Comparative Analysis To The Estimate Methods Of Treatment Effects

Program evaluation is a formalized approach to studying the goals, processes and impacts of projects, policies and programs. And treatment effects studies are more interesting in program evaluation. The central problem of evaluation research is that for individual we observe either the potential outcomes of receiving the treatment or not. Because of this missing data problem, individual treatment effects can not be observed, attention has been focused on some parameters which can summarize the impact of the problem in a meaningful way. Usually, the parameters of interest are average treatment effects and quantile treatment effects. This paper presents the evaluate methods in the literature. To solve the evaluate problem, we maintain the assumption (Rosenbaum and Rubin(1983)). Under the assumption, the role of the propensity score is significant. Because the distribution of potential outcomes is not parametrically specified, the semiparametric efficiency bounds was introduced in asymptotic normality. In the final part of the paper, the statistical simulation studies show the difference to these methods under finite different sample size.