a) What is a counterfactual, and what is the fundamental problem of causal inference? How does random assignment of a treatment address this problem?
b) Describe David Freedman’s views on whether applying econometric techniques (e.g., regression analysis) to non-experimental (social science) data adequately addresses this problem. Include Freedman’s views on the role of “research design” and discuss how John Snow’s “cholera” study fits into Freedman’s critique of the overuse of statistical techniques in the social sciences.