Engagement with incivility in tweets from and directed at local elected officials

Although social media have created novel opportunities for the public and local elected officials representing city or state governments to interact, concerns have been raised about the tenor of their discourse. We used machine learning in this project to identify the presence of incivility in tweets (N = 38,218) made by and directed at local officials representing a metropolitan area over an 18-month period and examine its implications. Compared to tweets made by local officials, incivility was 2.20 times as likely in tweets from the public directed at officials. Incivility was more likely to generate audience engagement, however, when used by officials but not by the public. Whereas the rate of retweets received by local officials was 2.10 times as high when they used incivility compared to when they did not, the retweet rate for the public was 0.57 times as high when their tweets directed at officials contained incivility.