Robust Machine Learning for Treatment Effects in Multilevel Observational Studies Under Cluster-level Unmeasured Confounding

Robust Machine Learning for Treatment Effects in Multilevel Observational Studies Under Cluster-level Unmeasured Confounding
复制标题

集群级未测量混杂下多级观察研究中治疗效果的稳健机器学习

DOI:
10.1007/s11336-021-09805-x
复制
发表时间:
2020
期刊:
影响因子:
3
通讯作者:
Hyunseung Kang
Hyunseung Kang
中科院分区:
心理学4区
文献类型:
--
作者:
Youmi Suk;Hyunseung Kang

文献摘要

参考文献

被引文献

相似文献

最近,机器学习(ML)方法已被用于因果推理来估计治疗效果,以减少对模型错误指定的担忧。然而,许多ML方法要求测量所有混杂因素,以一致地估计治疗效果。在本文中,我们提出了一系列ML方法,用于在存在聚类水平的未测量混杂因素的情况下估计治疗效果,这是一种在每个聚类中共享的未测量混杂因素,在多水平观察性研究中很常见。我们通过模拟研究表明,我们提出的方法对于各种多水平观察性研究中未测量的集群级混杂因素的偏差具有鲁棒性。我们还研究了从幼儿纵向研究,一个多层次的观察性教育研究,使用我们的方法,采取代数课程对数学成绩的影响。所提出的方法在CURobustML R包中可用。
Recently, machine learning (ML) methods have been used in causal inference to estimate treatment effects in order to reduce concerns for model mis-specification. However, many ML methods require that all confounders are measured to consistently estimate treatment effects. In this paper, we propose a family of ML methods that estimate treatment effects in the presence of cluster-level unmeasured confounders, a type of unmeasured confounders that are shared within each cluster and are common in multilevel observational studies. We show through simulation studies that our proposed methods are robust from biases from unmeasured cluster-level confounders in a variety of multilevel observational studies. We also examine the effect of taking an algebra course on math achievement scores from the Early Childhood Longitudinal Study, a multilevel observational educational study, using our methods. The proposed methods are available in the CURobustML R package.
DOI: 10.1016/j.jclinepi.2009.11.020
发表时间: 2010-08
影响因子: 7.2
作者:
Westreich, Daniel;Lessler, Justin;Funk, Michele Jonsson
通讯作者: Funk, Michele Jonsson
DOI: 10.1198/016214508000000733
发表时间: 2008-12-01
影响因子: 3.7
作者:
Shadish, William R.;Clark, M. H.;Steiner, Peter M.
通讯作者: Steiner, Peter M.
DOI: 10.1515/jci-2017-0027
发表时间: 2018-09-01
影响因子: 1.4
作者:
Yang, Shu
通讯作者: Yang, Shu
DOI: 10.1080/01621459.2017.1407322
发表时间: 2019-01-02
影响因子: 3.7
作者:
Ding, Peng;Feller, Avi;Miratrix, Luke
通讯作者: Miratrix, Luke
固定效应面板数据模型的非参数估计和测试。
DOI: 10.1016/j.jeconom.2008.01.005
发表时间: 2008
影响因子: 6.3
作者:
Henderson,DanielJ;Carroll,RaymondJ;Li,Qi
通讯作者: Li,Qi