Zuo Shuozhi, Ghosh Debashis, Ding Peng, Yang Fan
Department of Biostatistics and Informatics, Colorado School of Public Health.
Department of Statistics, University of California, Berkeley.
J Am Stat Assoc. 2025;120(550):794-804. doi: 10.1080/01621459.2024.2359132. Epub 2024 Jun 26.
Mediation analysis is widely used for investigating direct and indirect causal pathways through which an effect arises. However, many mediation analysis studies are challenged by missingness in the mediator and outcome. In general, when the mediator and outcome are missing not at random, the direct and indirect effects are not identifiable without further assumptions. We study the identifiability of the direct and indirect effects under some interpretable mechanisms that allow for missing not at random in the mediator and outcome. We evaluate the performance of statistical inference under those mechanisms through simulation studies and illustrate the proposed methods via the National Job Corps Study.
中介分析被广泛用于研究效应产生的直接和间接因果路径。然而，许多中介分析研究受到中介变量和结果变量缺失值的挑战。一般来说，当中介变量和结果变量非随机缺失时，如果没有进一步的假设，直接效应和间接效应是无法识别的。我们研究了在一些可解释机制下直接效应和间接效应的可识别性，这些机制允许中介变量和结果变量非随机缺失。我们通过模拟研究评估了在这些机制下统计推断的性能，并通过国家职业培训团研究说明了所提出的方法。
