The primary aim of causalverse is to streamline the research process, particularly data analysis, for researchers working in causal inference. It offers a range of helper functions designed to minimize time spent on data analysis. The package includes methods such as regression discontinuity, difference-in-differences, synthetic control, instrumental variables, event studies, and more. Additional methods may be introduced in future updates.

How to cite this package

You can cite this package as follows: “we utilized the causal inference methodologies from the causalverse R package (Nguyen 2023)”. Here is the full bibliographic reference to include in your reference list (don’t forget to update the ‘last accessed’ date):

Nguyen, M. (2023). The causalverse Package: Causality in Clarity. Zendono. http://doi.org/10.5281/zenodo.8254063. Retrieved from https://github.com/mikenguyen13/causalverse.

Vignettes

All the vignettes can be accessed via the package’s website.

Installation

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

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

Example

This is a basic example which shows you how to solve a common problem:

library(causalverse)
## basic example code

Citation

Nguyen, M. (2023). The causalverse Package: Causality in Clarity. Zenodo. http://doi.org/10.5281/zenodo.8254063. Retrieved from https://github.com/mikenguyen13/causalverse.