科学网-[权威，观点，汇集] 传统的同行评审（同行评议）：它消极对待真正的创新思想，拒绝领域内具革命性的文章-杨正瓴的博文

（3）在试验获得资助之前收到的同行评审优先级分数与发表时间之间没有显着关联。 There was no significant association between the peer-review priority scores received before the trial was funded and the time to publication,

（5）中国科学报官方账号，2022-09-19，Science发文：论文署名改为诺奖得主，接收率飙升50%

https://baijiahao.baidu.com/s?id=1744348535463678717

https://weibo.com/1729014640/M6lTyBRG5

https://mp.weixin.qq.com/s/PaKuRAF8v4CjRwYmOLTf6Q

https://www.science.org/content/article/reviewers-award-higher-marks-when-paper-s-author-famous

这篇报道引述了一项最新研究结论：如果一篇论文的唯一作者是无名之辈，只有9.9%的评审人愿意接收；同样一篇论文，如果改成由诺贝尔奖得主署名后，被接收率竟飙升至59%。

Reviewers award higher marks when a paper’s author is famous

“Matthew effect” is powerful, unusually large study finds

13 SEP 20225:20 PM ETBYJEFFREY BRAINARD

doi: 10.1126/science.ade8721

https://www.science.org/content/article/reviewers-award-higher-marks-when-paper-s-author-famous

参考资料：

[1] 张晴丹，中国科学报，2022-09-19，Science发文：论文署名改为诺奖得主，接收率飙升50%

https://baijiahao.baidu.com/s?id=1744348535463678717&wfr=spider&for=pc

https://news.sciencenet.cn/htmlnews/2022/9/486444.shtm

[2] Jeffrey Brainard. Reviewers award higher marks when a paper’s author is famous [J]. Science, 2022, 377(6612): 1251

doi: 10.1126/science.ade8721

https://www.science.org/content/article/reviewers-award-higher-marks-when-paper-s-author-famous

[3] 杨正瓴. 同行评议的局限性和改进之策[J]. 科技中国, 2019, (11): 34-36. 日期：2019-11-19

http://www.casted.org.cn/channel/newsinfo/7562

回顾一下人类的科技史，不难发现：人类科技史上的3个“奇迹年Annus mirabilis”，作为现代科学诞生年的1543年(维萨里的《人体构造》和哥白尼《天体运行论》)、1666年(牛顿)和1905年(爱因斯坦)，科学巨人们所创造的科学奇迹，以及构成现代科技体系的大多数重大科技成果，都不是在同行评议下完成的。

以下对应英文的汉语，来自机器翻译。感谢您的批评指正！

附录：

2024-01-25, A New National Purpose: Leading the Biotech Revolution, Tony Blair Institute for Global Change 新的国家目标：引领生物技术革命，托尼·布莱尔全球变化研究所

https://www.institute.global/insights/politics-and-governance/a-new-national-purpose-leading-the-biotech-revolution#introduction

The organisation of UK R&D has changed little in 70 years. It is highly dependent on an academic model whereby principal investigators manage and mentor laboratories of temporary trainees. While the UK is fortunate to have some of the world's historically great universities, as well as many with unique specialisations, globally this model of organisation research has come under increasingly serious strain, as the previous New National Purpose reports highlighted.

The global challenges of this model are increasingly well recognised,[27]Link to footnote and are particularly pressing for biotechnology. There are at least three major and interrelated challenges.

Hypercompetition: The default academic career path is under strain due to the number of trainees exceeding available positions by much more than an order of magnitude. This has created an increasingly pyramidal career structure[28]Link to footnote in which senior academics use limited grants to fund an ever-increasing number of early-career researchers, despite a scarcity of long-term positions or prospects for career development. This has serious negative impacts on research culture, leading to an increasing exodus of the best up-and-coming talent.[29]Link to footnote

Poor talent utilisation: As biotech-related science has grown more complex, individual projects can require a combination of skills and methods from areas as diverse as AI, big data, materials science, genetics, virology, robotics and synthetic biology, with resulting growth of group sizes.[30]Link to footnote It is not possible for a single person, or even a small lab, to possess the required expertise in all of the necessary areas.

AlphaFold’s work in protein-structure prediction, for instance, required a team of individuals with not only cutting-edge machine-learning expertise but also the ability to pair that expertise with a deep understanding of structural biology. Few UK-based organisations and institutions offer training or research positions at the forefront of interdisciplinary work.

Furthermore, UK academia can often be so focused on studying biology as it is found in nature, rather than the design of new biological systems, that there are few people being trained in how to design and evaluate new biological designs at rapid scale. This can make it unattractive to start and scale a company working on the most novel biotechnologies in the UK, as opposed to in the US where the academic talent pool is more well balanced across the continuum of scientific discovery and technological invention.

Globally, these changes are motivating the creation of new types of research laboratory and support that differ markedly from conventional academia, a topic that has recently been well explored by Sam Rodriques.[31]Link to footnote The recently announced UK Research Ventures Catalyst is a step in the right direction but is small in scale.[32]Link to footnote

3. Large and atypical labs are increasingly the only ones able to consistently compete internationally: A prominent structural change in biotech-related academia has been the growth of large groups working for a single professor. We could not find clear data on this, but it appears to be particularly prominent in elite US academia, with individual professors having groups of 50 to 100 trainees. There are also now some examples of such labs in UK universities. These large groups allow pooling of resources to bring a critical mass and concentration of diversely skilled talent, which can be especially important in technology-intensive areas of research such as biotech.

However, such an approach to organising biotech research also fundamentally changes research in a way that cannot be sustainable nor to the collective benefit of science. Such a research environment can become increasingly hierarchical, with a career system to be gamed, incentivising resource acquisition and the exploitation of junior scientists at the expense of pushing the boundaries of human knowledge. It is doubtful, for example, that a professor with almost 100 people in their lab can be meaningfully involved in the mentoring or work underlying the research for which they take credit, collect prizes and hold patents. The system ends up rewarding administrators and empire-builders, not creative scientists actively engaged in research and mentoring. Gone are the days of the early LMB when Fred Sanger did the work for his second Nobel prize with his own hands at the bench.

70年来，英国的研发组织几乎没有变化。它高度依赖于一种学术模式，即主要研究人员管理和指导临时学员的实验室。尽管英国有幸拥有一些世界上历史上伟大的大学，以及许多具有独特专业的大学，但正如之前的《新国家目标》报告所强调的那样，在全球范围内，这种组织研究模式正面临越来越严重的压力。

这种模式的全球挑战越来越得到认可，[27]链接到脚注，对生物技术来说尤其紧迫。至少有三个相互关联的主要挑战。

过度竞争：由于受训人数远远超过可用职位的数量级，默认的学术职业道路正面临压力。这创造了一个日益金字塔形的职业结构[28]链接到脚注，其中高级学者使用有限的资助来资助越来越多的早期职业研究人员，尽管缺乏长期职位或职业发展前景。这对研究文化产生了严重的负面影响，导致最优秀的后起之秀人才越来越多地外流。[29]脚注链接

人才利用率低：随着生物技术相关科学变得更加复杂，单个项目可能需要来自人工智能、大数据、材料科学、遗传学、病毒学、机器人学和合成生物学等不同领域的技能和方法的结合，从而导致团队规模的增长。[30]链接到脚注一个人，甚至一个小实验室，都不可能在所有必要的领域拥有所需的专业知识。

例如，AlphaFold在蛋白质结构预测方面的工作需要一个不仅具有尖端机器学习专业知识，而且能够将这些专业知识与对结构生物学的深刻理解相结合的团队。很少有总部位于英国的组织和机构在跨学科工作的前沿提供培训或研究职位。

此外，英国学术界往往过于专注于研究自然界中的生物学，而不是设计新的生物系统，因此很少有人接受过如何快速设计和评估新生物设计的培训。这可能会使在英国创办和扩大一家从事最新生物技术的公司变得没有吸引力，而在美国，学术人才库在科学发现和技术发明的连续体中更加平衡。

在全球范围内，这些变化正在推动创建与传统学术界截然不同的新型研究实验室和支持，Sam Rodriques最近对此进行了很好的探索。[31]链接到脚注最近宣布的UK Research Ventures Catalyst是朝着正确方向迈出的一步，但规模较小。[32]脚注链接

3.大型和非典型实验室越来越成为唯一能够持续参与国际竞争的实验室：生物技术相关学术界的一个突出结构性变化是为一位教授工作的大型团体的增长。我们无法找到这方面的明确数据，但在美国精英学术界，这一点似乎尤为突出，个别教授有50到100名学员。现在英国大学也有一些这样的实验室的例子。这些大型团体可以汇集资源，带来足够数量和集中的多样化技能人才，这在生物技术等技术密集型研究领域尤为重要。

然而，这种组织生物技术研究的方法也从根本上改变了研究，既不可持续，也不利于科学的集体利益。这样的研究环境可能会变得越来越等级化，职业体系会被玩弄，以牺牲人类知识的边界为代价，激励资源获取和初级科学家的开发。例如，一位实验室里有近100人的教授能否有意义地参与指导或研究工作，并获得荣誉、奖项和专利，这是值得怀疑的。该体系最终奖励的是管理者和帝国缔造者，而不是积极参与研究和指导的创造性科学家。早期LMB的日子已经一去不复返了，弗雷德·桑格在替补席上亲手为他的第二个诺贝尔奖做了工作。

参考文献：

[1] Asher Mullard, special to C&EN, 2025-02-27, ‘Why is it that nobody can reproduce anybody else’s findings?’

https://cen.acs.org/policy/publishing/nobody-reproduce-anybody-elses-findings/103/web/2025/02

[2] Csaba Szabo M.D., Ph.D., D. Sci. Professor, Department Anesthsiology, The University of Texas Medical Branch at Galveston

https://www.utmb.edu/phtox/faculty-staff-students/faculty-bios/csaba-szabo

[3] “Unreliable” by Csaba Szabo – book review and excerpt

https://forbetterscience.com/2025/02/03/unreliable-by-csaba-szabo-book-review-and-excerpt/

[4] sleuth Pubpeer, 2025-03-09, 震惊！90% 科研文献不可重复，20%-30% 竟是编造，科学家新书揭秘惊人真相

https://mp.weixin.qq.com/s/w_GzeqPD2LvcnKsCNZMH8w

[5] 欧亚系统科学研究会， 2024-11-28，一半以上实验结果无法重复，学术界正在遭遇一场危机吗？

https://www.essra.org.cn/view-1000-6460.aspx

[6] 搜狐，2016-06-15，Nature：超过70%的科研人员都有过重复别人试验，并且失败的经历

https://www.sohu.com/a/83510052_183834

[7] 杨正瓴. 同行评议的局限性和改进之策[J]. 科技中国，2019(11): 34-36.

http://www.casted.org.cn/channel/newsinfo/7562

https://www.cnki.com.cn/Article/CJFDTotal-KJZG201911009.htm

https://d.wanfangdata.com.cn/periodical/kjzg201911009

https://lib.cqvip.com/Qikan/Article/Detail?id=7100179990&from=Qikan_Search_Index

[8] 假结果或夸大结果，如何对待和处置科学研究的不可重复结果？

科学网，2018-10-07：

http://news.sciencenet.cn/htmlnews/2018/10/418343.shtm

澎湃新闻：https://www.thepaper.cn/newsDetail_forward_2439849

而研究结果不能重复则甚至已经是今天科学研究中普遍存在的情况。早在2011年，美国弗吉尼亚大学开放科学中心的心理学教授布莱恩·诺赛克（Brian Nosek）等人牵头，招募全球250多名科学家参与，对2008年发表在3家顶级心理学期刊上的100项心理学研究结果进行重复研究，2015年该项重复性研究完成，结果仅有39项研究结果可以再现，61项研究结果无法重现，不可重复率达到61%。

此后，诺赛克等人又再次组织了由多个国家研究人员参与的一次重复性研究，对2010-2015年发表在《自然》和《科学》期刊上的21篇心理学论文的结论进行重复性研究，结果是，38%的结果不可重复。这一重复性研究结果已经在线发表在2018年8月27日的《自然》子刊《自然人类行为》期刊上。

不仅是心理学研究结果有很高的不可重复比率，就连生物医学研究也位居不可重复率的排行榜的高位。拜尔医学的研究人员对发表在世界著名科学期刊上的67个实验项目的数据进行重复研究，结果显示，仅有21%的项目（14个）可重复，高达65%的项目（43个）数据不能重复；另有7%的项目（5个）能重复主要数据，4%的项目（3个）可重复部分数据。

而且，上述研究都是发表在今天影响因子高的主流期刊上，包括《细胞》（Cell）、《自然》（Nature）、《美国科学院院刊》（PNAS）和《科学》（Science），即CNPS。不仅如此，不可重复的科学研究结果几乎遍布于所有学科。一项研究表明，今天在全球一流的学术期刊CNPS发表的论文的结论至少有一半不可重复和检验。

相关链接：

[1] 2025-03-28 22:49，[随感，同行评议] “评价”深度影响未来的“发展”？

https://blog.sciencenet.cn/blog-107667-1479770.html

[2] 2025-03-11 22:45，[打听，同行评议] 全球生物医学科：可能 90% 科研文献不可重复？

https://blog.sciencenet.cn/blog-107667-1477040.html

[3] 2025-03-12 22:42，[从前（8），同行评议] 从苏联 1952年的“科学作品寄存制度”到 2020年的“出版纸质《文明火花淘金》”

https://blog.sciencenet.cn/blog-107667-1477224.html

[4] 2025-03-11 22:45，[打听，同行评议] 全球生物医学科：可能 90% 科研文献不可重复？

https://blog.sciencenet.cn/blog-107667-1477040.html

[5] 2024-12-15 22:49，[趣闻，同行评议，偏见] 稿件的录用率：取决于你的名字

https://blog.sciencenet.cn/blog-107667-1464451.html

[6] 2021-10-15 19:00，[旧闻] 2014年 SCIENCE 杂志：“同行评议根本不能预测研究的成果。这令人非常不安。”