Comparative Causal Metrics

Welcome to Comparative Causal Metrics! (Work in Progress)

An Introduction to Regional Impact Evaluation

An introduction to regional impact evaluation using modern causal-inference methods implemented in R and rendered with Quarto. The resource covers quasi-experimental techniques for evaluating policy effects and interventions on regional outcomes, with worked examples and publicly available data for full reproducibility.

This work in progress book features:

• A comparative tour of methods — From interrupted time series and difference-in-differences to synthetic control, Bayesian structural time series, and modern panel-data estimators, all with a regional comparative focus.
• R + Quarto Notebooks — Reproducible chapters with collapsible code, ready to render locally or extend with your own data.

The book is organized in two parts:

• Part I — Single treated unit (Chapters 1–9) builds intuition with one running case study: California’s 1989 Proposition 99 cigarette tax.
• Part II — Staggered adoption (Chapters 10–12) moves to settings where many units adopt a policy at different times, using a Callaway–Sant’Anna minimum-wage county panel.

Chapters

Part I — Single treated unit

1. Introduction
2. Interrupted Time Series
3. Basic Differences-in-Differences
4. Classical Synthetic Control
5. Augmented Synthetic Control
6. Synthetic Difference-in-Differences
7. Structural Bayesian Time Series
8. Synthetic Control with Prediction Intervals
9. Bayesian Spatial Synthetic Control

Part II — Staggered adoption

10. Staggered Differences-in-Differences
11. Interactive Fixed Effects and Matrix Completion
12. Generalized Synthetic Control

Plus: References

Contribute and provide feedback at https://github.com/quarcs-lab/ccm.

Related project

Companion resource: Mastering Causal Metrics — an AI-powered Python study guide based on Angrist & Pischke’s Mastering ‘Metrics.