DRDID: An R package to compute doubly robust difference-in-differences estimators

[Official website] [CRAN] [Github]

The DRDID R package implements different estimators for the Average Treatment Effect on the Treated (ATT) in Difference-in-Differences (DID) setups where the parallel trends assumption holds after conditioning on a vector of pre-treatment covariates. The main estimators implemented here are the locally efficient, doubly-robust DID estimators proposed by Sant'Anna and Zhao (2020), "Doubly Robust Difference-in-Differences Estimators". The package covers both panel data and repeated cross-section data setups with two treatment groups (treated and comparison group) and two time periods (pre-treatment and post-treatment).

For detailed explanations on how to use the package, see the package website.

did: An R package for difference-in-differences with multiple time periods

[Official website] [CRAN] [Github]

The did R package provides tools to estimate and conduct asymptotically valid inference about average treatment effects in Difference-in-Differences models with multiple time periods and variation in treatment timing. In short, this package implements the causal inference tools proposed in Callaway and Sant'Anna (2020), "Difference-in-Differences with Multiple Time Periods."

For detailed explanations on how to use the package and/or for a general guidance on DiD methods, see the package website.

IPS: Covariate Distribution Balance via Integrated Propensity Scores


The IPS R package implements the different integrated propensity score (IPS) estimators proposed in Sant'Anna, Song and Xu (2019), "Covariate Distribution Balance via Propensity Scores", and also the inverse probabily weigthed (IPW) estimators for the average, quantile and distributional treatment effects that build on these IPS estimators.

The IPS is estimated by fully exploiting the covariate balancing of the propensity score, i.e., by maximing the entire covariate distribution balance between the treated, untreated, and combined groups. The IPS estimators are data-driven, do not rely on tuning parameters such as bandwidths, and admit an asymptotic linear representation, which, in turn, facilitates the statistical analysis of IPW estimators for the average, quantile and distributional treatment effects.

For explanations on how to use the package, see the package's Github website.

pstest: An R package to assess the goodness-of-fit of parametric propensity score models

[Github] [CRAN]

The pstest R package provides data-driven nonparametric diagnostic tools for detecting propensity score misspecification, that do not depend on tuning parameters, and do not suffer from the "curse of dimensionality''. In short, this package implements the class of specifications tests for the propensity score proposed in Sant'Anna and Song (2019), "Specification Tests for the Propensity Score".

For explanations on how to use the package, see the package's Github website.

kmte: An R package for treatment effects with censored outcomes


The kmte R package includes a variety of policy evaluation tools suitable for right-censored (duration) outcomes. The content includes estimators and tests related to average, quantile, and distributional treatment effects under difference identifying assumptions including unconfoundedness, local treatment effects, and nonlinear difference-in-differences. In short, this package implement all estimators proposed in Sant'Anna (2016), "Program Evaluation with Right-Censored Data", and all tests proposed in Sant'Anna (2020), "Nonparametric tests for Treatment Effect Heterogeneity with Duration Outcomes".

The kmte package is still under development, so it is not yet on CRAN. Nonetheless, you can install its most recent version from Github, using devtools::install_github("pedrohcgs/kmte"). Additional documentation will be provided very soon.