In October 2023, shortly after the war in Gaza began, scientist Michael Eisen shared an article from the satirical news website, The Onion, on Twitter. Entitled “Dying Gazans Criticized For Not Using Last Words To Condemn Hamas,” Eisen retweeted and added: “The Onion speaks with more courage, insight and moral clarity than the leaders of every academic institution put together.”
At the time, Eisen was editor-in-chief of an open-access science journal called eLife. Ten days later, he was fired. After five of the journal’s editors resigned in protest, eLife’s board of directors released a statement:
Mike has been given clear feedback from the board that his approach to leadership, communication and social media has at key times been detrimental to the cohesion of the community we are trying to build … It is against this background that a further incidence of this behavior has contributed to the board’s decision.
This happened more than two years ago. At the time, it received extensive media coverage. Some news articles celebrated the decision, while others quoted sources who criticized the journal for their “irrational attack” on Eisen’s freedom of speech. But never explained — at least not by the mainstream press — were details about the precise events and academic in-fighting that preceded Eisen’s ousting.
Eisen was not fired because of a tweet, says Prachee Avasthi, who served on eLife’s board of directors. Rather, tensions had been mounting for months between eLife’s leadership team and its editors and readers. The journal had spent years pushing the boundaries of both publishing and peer review. eLife first required authors to publish preprints before submitting to the journal, and then they got rid of accept-reject decisions entirely. Eisen increasingly found his decisions at odds with the norms of the scientific community he was trying to reform. So when Eisen sent out his tweet, says Avasthi, the board just had a convenient excuse to get rid of him.
The whole story is quite strange, especially given that the people involved — not only Eisen and Avasthi, but also former editors at the journal — still regularly cross paths in San Francisco’s open-science community. Even as eLife fractured, the same ideas and people reassembled elsewhere, carrying forward pieces of the original reform efforts. Conversely, the journal has quietly retreated from parts of the vision Eisen laid out for it.
Origins of eLife
eLife was designed as an experiment in removing gatekeepers from scientific publishing.
Founded in 2012, eLife quickly joined the ranks of “prestige” journals because three famous research charities — the Howard Hughes Medical Institute, the Max Planck Society, and the Wellcome Trust — agreed to fund a large portion of its operations. From the start, this support meant that eLife didn’t have to worry about the financial pressures that often plague academic journals, which otherwise rely on large subscription and article-processing fees for survival.
The journal’s first editor-in-chief was Nobel Laureate Randy Schekman. In 2013, Schekman criticized Nature, Science, and Cell as “luxury journals,” comparing their low acceptance levels with high-end “fashion designers” who deliberately inflate demand due to perceived scarcity. Just 8 percent of papers submitted to Nature are eventually accepted.1 (eLife’s acceptance rate in 2025, for comparison, was 15.4 percent.)
Under Schekman, the journal began implementing reforms to remove gatekeepers and give authors more control over decisions. In mid-2018, eLife began requiring that authors post their manuscripts on preprint servers, such as bioRxiv, before submitting to the journal. The goal, as Eisen explained, was to show that journals could act as reviewers rather than as judges. “The main reason for doing that,” Eisen says, “was to show that publishing wasn’t our job. We were reviewing papers that authors had already published themselves.”
The second major reform at eLife was to do away with accept-reject decisions entirely, thus making editors more like academic collaborators than gatekeepers.
This was a big change, at least compared to the conventional publishing model. When scientists submit a paper to Nature, say, it is first assigned to an in-house editor, who decides whether a submission meets the journal’s standards. According to Nature’s website:
The criteria for a paper to be sent for peer-review are that the results seem novel, arresting (illuminating, unexpected or surprising), and that the work described has both immediate and far-reaching implications.
These criteria don’t leave room for null results, incentivizing authors to overstate the merits of their work. If the editor thinks that a paper fails to meet this bar, they can unilaterally reject it. If the editor thinks that it does meet this standard — and there is usually a bit of politicking involved — then the paper is sent to two or three peer reviewers.
The reviewers usually take 2-3 weeks to read the paper and write feedback. Their reports return to the editor, who decides whether the manuscript should be rejected, revised, or accepted. Most papers go through at least two rounds of review. The full process can take months or years.
Reviewers may give differing opinions about a paper, too. Reviewer #1 might ask the scientists to repeat an experiment, while Reviewer #2 commends it and tells the editor to accept the paper as-is. This can get quite confusing, of course. But at most journals, the editor compiles and returns all of the reviewers’ feedback — conflicting or not — to the authors, who must then decide whether to revise their manuscript or send it elsewhere. The editors usually wield final authority over which studies appear in the journal.
But since its inception, eLife — seeking to improve this peer review process — had adopted something called “consultative peer review,” meaning reviewers and editors talked to each other before sending back a single set of non-conflicting comments to the authors. Unlike most other journals, in a show of transparency, eLife also openly published the decision letters and reviewer reports for all accepted articles. (In a 2016 survey, 95 percent of reviewers said that “that the consultation process at eLife adds value for authors.”)
In mid-2018, under Schekman’s tenure, eLife launched the Triage Trial, an experiment that removed accept-reject decisions for roughly 300 papers. Editors and reviewers still gave authors feedback, but eLife then published those comments whether or not the authors revised the manuscript or resubmitted it to a different journal. In other words, eLife became a peer review platform, rather than a typical publisher with yes-no decisions. “If we publish reviews for all papers, then why do we need an accept or reject decision at all?” Eisen says. “There was no good argument for only publishing the reviews of accepted papers.”
Each of these decisions — from the shift in peer review, to requiring that authors post their studies on preprint servers, to eliminating accept-reject decisions — moved eLife closer to its ultimate aim of putting authors, rather than editors, in control of publishing.
Schekman resigned from his position in early 2019, and was replaced that February by Michael Eisen, a geneticist at UC Berkeley. Eisen had previously founded the first open-access journal, the Public Library of Science (PLOS), in 2001. The board interviewed five candidates, says Avasthi, and the finalists were Eisen and a “more moderate” choice. But Eisen received unanimous support among the board’s selection committee.
At the same time, everybody at eLife knew that Eisen was not a politically neutral choice. Eisen has urged scientists to access papers via SciHub, “a shadow library that provides free access to millions of research papers, regardless of copyright,” according to Wikipedia. (Many scientists do this anyway, but it’s kind of a “you’re not supposed to talk about it” situation within academic circles). In 2018, Eisen also ran for a U.S. Senate seat in California, tweeting an image of Donald Trump with the caption, “What a fucking asshole.”
In October 2022, based on results from the Triage Trial, eLife — now under Eisen — announced that they would scrap accept-reject decisions across the entire journal. From now on, editors would only screen submissions for serious flaws, and then pass them to outside experts for review. These reviewers would write a public report and, with the editor, an accessible description of what the study showed. eLife would then post both the paper and these reviews — even scathing ones — online as a “Reviewed Preprint.” Authors could then revise and repost their paper through eLife or resubmit elsewhere.
“By relinquishing the traditional journal role of gatekeeper and focusing instead on producing public peer reviews and assessments, eLife is restoring control of publishing to authors,” Eisen wrote in a public letter. Avasthi, a member of eLife’s board of directors, was fully supportive of Eisen’s vision. Rejecting papers often felt arbitrary, she says, since those same studies would later appear in other journals anyway.2 “So why reject it? We realized these processes just slow things down and remove author agency.”3 (For a 2012 paper, researchers found that about 75 percent of all submitted papers are published in their first-choice journal. Of those papers that do get rejected, a majority are later published elsewhere.)
At the same time, eLife changed its revenue model. Previously, authors paid $3,000 only if their paper was accepted. Most other journals also charge authors only upon acceptance, meaning that the more selective they are, the more money they lose reviewing papers. Instead, eLife began charging a flat fee of $2,000 for all submissions. Critics saw this as opportunistic, but it directly tied the fee to a service: namely, peer review.
With each change, eLife moved closer to Eisen and Avasthi’s ultimate goal of “a world where journals might not even need to exist.” Scientists would get to choose when and how to share their work, but without being able to hide criticisms by quietly re-submitting rejected work to other journals. The scientific community would then organize around studies or reject them conceptually, based on these public reviews, rather than promoting studies merely because they appeared in a “prestigious” journal.
But then tensions started mounting.
The Gatekeepers’ Demise
For a while, eLife appeared to navigate this publishing shift successfully. Initial reactions were positive, submissions were steady, and the leadership team felt optimistic that other journals might follow their example.
But then, in early 2023, with eLife poised to fully implement their new policies, a group of prominent editors — including Schekman — began voicing concerns, according to reporting in Nature. In private letters, nearly 30 senior editors threatened resignation, arguing that removing the accept-reject decision would undermine the journal’s prestige and compromise peer review standards. The board also began to question their ideas, then quietly postponed their scrapping of accept-reject decisions.
Stephen Heard, an evolutionary biologist and writer at the University of New Brunswick, was one of eLife’s most vocal critics. Heard argued that eLife’s policy wasn’t even particularly radical; the journal still charges fees, screens submissions, and “is mostly a journal — just one with a 0 percent rejection rate for manuscripts that make it to the peer-review stage.”
Heard also suggested that publishing “unrevised” papers would shift the function of scientific quality-control onto readers. By removing power from editors, readers would need to evaluate the merits of papers themselves. This could be problematic, he claimed, because not all readers have the time or expertise needed to judge a paper’s quality.
Mark Hanson, professor at the University of Exeter Penryn, applauded eLife’s courage but thought the move was “bad for the health of science … a push towards the death of expertise.” In his view, eLife hadn’t killed gatekeeping, but had instead swapped “hard” editorial power for “soft” influence:
Before, editors gave a binary accept/reject. Now they give an implicit accept/reject. I doubt authors will actually publish articles as final version of record [sic] if they’re totally trashed in the editor statement. But it frees up authors in that grey area to publish anyways, even if the editor and reviewers aren’t fully endorsing the article …
eLife remained publicly supportive of Eisen. In an editorial, the board and a few editors urged researchers to give the model a chance. “It is a very exciting time at eLife as we try to push the frontiers of publishing and navigate the challenges along the way,” wrote Deputy Editor, Tim Behrens.
But scientists continued to worry about submitting manuscripts to the journal. They wondered if papers published there would “count” for their career. If eLife was no longer accepting or rejecting articles, would universities and hiring committees treat them in the same way as other published articles? There was concern that papers in eLife would be viewed as “lesser than” articles published in standard journals.
In March 2023, twenty-nine eLife editors and Schekman, the former editor-in-chief, wrote a letter to the executive editor of eLife’s non-profit owner, Damian Pattinson, urging him to replace Eisen “immediately,” according to reporting in Nature:
They added that they had no confidence in Eisen’s leadership because he had dismissed their concerns and had not considered compromise positions. One of the journal’s five deputy editors had already stepped down from that leadership position, and ‘significant numbers’ of reviewers and senior editors were ‘standing ready to resign.’ they wrote.
The eLife board published an open letter in response, reiterating support for the new model. But despite this public show of support, members of the board privately worried.
In group chats and emails, Avasthi said, several board members were discussing the blowback from scientists “on a daily basis” and fielding “daily complaints from powerful scientists.” Avasthi, who was then Chair of the board, felt like her colleagues were not doing enough to support Eisen, especially given the fact he was merely championing decisions that the board, themselves, had already made. In late March, Avasthi resigned from her position. A few months later, Eisen wrote the infamous tweet that got him sacked. “I think the truth is that they had grown sick of me,” he says:
[The board] didn’t want to deal with the reality of what [reform] actually looks like, which is that people were going to get upset. People were going to say, ‘I am never publishing in eLife anymore.’ People were going to accuse us of ruining science. People were going to attack us. So the practical consequences of trying to actually do something, as opposed to pretending to do something, is not easy and it’s not pleasant.
End of the Impact Factor
After Eisen’s firing, eLife tapped its two deputy editors, plant biologist Detlef Weigel and neuroscientist Tim Behrens, to run the journal through 2024 while the board looked for a permanent leader. Weigel and Behrens’ first task was crisis control: reassuring skittish editors, calming authors, and keeping the accept-reject system alive.
In October 2024, however, a company called Clarivate announced that eLife would lose its Impact Factor, a metric invented in the 1960s to help librarians select subscriptions from a growing number of scientific journals. In the last couple decades, universities have increasingly treated Impact Factors as a shorthand for prestige. Most scientists call the metric “stupid” or “arbitrary,” yet still try to publish in journals with high numbers because university hiring committees — as well as grant reviewers at major institutes — care about it.
An Impact Factor is calculated by taking the number of citations on papers published by the journal in a given year, divided by the total number of papers published. If Nature published 100 papers that collectively netted 2,000 citations in a year, for example, then its Impact Factor would be 20. In 2023, eLife’s impact factor was 6.4.
Clarivate’s decision to remove the journal’s Impact Factor hinged on a simple concept: If the journal was not making accept-reject decisions, then there was no way to tag a paper as officially being “in” the journal, and so Clarivate could not fairly calculate the metric. Even before Clarivate made this decision public, the company had flagged eLife’s listing as “On Hold;” university librarians also noticed that some eLife papers had quietly disappeared from Web of Science, a paper indexing platform created by Clarivate. Authors began to worry that their eLife papers would become hard to find and, therefore, difficult to cite.
In response, the board made a partial compromise: The journal would send papers above a certain quality threshold (submissions designated as “solid” or higher in the eLife editorial assessments) to be indexed in Clarivate’s Web of Science. But publicly, the journal protested, arguing that Clarivate’s move would stifle “attempts to show how publishing and peer review can be improved using open-science principles.” Critics, including Avasthi, viewed eLife’s decision as a capitulation; as a way of bringing back accept-reject decisions, albeit in a different form.
By early 2025, Behrens — now permanent Editor-in-Chief — insisted eLife would press on even if it lost its Impact Factor entirely. “We want to prove you can succeed without that number,” Behrens wrote to staff. But submissions to the journal have dipped significantly, especially from scientists in geographic regions where metrics are a deeply entrenched part of academic evaluation, such as in China and the United States. (Submissions from Europe dipped only slightly, according to Behrens.) eLife fully lost its impact factor in June 2025.
Even so, eLife refused to return to the old system. “Our job is to take the risks now so other journals can copy what works, with much lower risk,” says Behrens. “Any experiment that plays by different rules from Clarivate will hit the same breakpoint” as his journal experienced.
Instead, the journal would lobby universities, funders, and scientific societies to state publicly that eLife papers (and papers published in other journals that follow their model) will count just as much as Impact Factor papers. On 8 May 2025, eLife publicly stated on their website:
We’ve spoken to funders and institutions around the world and found that more than 100 (over 95% of respondents) still consider eLife papers in research evaluation despite eLife’s exclusion from the journal Impact Factor.
“We’ve conflated sharing science with judging scientists,” says Behrens. “Decoupling them is the whole point.”
Conclusion
Despite the upheaval, eLife remains a great journal. Its decision to get rid of accept-reject decisions did not condemn it to obscurity. The journal still publishes thousands of articles yearly and is widely respected amongst scientists. Articles average a speedy 95 days between submission and publication.
But still, as the journal decided first to require preprints and then to remove power from its editors, scientists wondered: Why not just do these experiments at another journal? Why “tank” the prestige of eLife itself?
The reason, says Eisen, is that the journal carried weight; it was considered prestigious. “The only way to address the reliance on a journal’s brand as a proxy for quality was to take a recognized signifier and destroy that signification,” he says. “We wanted to remove reliance on that brand.”
A spin-off journal, lacking the same measure of perceived prestige, would be ignored by the scientific community. The whole point was to run this experiment in a big journal, in other words, so that people would pay attention.
eLife’s decision to remove accept-reject decisions, and the backlash that ensued, also reveals the difficulties with metascience reforms as a whole. It shows how everything in academia is intertwined. When one pulls a single thread, like accept-reject decisions, the whole fabric of how we do science begins unraveling, and one sees how tightly a journal’s “prestige” is linked to hiring decisions and grant funding. Scientists may cheer reform in principle, but big actions usually fail because nobody in the collective wants to go first.
The story of eLife, then, is a test case for the entrenched incentives in science, many of which are governed by anointed “overseers.” Isn’t it bizarre that a for-profit company, Clarivate, is able to set and control a metric that has become so critical for tenure, grants, and publishing? Isn’t it odd that editors wield such extraordinary control over which papers get accepted or rejected in their journal?
Scientists who publish in Nature, Cell, or Science are far more likely to win big grants, given the perceived prestige of those journals. In many ways, then, a scientist’s success stems from their relationships with editors at those journals. “The editors of Nature have more influence over funding than the head of the NIH,” says Behrens. “That’s absurd.”
And then there are the papers themselves, which aren’t a logical way to convey science in the first place. The work of a laboratory isn’t “done” when a paper goes out and gets published. Academic papers present experimental results as a linear, neatly-packaged “story,” when in reality, science does not resemble this in practice. A better way to publish would be to keep notebooks which get updated in real-time and clearly explain both the successes and failures of a given project.
“If you ask me what’s the absolute worst thing about the current system of publication,” says Eisen, “it’s that we decide, at some point, that we have all the information that we need in order to show others that [our science is] valid and important.”
Still, the culture of scientific publishing is slowly shifting. Thousands of researchers have posted preprints and written public reviews for eLife. The journal is working hard to scale-up their efforts. Behrens plans to give away eLife’s software, data pipelines, and even financial information so other journals can try out the same model.
At the end of the day, the story of eLife is not the story of Eisen, or of his firing, or of free speech. It’s about what happens to those who try to change the incentive structures of science. eLife itself is just a journal — “one journal of thousands,” as Avasthi says — in a sea of other journals. Its rise, fall, and continued existence is arbitrary, as is so much else about how we do science.
- These numbers are misleading. Authors who “know” their paper won’t make it into Nature usually don’t submit it at all, because it wastes time. Instead, they often send it to a journal with a higher acceptance rate or a journal more suited to their particular field, like Nature Microbiology or Nature Biotechnology. ↩︎
- eLife actually collected data on this to verify the claim; so it’s neither speculative nor anecdotal. ↩︎
- This often has the effect of hiding the feedback that resulted in the rejection anyway, since that may or may not surface upon review elsewhere. Authors could ignore that feedback and publish the work as-is elsewhere, but readers suffer by not having reviewer concerns revealed. ↩︎
