Lectures: This course introduces key ideas in causal inference, experimentation, and design based inference. We introduce the potential outcomes framework, methods for random assignment, and strategies for inference that are justified by assignment. The course also touches on practical aspects of experimental design including ethical issues, transparency issues, and design formalization and diagnosis. Lectures are based on these slides.
Books: There will be no textbook, but the best book to read for preparation is Gerber and Green's Field Experiments. As well as going through details of design and analysis of experiments it gives an excellent introduction to causal inference. Two complementary books are Imbens and Rubin which goes further on the statistics, and Glennerster and Takavarasha which goes further on the design and implementation side.
Articles: Read these:
- Lewis Counterfactual dependence and time's arrow
- Holland: Statistics and Causal Inference
- Imai, King & Stuart Misunderstandings between experimentalists and observationalists about causal inference
Programming: I would like you to know some R. You can get by without knowing R but knowing a little R will help a ton. R is wonderful for analysis but it is also wonderful for simulation. I will try to make sure all the figures and examples are coded up in R which will make it easy for you to play with them. I am also hoping that we can fire up DeclareDesign and use it to demonstrate a set of design and analysis principles.
- Resources for learning R: http://www.r-bloggers.com/how-to-learn-r-2/
- DeclareDesign (Installing DeclareDesign -- make sure your R is up to date)
Duration: This short course is taught over four 2-3 hour sessions with problem sets. Prepared originally for the Essex summer school on experiments: