同行评审如何运作（以及如何核查期刊）

官方定义

Peer review is the critical assessment of manuscripts submitted to journals by experts who are usually not part of the editorial staff.

真正做决定的人

A peer-reviewed journal is under no obligation to send submitted manuscripts for review, and under no obligation to follow reviewer recommendations, favorable or negative.

• 在 PDF 上确认条目类型：原创研究、系统综述或方法论文——不是来信、新闻或特邀评论。
• 打开期刊的 “About”、“Instructions for authors” 或 “Editorial policies” 页面，阅读其所称的评审模式（匿名、开放或透明）。
• 问清楚期刊在判断什么。Nature 按兴趣与推进筛选；PLOS ONE 按技术严谨筛选。“同行评审”并不等于同一道门。
• 对生物医学刊名，核对其是否目前被 MEDLINE 收录。MEDLINE 会问同行评审过程是否 “explicit and sufficiently detailed”。
• 对你不认识的开放获取刊名，运行 Think. Check. Submit. 期刊核对表（英文文本最近更新于 2023 年 6 月），不要相信主页上的 “peer-reviewed” 徽章。
• 不要把 arXiv、bioRxiv、medRxiv 或 SSRN 上的预印本当作已经评审。ICMJE §III.D.3.a 要求作者选择会 “Clearly identify preprints as work that is not peer reviewed.” 的存档。
• 若文章公布了同行评审文件，打开它。Nature Portfolio 现在会把该文件与 2025 年 6 月 16 日及之后投稿的原创研究一并发表。

这一期仍由编辑做主

The editor of a journal is ultimately responsible for the selection of all its content, and editorial decisions may be informed by issues unrelated to the quality of a manuscript, such as suitability for the journal.

期刊必须描述这一过程

For this reason and as a service to authors, journals should publish a clear, transparent description of their peer-review process for all types of manuscripts.

最低评审

At least two independent reviewers must review each article, with an exception for arts & humanities journals (for which editorial review must be handled with a minimum of two editors).

谁坐在评审组里

The first level of review is carried out by a Scientific Review Group (SRG), also referred to as a study section, composed primarily of non-federal scientists who have expertise in relevant scientific disciplines and current research areas.

评分标尺

The NIH utilizes a 9-point rating scale (1 = exceptional; 9 = poor) in whole numbers (no decimals).

三个因素，不是五个单独分数

The Simplified Framework for NIH Peer Review Criteria retains the five regulatory criteria (Significance, Investigators, Innovation, Approach, Environment) but reorganizes them into three factors — two receive numerical criterion scores and one is evaluated for sufficiency.

Factor 1 — Importance of the Research:

Significance 与 Innovation，按 1–9 计分。

Factor 2 — Rigor and Feasibility:

Approach（在适用处包括 inclusion 与临床试验时间表），按 1–9 计分。

Factor 3 — Expertise and Resources:

Investigators 与 Environment 标为充分或不充分。缺口需要解释。对研究者的一般名声没有单独的 1–9 分。

内部筛查

All submitted manuscripts are read by the editorial staff. To save time for authors and peer-reviewers, only those papers that seem most likely to meet our editorial criteria are sent for formal review.

几名审稿人

Manuscripts judged to be of potential interest to our readership are sent for formal review, typically to two or three reviewers, but sometimes more if special advice is needed (for example on statistics or a particular technique).

• Accept, with or without editorial revisions
• Invite the authors to revise their manuscript to address specific concerns before a final decision is reached
• Reject, but indicate to the authors that further work might justify a resubmission
• Reject outright, typically on grounds of specialist interest, lack of novelty, insufficient conceptual advance or major technical and/or interpretational problems

不是选票

Editorial decisions are not a matter of counting votes or numerical rank assessments, and we do not always follow the majority recommendation.

审稿人是做什么的

Peer review is designed to select technically valid research of significant interest. Referees are expected to identify flaws, suggest improvements and assess novelty.

审稿人看不见什么

Journal editors do not expect peer review to ferret out cleverly concealed, deliberate deceptions. A peer reviewer can only evaluate what the authors chose to include in the manuscript.

期刊信息

We evaluate research on the basis of scientific validity, strong methodology, and high ethical standards—not perceived significance.

审稿人指南

Unlike many journals which attempt to use the peer review process to determine whether or not an article reaches the level of 'importance' required by a given journal, PLOS One uses peer review to determine whether a paper is technically rigorous and meets the scientific and ethical standard for inclusion in the published scientific record.

这种格式

Registered Reports is a publishing format that emphasizes the importance of the research question and the quality of methodology by conducting peer review prior to data collection. High quality protocols are then provisionally accepted for publication if the authors follow through with the registered methodology.

期刊政策里必须有什么（COS 最低要求）：

同行评审发生在观察结果之前。通过 Stage 1 的稿件获得 in-principle acceptance，不会因为结果令人失望而被撤销——只会因质量失败、偏离方案或无法解决的报告问题而被撤销。

1953 年的来信（真实书目数据）：

Watson, J. D., & Crick, F. H. C. (1953). Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid. Nature, 171(4356), 737–738. https://doi.org/10.1038/171737a0

这个案例教什么：

一篇论文可以是奠基性的，却仍跳过了你的教学大纲现在要求的过程。“发表在 Nature”本身并不是外部同行评审的证据——1953 年不是，今天对 Nature Portfolio 期刊的每一个栏目也不是。核对文章类型和该刊当前政策。

APA 7（version of record）：

Luebber, F., Krach, S., Paulus, F. M., Rademacher, L., & Rahal, R.-M. (2025). Lottery before peer review is associated with increased female representation and reduced estimated economic cost in a German funding line. Nature Communications, 16, Article 9824. https://doi.org/10.1038/s41467-025-65660-9

这篇论文是什么（以免误引）：

关于 Stiftung Innovation in der Hochschullehre 的 Freiraum 资助线的实证报告。它是关于一项拨款分配实验的期刊同行评审。它本身不是 NIH study-section 分数，已发表的评审文件也不是 grant summary statement。

评审文件不会显示什么：

给编辑的保密意见，以及编辑的私下讨论。Nature Portfolio 说明，已发表文件是对过程的一瞥，不是全部决定记录。

Single anonymized

In this type of review, the names of the reviewers are hidden from the author. This is the traditional method of reviewing and is the most common type by far.

Double anonymized

Both the reviewer and the author are anonymous in this model.

2025 年 6 月 16 日起的默认

For submissions received on or after 16th June 2025, this peer review file will be published alongside all our original research articles.

那份文件是什么：

给作者的审稿意见、作者答辩信，以及在某些情况下编辑决定信，作为补充的同行评审文件发布。该日期之前投稿的稿件仍可选择退出。

那份文件不是什么：

决定的完整逐字记录。从另一本 Nature Portfolio 期刊转入时，不会重新发表先前期刊的报告。

标题

Is the journal clear about the type of peer review it uses?

速度测试

The journal doesn’t guarantee acceptance or a very short peer review time.

同一核对表上的后续问题：

Does the website mention whether the process involves independent/external reviewers, how many reviewers per paper? Is the publisher offering a review by an expert editorial board or by researchers in your subject area?

顾问寻找什么

Is the peer review process explicit and sufficiently detailed, including information on the type of peer review and the number of reviewers typically assigned to a manuscript?

“predatory” 或 “pseudo-journals” 做什么

These journals (“predatory” or “pseudo-journals”) accept and publish almost all submissions and charge article processing (or publication) fees, often informing authors about this after a paper’s acceptance for publication. They often claim to perform peer review but do not and may purposefully use names similar to well-established journals.

限度

Peer-reviewed work isn’t necessarily correct or conclusive, but it does meet the standards of science.

影响因子

The journal impact factor is widely misused as a proxy for research and journal quality and as a measure of the importance of specific research projects or the merits of individual researchers, including their suitability for hiring, promotion, tenure, prizes, or research funding.

ICMJE §II.C.3

Reviewers must request permission from the journal prior to using AI technology to facilitate their review.

ICMJE 论人工智能输出

Reviewers should be aware that AI can generate authoritative-sounding output that can be incorrect, incomplete, or biased.

ICMJE Section V（2026 年 1 月）——保密

Editors, reviewers, or publishers should not upload submitted manuscript into AI systems where confidentiality cannot be assured without authors’ explicit permission.

Nature Portfolio，“AI use by peer reviewers”

If any part of the evaluation of the claims made in the manuscript was in any way supported by an AI tool, we ask peer reviewers to declare the use of such tools transparently in the peer review report.

DOAJ Guide to applying——人工智能评估

Peer reviewers and editors should not use generative AI to create their assessments, due to risks such as breaches of confidentiality, superficial and non-specific feedback, bias, hidden prompts, and false information such as fake references; editing and rewriting may be acceptable if disclosed.