(iPoster) Municipal Structure and Civic Engagement: Evidence from Government Social Media

• In Event: Poster Sessions: Public Administration & Political Organizations and Parties
  In Poster Session: Poster Session: Public Administration

Abstract

Civic engagement is a cornerstone of democratic governance, and in the digital age, local governments increasingly rely on social media and other online platforms to inform, consult, and collaborate with citizens. While existing research has explored contextual factors such as platform use, messaging strategies, and administrative capacity, far less attention has been paid to how formal institutional arrangements of municipal governments shape their proactive civic engagement behaviors online. As the leadership team responsible for making and implementing public decisions, the structure of the governing body, including form of government (mayor-council vs. council-manager), council size, and council composition (e.g., gender representation), may systematically influence how governments use social media to engage residents. Moreover, existing literature has mainly used self-reported surveys or simple interaction metrics, such as counts of likes, retweets, and replies, as proxies for engagement levels, which offer limited insight into the content and quality of engagement.

To address these gaps, this study investigates the intersection of municipal institutional structures and online engagement practices. It focuses on two key research questions: 1) How does municipal structure, including its form, size, and gender composition, influence local electronic civic engagement? 2) How to empirically assess governments’ civic engagement efforts through social media content? I will leverage a novel, multi-source dataset covering 1,248 U.S. cities with populations over 30,000 (2021 U.S. Census) that integrates individual-level information on U.S. municipal officials collected from city websites, verified city posts on X (formerly Twitter), and city-level demographic variables. Specifically, to measure local governments’ civic engagement efforts online, I scraped over 400,000 X posts from verified city government X accounts during two periods (July–Dec. 2022 and Jan.–June 2025), before and after the federal presidential transition. BERTweet, a pre-trained machine learning language model for English Tweets, is then employed to conduct a partly automated content analysis to identify government Tweets that seek citizen inputs, online dialogues, and offline collaborations.

Findings from this study help deepen our understanding of how the form, size, and composition of municipal governments shape their citizen engagement efforts and further enhance democracy and political trust. This paper also makes a methodological contribution by demonstrating the application of natural language processing methods as a new tool for public administration scholars and policymakers to extract valuable data from social media posts.

Author

• Unregistered Participant