# README for Notebooks

## Overview

This folder contains R notebooks used for replication and analysis tasks in the **OutcomeWeights - Extending Treatment Effect Estimation** project. The notebooks demonstrate the application of the OutcomeWeights package on 401(k) data and serve as a foundation for extending the package's functionality.

## Contents

### Notebooks

1. **Notebook_Average.Rmd**

   - Focuses on estimating the **Average Treatment Effects (ATE)** using the OutcomeWeights package.

   - Includes replications of average effect calculations and validation with the 401(k) dataset.

2. **Notebook_Heterogeneous.Rmd**

   - Examines **Heterogeneous Treatment Effects (HTE)** based on the 401(k) dataset.

   - Demonstrates the use of causal forests to estimate subgroup effects.

### Supporting Materials

- **Dataset**: The notebooks use the 401(k) dataset bundled within the OutcomeWeights package.

- **Outputs**: Graphs and tables generated during the replication are saved in the `outputs/figures` directory for reference.

## Installation and Setup

To run the notebooks:

1. Ensure the required R packages are installed:

   ```R

   install.packages(c("grf", "hdm", "OutcomeWeights"))

   ```

2. Open the `.Rmd` files in RStudio or any compatible R Notebook interface.

3. Run the code chunks sequentially to replicate the results.

## Usage

- **Notebook_Average.Rmd**:

  - Explore how to compute ATE using the OutcomeWeights package.

  - Validate results using the built-in functions and manual calculations.

- **Notebook_Heterogeneous.Rmd**:

  - Investigate heterogeneous effects using causal forests.

  - Analyze subgroup-level treatment effects and compare them with average effects.

## Notes

- Ensure a working internet connection for downloading required packages if not already installed.

- Use the provided `requirements.R` file in the root directory to check for dependencies.

## References

- [OutcomeWeights Documentation](https://github.com/MCKnaus/OutcomeWeights)

- [Generalized Random Forests Documentation](https://github.com/grf-labs/grf)

- [Research Paper: "Treatment Effects as Weighted Outcomes"](https://doi.org/10.48550/arXiv.2411.11559)

---

For further queries, please refer to the main `README.md` file.