Mediation Analysis with Binary Mediators: a New Parametric Method and R Programs

• Activity Number: 255 - Contributed Poster Presentations: Section on Statistical Computing
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Computing
• Author(s): Yujiao Mai* and Deo Kumar Srivastava and Hui Zhang
• Companies: St. Jude Children's Research Hospital and St. Jude Children's Research Hospital and St. Jude Children's Research Hospital
• Keywords: mediation effect; binary data; nonlinear effect; nonnormality

Abstract

Mediation analysis revealing whether an exposure affects a response via a mediator is scientifically and practically important. Analyzing binary mediators has remained a challenge for decades. Mediation analysis within structural equation modeling (SEM) framework is the most discussed parametric approach, but its estimation and significance testing are technically and practically challenging to most researchers. In addition, the current methods within SEM framework cannot handle nonnormally distributed residuals. To address these issues, this study first introduces a joint-distribution based algorithm for effect estimation and significance testing, then implements the proposed algorithm in R programs, and finally illustrates the application by an example with real data. Moreover, preliminary results of simulation are presented for evaluating the proposed method.