Recently, there has been great interest in estimating the conditional average treatment effect using flexible machine learning methods. However, in practice, investigators often have working hypotheses about effect heterogeneity across pre-defined subgroups of individuals, which we call the groupwise approach. The paper compares two modern ways to estimate groupwise treatment effects, a nonparametric approach and a semiparametric approach, with the goal of better informing practice. Specifically, we compare (a) the underlying assumptions, (b) efficiency characteristics, and (c) a way to combine the two approaches. We also discuss how to obtain cluster-robust standard errors if the study units in the same subgroups are not independent and identically distributed. We conclude by reanalyzing the National Educational Longitudinal Study of 1988.
翻译:最近,人们对使用灵活的机器学习方法来估计有条件平均治疗效果极感兴趣,但在实践中,调查人员往往对预先界定的个人分组之间的效果差异有工作假设,我们称之为小组方法,该文件比较了两种现代方法,即群体处理效果、非对称方法和半对称方法,目的是更好地说明做法,具体地说,我们比较了(a) 基本假设,(b) 效率特点,(c) 两种方法相结合的方法。我们还讨论了如果同一分组的研究单位不独立和分布相同,如何获得集群-野蛮标准错误。我们最后对1988年国家教育纵向研究进行了重新分析。