Mixing Methods: A Bayesian Approach for Integrating Inferences from Qualitative and Quantitative Research
15. Juli 2015
IACM Lecture Series
Prof. Macartan Humphreys (Department of Political Science, Columbia University)
Some scholars have proposed that quantitative and qualitative approaches to causal inference can be unifed under an essentially correlational framework. Other scholars have argued that such a conceptualization does not exploit the non-correlational information that is central to process tracing.
We formalize an alternative unification – Bayesian Integration of Quantitative and Qualitative data (BIQQ) – that allows researchers to draw causal inferences from combinations of within- and cross-case observations given prior beliefs about causal effects, assignment propensities, and the informativeness of different types of within-case observations.
We illustrate with two applications to substantive issues that have received significant quantitative and qualitative treatment in political science: the origins of electoral systems and the causes of civil war. Finally, we demonstrate how the framework can yield guidance on research design, presenting results on the optimal combinations of within-case and cross-case data collection under different research conditions.
Macartan Humphreys and Alan Jacobs (forthcoming): Mixing Methods: A Bayesian Approach, American Political Science Review
Wed, 15 July 2015, 5 pm
University of Konstanz, V 1001 (Senatssaal)
Contact
Dr. Martin Welz iacm[at]uni-konstanz.de