科技论文数据造假的核查策略和统计学方法验证  

Strategies for discerning data fraud in scientific papers and statistical verification

作　　者：余菁 邬加佳 孙慧兰[1] 祁方昉 徐杰[1] 莫传伟[3] YU Jing;WU Jiajia;SUN Huilan;QI Fangfang;XU Jie;MO Chuanwei(Editorial Department of Journal of Sun Yat-sen University(Journal Management Center of Sun Yat-sen University),74 Zhongshan RoadⅡ,Guangzhou 510080,China;Editorial Office of Organ Transplantation,Center for Journal Publishing of the Third Affiliated Hospital,Sun Yat-sen University,600 Tianhe Road,Guangzhou 510630,China;Department of Biostatistics and Preventive Medicine,School of Basic Medical Sciences,Guangzhou University of Chinese Medicine,12 Jichang Road,Guangzhou 510405,China)

机构地区：[1]中山大学学报编辑部(期刊管理中心),广东广州510080 [2]中山大学附属第三医院期刊中心《器官移植》编辑部,广东广州510630 [3]广州中医药大学基础医学院医学统计学教研室,广东广州510405

出　　处：《中国科技期刊研究》2021年第6期770-776,共7页Chinese Journal of Scientific and Technical Periodicals

基　　金：2020年广东省科技期刊编辑学会基金一般项目“采用统计学原理分析和界定科技论文中数据造假”(202045)。

摘　　要：【目的】为帮助期刊编辑使用有效的手段核查论文数据真伪,提出防范论文数据造假的策略。【方法】运用统计学方法和原始数据核查,从数据一致性检测、t检验、F检验、卡方检验和生存分析等案例,分析和判断论文数据造假。【结果】期刊编辑应形成防范学术不端论文审核要点,建立初审工作流程,规范统计学审核步骤,综合判断论文数据真实性。【结论】科技期刊应建立健全稿件审核流程,强化编辑审稿过程中的责任意识,避免学术不端论文见刊。[Purposes]To help journal editors efficiently discern data fraud in scientific papers,this paper put forward some strategies. [Methods]Through statistical methods and data tracing,this paper analyzed the cases of identifying data fraud from the perspectives of data consistency test,t-test,F-test,Chi-square test,and survival analysis. [Findings]We should summarize the key points for discerning academic misconducts,establish the process of preliminary examination,and standardize the statistical review,thereby comprehensively preventing data fraud in scientific papers. [Conclusions] Scientific journals should establish a standard process for manuscript review and strengthen editors’ sense of responsibility to avoid academic misconducts.

关 键 词：学术诚信 统计学方法 数据造假 审核 验证 

分 类 号：G232[文化科学]