ATbounds is an R package that provides estimation and inference methods for bounding average treatment effects (on the treated) that are valid under an unconfoundedness assumption. The bounds are designed to be robust in challenging situations, for example, when the conditioning variables take on a large number of different values in the observed sample, or when the overlap condition is violated. This robustness is achieved by only using limited “pooling” of information across observations.
For more details, see Lee and Weidner, “Bounding Treatment Effects by Pooling Limited Information across Observations,” forthcoming at the Journal of Econometrics and available at arXiv:2111.05243 [econ.EM].
You can install the released version of ATbounds from CRAN with:
install.packages("ATbounds")Alternatively, you can install the development version from GitHub with:
# install.packages("devtools") if devtools is not installed
devtools::install_github("https://github.com/ATbounds/ATbounds-r")The package vignette provides worked examples using the NSW/Dehejia-Wahba, right heart catheterization, and electronic fetal monitoring datasets. After installing the package, the vignette can be opened in R with:
vignette("ATbounds_vignette", package = "ATbounds")The companion repository ATbounds/ATbounds-replication contains the replication materials for Lee and Weidner, “Bounding Treatment Effects by Pooling Limited Information across Observations,” forthcoming at the Journal of Econometrics.
The replication repository contains the scripts and generated outputs used to reproduce the paper’s numerical results. It is separate from this package repository: the vignette illustrates how to use the ATbounds package, while the replication repository reproduces the paper tables and main-text numerical results.
Lee, S. and Weidner, M. (forthcoming). “Bounding Treatment Effects by Pooling Limited Information across Observations.” Journal of Econometrics. arXiv:2111.05243 [econ.EM].