Title: Robust Mediation Analysis: From Methodological Advancements to Easy and
Reproducible Reporting

Authors: Andreas Alfons, Aurore Archimbaud, Patrick Groenen, Nufer Y. Ates

Abstract:
Mediation analysis is one of the most widely used statistical techniques in
the behavioral sciences. In its simplest form, a mediation model allows to
study how an independent variable (X) affects a dependent variable (Y) through
an intervening variable called a mediator (M). We developed a robust bootstrap
test for the indirect effect of X on Y through M that yields reliable results
even in the presence of outliers, heavy tails, or skewness. We implemented
this robust bootstrap test in the R package robmed together with other methods
such as the standard bootstrap test based on ordinary least squares
regressions.

Building on package robmed, our companion package robmedExtra provides
additional tools for easy reporting, such as (i) functions to convert results
from mediation analysis into a nicely formatted flextable, LaTeX table, or
Microsoft Word table, and (ii) a function to generate a model diagram for the
conducted mediation analysis. Moreover, package robmedExtra a graphical user
interface that allows researchers without prior programming experience to
access the functionality of packages robmed and robmedExtra. To ensure
reproducibility of the conducted mediation analysis, the graphical user
interface lets users download replication materials in the form of an R
script.

Paper on the robust bootstrap test: https://doi.org/10.1177/1094428121999096
Paper on the R package robmed: https://doi.org/10.18637/jss.v103.i13