This is Not US: How to Measure Affective Polarization in Multiparty Systems

In the last two decades, polarization has become an increasingly hot topic in political science. Besides discussing how much voters are polarized (if at all) (Fiorina & Abrams, 2008), a considerable part of the current debate is focusing on whether citizens are ideologically polarized, or whether the heart of the matter lies somewhere else. Recently it has been argued that an alternative indicator of mass-polarization is the extent to which partisans view each other as a disliked out-group (Iyengar, Sood, & Lelkes, 2012). The concept of affective polarization finds its roots in social psychology, and in the idea that humans have two competing social needs: inclusion and differentiation, meaning that typically an individual wants to fit in but, at the same time, does not want to disappear within the group. The direct consequence is that humans are motivated not only to form groups, but to form exclusive groups (Brewer, 1991): us and the others. This reasoning is easily exported into politics. One could argue that citizens are not polarizing ideologically and becoming more extreme in their opinions, but that they are simply processing a differentiation, becoming more extreme in their attitudes towards those from the other side of the political spectrum. Despite being based on an intrinsic human inclination, this attitude can be harmful for democracy if it becomes so drastic that sentiments of distrust and dislike hamper the dialogue and undermine even basic acceptance and compromise. It appears clear that to have a precise picture of this phenomenon is pivotal. However, most of the current research on affective polarization has taken place in the United States, a society with a very different social setting from Europe, and a different sense of in- and out-groups, heavily influenced by having a two-party system. In multiparty systems, however, people can be more polygamous in their party-allegiance, since there is not such a strong, obvious division between two groups. When translating this at the representational level the difference becomes clear: in multiparty systems cooperation among sometimes very different parties is often necessary to form coalitions, and different parties might be ideologically close and/or have similar opinions on certain issues. Affective polarization might, therefore, show different patterns in multiparty systems. Currently, different instruments and operationalizations are employed to measure affective polarization, but all relying on US-based context. In this paper, we aim at developing strategies and measurements that can effectively address the peculiarities of multiparty systems. In this pioneering study, we use the case of the Netherlands to test multiple measures of affective polarization. We also use these tools to substantively investigate the presence of different in- and outgroups and to analyze the dependence of affective polarization on ideology.