How policy-oriented emotions shape political preferences: comparing survey and brain-imaging methodology

For several decades, research about voting behaviour largely ignored the role of emotions. Instead, studies focused strongly on factors like socio-demographic characteristics, ideology, issue positions partisan identities and leader images. In the past two decades, emotions have gained much more attention. Although valuable insights have been obtained, there are several ways in which our understanding of voting behaviour – and political behaviour more broadly – can be further strengthened. The aim of this paper is to contribute to this both theoretically and methodologically. Studies of voting that focus on the impact of emotions typically analyse emotions that politicians and political parties evoke. In other words, they conceptualise emotions in terms of the political actors that are the object of emotions experienced by citizens. Emotions with respect the substance of politics – that is, policy issues that are central in political debates and decision-making – seldom feature prominently. Building on theories that concern the role of policy issues in voting behaviour (issue ownership, directional theory of issue voting) on the one hand, and psychological theories about emotions (appraisal theory, conceptualisation of emotions) on the other hand, we discuss how policy issues and emotions are central in the formation of political attitudes and guide political behaviour. Our understanding of voting behaviour can be strengthened by not only adding such a different theoretical angle, but also by taking advantage of methodological shifts. Since the introduction of the survey method, the use of large-scale questionnaires has been the primary foundation for electoral research. Especially for measuring and analysing so-called emotion episodes, the survey method has limitations. The use of brainimaging technology is one way to delve further into the physiological side of emotional responses and expand insight in the role of emotions. In order to test those ideas, we conducted an experiment in which participants watched short videos about three different policy issues (climate, immigration, health care). For each policy issue, three versions of the video were created (3x3 design). One version emphasised threat (stimulating anxiety), another emphasised blame (stimulating anger) plus a control condition. The imagery was taken from news and current affairs on television and online video material. The sound consisted of a carefully scripted text read by a professional voice actor. In the first stage, we pilot-tested the stimulus materials in a survey among a nationwide sample of voters (N=1.872), where emotional responses were measured with a 12-item PANAS. In the second stage, we mapped the physiological responses in two studies, which made use of EEG (N=) and fMRI (N=31), respectively. The empirical analyses show that participants reported different levels of anxiety and anger across the three conditions, while the nature of the policy issues also had an impact on those effects. Furthermore, several of our expectations (but not all) about differences in effects for different political parties were confirmed. These findings provide support for the proposition that the suggested theoretical and methodological advancements can contribute to our understanding of the way policy issues, emotions and party preferences are connected.