arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2306.03363econ.EM
收藏

使用机器学习对实验中处理效应方差进行稳健推断

Robust inference for the treatment effect variance in experiments using machine learning

Alejandro Sanchez-Becerra

AI总结:

本文针对 VCATE 在零边界处传统推断失效的问题,为带预测第一阶段的多步机器学习估计量提出考虑高阶项的有效置信区间,并揭示其与处理分配决策收益界的联系。

摘要:

实验人员经常收集基线数据以研究异质性。我提出了首个针对 VCATE(即可观测变量所解释的处理效应方差)的有效置信区间。当 VCATE 为零时,传统方法会产生不正确的覆盖率。因此,即使不存在异质性,实践者也可能倾向于检测到异质性。在边界处覆盖率恶化的原因在于,所有有效估计量都具有局部退化的影响函数,并且可能不满足渐近正态性。我针对一大类具有预测性第一阶段的多步骤估计量解决了这一问题。我提出的置信区间考虑了极限分布中的高阶项,并且计算速度很快。我还发现了 VCATE 与决定对谁进行处理这一问题之间的新联系。定向处理的收益被(严格地)界定为 VCATE 平方根的一半。最后,我记录了其在模拟中的出色表现,并对一项来自马拉维的实验进行了重新分析。

英文摘要:

Experimenters often collect baseline data to study heterogeneity. I propose the first valid confidence intervals for the VCATE, the treatment effect variance explained by observables. Conventional approaches yield incorrect coverage when the VCATE is zero. As a result, practitioners could be prone to detect heterogeneity even when none exists. The reason why coverage worsens at the boundary is that all efficient estimators have a locally-degenerate influence function and may not be asymptotically normal. I solve the problem for a broad class of multistep estimators with a predictive first stage. My confidence intervals account for higher-order terms in the limiting distribution and are fast to compute. I also find new connections between the VCATE and the problem of deciding whom to treat. The gains of targeting treatment are (sharply) bounded by half the square root of the VCATE. Finally, I document excellent performance in simulation and reanalyze an experiment from Malawi.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑