Playing Nice: Modeling Civility in Online Political Discussions

Abstract

This article explores a way to promote civility in online political discussions through modeling behaviors. An online experiment (N = 130) was conducted to investigate the influence of civil and uncivil political discourse on participants’ mode of discussion and participatory intentions. Results show that those who were exposed to civil discussion demonstrated a civil mode of discourse in their own comments and reported a higher level of willingness to participate in discussions, compared to those who were exposed to uncivil discussion. This study illuminates the possibility that civil participants may have more control over the online environment than they think.

Keywords:

• Civility
• Political Discourse
• Social Cognitive Theory

Notes

The issue of gun control was chosen as a topic of discussion due to a growing national debate at the time of this study following the mass shooting at Newtown, Connecticut, in December 2012.

In terms of political ideology, 9% of participants identified themselves as very conservative, 31.5% as conservative, 35% as middle-of-the road, 13.8% as liberal, 0.8% as very liberal, and 10% as other.

The experimental stimuli used in this study are available upon request.

The civil discourse version was consistently perceived as more respectful (t = 6.25, p < .001), calm (t = 5.70, p < .001), and friendly (t = 5.09, p < .001), while the uncivil version was perceived as more rude (t = 10.10, p < .001) and emotional (t = 3.58, p < .001) by participants.

Level of civility examines the general tone of participants’ comments in three levels—uncivil, neutral, and civil. Comments that include an element of incivility (i.e., name-calling, dismissive language, and insult) were coded as “uncivil,” comments that are devoid of incivility were coded as “neutral,” and comments that acknowledge different viewpoints in a respectful manner were coded as “civil.” If comments contained elements of both civility and incivility, they were coded as “uncivil” due to the presence of incivility. Interactivity indicates if participants referred to other discussants by name in their comments. Comments that address a specific discussant by name without an element of incivility were coded as 1; all others were coded as 0. Political label illustrates the use of political labels by participants. Comments that include the terms liberal or conservative were coded as 1; comments without political labels were coded as 0. The level of civility examines the general model-following behaviors of participants. Interactivity measures the use of a specific language marker that appeared only in the civil experimental stimulus. Likewise, political labels indicate the use of a specific language marker that appeared only in the uncivil experimental stimulus.

Meta-communication investigates whether participants commented on the overall tone of the discussion. Comments that address the tone of the discussion were coded as 1, all others as 0. Expression of agreement illustrates whether participants explicitly expressed agreements with other discussants. Comments that clearly indicate agreements with others were coded as 1; all others were coded as 0. Additional perspectives examines whether participants offered an additional perspective on the issue that was not discussed in the experimental stimuli. Comments that offer an additional perspective were coded as 1; all others were coded as 0.

Inter-coder reliability (Cohen's Kappa) for each variable: level of civility = .80, interactivity = .97, political labels = 1.0, meta-communication = .81, expression of agreement = 1.0, additional perspective = .95.

Additional information

Notes on contributors

Soo-Hye Han

Soo-Hye Han (PhD, University of Texas at Austin, 2008) is an assistant professor of Communication Studies at Kansas State University.

LeAnn M. Brazeal

LeAnn M. Brazeal (PhD, University of Missouri-Columbia, 2002) is an assistant professor of Communication at Missouri State University.