STA 640: Causal Inference
Fan Li
Department of Statistical Science, Duke University
Meeting times Spring 2024. Tuesday and Thursday 1:252:40pm, Old Chem 201.
Instructor Fan Li, Professor, Statistical Science, Email: fl35@duke.edu. Office hours (Old Chem 122): Thursday 1011am; Friday 2:303:30pm
Teaching Assistant Yueqi Guo (yueqi.guo@duke.edu): Monday 12noon 1pm; Hao Wang, hao.wang@duke.edu, Wednesday 12:301:30pm. All TA office hours will be in Old Chem 203B.
Textbooks No specific textbook, mostly based on the lecture notes and many papers. I highly recommend to read Peng Ding's textbook [A first course in causal inference], which follows a similar structure as the course, but with more contents, details, proofs, and importantly, code and data.
Evaluation Six problem sets and a final project. HWs are posted on Canvas.
Final Project Two options: (1) Conduct an independent project on causal inference, which can be theory, method or application; (2) Review two papers on a topic of your choice that is related to the material covered in the class. In both cases, you need to write a 5page (max) report, make slides, and upload a 5min lightening talk.
Statements (1) The material presented in the lecture notes reflect my own view and knowledge of the vast field of causal inference, which is by no means complete. All mistakes are mine. (2) If you use part of the material posted here for teaching or lecturing, please give proper acknowledgement. (3) The lecture notes will be periodically updated to reflect the trend of the field.
Acknowledgements I am grateful to Peng Ding, Alan Zaslavsky, Laine Thomas, Joey Antonelli, Fabrizia Mealli, Alessandra Mattei, Constantine Frangakis, and Georgia Papadogeorgou for sharing valuable ideas and material. In particular, I thank Fan Li at Yale Biostatistics (no joking, we have the same name :)) for writing parts of the slides on several lectures, including those on covariate adjustment, doublerobust estimators, sensitivity analysis, and sequential treatments.
Lecture Notes

Chapter 1. Introduction [slides]

Chapter 2. Randomized experiments
 Chapter 2.1: Fisher's and Neyman's mode of inference [slides]
 Chapter 2.2: Covariate adjustment in RCT [slides]

Chapter 3. Observational studies with ignorable assignments: singletime treatments
 Chapter 3.1. Outcome regression [slides]
 Chapter 3.2. Covariate balance, matching, stratification [slides]
 Chapter 3.3. Propensity score [slides]
 Chapter 3.4. Propensity score weighting: inverse probability weighting and overlap weighting [slides]
 Chapter 3.5. Augmented weighting and doublerobust estimators [slides]
 Chapter 3.6. Causal inference with multiple or continuous treatments [slides]

Chapter 4. Heterogenous treatment effects and machine learning
 Chapter 4.1. Subgroup analysis [slides]
 Chapter 4.2. Machine Learning [slides]
 Chapter 5. Sensitivity analysis [slides]

Chapter 6. Instrumental variable and principal stratification
 Chapter 6.1. Instrumental variable (IV), noncompliance in RCT [slides]
 Chapter 6.2. Posttreatment confounding: Principal Stratification [slides]
 Chapter 7. Regression discontinuity design (RDD) [slides]

Chapter 8. Multiple periods: Differenceindifferences (DID) and Synthetic control (SC) [slides]

Chapter 9. Sequentially ignorable assignments: timevarying treatments [slides]
 Chapter 10. Bayesian causal inference [slides] (a longer version can be found in the [short course])
 Chapter 11 (additional). Causal survival analysis [slides]
Some labs

Lab 1. Covariate adjustment [slides][data]

Lab 2. Propensity score analysis  binary treatments [slides][data]

Lab 3. Propensity score analysis  multiple treatments [slides][data]

Lab 4. Principal Stratification [slides][data]

Lab 5. Propensity score analysis with survival outcomes [slides][data][R functions]