A Brief History of Causality

We never see what would have happened. This book is about how science learned to know it anyway.

In 1739 David Hume argued that no one has ever observed a cause, only one thing following another. Today economists, statisticians and data scientists routinely estimate what a policy, a medicine or a training programme actually does. A Brief History of Causality tells the story of how we got from there to here, and teaches the methods along the way.

Every method enters the book the way it entered history: as the answer to a problem the previous generation could not solve. You learn why the randomised experiment was invented, why instrumental variables were needed, why difference-in-differences broke down with staggered timing and how it was rebuilt. Each idea is built up first from intuition, then from pictures, and only then from the mathematics.

Inside you will find:

• 35 chapters running from Hume, Galton and Fisher to causal forests and debiased machine learning
• 71 figures and diagrams, drawn for this book from simulations whose numbers you can check against the text
• 105 questions to test your intuition, with worked answers at the back
• Worked examples you can follow with a pencil, and debates that are still open
• A field guide for choosing a research design and judging any causal claim

Methods covered: randomised experiments, potential outcomes, regression and omitted variable bias, instrumental variables and LATE, matching and propensity scores, causal graphs and DAGs, difference-in-differences and event studies, regression discontinuity, synthetic control, panel data and fixed effects, g-methods for treatments that change over time, structural models, the lasso and double selection, double/debiased machine learning, causal forests, spillovers, bounds and sensitivity analysis, external validity, and the replication crisis.

For graduate students, applied economists, data scientists and researchers in public health, political science and sociology who have met these methods and want to understand them from the ground up: not only how they work, but why they exist.