We implement multisite causal mediation analysis using the methods proposed by Qin and Hong (in press). It enables causal mediation analysis in multisite trials, in which individuals are assigned to a treatment or a control group at each site. It allows for estimation and hypothesis testing for not only the population average but also the between-site variance of direct and indirect effects. This strategy conveniently relaxes the assumption of no treatment-by-mediator interaction while greatly simplifying the outcome model specification without invoking strong distributional assumptions.
| Version: | 0.0.1 |
| Depends: | R (≥ 3.1.0), lme4, statmod, Matrix |
| Published: | 2017-02-26 |
| Author: | Xu Qin, Guanglei Hong |
| Maintainer: | Xu Qin <xuqin at uchicago.edu> |
| BugReports: | https://github.com/Xu-Qin/MultisiteMediation/issues |
| License: | GPL-2 |
| URL: | https://github.com/Xu-Qin/MultisiteMediation |
| NeedsCompilation: | no |
| CRAN checks: | MultisiteMediation results |
| Reference manual: | MultisiteMediation.pdf |
| Package source: | MultisiteMediation_0.0.1.tar.gz |
| Windows binaries: | r-devel: MultisiteMediation_0.0.1.zip, r-release: MultisiteMediation_0.0.1.zip, r-oldrel: MultisiteMediation_0.0.1.zip |
| OS X El Capitan binaries: | r-release: MultisiteMediation_0.0.1.tgz |
| OS X Mavericks binaries: | r-oldrel: MultisiteMediation_0.0.1.tgz |
Please use the canonical form https://CRAN.R-project.org/package=MultisiteMediation to link to this page.