图书情报知识 ›› 2021, Vol. 38 ›› Issue (3): 165-176.doi: 10.13366/j.dik.2021.03.165

• 研究方法 • 上一篇    

图情档研究中的回归分析:常见问题

梁兴堃   

  • 出版日期:2021-05-10 发布日期:2021-07-03

Common Problems of Regression Analysis in Library, Information and Archives Management

  • Online:2021-05-10 Published:2021-07-03

摘要: [目的/意义]总结图情档研究中应用回归分析的常见问题并有针对性地提出解决途径。[研究设计/方法] 以文献综述法对图情档研究中使用回归分析的典型问题进行梳理和归纳,以案例分析法说明这些典型问题产生的原因和解决的途径。[结论/发现]图情档研究中在使用回归分析时,容易产生研究问题不适合、忽视研究假设、研究设计有遗漏、数据收集偏差欠考虑、回归模型设定不合理,忽视共线性、异方差和内生性等统计稳健性,结果阐释不合理等问题。[创新/价值]为相关研究者使用回归分析、评审人评阅回归分析的论文提供了参考依据。

关键词: 图情档研究, 回归分析, 共同方法偏差, 稳健性, 异方差, 共线性, 内生性

Abstract: [Purpose/Significance]This paper aims to summarize common problems of using regression analysis in library, information and archives management, and provide countermeasures accordingly.[Design/Methodology]These common problems were summarized from reviewing literature using regression analysis in library, information and archival studies. A case study was used to demonstrate the potential causes and solutions to these common problems.[Findings/Conclusion]These common problems include inappropriate research question, ignorance of hypothesis, drawbacks in research design, common method bias in data collection, unreasonable model specification and ignorance of statistical robustness, i.e., multicollinearity, heteroscedasticity and endogeneity, and inappropriate interpretation of results. [Originality/Value]This paper contributes to serving as a reference for researchers and reviewers in relevant fields to deal with papers with regression analysis.

Key words: Library, information and archival studies, Regression analysis, Common method bias, Robustness, Multicollinearity, Heteroscedasticity, Endogeneity