Skip to content

README.md

Latest commit

 

History

History
41 lines (28 loc) · 1.54 KB

README.md

File metadata and controls

41 lines (28 loc) · 1.54 KB

Omitted Variable Bias & Propensity Score Matching (PSM) Demo

This repository contains a simple, interactive web demo that illustrates omitted variable bias in regression analysis using Propensity Score Matching (PSM). It allows users to explore how excluding important covariates can bias treatment effect estimates.

📁 Contents

  • DCL Final.html – Main interactive HTML page containing the demo logic with checkboxes and Plotly visualizations.
  • DCL.qmd – Source file likely written in Quarto Markdown used to generate the HTML version (not viewable directly here).
  • styles.css – External CSS styles for UI enhancements.

🚀 Features

  • Interactive variable selection for matching covariates.
  • Dynamic visualization of regression estimates:
    • Biased model (omitting a confounding variable).
    • True model (including confounding variable).
  • Educational use case demonstrating how omitting variables can distort causal inference.

📦 Usage

To run the demo locally:

  1. Clone this repository: 
    git clone https://github.com/your-username/psm-omitted-variable-demo.git
cd psm-omitted-variable-demo
  2. Open the DCL Final.html file in your browser.

Note: Ensure you have an internet connection as the demo uses the Plotly CDN.

🧠 Educational Context

This tool is great for:

  • Econometrics or causal inference classes.
  • Demonstrating the intuition behind omitted variable bias.
  • Explaining the benefits of PSM in observational data analysis.

📄 License

MIT License. Feel free to use and modify.