What 110,000 Rejected Papers Reveal About Science’s Biggest Journals

Study Suggests Prestige, Not Just Peer Review, Shapes Who Gets Published

Getting a paper published in one of the world’s most selective science journals is supposed to be a pure meritocracy, the best science wins, full stop. But a sweeping new analysis of 110,303 manuscript submissions reveals that an author’s institution and team size carry enormous weight in acceptance rates, weight that dwarfs almost every other measured factor, gender included.

Researchers examined five years of submission data from Science and Science Advances, two of the most influential scientific publications on the planet, covering 2015 through 2020. What they found is a detailed, sometimes uncomfortable picture of how gatekeeping actually works at the top of science publishing, with researchers based in China facing a significant disadvantage at nearly every step, at both journals.

Journals typically keep review data confidential, so information on who submits, who gets reviewed, and who gets published rarely reaches outsiders. The study, published in Science Advances, stands out because the authors released a deidentified version of the dataset publicly.

The Two-Stage Filter at Elite Science Journals

Most people picture peer review as a single process: scientists submit a paper, outside experts weigh in, editors decide. At elite journals like Science, there are actually two stages, each its own filter. Editors, aided by an outside panel called the Board of Reviewing Editors, screen submissions before any formal expert sees the paper. Editors rejected 28.1% of the roughly 68,000 submissions outright, and of those sent to the panel, about 75% were turned down before formal review.

Only then does peer review begin, with acceptance rates of 35.3% at Science and 40.8% at Science Advances, far higher than the rates across the full editorial stage. This matters because the factors shaping a paper’s fate hit hardest at that first door, not the second: associations are generally stronger in editorial review than peer review, suggesting the initial screening is where the most consequential judgments happen.

Prestige and Team Size: The Numbers Behind Publication Success

At Science, acceptance rates varied sharply by institution and team size. The most prestigious quintile of institutions had an acceptance rate of 11.6%, compared to 3.4% for the two least prestigious quintiles, a gap of more than three to one; Science Advances showed a broadly similar pattern, though exact numbers differed. Large teams of ten or more authors were accepted at 9.9%, versus 3.3% for small teams of five or fewer, a gap that may reflect a genuine difference in the work big teams produce, or it may simply be that editors are reluctant to reject a paper backed by dozens of established names.

China-based corresponding authors faced a significant and persistent disadvantage at both journals, accepted at less than a third the rate of authors based in the United States or Canada. Both Chinese affiliation and having a Chinese last name were independently linked to worse outcomes at nearly every step. This pattern may possibly reflect a region-specific pressure to submit to high-impact journals regardless of fit, sometimes linked to financial incentives tied to prestigious publication, though the study does not demonstrate this explains the gap.

Gender Gap: Real, but Smaller Than Expected

One closely watched question is whether women face bias in peer review. The study found a real but modest gap. At Science, papers with women as corresponding authors were accepted at 5.3%, compared to 6.4% for men, a 1.1 percentage point difference, statistically meaningful but far smaller than the gaps tied to prestige or team size.

Reviewer sentiment showed a similar tilt. Both men and women reviewers gave modestly more positive reviews to manuscripts with men as corresponding authors at Science, consistent with reviewers slightly favoring men, though this does not necessarily demonstrate bias; it could reflect unmeasured seniority effects, since senior scientists tend to receive more favorable treatment and are statistically more likely to be men. The pattern appeared at Science but not Science Advances. Researchers estimate that if manuscripts with women corresponding authors had fared as well as those with men, Science could potentially have published 18.4 ± 4.9 additional papers per year with women as corresponding authors, roughly a 12 ± 3% increase, an estimate rather than a confirmed count.

A Paper’s Subject Was Also Closely Linked to Its Chances

What a paper is about plays a large role in whether it gets published. The study grouped submissions into ten broad topics and found acceptance rates at Science ranging from 1.7% to 9.6%, a nearly sixfold difference. Research on viruses, RNA, immune response, and infection fared best; research on politics, economics, gender, and the societal dimensions of COVID fared worst, with disparities at every stage. This could reflect genuine editorial preferences, differences in manuscript fit across fields, or an oversupply of submissions in some areas.

None of this proves deliberate bias drives these patterns. The data reveal disparities but cannot pin down their causes: prestige and team size might partly reflect genuine differences in research quality and resources. Documenting the patterns still matters, the authors argue, the same way documenting pay gaps matters for society before causes are understood. Editors largely follow reviewers’ advice, but papers from prestigious institutions are far more likely to make it through the first door long before reviewers ever see them.

---

Disclaimer: This article summarizes findings from a peer-reviewed study and is intended for general informational purposes. It does not represent an evaluation of any individual journal, editor, reviewer, or researcher, and it should not be read as an accusation of bias or wrongdoing by any party named or referenced in the underlying research.

---

---