Incivility, What Works? A Conjoint Experiment on the Effects of Incivility

• In Event: Incivility: Understanding and Fixing Disordered Discourse

Abstract

The political rise of politicians whose style is quintessentially uncivil has inspired a growing body of research assessing whether political incivility – i.e., engaging with political opponents with disrespect – is an effective political strategy. However, this literature is only emerging; it still provides mixed findings and is primarily US-centric. We regard incivility as a perceived norm violation sensitive to the flexible nature of social interactions. We argue that to draw conclusions on its consequences, we must consider the joint role of incivility’s audience, sponsor, and contextual characteristics. Based on this, we employ a conjoint experimental design on data collected from online samples of US (N = 2000; February 2023) and Dutch (N = 2000; April 2023) citizens to test which set of attributes – i.e., audience, sponsor, and contextual characteristics – leads to positive evaluations of uncivil candidates. After answering questions about their profile, respondents are presented with a fictitious press release about a debate between two candidates. During the debate, candidate A employs incivility against candidate B. This press release is drafted in such a way that nine attributes are manipulated: (i) the type of incivility (insults, sarcasm, shouting), (ii) the arena (parliament, television, social media, political rally), (iii) the direction of the uncivil attack (ingroup towards outgroup vs. outgroup towards ingroup), (iv) the sponsor and (v) target’s gender (male vs. female), (vi) the sponsor’s extremism (moderate vs. extremist), (vii) the issue being discussed (most vs. least important issue; assigned based on respondents’ previous answers), (viii) the injunctive normative context (incivility is socially approved vs. socially disapproved), and (ix) the descriptive normative context (incivility is common vs. uncommon). After exposure to the treatment, respondents are asked to evaluate candidate A in terms of their overall likeability, and their agency (competence and assertiveness) and communion (warmth and morality) attributes. Each respondent repeats this task three times with different but comparable versions of the same treatment. The random assignment of attributes’ levels in each task and for each respondent will allow us to assess the absolute effects of each attribute independent of each other and the relative effect of their interactions.

Authors

• Chiara Vargiu, Univertiy of Lausanne, Institute of Political Studies

• Alessandro Nai, University of Amsterdam

• Diego Garzia, University of Lausanne

• Frederico Ferreira da Silva, University of Lausanne