Abstract
The concept of party identification is widely thought to be of limited utility outside the United States, where partisan attachments are regarded as unstable. The authors argue that estimating the stability of party identification outside the United States requires attention to problems of dimensionality and measurement error. The authors develop a model for estimating the stability of partisanship that addresses these problems, and they apply the model to eight panel surveys drawn from Great Britain, Canada, and Germany. The results suggest that partisanship has been extremely stable in each country, with the exception of recent developments in Canada. The model and findings presented here suggest the need for refinement in the way partisanship is measured, and partisan stability assessed, in multiparty systems.
Get full access to this article
View all access options for this article.
