Skip to contents

causalverse

causalverse is the all-in-one R toolkit for causal inference — the tidyverse of causal analysis. Version 1.1.0 provides 80+ utility functions, 13 in-depth vignettes, and seamless integration with 100+ backend packages, covering the full pipeline from experimental design through publication-ready reporting.

The package covers 13 major method families:

Method Key Functions Backend Packages
Synthetic DID synthdid_est(), synthdid_est_ate(), sc_gap_plot() synthdid
Difference-in-Differences did_event_study(), staggered_summary(), bacon_decomp_plot(), did_power_analysis() fixest, did, did2s, bacondecomp, HonestDiD
Randomized Experiments mde_calc(), dr_ate(), placebo_test(), attrition_analysis(), power_sim() estimatr, grf, DoubleML
Regression Discontinuity rd_plot(), rd_assumption_tests(), rd_bandwidth_sensitivity() rdrobust, rddensity, rdpower
Synthetic Controls panel_estimate(), sc_gap_plot() Synth, augsynth, gsynth
Instrumental Variables iv_diagnostic_summary(), iv_sensitivity() ivreg, ivmodel, hdm, MendelianRandomization
Finance Event Studies event_study_finance(), plot_event_coefs() fixest, estudy2
Matching & Weighting balance_plot(), love_plot(), overlap_weights(), covariate_summary() MatchIt, cobalt, WeightIt
Heterogeneous Effects blp_analysis(), qini_curve(), causal_forest_summary(), heterogeneity_plot() grf, policytree
Mediation Analysis mediation_analysis(), lee_bounds() mediation, lavaan
Panel Data panel_diagnostics() plm, tseries
Causal DAGs dag_plot(), dag_adjustment_sets(), dag_test_implications() dagitty, ggdag
Sensitivity & Robustness multiverse_analysis(), spec_curve(), sensitivity_plot(), pretrend_sensitivity() sensemakr, EValue, HonestDiD

How to cite this package

You can cite this package as follows: “we utilized the causal inference methodologies from the causalverse R package (Nguyen 2026)”. Here is the full bibliographic reference to include in your reference list (don’t forget to update the ‘last accessed’ date):

Nguyen, M. (2026). causalverse: The All-in-One Causal Inference Toolkit (Version 1.1.0). Zenodo. https://doi.org/10.5281/zenodo.8254063. Retrieved from https://github.com/mikenguyen13/causalverse.

Vignettes

13 comprehensive, journal-quality tutorials covering every major causal inference method:

Installation

You can install the development version of causalverse from GitHub with:

# install.packages("pak")
pak::pkg_install("mikenguyen13/causalverse")

Or using devtools:

# install.packages("devtools")
devtools::install_github("mikenguyen13/causalverse")

To install all optional backend packages for full functionality:

library(causalverse)
install_backends()          # install everything
install_backends("did")     # install just DID packages
install_backends("rd")      # install just RD packages

Use this package in your AI assistant

The documentation of causalverse, including its vignettes and function reference, is indexed by the Research Methods Library, a Model Context Protocol (MCP) server that makes it available to AI assistants. Given a research question, an assistant connected to the server recommends candidate study designs, states their identifying assumptions, returns runnable R and Python code, and cites the sections of this package’s documentation and of the author’s other works on which each answer draws. Access is free and requires no account or API key.

Endpoint: https://research-methods-mcp.vercel.app/mcp

  • Claude (web, desktop, and mobile): Settings → Connectors → Add custom connector; enter the endpoint.
  • Claude Code: claude mcp add --transport http research-methods https://research-methods-mcp.vercel.app/mcp
  • ChatGPT: Settings → enable developer mode → add a custom connector of type MCP server with the endpoint and no authentication.
  • Cursor and Visual Studio Code: Add to Cursor or Install in VS Code.

Configuration for other MCP clients is described on the server’s website.

Quick Examples

Synthetic Difference-in-Differences 

library(causalverse)

# Estimate treatment effect using multiple methods
setup <- synthdid::panel.matrices(synthdid::california_prop99)
estimates <- panel_estimate(setup, c("sc", "sdid", "did", "sc_ridge"))
process_panel_estimate(estimates)

# Staggered adoption with ATE
ate <- synthdid_est_ate(data = fixest::base_stagg, method = "sdid")
synthdid_plot_ate(ate)

Event Study with Publication-Ready Plots 

library(causalverse)

# Sun & Abraham event study
model <- feols(y ~ sunab(year_treated, year) | id + year, data = base_stagg)

# Plot with causalverse theming
plot_event_coefs(model, ref_period = -1) + ama_theme()

RD Robustness Checks 

library(causalverse)

# Bandwidth sensitivity analysis
rd_bandwidth_sensitivity(data = df, y = "outcome", x = "running_var", c = 0)

# Placebo cutoff tests
rd_placebo_cutoffs(data = df, y = "outcome", x = "running_var", true_cutoff = 0)

# Covariate balance at the cutoff
rd_covariate_balance(data = df, covariates = c("age", "income"), x = "running_var", c = 0)

Specification Curve Analysis 

library(causalverse)

# Run all specifications and visualize
results <- spec_curve(
  data = df,
  y = "outcome",
  x = "treatment",
  controls = list(c("age"), c("age", "income"), c("age", "income", "educ")),
  fixed_effects = list(NULL, "year", c("year", "state"))
)

AMA-Style Publication Plots 

library(causalverse)

ggplot(mtcars, aes(wt, mpg)) +
  geom_point() +
  geom_smooth(method = "lm") +
  ama_theme() +
  ama_labs(x = "Weight", y = "MPG")

# Export for journal submission
ama_export_fig(filename = "figure1", filepath = "output/")

Citation

Nguyen, M. (2026). causalverse: The All-in-One Causal Inference Toolkit (Version 1.1.0). Zenodo. https://doi.org/10.5281/zenodo.8254063. Retrieved from https://github.com/mikenguyen13/causalverse.