ATE Covariate Selection

In this project, we tackled the challenge of accurately estimating the ATE while selecting covariates that enhance ATE assessment. We underscored the significance of the overlap and ignorability assumptions, demonstrating how these two assumptions can conflict. Specifically, including a confounder may disrupt the overlap condition, while excluding it could violate the ignorability assumption. To address this issue, we proposed a method based on Cohen's d metric to determine whether a confounder should be included. Through multiple simulations, we illustrated our findings and highlighted the potential of our method for effective covariate selection.

David Cohen, Harel Mendelman

Installing

You can install all the required packages by executing the following command:

pip install -r requirements_python_3_08.txt

Project files

Covariate_Selection/
├── ate_estimate.py                        # Script for ATE estimation
├── ATE_folds.py                           # Script for simulation 1
├── ATE_folds_simple.py                    # Script for simulation 2
├── data_generating_process.py             # Script to generate synthetic data
├── toy_problem.py                                # Script for toy problem simulation
├── plot_propensity_score_distribution.py  # Script to plot the distribution of propensity scores
├── README.md                              # Readme file explaining the project
├── requirements_python_3_08.txt           # File listing project dependencies
└── utilities.py                           # Utility functions used across the project