**Principal investigator(s):**
**Bryan T. Gervais**
University of Texas at San Antonio
Email: bryan.gervais@utsa.edu
Homepage: http://colfa.utsa.edu/polisci-geography/faculty/gervais
**Sample size:** 2275
**Field period:** 06/23/2021-11/22/2021
**Abstract**
Demonization of the partisan out-group is commonly identified as a source of partisan polarization. Support for this claim can be found in the scholarly literature, which links exposure to elite incivility to affective polarization (Mutz 2015, 65-70; Skytte 2020). Yet another set of findings suggests we should be skeptical of the claim that uncivil elite rhetoric intensifies the gap in feelings for in- and out-group, demonstrating that in-elite incivility can depolarize partisans (Costa 2021; Druckman et al. 2018; Frimer and Skitka 2018). I argue that whether in-elite political incivility aggravates or lessens affective polarization depends on two factors: strength of partisan identity and the presence of threat to group status. While in-elite incivility may aggravate affective polarization among strong identifiers, it may have a depolarizing effect on weak partisans. Moreover, partisan identifiers may polarize in response to in-elite political incivility when faced with a threat to group status, but these effects should be weaker absent group threat. I evaluate two hypotheses: (H1) the effect of in-elite political incivility on affective polarization is moderated by strength of partisan identity and (H2) in-elite political incivility aggravates affective polarization when the partisan out-group threatens in-group status. I employ a 2 (no threat/threat to group status) x 2 (in-elite civil negativity/in-elite incivility) survey experiment with a control group to test these claims. While the group threat manipulation failed, using an alternative measure of status threat (state-level partisan competition), I find that in-elite incivility depolarizes people high on the expressive partisanship scale in competitive contexts.
**Hypotheses**
* (H1) the effect of in-elite political incivility on affective polarization is moderated by strength of partisan identity
* (H2) in-elite political incivility aggravates affective polarization when the partisan out-group threatens in-group status.
* (RQ1) Does the effect of political incivility on affective polarization depend on whether an election is locally competitive (i.e., at the state-level)?
**Experimental Manipulations**
Subjects (excepting those in the control group) were exposed to an (invented) excerpt from a news article. In the civil negative condition, an in-group elite (an unnamed “top” Democrat or Republican) critiques the partisan out-group. In the uncivil version, several forms of incivility are added to the out-group critique; the uncivil stimuli are based on messages previously validated. In the “no threat” condition, subjects read about how campaigns are preparing for the 2022 midterm elections. In the group threat condition, elements adapted from the stimuli used by Huddy et al. (2015) are leveraged, in which an out-group partisan crows about his party’s advantages heading into the election, insulting subjects’ in-group in the process. That the in-party will lose seats in the midterms and will subsequently be dominated by the out-party, is presented as a likely scenario. The difference between the group stimuli used in this study and the original Huddy et al. treatment is that the group threat language has been strengthened.
**Outcomes**
Subjects were asked to rate both the Democratic and Republican parties on feeling thermometers (0-100) in the posttreatment stage. By taking subject’s self-reported partisanship into consideration, a measure of out-party feeling and in-party feeling was generated. The primary measure of affective polarization is the difference between out-party feeling scores from in-party feeling scores.
**Summary of Results**
The group threat manipulation failed. The failure of the manipulation may stem from the study taking place after electoral competition has passed. However, expressive partisanship and state competition moderate the effect that incivility has on affective polarization (interaction p<0.01). In low competition states among folks with weak expressive partisan identities, and in high competition states among folks with strong expressive partisan identities, incivility has a depolarization effect: it reduces the gap between feelings felt for the in-group and out-group.
That weaker partisans might depolarize when exposed to incivility and when partisan competition is less evident is not all that surprising. The most surprising result—one that is ostensibly contrary to theoretical expectations—is that affective polarization decreases among strong partisans when exposed to incivility targeting the out-group and when they are living in states with a recent history of intense and salient partisan electoral competition.
While group comparisons continue once salient electoral competition passes, the most immediate threat to group status is the perception that the in-group lacks moral superiority. When it appears that it is the in-group that is violating norms and initiating conflict, rather than reducing affect towards the out-group, attacks induce reparative actions (Maitner et al. 2006). The results suggest that when the immediacy of electoral loss is absent, the stronger one’s partisan identity and the more salient the partisan rivalry, the larger the reparative action; in-elite incivility leads the strongest of partisan identifiers to have warmer feelings towards the out-group in these contexts.
**References**
Gervais, Bryan T. “In-Group Elite Incivility and Affective Polarization: Considering the Role of Intergroup Dynamics,” presented at the American Political Science Association Annual Meeting, September 15-18, 2022, Montreal, QC.
Gervais, Bryan T. “Does Political Incivility Polarize? Considering the Role of Intergroup Dynamics,” presented at the Midwest Political Science Association Annual Meeting, April 6-10, 2022, Chicago, IL.