报告题目：Selecting polychoric instrumental variables in factor analysis: An alternative specification test and effects of instrumental variables
报告人：Dr. Shaobo Jin（瑞典乌普萨拉大学）
报告时间：2018年11月19日(星期一)10:00-11:00
报告地点：尚贤楼706
主持人：曹春正教授
欢迎广大师生踊跃参加！
报告内容简介：
Factor analysis is widely used to analyze questionnaire data. The traditional approaches are full information-based, which is prone to model specification bias. Commonly used model modification techniques are sequentialand are known to be unstable to small perturbation of the data. Polychoric instrumental variable (PIV) approach is a recently proposed method to fit a factor analysis model with ordinal data. In this paper, we examine the properties of the specification testsfor testing the validity of instrumental variables (IVs). Our results show that the Sargan’s chi-square test derived for continuous data is extremely oversized when applied to ordinal variables. Modifications to the Sargan’s chi-square test for ordinal dataare proposed in the present study. Simulation results show that the modified specification tests with all available IVs perform well and are able to detect model misspecification.
专家简介
Dr. Shaobo Jin，2015年毕业于瑞典乌普萨拉大学统计系，2017年至今在乌普萨拉大学统计系任Associate Senior Lecturer及博士生导师，曾为伦敦政治经济学院统计系和首尔国立大学统计系访问学者。主要研究领域包括潜变量建模和推断，基于结构方程模型和项目反映理论的心理测量学，以及非贝叶斯方法的模型平均，曾于2015年被提名为国际心理测量协会年度最佳毕业论文候选。
数学与统计学院
江苏省统计科学研究基地
2018年11月16日