A First Course in Causal Inference
This paper presents a comprehensive textbook introducing causal inference, covering key concepts such as causal diagrams, potential outcomes, identification, and estimation methods. It is designed for graduate students and researchers with a background in statistics or machine learning, providing both theoretical foundations and practical applications.
Background
- This is an advanced textbook (arXiv:2305.18793) on **causal inference** — a branch of statistics that tries to answer "what would happen if we changed X?" from observational data, not just "is X correlated with Y?"
- The author, **Peng Ding**, is a professor of biostatistics at UC Berkeley and a leading figure in causal methodology.
- The book covers core frameworks: **Potential Outcomes** (Rubin Causal Model) and **Directed Acyclic Graphs** (DAGs), plus topics like instrumental variables, mediation, sensitivity analysis, and experiments.
- Causal inference is central to evidence-based policy, medicine, economics, and tech (e.g., A/B testing at scale). Without it, correlations can mislead — e.g., "ice cream sales cause drowning" (both rise in summer). The field provides formal tools to avoid such mistakes.