
[Update 9/4: At request from Clarivate, I have corrected a reference to the organization’s treatment of eLife, which is partially indexed in Web of Science’s “Emerging Sources Citation Index.”]
As Peer Review Week nears (September 15-19), I’ve been reflecting on this process that has become so ingrained in the contemporary production and communication of scholarship. What follows are some of my meandering thoughts on the current state of peer review, about which I certainly do not claim to be the foremost expert. Many smart people have been wrestling with peer review for quite a while; I’m just another librarian-blogger weighing in.
Issues
Peer review is a critical component of today’s scholarly publishing process. The validation of information via peer review is perhaps the main value scholarly publishers purport to provide (in addition to infrastructure, like persistent identifiers and metadata). In an age of rampant mis- and disinformation, when most research is published in countries whose democracies and press freedoms are under threat, a process that validates information credibility is undeniably and incredibly valuable—when it works.
But sometimes it doesn’t work as well as we’d like. As of August 26, 2025, Retraction Watch’s database contains records for over 66,000 articles that have been retracted or marked as concerning since they began tracking in 2017. Over 43,000 of those articles were published in the last 10 years, between 2014-2025.
We can’t lay blame at the feet of individual peer reviewers for a problem so systemic and complex and persistent. A (literally) large part of the issue is the incomprehensible amount of scholarship flowing through the system: in 2022, just 10 of the largest publishers produced around 2.82 million journal articles. Retraction Watch reports 9,973 retracted or markedly concerning articles by those same publishers in 2022[1], meaning the rate was around 0.35%, or about 1 in every 282 articles. Even if that might otherwise seem like a permissibly low error rate, the sheer quantity of submissions in the system still meant nearly 10,000 deeply flawed or fraudulent articles made it through peer review in 2022, many of which appeared in journals owned by well-known and established publishers. That is a startling number and a direct consequence of the immense strain on the scholarly journal publishing system.
Even while acknowledging those conditions, the blind method of peer review also deserves some scrutiny. Double-blind peer review is the particular flavor I can speak to from experience, during which neither the authors nor the reviewers are privy to each other’s identities (you’ve likely heard of/encountered the oft-maligned Reviewer 2). Review interactions are de-identified and kept private from readers. The main rationale behind this approach to peer review is that anonymity (such that it can be achieved) protects both authors and reviewers from bias, whether positive or negative, and from potential retribution. These are, without doubt, important considerations that should not be dismissed lightly. Many disciplines have relatively small communities of scholars, and particular specialties within disciplines are shared by even smaller groups who might often be called upon as experts to review each other’s work. However, the need to invite niche experts as peer reviewers may result in unavoidable opportunity for bias, as a new report suggests peer reviewers are more likely to recommend studies for publications that cite their own work.
Involving disciplinary experts is necessary because peer review is meant to do more than check citation accuracy and assess methodology and catch evidence of generative AI authorship. It is also meant to help authors address valid questions about their work and communicate new insights more clearly via constructive feedback. This begs the question: is the blind peer review process constructive? Again, I can only speak to my experience participating in a few double-blind peer review processes for journal articles as both author and reviewer. I do think those reviews improved the clarity and quality of the ultimate products, but although I appreciated that improvement as both author and reviewer, the interactions between authors and reviewers didn’t feel particularly constructive. I recall them feeling somewhat rude, combative, and dismissive, despite the manuscripts’ eventual acceptance. This is all anecdotal, of course, and I could have been unlucky (and probably sound whiny). I’m curious to hear about other folks’ experiences. Still, I’ve come away from those interactions further convinced that anonymity does not usually bring out the best in folks, especially when they’re likely overworked. Overworked reviewers might even be tempted to make undisclosed use of Generative AI themselves for a little relief, to which authors are now responding by hiding prompts in their manuscripts to encourage positive reviews. In sum, one of blind peer review’s critical flaws is the lack of reviewer accountability, in both the spirit and integrity of the review.[2]
Another part of the peer reviewers’ role is to recommend the article for acceptance or rejection, so some level of judgement is a necessary part of the system. However, in a system where acceptance/rejection rates are a badge of honor for journals, one starts to wonder whether the current mode of peer review is still founded in the spirit of progressing knowledge, or if it’s been turned into a zero-sum game for publishers’ private gain.
In the same vein, the behind-the-curtains nature of blind peer review enables the devaluation of reviewers’ labor even though it is the core service publishers (or rather, editorial boards) provide to the scholarly communication system. The flip side of public accountability is public recognition, which is also lost in the blind peer review process. Scholars are expected to volunteer their peer review services as part of their contribution to the progress of knowledge in their field, while for-profit publishers just so happen to enjoy the material profits of that uncompensated labor. There are some exceptions to this rule as more journals work with services like ORCID to formally acknowledge peer reviewers. There are also strong opposing arguments, specifically for paying peer reviewers. Regardless, it’s not a big leap to connect the devaluation of peer reviewers’ labor to longer review timelines, difficulty in finding willing peer reviewers, and the increasing number of retracted journal articles whose peer reviewers missed major flaws and/or falsified information.
Libraries
How does all this impact academic libraries? To start, our collection development responsibilities position us as arbiters of information legitimacy. If something is in our collection, we’ve symbolically stamped it with our seal of approval to some degree. Of course, we can’t review and fact check of every piece of content in our collections, and even if we had the time, we wouldn’t have the expertise to do so across disciplines. We rely on “reputable” sources to facilitate this validation of credible scholarship, often via blind peer review. But if the system we’ve relied on to legitimize scholarship isn’t working as well as it could (in many ways), then as information professionals, we necessarily should be interested in potential solutions and alternative approaches. Further, considering our goals to promote information literacy, I’d argue that blind peer review obscures from readers the constructed and contextual nature of authority, the process of information creation, and the layers of conversation that underpin a specific piece of scholarship. Blind peer review is simply not an ideal fit in an information age when transparency and trustworthiness feel so imperiled.
So, are there viable alternatives to blind peer review?
Again, smart folks have been wrestling with peer review for quite a while. Below are a few interesting approaches that have been/are currently being tested. Many of these approaches would fall under the general category of “open peer review,” but they use the term “open” to mean different things about the review process, so I’m going to distinguish them more explicitly, if possible, although they do share an openness as compared to blind peer review.
Small Changes
The most modest adjustment to blind peer review is simply to make both authors and reviewers aware of each other’s identities. Let’s call this “pre-publication open identities” peer review. Although this approach may encourage more humanity between its participants, it leaves other issues (e.g. reviewer recognition, context) unaddressed.
For those who want to retain the current review workflow but don’t think pre-publication open identities review moves the needle much, “transparent” peer review would be the next step to consider. The basic premise is that some amount of the review interaction is included as part of the final published product. This alternative seems palatable to some of the larger publishers (like Wiley and IEEE, for example), as it doesn’t threaten much about their workflow or revenue streams.
A Big Change
To this last point: In a 2024 interview with the Scholarly Kitchen, Nici Pfeiffer, Chief Product Officer at the Center for Open Science (known most for the Open Science Framework), offered this response to a question about the biggest opportunities facing the scholarly communication community:
“The biggest opportunities that lay before us are the separation of peer review from the publication of scholarship, and more assessment tools to evaluate the quality, rigor, reproducibility, robustness, etc. of research outcomes. To realize these opportunities, we will need to embrace new business models for publication, how we reward and incentivize getting it right over getting it published, and putting research evaluation in the hands of the communities that practice it.”
If we were to rethink the publication workflow to achieve a more democratized and community-governed peer review process, what would that look like, exactly?
Both compelling and provocative, one approach that’s gaining traction is “preprint review,” also dubbed “publish-review-curate” by MetaResearch Open Review, one of its adopters (along with eLife, Open Research Europe, and others). In these preprint review/publish-review-curate models, authors post their preprint (i.e. initial draft) in an open repository and provide a link to the relevant review organization (e.g. MetaROR), who facilitates peer review, publishes those reviews and an editorial assessment, provides a platform for public comment from readers, and categorizes the review materials in a disciplinary collection on their site. In MetaROR’s case, the review service does not publish the author’s work, and the author retains copyright and is able to do what they wish with the work at that point, including taking it to a traditional publisher if they so choose. F1000 and eLife serve as both the preprint repository and the review service to consolidate all elements in one location (here’s an example of what this looks like). Some of these review services are free for authors and some are not; some allow reviewers to remain anonymous and some do not. To me, the ideal version would not allow anonymous reviews and would be free to use, although that may be hard to achieve.
The scholarly publishing establishment has not been particularly receptive to the preprint review model, most notably via Clarivate’s decision to temporarily suspend eLife from its Web of Science database and eventually deem it ineligible for an impact factor, because eLife publishes all preprints on its platform that are accepted into the review process, regardless of positive or negative review [this sentence was updated on 9/4 by request from Clarivate, as eLife is now partially indexed in their “Emerging Sources Citation Index .”] In contrast, the European Commission’s state-backed commitment to this model and the Gates Foundation’s requirement of preprint submission both serve as strong evidence of preprints’ potential as a radical alternative to the traditional blind peer review publishing workflow.
To me, the transparency of preprint review seems to encourage accountability, effort, and fairness from reviewers; allows for their labor to be incentivized; enriches scholarship with context that supports information literacy; and promises to expedite the entire review process. As Christopher Steven Marcum notes in his excellent essay on the firehose problem in scholarly communication, preprint review may help “…shift peer review from a monoculture maintained solely by publishers into an entire ecosystem largely maintained by researchers, their institutions, and their funders.” And although it might not solve the overload problem itself, it does begin to incentivize other contributions to the field aside from the publish-or-perish pressures that burden so many academics, and gives their institutions the chance to recognize credited peer review in tenure portfolios.
What’s your review?
If you’ve read this long, thank you, and I’m curious about your thoughts on peer review. Does it need fixing? If so, what pathway do you find most promising? As information professionals, we belong in these conversations. Thanks for your thoughts.
[1] To match the publishers studied by Hanson et al. (2024), I used the following search string in the Publisher field of Retraction Watch’s database, filtered for papers published between 01/01/2022–12/31/2022: [BioMed Central (BMC) OR Elsevier OR Frontiers OR Hindawi OR MDPI OR Nature Publishing Group OR PLoS OR Springer OR Springer – Nature Publishing Group OR Taylor and Francis OR Wiley OR Wiley-Blackwell].
[2] Single-blind peer review is another method, in which the authors are blinded to the reviewers’ identities, but the reviewers are aware of the authors’ identities. To me, this approach manages to both a) not hold reviewers accountable for their work and b) guarantee the chance for bias in the review process. Without any obvious benefits, and I don’t understand why a single-blind approach would be preferable.


