【第712期】Marginal Treatment Effects, Randomized Trials, and Understanding the Relationship between Selection and Treatment Effects
【主 讲】Edward Vytlacil (Yale University)
【时 间】2016年7月8日(周五)09:30-11:30
【地 点】上海财经大学经济学院楼701室
【语 言】英文
【摘 要】The goal of this research is to
combine results from Randomized Control Trials (RCTs) and nonexperimental methods on observational data, focusing on the use of instrumental variables and nonparametric selection model methods on the observational data while allowing for rich heterogeneity in treatment effects and for agents to act based on some knowledge of their own treatment effect. The research proceeds in two steps, first testing compatibility of the RCT data and the observational data, and, if found to be compatible, combining results from both methods. Combining results from both methods provides greater external validity than would be possible with the RCT alone, and provides insights into the connection between selection and individual effects that would not be possible from the RCT alone, while mitigating many of the limitations of the nonparametric selection model approach if only using observational data.
友情链接