factor is flawed and they didn’t want one, so this is included for accuracy rather than a statement on the impact factor itself.
Conclusion
Ultimately, the PRC model does not appear to be a genuine attempt to re-imagine publishing. Instead it is an effort focussed on creating a financial business model for preprint peer review services and diamond OA journals. As an experiment, that should proceed (with greater transparency and honesty) but PRC is not currently being positioned as an experiment. Instead it is being forced as the next direction for preprints by a very small minority, not all of whom have the expertise or experience to dictate such directions to the life sciences.
The, at times, active misleading communications and advocacy of the PRC effort/alliance is highly concerning and demands greater accountability from the open science space. Unfortunately, given the significant overlap of the same faces across open science efforts, this is not going to happen.