ÐÂÏã¸ÛÁùºÏ²Ê¿ª½±½á¹û

The module's main objective is to provide students with an introduction to the rapidly growing field of causal inference. Increasingly, social scientists are no longer willing to establish correlations and merely assert that these patterns are causal. Instead, there is a new focus on design-based inference, designing research studies in advance so that they yield causal effects. This module discusses the nature of causation in the social sciences, and goes on to look at some of the most popular research designs in causal analysis, including experiments (also known as randomised control trials), natural experiments that we can analyse with instrumental variables and regression discontinuity techniques, as well as causal inference over time using the methods of difference-in-differences and synthetic control. We will also evaluate ‘observational’ methods -- regression and the closely related technique of matching -- from the standpoint of causal inference. This module has a hands-on, practical emphasis. Students will learn to design effective studies and implement these methods in R, and will become critical consumers and evaluators of cutting-edge research, able to read and evaluate original journal articles. Examples will be drawn from economics, political science, public health and public policy.

Prerequisites (for PIR who have completed POLS0083):
- be reasonably confident with R (e.g., you understand how R works, know how to load data, run regressions in R)
- be familiar with concepts and terms such as statistical and substantive significance, hypothesis tests, p-values, potential outcomes
- be comfortable interpreting regression coefficients
Ìý