A sentiment analysis of U.S. local government tweets: The connection between tone and citizen involvement

https://doi.org/10.1016/j.giq.2015.03.003Get rights and content

Highlights

  • Local governments still push information to followers on social media.

  • City governments are likely to adopt a neutral tone and use a push style of information sharing.

  • City governments that have a positive tone are likely to encourage citizen participation on social media.

  • Positive tone needs to be coupled with photo sharing, exclamation points, and retweets to further encourage participation.

Abstract

As social media tools become more popular at all levels of government, more research is needed to determine how the platforms can be used to create meaningful citizen–government collaboration. Many entities use the tools in one-way, push manners. The aim of this research is to determine if sentiment (tone) can positively influence citizen participation with government via social media. Using a systematic random sample of 125 U.S. cities, we found that positive sentiment is more likely to engender digital participation but this was not a perfect one-to-one relationship. Some cities that had an overall positive sentiment score and displayed a participatory style of social media use did not have positive citizen sentiment scores. We argue that positive tone is only one part of a successful social media interaction plan, and encourage social media managers to actively manage platforms to use activities that spur participation.

Introduction

Scholars and practitioners agree that providing citizens with the opportunity to participate in governance practices can increase the legitimacy of the administrative state (Arnstein, 1969, King et al., 1998, Stout, 2013, Waldo, 2007) while encouraging administrators themselves to adopt a democratic-minded ethos that appreciates citizen participation despite inherent challenges (Irvin and Stansbury, 2004, Rawlings and Catlaw, 2011). Responsive governance embodies processes, politics, and partnerships that subsequently enhance administrative decision-making through transparency and citizen engagement. In contemporary times, information communication technologies (ICTs) – especially social media – are means through which administrators can increase citizens' access to government agencies and programs (Mergel, 2013a, Mergel, 2013b).

Originally, static, one-way websites were platforms where people could engage in transactional relationships with governments, such as paying bills or filing forms (West, 2004). These one-way uses often are grouped together under the Web 1.0 moniker (Bryer & Zavattaro, 2011). More recently, however, social media technologies have emerged as key means through which government agencies at all levels are opening the doors of government, at least metaphorically, 24 h a day, seven days a week (Bryer and Zavattaro, 2011, Mergel, 2013a). Social media come in myriad forms but have in common capabilities such as instant information gathering and sharing, potential for networking, knowledge co-creation, and interactivity (Bryer and Zavattaro, 2011, Mergel, 2013a, Mergel, 2013b). As a result of these innovations, the opportunity exists to engage a significant number of individuals with varying interests in governmental affairs. Early government adopters, however, might not be taking full advantage of these interactional capabilities and thus are only increasing capacity for participation rather than meaningful citizen participation and engagement (Brainard and Derrick-Mills, 2011, Brainard and McNutt, 2010, Bryer, 2013, Hand and Ching, 2011, Mergel, 2013a, Rishel, 2011, Zavattaro and Sementelli, 2014).

As social media tools grow in popularity, it becomes important to understand how they can encourage meaningful citizen interaction (Brenner and Smith, 2013, Lutz et al., 2014). This research builds upon the emerging literature that examines social media use at the local government level (Hand and Ching, 2011, Mossberger et al., 2013, Oliveira and Welch, 2013) by incorporating a construct within technology use: sentiment analysis (Dardenne et al., 2011, Steiglitz and Dang-Xuan, 2013). The purpose of this analysis is to determine how sentiment of local government social media posts influences citizen involvement on Twitter. To do this, Mergel's (2013b) framework for social media evaluation is coupled with machine-learning sentiment analysis.

In line with extant literature (Brainard and McNutt, 2010, Mergel, 2013a, Mergel, 2013b), our results indicate that government agencies are adopting an overall neutral, informative tone via social media. We also found, however, that agencies that adopt a positive tone – and undertake activities such as retweeting information from other local agencies, responding directly to citizens on Twitter, sharing photos, and using exclamation points – are more likely to encourage citizen participation on Twitter than cities that simply share information in a push manner. In other words, simply using happy words or exclamation points are not enough in and of themselves to create citizen participation; a mix is necessary. Our findings show that administrators are not using the platforms to their full dialogic capabilities and are stopping short of creating what Bryer (2013) calls the empowered citizen, whereby administrators and citizens have equal chances to contribute ideas via social media. Put simply, government use of social media still has progress to make when it comes to creating meaningful collaboration, but our findings can show at least how to create environments for participation, which can bring us closer to collaboration.

Section snippets

Social media in public administration

Public administrators necessarily concern themselves with fostering meaningful and effective citizen participation and engagement (Arnstein, 1969, King et al., 1998, McGuire, 2006, Yang, 2005). Citizen participation is understood as citizen involvement with the administrative apparatus of government (Yang & Pandey, 2011), while engagement aligns with King et al.'s (1998) authentic participation, whereby there is “continuous involvement in administrative processes with the potential for all

Machine-learning sentiment analysis

As noted in the Introduction section, the purpose of this research is to see how, if at all, sentiment influences citizen–government involvement on social media platforms at the local government level. To do this, we couple Mergel's (2013b) framework of social media transparency, participation, and collaboration with machine-learning sentiment analysis. This section will briefly detail sentiment analysis, while Section 4 highlights Mergel's framework as a means to address the research question

Mergel's framework of social media interaction

The final building block for this study is Mergel's framework for social media interaction (Mergel, 2013b). Mergel developed the framework based upon interviews with social media directors at the U.S. federal government level, and the logic is applied herein to municipal government use of social media. We hand-coded city Twitter feeds to determine what social media style was used, then this hand coding was coupled with the machine-learning sentiment analysis to reveal the findings presented

Methods

Results of this study come from analyzing tweets to and from U.S. local government agencies. To generate a database of cities, a systematic random sample was conducted utilizing the International City/County Management Association Municipal Yearbook 2012. Starting with a random number, every 25th city was counted until the database total reached 750. Cities are from each of the U.S. Census Bureau's geographic regions to ensure representation from a broad variety of municipalities in the sample.

Findings

Proposition 1

Government social media managers will likely adopt a push strategy to share information in a one-way manner.

This first proposition aimed to examine the overall strategies public administrators were using on Twitter. Our findings confirm this proposition. To code for this, Mergel's (2013b) framework was used as detailed above. When examining the Twitter pages of all 125 cities in the database, the researchers looked for characteristics such as color scheme, language used, retweets, photos,

Discussion and conclusion

The trend toward transparency and push styles of social media use seems to be preventing government agencies from fully taking advantage of social media's tremendous interactive abilities. In this research, we were curious to understand how tone influences citizen participation on social media. While we found that an overall positive tone does more to encourage participation than a neutral or negative tone, this was not always a perfect relationship, meaning that citizen sentiment scores were

Acknowledgments

Special thanks is given to the Social Science Research Center at Mississippi State University for their cooperation and input into this research. We also thank Dr. Ines Mergel for her insightful, helpful feedback. Finally, we thank the referees for their suggestions that have improved this manuscript.

Dr. Staci M. Zavattaro is an Assistant Professor of Public Administration at the University of Central Florida. Her book, Cities for Sale: Municipalities as Public Relations and Marketing Firms, was released by SUNY Press in 2013. Her work appears in journals including Place Branding and Public Diplomacy, Journal of Place Management and Development, Administrative Theory & Praxis, and Administration & Society. Her research interests include administrative theory, place marketing, and

References (64)

  • S.M. Zavattaro et al.

    A critical examination of social media adoption in government: Introducing omnipresence

    Government Information Quarterly

    (2014)
  • S.R. Arnstein

    A ladder of citizen participation

    Journal of the American Institute of Planners

    (1969)
  • Y. Bae et al.

    Sentiment analysis of Twitter audiences: Measuring the positive or negative influence of popular Twitterers

    Journal of the American Society for Information Science and Technology

    (2012)
  • E. Boiy et al.

    A machine learning approach to sentiment analysis in multilingual web contexts

    Information Retrieval

    (2009)
  • L.A. Brainard et al.

    Electronic commons, community policing and communication: On-line police –citizen discussion groups in Washington D.C.

    Administrative Theory & Praxis

    (2011)
  • L.A. Brainard et al.

    Virtual government–citizen relations: Informational, transactional or collaborative?

    Administration and Society

    (2010)
  • J. Brenner et al.

    72% of online adults are social networking site users

    (2013)
  • T.A. Bryer

    Designing social media strategies for effective citizen engagement: A case example and model

    National Civic Review

    (2013)
  • T.A. Bryer et al.

    Social media and public administration

    Administrative Theory & Praxis

    (2011)
  • B.Y. Clark et al.

    Coproduction of government services and the new information technology: Investigating the distributional biases

    Public Administration Review

    (2013)
  • B.A. Conway et al.

    Twitter use by presidential primary candidates during the 2012 campaign

    American Behavioral Scientist

    (2013)
  • B. Dardenne et al.

    Mood and positive testing in social interaction

    European Journal of Social Psychology

    (2011)
  • K. Emerson et al.

    An interactive framework for collaborative governance

    Journal of Public Administration Research and Theory

    (2012)
  • A. Hamouda et al.

    Building machine learning based senti-word lexicon for sentiment analysis

    Journal of Advances in Information Technology

    (2011)
  • L.C. Hand et al.

    You have one friend request: An exploration of power and citizen engagement in local governments' use of social media

    Administrative Theory & Praxis

    (2011)
  • T. Heintze et al.

    Information technology and restructuring in public organizations: Does adoption of information technology affect organizational structures, communications, and decision making

    Journal of Public Administration Research and Theory

    (2000)
  • R.A. Irvin et al.

    Citizen participation in decision making: Is it worth the effort?

    Public Administration Review

    (2004)
  • S. Kim et al.

    E-participation, transparency, and trust in local government

    Public Administration Review

    (2012)
  • C.S. King et al.

    The question of participation: Toward authentic public participation in public administration

    Public Administration Review

    (1998)
  • H. Kwak et al.

    What is Twitter, a Social network or news media?

  • O. Kwon et al.

    Factors affecting the intensity of emotional expressions in mobile communications

    Online Information Review

    (2013)
  • K. Lee et al.

    Social media for socially responsible firms. Analysis of Fortune 500's Twitter profiles and their CSR/CSIR ratings

    Journal of Business Ethics

    (2013)
  • Cited by (126)

    • Sentiment analysis of public services for smart society: Literature review and future research directions

      2022, Government Information Quarterly
      Citation Excerpt :

      Active participation of citizens on social media platforms like Twitter and Facebook allows the government to gauge the public mood and strategize action plans for society's overall welfare. Sentiment valence is key to devising a media interaction plan, and the tone of expressed communication determines the direction of collaborative communication between citizens and government (Zavattaro et al., 2015). The level of citizens' participation in local governance indicates the administrative state (Stout, 2013).

    • Textual variations affect human judgements of sentiment values

      2022, Electronic Commerce Research and Applications
    View all citing articles on Scopus

    Dr. Staci M. Zavattaro is an Assistant Professor of Public Administration at the University of Central Florida. Her book, Cities for Sale: Municipalities as Public Relations and Marketing Firms, was released by SUNY Press in 2013. Her work appears in journals including Place Branding and Public Diplomacy, Journal of Place Management and Development, Administrative Theory & Praxis, and Administration & Society. Her research interests include administrative theory, place marketing, and organizational communication. Palgrave Macmillan published her book regarding cities through phases of the image in 2014.

    Dr. P. Edward French is a Professor of Public Administration at Mississippi State University and serves as a Director of the Stennis Institute of Government. He also currently serves as Editor-in-Chief of Public Personnel Management. He received his Ph.D. in Public Policy and Administration from Mississippi State University, a Master's degree in Education from The University of Virginia, a Master's degree in City Management from East Tennessee State University, and a Bachelor of Arts degree in Political Science from The University of Tennessee.

    Dr. Somya D. Mohanty received his Master's degree in Computer Science from Florida State University and his doctorate from the Department of Computer Science and Engineering at Mississippi State University. His doctoral research focuses on designing security kernels for distributed applications. Somya is currently working as a Data Scientist/Systems Architect on the Social Media Tracking and Analysis System (SMTAS) project with the Innovative Data Laboratory at the Social Science Research Center. He designs system architectures capable of handling Big Data and develops algorithms to gain insights from the data in real-time. He also contributes to the server architecture design with the use of dynamic scalable components capable of handling large data influx (Big Data). The work involved in SMTAS also includes researching current advances in social media and data mining technologies. Somya's other research interests include information/network security, cryptographic protocols, content analysis, machine learning and distributed storage architectures.

    View full text