How can spatial models of vote choice contend with rise of populism? The previous decades have seen the rise of ideologically agnostic politicians employing political strategies deliberately aimed at inducing emotional reactions, most often anger. The growth of global populism is reflective of this change. Many modern politicians eschew policy discussions in favor of demonizing their political opponents and supporters. Political success is often determined by how virulent, and thus mobilized, the candidate’s base is, not on effective governance or policy outcomes. Factors such as rhetorical style and candidate attributes are increasingly decisive in contemporary elections. These trends pose an intractable problem to existing spatial models of vote choice. Building on the existing spatial modeling literature, I develop a revised spatial model of politics that incorporates non-policy considerations, the credibility of party offerings, and a novel treatment of issue importance. I test this model on recent election results from Argentina, Australia, and the United States through Bayesian logistic and multinomial logistic approaches. I demonstrate the comparative utility of this approach in predicting vote choice over partisan identification and traditional spatial models, and the comparable performance to powerful indicators such as feeling thermometers and approval ratings. These results suggest an improvement in our understanding of the determinants of vote choice and may yield strategies to bolster democracy against populist threats in an increasingly unstable era.