mediationverse

Unified ecosystem for mediation analysis in R

Overview

The mediationverse is a collection of R packages for mediation analysis, providing a unified ecosystem for:

Quick Start

# Install mediationverse
pak::pak("Data-Wise/mediationverse")

# Load foundation package + packages you need
library(mediationverse)  # Loads medfit (foundation)
library(RMediation)      # For confidence intervals

# Fit mediation models
fit_m <- lm(M ~ X + C, data = mydata)
fit_y <- lm(Y ~ X + M + C, data = mydata)
med_data <- extract_mediation(fit_m, model_y = fit_y,
                               treatment = "X", mediator = "M")

# Get confidence intervals
ci_result <- ci(med_data, type = "dop")

Verify your installation

# Prints a situation report of all ecosystem packages, their versions,
# and installation sources
mediationverse_sitrep()

Ecosystem Status Dashboard

πŸ“Š View Full Dashboard | πŸ—ΊοΈ Development Roadmap

Package Overview

CRAN and Build columns are live badges (shields.io / GitHub Actions) β€” they reflect real-time status and update automatically as packages are submitted, accepted, or re-checked. No manual edits needed here when a package's CRAN state changes. (Deliberately using img.shields.io/cran/v/<pkg>, not r-pkg.org/badges/version/<pkg> β€” the latter 404s with a broken HTML error page, not a valid SVG, for packages not yet on CRAN.)

Package CRAN Build Website Role
medfit CRAN Build Docs Foundation
RMediation CRAN Build Docs Confidence Intervals

Package Ecosystem

The mediationverse follows a modular architecture with selective loading inspired by tidyverse and easystats:

Data Flow

User Data
   β”‚
   β–Ό
fit_mediation() or lm()/glm()/lavaan::sem()
   β”‚
   β–Ό
extract_mediation()  ──────►  MediationData object
   β”‚                                  β”‚
   β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
   β”‚              β”‚
   β–Ό              β–Ό
RMediation::   medfit::
medci()        bootstrap_mediation()
   β”‚              β”‚
   β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜
          β”‚
          β–Ό
  Results & Inference

Core Packages

medfit: Foundation Infrastructure

Purpose: Unified S7-based infrastructure for mediation analysis Status: Phase 4 in progress Provides:

Links: Website | GitHub | Issues

RMediation: Confidence Intervals

Purpose: Asymmetric confidence intervals via Distribution of Product Status: Stable on CRAN Provides:

Links: CRAN | Website | GitHub

Installation

From r-universe (recommended)

The whole ecosystem is available as pre-built binaries at the Data-Wise r-universe β€” no compiler needed, and GitHub-only packages resolve automatically:

install.packages(
  "mediationverse",
  repos = c("https://data-wise.r-universe.dev", "https://cloud.r-project.org")
)

From GitHub (development)

# Install via pak (handles CRAN + GitHub deps automatically)
pak::pak("Data-Wise/mediationverse")

From CRAN (future)

# Planned β€” once mediationverse itself is submitted to CRAN
install.packages("mediationverse")

Usage

Loading Packages

The mediationverse uses selective loading: only the foundation package (medfit) is loaded by default. Load other packages as needed for your analysis.

library(mediationverse)
#> ── Attaching mediationverse 0.1.1 ──
#> βœ” medfit 0.3.2 (foundation package)
#> β„Ή Use library(RMediation) for DOP/MBCO inference
#> ──────────────────────────────────────────────────────

# Load additional packages as needed
library(RMediation)   # For confidence intervals

Why selective loading?

Package Management

# List installed packages and versions
mediationverse_packages()

# Update all packages
mediationverse_update()

# Check for function conflicts
mediationverse_conflicts()

Complete Analysis Workflow

# Load foundation and packages you need
library(mediationverse)  # Loads medfit (foundation)
library(RMediation)      # For confidence intervals

# 1. Fit mediation models (medfit - already loaded)
fit_m <- lm(M ~ X + C, data = mydata)
fit_y <- lm(Y ~ X + M + C, data = mydata)

# 2. Extract mediation structure (medfit)
med_data <- extract_mediation(fit_m, model_y = fit_y,
                               treatment = "X", mediator = "M")

# 3. Bootstrap inference (medfit)
boot_result <- bootstrap_mediation(med_data, n_boot = 2000)

# 4. Get confidence intervals (RMediation)
ci_result <- ci(med_data, type = "dop")

Design Philosophy

Following principles from tidyverse and easystats:

  1. Foundation Package: medfit provides shared S7 infrastructure
  2. Specialized Packages: Each package focuses on one methodological contribution
  3. Consistent API: Unified interfaces across all packages
  4. Type Safety: S7 classes ensure data integrity
  5. Low Dependencies: Minimal external dependencies for stability
  6. Comprehensive Testing: >90% code coverage across packages

Documentation

Getting Started

Status & Planning

Package Documentation

Contributing

The mediationverse is in active development. Contributions are welcome!

Development Workflow:

  1. Fork the repository
  2. Create a feature branch from dev
  3. Make your changes with tests
  4. Submit a pull request to dev

See CONTRIBUTING.md for detailed guidelines.

Citation

If you use packages from the mediationverse in your research, please cite the individual packages:

Tofighi, D. (2026). medfit: Infrastructure for Mediation Model Fitting and
Extraction. R package version 0.3.2. https://CRAN.R-project.org/package=medfit

Tofighi, D., & MacKinnon, D. P. (2011). RMediation: An R package for
mediation analysis confidence intervals. Behavior Research Methods, 43,
692-700. https://CRAN.R-project.org/package=RMediation

Full citations available via citation("packagename").

Related Ecosystems

The mediationverse draws inspiration from:

Code of Conduct

Please note that the mediationverse project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.

License

GPL (>= 3)

Contact

Maintainer: Davood Tofighi (dtofighi@gmail.com) ORCID: 0000-0001-8523-7776

Sources: