基于分层有序Probit模型研究教育对自评健康切点影响  

Study on the Effect of Education on Self-rated Health Cut-points using Hierarchical Ordered Probit Model

作　　者：郭恺 吴明[1,2] Guo Kai;Wu Ming(Department of Health Policy and Management,Peking University School of Public Health(100191),Beijing)

机构地区：[1]北京大学公共卫生学院卫生政策与管理系,100191 [2]北京大学医学部卫生政策和技术评估中心

出　　处：《中国卫生统计》2021年第1期6-9,共4页Chinese Journal of Health Statistics

摘　　要：目的分析教育对自评健康评价切点的影响。方法利用2014年中国家庭追踪调查19~51岁成人数据,使用虚拟情境锚定法校正切点位移偏倚,采用分层有序Probit模型研究教育对健康评价标准的影响。结果本次研究共纳入研究对象11687人,自评健康为比较健康者占35.50%,情境1健康评价为很健康者占37.07%,情境2健康评价为不健康者占59.92%。分层有序Probit模型回归结果显示,随着受教育年限的提升,"不健康-一般"、"一般-比较健康"切点值降低,"比较健康-很健康"、"很健康-非常健康"切点值升高。结论研究者应重视自评健康的切点位移偏倚,教育会对人群自评健康评价切点产生影响,造成自评健康评价结果趋中。Objective To analyze the effect of education on self-rated health cut-points.Methods Based on the data of 19-51 year-old adults in the 2014 China Family Panel Studies,anchoring vignettes method was used to correct the cut-point shift bias,and hierarchical ordered probit(HOPIT)model was used to study the effect of education on health evaluation standards.Results A total of 11687 respondents were included in this study,35.50%of them rated their health as good,37.07%of them rated vignettes 1 as very good,59.92%of them rated vignettes 2 as poor,and the HOPIT model showed that with the increase of education years,the cut-point of“poor-fair”and“fair-good”decreased,and“good-very good”and“very good-excellent”increased.Conclusion Researchers should pay attention to the effect of the cut-point shift bias in self-rated health.Education have a effect on the self-rated health cut-points,leading to the results of self-rated health tend towards the middle.

关 键 词：自评健康 切点位移偏倚 切点 分层有序Probit模型 

分 类 号：R195.1[医药卫生—卫生统计学]