如何用锚点情境法降低自评健康的回答偏误?——一项基于CFPS2012数据的实证分析  

How to Apply the Anchoring Vignette Method to Reduce the Response Bias of Self-rated Health——An Empirical Analysis Based on CFPS2012 Survey Data

作　　者：吴菲[1] WU Fei(Department of Sociology,Fudan University,Shanghai 200433,China)

机构地区：[1]复旦大学社会学系,上海200433

出　　处：《人口与发展》2019年第2期117-128,共12页Population and Development

基　　金：复旦大学文科"双一流"建设项目"社会学理论与社会结构转型研究"创新团队的一项成果

摘　　要：健康不平等是人口学研究中的核心议题之一。然而,以自评健康为代表的主观健康测量因受回答异质性问题的困扰,可能会对研究结论的可信性造成威胁。锚点情境法(anchoring vignette method)作为一项有效处理回答异质性的方法在西方学术界已被广泛使用。介绍了该方法的基本原理、核心假设及其验证方法。使用2012年中国家庭追踪调查(CFPS2012)数据,应用锚点情境法对自评健康(self-rated health)进行调整,发现不同群体在面对同样的自评健康选项时,的确存在回答异质性的问题,如未加调整,所得的年龄、迁移状态以及社会经济地位对于自评健康的影响会产生实质性的偏误。Health inequality is one of the core issues in demographic studies.However,the reliability of conclusions drawn from self-rated health have been seriously handicapped by the problem of response heterogeneity.This review firstly introducesthe widely-used anchoring vignette method to address this problem:its core principle,basic measurement assumptions and related way for validation.Furthermore,based on CFPS2012,we demonstrate the power of anchoring vignette on measuring and correcting the response heterogeneity in self-rated health.Our analyses suggest that the health effects of age,migration status and socioeconomic status would be biased substantially if not being adjusted for the response heterogeneity problem.

关 键 词：锚点情境法 自评健康 回答异质性 情境等价性假设 回答一致性假设 

分 类 号：C91-03[经济管理]