Advanced Methods for The Rest of Us
A Friendly Introduction to 'New' Tools in Quantitative Political Science
Empirical political science has become increasingly quantitative and technical, and the profusion of “advanced quantitative methods” can look impossible to parse, let alone replicate. The aim of this course is to show that this appearance is wrong. Students learn to understand what these models are trying to do, replicate existing empirical work, and apply the tools to their own questions.
The course covers experimental methods, regression discontinuity designs, instrumental variables, and panel and multilevel models, grounded in the fundamentals of causal inference (directed acyclic graphs and the potential outcomes framework) and current best practice (preregistration, power analysis, simulation). A running applied example uses European Social Survey data on inequality and disadvantage in Europe.
The course is estimand-first: question, theory, and DAG come before any statistical procedure, simulation is used as a universal tool to understand what each method can and cannot recover, and every method is treated as a tool with specific strengths and weaknesses rather than a default.