Rival Strategies of Validation: Tools for Evaluating Measures of Democracy

Abstract

The challenge of finding appropriate tools of measurement validation is an abiding concern in political science. This article considers four traditions of validation, using examples from comparative research on democracy: the levels-of-measurement (LoM) approach, structural-equation modeling with latent variables (SEM-L), the pragmatic tradition, and the case-based method. Methodologists have sharply disputed the merits of alternative traditions. We encourage scholars — and certainly analysts of democracy — to pay more attention to these disputes and to consider strengths and weaknesses in the validation tools they adopt. An online appendix summarizes the evaluation of six democracy data sets from the perspective of alternative approaches to validation. The overall goal is to open a new discussion of alternative validation strategies and their role in evaluating measures of democracy.

Keywords: validity, measurement, democracy, methodology, multi-method, cross-national research, comparatice case studies, structural equation modeling, level of measurement

JEL Classification: P50, C10, C80, C81, C82