Restructuring Issue Space: Unifying Proximity and Directional Models

Abstract

In this paper I unify the proximity and directional models. The controversy surrounding these two models is ongoing. To resolve this controversy, I propose a new unified model, Elastic Proximity Model. This model does not attempt to adjust utility functions or candidates'/parties' positions as existing unifying models, but it restructures voters' issue spaces at the recognition level. The previous models assume that distances between two adjoined points are constant, but that of the new issue space are elastic. To confirm the elasticity, I estimate elasticity parameters using data from a survey experiment. The results indicate that voters have another issue space, which is different from an issue space measured in observational data. Furthermore, I show that the new unified model performs better than not only the traditional models but also the existing unified models.

Keywords: Issue Voting, Proximity Model, Directional Model, Elastic Proximity Mode, Experimental Method