Jason Seawright, Northwestern University
This course extends the discussion of multi-method research designs to include contemporary tools for causal inference, involving combinations of matching methods, natural experiments, and randomized experiments with qualitative methods. We will review the mechanics and assumptions involved in each statistical tool. We will then discuss efficient designs that use case studies to test assumptions and fill inferential gaps in a causal inference justified by these methods for causal inference.
For each quantitative method, we will analyze optimal case-selection strategies, discuss special process-tracing designs for testing assumptions distinctive to that method, and explore special considerations that arise in embedding that method as a step in a process-tracing argument. We will carry out practical exercises reviewing each method, and will discuss optimal designs for a variety of real-life inferential problems.
A solid understanding of regression analysis and basic qualitative methods. Students should also have familiarity with natural experiments, instrumental variables, matching and randomized experiments, as well as familiarity with statistical packages such as R and/or Stata.