Bounds on Direct and Indirect Effects of Treatment on a Continuous Endpoint | |
Luo, Peng ; Geng, Zhi | |
2016 | |
关键词 | Causal inference direct and indirect effects mediation analysis bound CONFOUNDED INTERMEDIATE VARIABLES MODELS |
英文摘要 | Direct effect of a treatment variable on an endpoint variable and indirect effect through a mediate variable are important concepts for understanding a causal mechanism. However, the randomized assignment of treatment is not sufficient for identifying the direct and indirect effects, and extra assumptions and conditions are required, such as the sequential ignorability assumption without unobserved confounders or the sequential potential ignorability assumption. But these assumptions may not be credible in many applications. In this article, we consider the bounds on controlled direct effect, natural direct effect, and natural indirect effect without these extra assumptions. Cai et al. [2008] presented the bounds for the case of a binary endpoint, and we extend their results to the general case for an arbitrary endpoint.; NSFC [11171365, 11331011]; 863 Program [2015AA020507]; 973 Program [2015CB856000]; Center for Statistical Science, Peking University; SCI(E); EI; ARTICLE; luopeng@szu.edu.cn; zhigeng@pku.edu.cn; 2,SI; 7 |
语种 | 中文 |
出处 | SCI ; EI |
出版者 | ACM TRANSACTIONS ON INTELLIGENT SYSTEMS AND TECHNOLOGY |
内容类型 | 其他 |
源URL | [http://hdl.handle.net/20.500.11897/435686] |
专题 | 数学科学学院 |
推荐引用方式 GB/T 7714 | Luo, Peng,Geng, Zhi. Bounds on Direct and Indirect Effects of Treatment on a Continuous Endpoint. 2016-01-01. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论