Getting Started with RobustMediate"

In RobustMediate: Causal Mediation Analysis with Diagnostics and Sensitivity Analysis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment  = "#>",
  fig.width  = 7,
  fig.height = 4.5
)

Overview

RobustMediate estimates natural direct and indirect effects (NDE/NIE) for continuous treatments using inverse probability weighting (IPW). It embeds three publication-ready visualisations and a paper-writing helper (diagnose()) directly into the package workflow.

Simulated data

set.seed(42)
n  <- 600
Z1 <- rnorm(n)
Z2 <- rbinom(n, 1, 0.5)
X  <- 0.5 * Z1 + 0.3 * Z2 + rnorm(n)  # continuous treatment
M  <- 0.4 * X  + 0.2 * Z1 + rnorm(n)  # mediator
Y  <- 0.3 * X  + 0.5 * M  + 0.1 * Z1 + rnorm(n)  # outcome
dat <- data.frame(Y, X, M, Z1, Z2)

Fitting the model

library(RobustMediate)

fit <- robustmediate(
  treatment_formula = X ~ Z1 + Z2,
  mediator_formula  = M ~ X + Z1 + Z2,
  outcome_formula   = Y ~ X + M + Z1 + Z2,
  data    = dat,
  R       = 200,   # use >= 500 in practice
  verbose = FALSE
)
print(fit)

1. Love Plot — plot_balance()

The love plot shows standardised mean differences (SMDs) before and after IPW weighting for both the treatment and mediator pathways. Reviewers require |SMD| < 0.10; the dashed lines mark this threshold.

plot_balance(fit)

2. Dose-Response Curve — plot_mediation()

NDE and NIE as smooth curves over the full treatment range, with pointwise bootstrap confidence bands. The dotted vertical line marks the reference dose.

plot_mediation(fit, estimands = c("NDE", "NIE", "TE"))

3. Sensitivity Contour — plot_sensitivity()

The novel bivariate robustness map. The x-axis is the E-value (how strongly an unmeasured confounder would need to be associated with both treatment and outcome to explain away the NIE). The y-axis is Imai's ρ (sequential ignorability violation). The bold dashed contour marks where the NIE = 0.

plot_sensitivity(fit)

4. Diagnose — diagnose()

Prints a formatted report with a ready-to-paste Results paragraph.

diagnose(fit)

Clustered data

Supply cluster_var to account for clustering:

fit_clus <- robustmediate(
  treatment_formula = X ~ Z1 + Z2,
  mediator_formula  = M ~ X + Z1 + Z2,
  outcome_formula   = Y ~ X + M + Z1 + Z2,
  data        = dat,
  cluster_var = "school_id",
  R           = 500
)

Theoretical note on the sensitivity contour

The E-value × ρ surface is a bivariate robustness display, not a joint causal model. For each pair (E-value, ρ), the surface shows the estimated NIE if that single violation were present. The dimensions should be interpreted separately: researchers can ask how large each violation would need to be—on its own—to overturn the finding.

RobustMediate documentation built on April 16, 2026, 5:08 p.m.