Between 2013 and 2015, I worked with Jim Speckart and the Social Science Research Institute (SSRI) at Duke to create a series of videos on causal inference. These are nontechnical explanations of the basic methods social scientists use to learn about causality. They're aimed at high school seniors or 1st year undergraduates, and are quite short---around 2 to 5 minutes on average.
Click on any topic below to expand and view the video. Here I give my recommended viewing order, but you can also pick and choose whatever sounds interesting. These are free for everyone to use, so please check them out!
1. Introduction to causality
1. Introduction 2. Measurement 3. Describing data 4. Correlation versus causation 5. Average treatment effects 6. Unit level treatment effects 7. Conditional average treatment effects 8. Counterfactuals 9. Confounders 10. Statistical versus causal inference 11. How to read empirical papers
2. Experiments
1. Controlled experiments 2. Randomized experiments 3. Design of the Oregon Health Experiment 4. Noncompliance in experiments 5. Reading ATEs and CIs: Depression in the Oregon Health Experiment 6. Cholesterol in the OHE 7. Survey nonresponse 8. Survey nonsponse in OHE 9. Introduction to Perry Preschool 10. Perry Preschool: Educational Attainment 11. Perry Preschool: Effects on crime 12. Perry Preschool: Lifetome cost of crime 13. Multiple testing and sample size in Perry Preschool 14. The point of statistical inference 15. Randomized controlled trials 16. Important experimental issues we've ignored 17. Common issues in experiments 18. Difficulties in implementing experiments 19. Lifetime outcomes of treatments 20. Spillovers in experiments 21. Two kinds of natural experiments 22. Benefits of natural experiments 23. Finding data from natural experiments 24. Justifying as-if randomization 25. Analyzing natural experiments 26. Analyzing natural experiments: Effect of property rights on child health 27. Effect of property rights on teenage pregnancy 28. London Cholera outbreak: Introduction 29. London Cholera outbreak: A natural experiment 30. Bounds analysis for missing data 31. Causal inference issues in lab experiments
3. Regression and causality
1. Introduction to regression 2. Basic elements of a regression table 3. Economic development and property rights 4. Using regression to get causal effects: Unconfoundedness 5. Ordinary least squares (OLS) 6. Defining the average effect of treatment on the treated (ATT) 7. How to compute ATE under unconfoundedness, and what not to do 8. Matching methods 9. The lifetime earnings of veterans and nonveterans 10. The unconfoundedness assumption and lifetime earnings of veterans 11. The effect of military service on lifetime earnings: Results 12. The credibility of the unconfoundedness assumption
4. Instrumental variables
1. The logic of instrumental variables 2. The visual logic of instrumental variables 3. IV in action: Education and wages (graphs) 4. IV in action: Education and wages (tables) 5. Some IV terminology 6. The three IV assumptions 7. Refutability and non-refutability of the IV assumptions 8. Indirect inference 9. Two stage least squares (2SLS) 10. Using IVs to solve noncompliance in experiments 11. Using IVs in the Oregon Healthcare Experiment 12. Weak instruments 13. Property rights and economic development: IV analysis 14. The settler mortality instrument 15. Where do instruments come from? 16. Property rights and market beliefs 17. What are we actually getting with IV? 18. Defining LATE: The local average treatment effect 19. The no defiers assumption 20. Computing LATE, part 1 21. Computing LATE, part 2 22. Computing LATE, part 3 23. The pros and cons of LATE 24. ATEs, CATEs, and LATes: What's the difference?
5. Panel data
1. Does the death penalty reduce homicides? 2. The common trends assumption 3. The effect of immigration on employment 4. Graphical analysis of common trends 5. The effect of the minimum wage on employment 6. The effect of the minimum wage on employment: Main results 7. The effect of the minimum wage on employment: Additional results 8. The effect of the minimum wage on employment: Even more results 9. Panel data terminology 10. Three kinds of panel data 11. Individual fixed effects and time varying treatments 12. The random assignment of changes 13. Does posting calorie counts lower calorie consumption? 14. Heterogeneosu treatment effects: The switchers
6. Regression discontinuity
1. Introduction to RD 2. Thistlethwaite and Campbell continued 3. Computing treatment effects at the cutoff 4. School quality and house prices 5. School quality and house prices: Results 6. School quality and house prices: Economic interpretation 7. Fuzzy regression discontinuity 8. Fuzzy RDD and Swiss relgion
7. Modeling
1. Basics of modeling behavior 2. Discrete choice analysis 3. Confounders in discrete choice analysis 4. How do people choose healthcare plans? 5. Senior citizens' preferences for health insurance plans 6. Human produced CO2 and climate change: When experiments are impossible 7. The impact of government stimulus
8. Conclusion
1. Recap 2. The power of assumptions 3. Balacing data and assumptions 4. Topics we skipped 5. Which causal inference method is the best?