Comparing Methods: Possible Solutions for the Small-N Problem

Abstract

There is a considerable disagreement in the methodology community about the appropriate use of methods and modelling in social science research. The two poles of the debate are represented by the seminal work of King et al. (1994) and Ragin (1987, 2000). At the heart of this debate is the so called small-n problem. Much comparative research suffers from the fact that there are too many cases for in-depth case studies, yet not large enough to apply advanced statistical methods commonly used in large-n research designs. This panel shall contribute to this ongoing debate by comparing methods that have been used to solve the small-n problem. More specifically: • How can the methodological gap between King et al. (1994) and Ragin (1987, 2000) be bridged? What kind of diagnostic tools exist to detect configurational or multiplicative relationships? • What kind of advanced statistical methods exist that try to remedy the conceptual problems of linear additive regression analysis? • What kind of advanced statistical methods exist that try to solve the small-n problem? • How can the fs/QCA methodology be improved to contribute to small-n research questions when there are many variables? • How can the fs/QCA methodology be applied to midsize-n research designs?

Title	Details
What does the Government say, and why? A probabilistic QCA-based study	View Paper Details
Large-n QCA or logistic regression? A methodological experiment to explain formation of ethnic minority parties	View Paper Details
Measuring Policy Effects - Multilevel Analysis in Comparative Political Research	View Paper Details