Information and Ideological Structure in Spatial Voting

Abstract

This article aims at unifying the theory of spatial voting and the theory that is variously called conceptualization, information, or sophistication. Following Downs's early insights on uncertainty as well as recent developments in both literatures, I argue that it is of critical importance that spatial voting models explicitly incorporate information effects. For this purpose, I develop a heteroskedastic probit model that allows for the specification of information heterogeneity. This model is applied to the Taiwan Election and Democratization Study's 2004 post-presidential election survey data. In 2004, Taiwan’s political landscape was dominantly defined by the Green vs. Blue ideological cleavage, and the candidates were perceived as taking divergent positions. This article investigates the effects of information and activism on the spatial structure and their implications on candidates' strategies. My findings confirm the existence of these effects on voter uncertainty in the framework of spatial analysis.