摘要：

【目的】结合大数据相关的理论及技术，利用科技期刊数据构建科研诚信记录指标体系，并提出科研诚信记录体系框架，为推进科研诚信建设提供参考及依据。【方法】 采用文献调研法和比较分析法梳理国内外研究现状，结合大数据相关技术，设计评价指标体系。通过实例分析，验证指标体系的可行性，在此基础上提出科研诚信记录体系框架。【结果】 实例分析中，对5组被通报科研人员和未被通报的科研人员的科研诚信记录数据进行两两对比。卡方检验结果显示，5组科研人员的期刊撤稿、更正、同行质疑、预警期刊发文以及表达关切数与SCI收录论文数的比值等评价指标存在显著差异。【结论】 通过科研诚信记录指标体系中的期刊撤稿数、更正数、同行质疑数、预警期刊发文数以及表达关切数与SCI收录论文数的比值，能够对科研人员的科研诚信状况进行判定，可以将指标体系作为科研诚信记录体系框架构建的依据。

Abstract:

[Purposes] Leveraging big data theory and related technologies, the scientific journals' data are utilized to construct an indicator system for recording scientific research integrity. The proposed framework aims to provide a reference for promoting scientific research integrity. [Methods] Adopting literature review and comparative analysis, we organized current research status domestically and internationally, integrated relevant big data technologies, and designed an evaluation indicator system. Feasibility of the system was validated through case analysis, forming the basis for proposing a framework for scientific research integrity records. [Findings] In the case analysis, research integrity records of five groups of notified researchers and non-notified researchers are compared pairwise. Chi-square tests reveal significant differences in evaluation indicators such as the ratios of journal retractions, corrections, peer challenges, flagged journal publications, and expressions of concern to SCI-indexed papers. [Conclusions] Indicators such as the ratios of the total number of journal retractions, corrections, peer challenges, publications in flagged journals, and expressed concerns to SCI-indexed papers can effectively evaluate researchers' scientific research integrity. This indicator system provides a basis for constructing a framework for scientific research integrity records.

中图分类号: 

• 科研诚信记录

• 指标体系

• 科技期刊数据

• 数据挖掘

• Research integrity record

• Indicator system

• Scientific journal data

• Data mining