| 曾喜平,俞立平.多属性学术评价值的误差形成机制及对策研究[J].情报工程,2026,12(4):013-024 |
| 多属性学术评价值的误差形成机制及对策研究 |
| Research on Error Checking and Generation Mechanism of Multi-attribute Academic Evaluation Results |
| |
| DOI： |
| 中文关键词: 学术评价；多属性评价；指标体系；评价结果；系统误差 |
| 英文关键词: Co-word Analysis; Social Network Analysis; Medical Subject Headings (MeSH); Theme Structure; Theme Evolution |
| 基金项目:国家社会科学基金项目“学术期刊评价——指标创新与方法研究”（21FTQB016）。 |
| 作者 | 单位 | | 曾喜平 | 1. 上海科学技术职业学院 商贸管理学院 上海 201800
| | 俞立平 | 2. 浙江工商大学 统计与数学学院 杭州 310018 |
|
| 摘要点击次数: 7 |
| 全文下载次数: 6 |
| 中文摘要: |
| [目的/意义]多属性学术评价值的误差影响学术评价质量以及评价结果的运用，从评价结果角度开展的检查和深度分析研究一直较为缺乏。[方法/过程]将误差理论引入到多属性学术评价值的问题分析中，认为多属性学术评价值主要存在系统误差，在对多属性学术评价利益相关方分析的基础上，探讨了多属性学术评价值的误差产生机制，并提出了相应对策。[结果/结论]多属性学术评价值系统误差的产生环节包括评价值不合理、只有排序评价结果、评价值绝对优劣价值判断和比较错误、跨年评价值比较错误、评价结果数据分布与偏倚问题等。评价值的系统误差问题既是评价者的问题也是评价结果应用者问题。通过加强学术评价过程控制、采取定值评价方法、利用 sigmoid 函数处理评价指标或评价结果、慎重进行跨期比较等手段可有效降低系统误差。评价机构、政府与学术界在评价质量控制中应发挥好各自作用。对于多属性学术评价值的质量问题需要采取一套标准的范式进行检查。 |
| 英文摘要: |
| [Objective/Significance] The error of multi-attribute academic evaluation results affects the quality of academic evaluation and the application of evaluation results. Research on inspection and in-depth analysis from the perspective of
evaluation results has been lacking. [Methods/ Processes] Introducing the error theory into the problem analysis of multi-attribute academic evaluation results, it is considered that there are mainly systematic errors in the multi-attribute academic evaluation results. Based on the analysis of the multi-attribute academic evaluation stakeholders, this paper explores the error generation mechanism of multi-attribute academic evaluation results and put forward corresponding countermeasures. [Results/Conclusions] The study found that the generation of systematic errors in multi-attribute academic evaluation results includes unreasonable evaluation values, only ranking evaluation results, judgment and comparison errors of absolute advantages and disadvantages of evaluation values, errors in comparison of evaluation values across years, data distribution and bias of evaluation results, etc. The systematic error of the evaluation value is not only the problem of the evaluator, but also the problem of the applicator of the evaluation results. Systematic errors can be effectively reduced by strengthening the control of the academic evaluation process,
adopting fixed value evaluation methods, using sigmoid function to process evaluation indicators or evaluation results, and prudently making intertemporal comparisons. Evaluation institutions, government and academia should play their respective roles in evaluation quality control. For the quality of multi-attribute academic evaluation values, a set of normative paradigms should be adopted for inspection. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
