Perceived Party Placements and Uncertainty on Immigration in the 2017 German Election.

Abstract

Almost all national election studies contain policy scales that are intended to measure where respondents perceive parties or candidates on central campaign issues. These placements form the basis for models of survey responses, party perceptions, and voter choice. It is well known that the placements might be affected by uncertainty. We use the finite mixture model ‘BetaBin’ to study response patterns to party placements on policy issues. The model consists of a placement part and an uncertainty part. Whereas the placement part of the model accounts for lower and higher placements on the ordinal scales, the uncertainty component accounts for tendencies to locate the parties on the middle or at the extremes of the policy scales. We use the 2017 German national election and apply the model to the immigration issue. Our results demonstrate that uncertainty strongly influences the respondents’ perceptions of most parties. Neglecting this structure leads to worse models as indicated by performance measures.