Foundations of Causal Inference

This course builds the conceptual foundation for causal inference. The emphasis is on learning to state the question before estimating anything: what action is being considered, what counterfactual comparison is required, who the estimand is about, and which assumptions connect the observed data to the desired causal claim.

The objective is to make causal reasoning explicit enough that a project can be audited. By the end of the course, a reader should be able to distinguish prediction from intervention, define ATE/ATT/ATC/CATE-style estimands, explain the role of identification assumptions, use DAGs to communicate causal structure, and recognize when adjustment choices change the causal question.

Figure: A ranking-policy imbalance plot used to make confounding visible before any estimator is introduced (adapted from Lecture 08: Confounders, Mediators, Colliders, and Selection Bias).

Lecture Sequence

01. What Is a Causal Question?

This lecture connects What Is a Causal Question? to estimands, assumptions, counterfactual reasoning, and common design failures.

02. Prediction vs Causal Inference

This lecture develops Prediction vs Causal Inference with examples that make assumptions, diagnostics, and interpretation visible.

03. Potential Outcomes and Counterfactuals

This lecture uses Potential Outcomes and Counterfactuals to clarify the analyst’s question, evidence, assumptions, and decision implications.

04. Treatment, Outcome, Estimand, Population

This lecture clarifies treatment, outcome, estimand, and population before any estimator enters the workflow.

05. Average Treatment Effects: ATE, ATT, ATC, and CATE

This lecture frames Average Treatment Effects: ATE, ATT, ATC, and CATE as a decision problem and asks what evidence can be trusted, challenged, and communicated.

06. Core Identification Assumptions

This lecture builds intuition for Core Identification Assumptions and ties the result to model choice, uncertainty, and action.

07. Causal DAGs and Graphical Assumptions

This lecture applies Causal DAGs and Graphical Assumptions with emphasis on diagnostics, tradeoffs, and evidence limits.

08. Confounders, Mediators, Colliders, and Selection Bias

This lecture develops Confounders, Mediators, Colliders, and Selection Bias as a practical pattern for analysis, diagnostics, and decision support.