Designing Multi-Method Research
Wednesday, September 2, 1:30-5:30 p.m.
Hilton Franciscan C
This course provides students with an introduction to research designs that combine a qualitative and a quantitative component in the service of a single causal inference: multi- or mixed-method designs. We will discuss older “triangulation” ideas about multi-method design but focus on the newer “integrative” approach that uses one method to test the assumptions of the other. We will explore motivating ideas about causation, causal inference, and the strengths of various methods. However, the center of gravity is on considering formal multi-method research designs combining case studies with regression, matching, natural experiments, and randomized experiments. We begin with key ideas about causation and causal inference that drive contemporary statistical and multi-method thinking, centrally including the potential outcomes framework. We will discuss that framework, considering what it captures and omits from other ideas about causation. Centrally, we will discuss the way that the potential outcomes framework opens opportunities for multi-method research by specifying the assumptions needed to get causal results out of regression analysis. We then move to the central question in most discussion of multi-method research: how to combine regression-type studies with case studies. Optimal case-selection strategies will be analyzed. We conclude by considering multi-method designs that include more recent, and sometimes more credible, quantitative components: matching, natural experiments, and randomized experiments. For each design, we will look at the assumptions needed for causal inference, identify relevant case-study designs, and explore case selection.