Outliers: Concepts, Treatments, and Uses in Different Methodological Approaches

Abstract

Recent assessments of the vices and virtues of large N quantitative vis-à-vis small N qualitative methodological approaches (e.g. King/Keohane/Verba 1994, Brady/Collier 2004) have paid relatively little attention to the concept of outliers, or, deviant cases. This is despite the fact that outliers are at the core of seemingly unrelated topics, such as selection bias, residual diagnostics and outlier correction procedures, typologies of (comparative) case studies, concept formation, and measurement. While progress has been made in developing refined techniques and/or vocabulary to come to grips with the outlier phenomenon within different methodological camp, what is missing is an inter-methodological approach to the topic of outliers. This panel invites contributions that address questions such as: What are outliers in different methods? How should they be conceptualised and detected in methodological approaches such as QCA in which equifinality is assumed? What are their contribution to theory building, concept formation, and measurement? Discussions of specific statistical techniques, the same as overview articles are welcome. The overall aim of the panel is to assess whether there can be a unified approach to outliers across different methods or whether, instead, the concept and treatment of outliers represents one of the (few) irreducible differences between qualitative and quantitative approaches?

Title	Details
Making Sense of outliers	View Paper Details
Interpreting Fat Tails: Qualitative analysis of regression residuals	View Paper Details
Not Just Outliers: Making Sense of Exceptionalism in Cross-national Comparisons	View Paper Details