Party-System Characteristics and People’s Positioning on the Left-Right Spectrum

• In Event: Poster Session: Public Opinion and Comparative Politics

Abstract

How useful are survey questions that ask respondents to place themselves on a left-right political spectrum? Do these questions resonate with everyone equally? Do they make sense in places without a tradition of politics framed in terms of left and right? Previous authors have tackled similar questions by predicting people's placement on the scale via linear regression, and then focusing on the variance of the unexplained residuals as an outcome of yet another model, a method known as heteroscedastic regression. We improve on previous work in several ways. First, we have a broader coverage of countries and time periods, as well as much more fine-grained data on party system structuration. Second, rather than specifying a model upfront, we use a random forest model to identify the factors correlated with left-right placement in each country. Finally, given the downsides of heteroscedastic regression, we suggest alternative empirical strategies for understanding response variability to this question.

Authors

• Pilar Manzi, Northwestern University

• Maximilian Weylandt, northwestern university