This lecture connects Mediation Analysis to mechanisms, assumptions, and uncertainty in causal explanation.
Advanced Topics in Causal Inference
Advanced Causal Inference
Lecture Notes
This course collects the advanced topics that separate routine effect estimation from mature causal practice. The focus is on the complications that appear once projects move beyond a clean single-treatment, single-outcome, no-missingness setting: mechanisms, principal strata, missing data, measurement error, generalization, interference, panels, discovery, Bayesian workflows, and AI-system complications.
The objective is to help readers diagnose where standard causal workflows bend or break. By the end of the course, a reader should be able to explain which complication is present, how it changes the estimand or identifying assumptions, which diagnostics are useful, and how to communicate uncertainty when the design remains credible despite imperfections.
Lecture Sequence
This lecture develops Principal Stratification with examples that make assumptions, diagnostics, and interpretation visible.
03. Missing Data and Causal Inference
This lecture uses Missing Data and Causal Inference to clarify the analyst’s question, evidence, assumptions, and decision implications.
This lecture studies measurement error as a threat to estimands, identification, and uncertainty interpretation.
05. Transportability and External Validity
This lecture frames Transportability and External Validity as a decision problem and asks what evidence can be trusted, challenged, and communicated.
06. Interference and Spillovers
This lecture builds intuition for Interference and Spillovers and ties the result to model choice, uncertainty, and action.
This lecture applies Panel Data Complications with emphasis on diagnostics, tradeoffs, and evidence limits.
08. Causal Discovery, With Caveats
This lecture develops Causal Discovery, With Caveats as a practical pattern for analysis, diagnostics, and decision support.
This lecture connects Bayesian Causal Inference to mechanisms, missingness, measurement, transport, interference, and uncertainty-aware reporting.
10. Causal Inference with LLM and AI Systems
This lecture develops Causal Inference with LLM and AI Systems with examples that make assumptions, diagnostics, and interpretation visible.