调查问卷的主观性问题数据测量误差探讨  

Discussion on Measurement Error of Subjective Question Data From the Questionnaire

作　　者：侯麟科[1] 林士新 于瀚 李亚飞 Hou Linke;Lin Shixin;Yu Han;Li Yafei(The Center for Economic Research,Shandong University;China Development Bank Shandong Branch,Jinan 250000,China;College of Economics and Management,China Agricultural University,Beijing 100083,China)

机构地区：[1]山东大学经济研究院 [2]国家开发银行山东省分行,济南250000 [3]中国农业大学经济管理学院,北京100083

出　　处：《统计与决策》2025年第4期57-62,共6页Statistics & Decision

基　　金：国家社会科学基金青年项目(23CJY022)。

摘　　要：主观性问题是社会科学研究的重点之一,学者们主要通过问卷调查的方式获得主观性问题数据,然而由于存在认知性误差、社会效应误差和态度性误差,因此主观性问题数据往往存在测量误差。文章通过理论分析发现,将测量误差较小的主观变量作为解释变量或控制变量时,对实证研究具有较强的解释力;然而,将带有测量误差的主观变量作为被解释变量时,则会产生潜在的内生性问题。以CFPS数据库为例,介绍了判断主观性问题数据是否存在测量误差的方法并评估了其对实证结果的影响。研究发现:主观性问题数据中存在占主导地位的客观信息,对客观变量具有一定的解释力,但其测量误差也具有不可忽视的影响。此外,讨论了减少主观性问题数据测量误差的方法:锚定法和工具变量法。Subjectivity problem is one of the priorities of social science research.Scholars mainly obtain data of subjective questions through questionnaire survey.However,due to cognitive errors,social effects errors and attitude errors,the data of subjective questions often have measurement errors.Through theoretical analysis,this paper finds that when subjective variables with small measurement errors are used as explanatory variables or control variables,they have strong explanatory power to empirical study.However,potential endogeneity problems arise when subjective variables with measurement errors are used as explained variables.Taking the CFPS database as an example,this paper introduces the method of judging whether there is measurement error in data of subjective questions,and evaluates its influence on empirical results.The study finds that there exists dominant objective information in data of subjective questions,which has a certain explanatory power to objective variables,but its measurement error also has a non-negligible impact.In addition,the paper also discusses the methods to reduce the measurement error of subjective question data:anchoring method and instrumental variable method.

关 键 词：主观变量 测量误差 内生性 

分 类 号：F062.6[经济管理—政治经济学]