Provides tools for causal mediation analysis with continuous
treatments. Estimates natural direct and indirect effects by g-computation
over a user-defined treatment grid, with bootstrap intervals that can
resample whole clusters, and supports nonlinear dose-response through
spline terms in the model formulas. Includes balance diagnostics based on
standardized mean differences and stabilized inverse probability
weights. Provides sensitivity analysis for
confounding of the mediator-outcome relationship using the error
correlation parameter of Imai, Keele, and Yamamoto (2010)
Causal mediation analysis for continuous treatments with embedded balance diagnostics, dose-response curves, and sensitivity analysis for mediator-outcome confounding.
| Function | What it gives you |
|---|---|
robustmediate() |
Fit treatment / mediator / outcome models, NDE/NIE/TE curves by g-computation with bootstrap CIs, and the sensitivity analysis in one call |
plot_balance() |
Dual love plot: covariate balance before/after weighting for both pathways simultaneously |
plot_mediation() |
Dose-response curves of NDE, NIE, TE with pointwise confidence bands |
sensitivity_paths() |
Robustness values for each path and for the indirect effect under joint-significance and Sobel tests (Cinelli & Hazlett, 2020) |
sensitivity_rho() |
Indirect effect as a function of the mediator-outcome error correlation rho (Imai, Keele & Yamamoto, 2010) |
plot_sensitivity() |
Plot of the indirect effect against rho |
diagnose() |
Formatted report with a paste-ready Results paragraph |
# Development version from GitHub
# install.packages("pak")
pak::pkg_install("subirhait/RobustMediate")
library(RobustMediate)
fit <- robustmediate(
treatment_formula = X ~ Z1 + Z2,
mediator_formula = M ~ X + Z1 + Z2,
outcome_formula = Y ~ X + M + Z1 + Z2,
data = mydata,
R = 500
)
plot_balance(fit) # love plot
plot_mediation(fit) # NDE / NIE dose-response curve
plot_sensitivity(fit) # indirect effect vs rho
diagnose(fit) # paste into Results section
cobalt and WeightIt give balance diagnostics for the treatment model only,
and mediation targets binary or categorical treatments. RobustMediate
puts balance checks for both pathways, dose-response curves for a continuous
treatment, and a sensitivity analysis into one workflow.
Bug reports and feature requests via GitHub Issues.
MIT