Resource Round-Up: Causal Inference

Free books, lectures, blogs, papers, and more for a causal inference crash course
causal
resources
Author

Emily Riederer

Published

January 30, 2021

Photo by Sharon McCutcheon on Unsplash

In my post on causal design patterns, I argue that these techniques are currently underutilized in industry because (at least, in part) they are so often couched in domain-specific language. Fortunately, the past few years have seen an explosion of fantastic resources and tools to help practitioners more readily learn and apply these methods. Below, I link some of my favorite readings. To help prioritize which resources (or which chapters within them) will be most relevant to you, please see my previous post for an “advertisement” and overview of some of the main techniques.

Free Books

  • Causal Inference: What If by Miguel Hernan and Jamie Robins
    • Written from epidemiology perspective
    • Nice “model free” section introduces causal intuition
    • Code supplement in R, python, SAS, Stata
  • Introduction to Causal Inference by Brady Neal
    • Written from ML perspective including “advanced” topics such as Bayesian networks, causal discovery
    • Builds strong theoretical basis with graphical and probabilistic proofs
    • Book complemented by video lectures
  • Causal Inference: the Mixtape by Scott Cunningham
    • Written from economics perspective
    • Provides great insights into the history and relevance of different methods in economics literature
    • Includes interactive R code chunks to run as you read
  • The Effect: An Introduction to Research Design and Causality by Nick Huntington-Kline
    • Written from the economics perspective
    • Takes a holistic approach to research design with rich examples from literature
  • Impact Evaluation in Practice by Gertler, Martinez, Premand, Rawlings, Vermeersch of the World Bank
  • Handbook of Field Experiments by Ahbijit Banerjee and Esther Duflo
    • Technically a type of experimentation not causal inference
    • However, the real world challenges of field (versus clinical) research creates some nice “blended” methodologies.
    • For example encouragement designs are closely related to instrumental variable methods. These may have been inadvertently conducted in your business strategy and be available in historical data.

Course Material

Survey Papers & Blogs

Surveys

Deeper Dives

Propensity Score Focused

Miscellaneous Advanced Topic Talks

Other Introductory Books