Partisan Hostility Inhibits Spatial Voting

A long literature on spatial voting argues that voters will choose the most ideologically proximate (viable) party. Pre-existing allegiances and partisan affect, however, lead voters to project parties they like (dislike) as ideologically close (distant). The literature remains unclear as to a) Which force is stronger: spatial voting or projection, and b) When projection occurs, whether it is driven more by liking or disliking a party. Most investigations rely on either cross-sectional surveys that make it impossible to disentangle these forces, or experiments that test only one path at a time, with limited external validity. In this paper, I make use of long-term panel data from the United Kingdom and Austria to test these questions. I find that the path from partisan affect to perceived ideological closeness is as strong, if not stronger, than the path from perceived ideological closeness to partisan affect. Strongly disliking a political party strongly drives projections of ideological dissimilarity, and results in significant inaccuracy in placing parties ideologically. Conversely, strongly liking a political party results in little to no projections of ideological similarity.