图书情报工作 ›› 2017, Vol. 61 ›› Issue (1): 134-152.DOI: 10.13266/j.issn.0252-3116.2017.01.016

• 综述述评 • 上一篇    下一篇

同行评议与文献计量在科研评价中的作用分析比较

万昊1,2, 谭宗颖2, 朱相丽2   

  1. 1. 中国科学院大学 北京 100191;
    2. 中国科学院文献情报中心 北京 100190
  • 收稿日期:2016-09-18 修回日期:2016-12-13 出版日期:2017-01-05 发布日期:2017-01-05
  • 作者简介:万昊(ORCID:0000-0002-4754-7148),博士研究生,E-mail:wanhao@mail.las.ac.cn;谭宗颖(ORCID:0000-0003-3945-7174),研究员,博士生导师;朱相丽(ORCID:0000-0002-4794-4530),副研究员。
  • 基金资助:

    本文系国家自然科学基金项目“力学学科发展态势评估研究:2006-2015”(项目编号:L1524029)研究成果之一。

Comparison of the Role of Peer Review and Bibliometrics in Research Evaluation

Wan Hao1,2, Tan Zongying2, Zhu Xiangli2   

  1. 1. University of Chinese Academy of Sciences, Beijing 100191;
    2. National Science Library, Chinese Academy of Sciences, Beijing 100190
  • Received:2016-09-18 Revised:2016-12-13 Online:2017-01-05 Published:2017-01-05

摘要:

[目的/意义] 科研评价实践中普遍使用的两种评价手段和量度工具是同行评议与计量为基础的评价(文献计量、经济计量甚至替代计量)。为了对比同行评议质量判断与计量为基础的评价结果的一致性,进而回答同行评议能否被新兴的计量评价方法取代。[方法/过程] 本文采用非实证的研究方式,收集前人在论文发表、项目资助、研发评价实践、个人颁奖以及Altmetrics计量评价领域应用两种评价方法对研发绩效测度的120多篇实证研究进行元分析并汇总成定量化综述,基于认知演化的视角进行系统记叙。[结果/结论] 传统同行评议为基础的决策判断始终是研发评价的首要量度工具且地位无法取代,它与引用为基础的计量评价结果在统计上仅存在适度的正相关性,Spearman秩相关系数r∈[0.24,0.88],地位上后者更应充当前者的补充而非替代,即建构在定量辅助基础上的知情同行评议模式不仅使评价结果更加客观透明,也添加额外的附加价值,通过数据来加深对问题的认识,为评价的黄金法则。

关键词: 研发评价, 同行评议, 文献计量, 量度, 相关性

Abstract:

[Purpose/significance] The two most commonly used measuring instruments for scientific assessment are peer review and metric-based scientific evaluation(Bibliometrics, Econometrics and Altmetrics).[Method/process] In order to explore the consistency between the above two measuring "tools" and answer the question about whether the peer review can be replaced by the emerging metric-based evaluation, this paper adopts a non-empirical research approach and collects more than 120 studies on the R&D performance measurements through the two methods at the level of institution, project, paper, and awarding for the quantitative review with meta-analysis, following the evolutionary perspective for systematic narration.[Result/conclusion] Peer review is still the most important tool for scientific assessment always and forever, but the potential of citation-based bibliometrics is huge. Quantitive evaluation should serve as the supplementary rather than substitutionof qualitative peer review judgement with moderateSpearman rank correlationcoefficient r∈[0.24, 0.88]. The complementary combination of the qualitative and quantitative evaluations will make the evaluation more objective and transparent with additional value through data manifestation to better understand the problem,which is called the "Golden Standard"of evaluation.

Key words: R&D evaluation, peer review, bibliometri, cmetric, correlation

中图分类号: