分析八万篇审稿意见，发现审稿语气受作者性别、族裔等影响，匿名未必更公平。
Disparities in Peer Review Tone and the Role of Reviewer Anonymity
- 通过大规模语言分析，检测审稿语调与作者身份的关系。
- 匿名审稿时，女性和少数族裔作者获得更负面评价。
- 揭示语言偏见机制，适合关注学术公平的研究者阅读。
同行评审常被视为科学诚信的守门人，但越来越多证据表明其存在偏见。尽管结构性不平等备受讨论，但语言如何隐性加剧差异仍较少被关注。本研究对超过8万份来自两大期刊的审稿意见进行迄今最全面的语言分析，运用自然语言处理与大规模统计建模，揭示了审稿语气、情感倾向及支持性语言在作者性别、种族和机构背景上的差异。结合包含匿名与署名审稿的数据集，研究还发现披露审稿人身份会改变评价语言。结果不仅暴露了审稿反馈中的隐蔽偏见，也挑战了匿名性有助于公平的传统假设。随着学术出版面临改革，这些发现促使人们重新思考评审政策如何影响职业发展与科学进步。
原文摘要 · Abstract (English)
The peer review process is often regarded as the gatekeeper of scientific integrity, yet increasing evidence suggests that it is not immune to bias. Although structural inequities in peer review have been widely debated, much less attention has been paid to the subtle ways in which language itself may reinforce disparities. This study undertakes one of the most comprehensive linguistic analyses of peer review to date, examining more than 80,000 reviews in two major journals. Using natural language processing and large-scale statistical modeling, it uncovers how review tone, sentiment, and supportive language vary across author demographics, including gender, race, and institutional affiliation. Using a data set that includes both anonymous and signed reviews, this research also reveals how the disclosure of reviewer identity shapes the language of evaluation. The findings not only expose hidden biases in peer feedback, but also challenge conventional assumptions about anonymity's role in fairness. As academic publishing grapples with reform, these insights raise critical questions about how review policies shape career trajectories and scientific progress.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。