۱- Introduction

۲- Study 1

۳- Study 2

۴- Discussion and conclusion

References

Abstract

Acts of political violence have negative consequences for intergroup relations, peaceful democratic participation, and increase mistrust and unrest between political groups and factions within society. A growing literature points to the role of uncertainty in driving political violence, but existing studies tend to rely on student and online convenience samples or macro-level indicators of uncertainty. This paper investigates the generalizability or the relationship between uncertainty and political violence and seeks to uncover whether this relationship is homogeneous in the population or contingent on individual differences in personality. In two large samples of the U.S. adult population (total n = 4806), the relationship between uncertainty and political violence is shown to depend on the trait of openness to experience. For those with low levels of openness, there is a strong and replicable relationship between uncertainty and political violence. This is not the case for those with high levels of openness. This interaction is robust to inclusion of a range of demographics factors, and shows how the combination of low openness and high uncertainty is a high-risk mix for political violence not only in a limited part of the population, but across groups and issue cleavages.

Introduction

In the past years, there has been an uptick in instances of political violence from left-wing, right-wing and religious extremist groups. These acts of political violence impel policymakers to act, design legislation and safeguards that are meant to reduce the risk of future violence (Davenport, 2007) and create an imperative for understanding the causes of political violence in the population. At the same time, developments in political psychology have led to an increased level of knowledge of individual factors involved in violence. A recent review of the empirical evidence of the processes leading to political violence (Gøtzsche-Astrup, 2018) found that uncertainty related to an individual’s self, place in the world and future is a key antecedent, as proposed by uncertainty-identity theory (Hogg, 2014). However, existing studies rely on non-representative student samples, which limits the generalizability to explain general developments in society. The use of student pools in psychological research has often been problematized since, as a group, students are more homogeneous and less attentive participants than the adult population (Hauser & Schwarz, 2016), although studies based on student samples often show similar effect sizes as those using nationally representative populations (Buhrmester, Kwang, & Gosling, 2011; Druckman & Kam, 2010). The relationships found in student samples do not necessarily generalize or may be contingent on differences in the population. Therefore, it is important that we replicate and expand such findings to population-representative samples. The purpose of this paper is twofold. First, it aims to assess the generalizability in the relationship between uncertainty and support for political violence in the adult U.S. population using samples with diversity in sociodemographic factors and individual dispositions. Second, the paper seeks to investigate whether this relationship is heterogeneous when conditioning on personality traits. One important way in which student samples might be more homogeneous than the general population is in terms of individual differences in personality. Since personality moderates a range of political and psychological mechanisms (Mondak, Hibbing, Canache, Seligson, & Anderson, 2010), heterogeneous psychological relationships in the wider population are likely in the case of political violence also.