tdid

These R functions calculate triple difference-in-differences (TDID) with controls using double-robust estimators, based on Leventer (2025).

Installation

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

devtools::install_github("dorleventer/tdid")

Example

An example, as well as replication of the graph from the paper, is given in the example vignette.

In Progress

Currently this is a bare-bones package repo for replication of the results in the paper. In the future will upgrade such that

• Handle different names from outcome, time, …
• Allow inputting covariate formulas
• Useage of different ML learners for the propensity score and outcome regression
• Cross-fitting of the propensity score and outcome regression

References

Leventer, D. (2025). Conditional Triple Difference-in-Differences. arXiv preprint arXiv:2502.16126.