How elite US and German universities “speak” to their stakeholders during crisis: Computer-aided text analysis of COVID-19 tweets

• In Event: Highlighted Session: Leading during a global crisis: Institutional responses to the COVID-19 pandemic

Proposal

Introduction
COVID-19 has presented the world with never-before seen global challenges, effecting all areas of our modern-day lives. Using computer-aided text analysis (CATA) methods with the Linguistic Inquiry and Word Count (LIWC) software, the current comparative study examines tweets of elite U.S. and German universities through the lens of agenda setting and corporate social responsibly (CSR) as a managerial tool in crisis communication for the period January – May 2020 to produce an understanding of organizational communication of how universities ‘speak’ to their stakeholders on both sides of the Atlantic.

Theoretical Framework

The media’s influence on the “focus of public attention” (McCombs & Shaw, 1972) consists of displayed content and attributes as a representation of “what journalists and, subsequently, members of the public have in mind when they think and talk about news objects” (McCombs & Valenzuela, 2007, p. 47). The theory of agenda-setting not only explains how the media sets the stage for public discourse, but also how it sets the tone.

The emergence of so-called stakeholder media, like blogs, websites, or social media (Voinea & Kranenburg, 2017) adds the dimension of “agendamelding”. Stakeholders as users are actively involved in the following, sharing and creating of agendas (McCombs et al., 2014). This presents organizations with the opportunity to directly communicate with their audience (Albalawi & Sixsmith, 2015).

A pandemic shutting down most non-virtual human interactions renders successful organizational communication crucial (Bates, 2015). With few studies on social media crisis communication (Chen et al., 2020; Damiano & Catellier, 2020; Rao et al., 2020; Yu et al., 2020), there is a dearth of studies focusing on higher education institutions (HEIs).

Twitter as one of the most-used forms of social media (Buschow et al., 2014; Cozma & Chen, 2013) and plays a significant role in organizational communication and agenda-setting (Hunter et al., 2013). Social media allows for “sharing opinions, insights, experiences and perspectives with others” (Marken, 2007, p. 10) on a global scale, which effects stakeholders’ perceptions of the organization. The rising interest in research on audience-tailored crisis communication (Pyle, 2018; Sellnow et al., 2015), brings about questions of the corporate social responsibility (Ham & Kim, 2019). Within CSR programs, public philanthropy is the dominating domain. Issues of health and safety as part of public philanthropy (Hou & Reber, 2011) become especially crucial in times of crisis. In order to developing a comparative understanding of the way organizations (HEIs) communicate with their stakeholders on Twitter during COVID-19 the current study will answer the following questions:

RQ1: How is COVID-19 presented in the tweets from US and German universities ?
RQ2: What is the tone of the tweets in the US and in Germany? Can we observe any isomorphism?
RQ3: Do the tweets differ from German and US universities over time (prior and during the pandemic)?
RQ5: What tweets from German and US universities received more engagement?

Methodology
Computer-aided text analysis (CATA) uses computer algorithms to measure “the beliefs, cognitions, and emotions of individuals as reflected in their narratives and written texts” (Short et al., 2018, p. 45). This study employs LIWC software (Pennebaker et al., 2015b) to analyze the official Twitter updates (tweets) of 100 elite universities (50 U.S. and 50 German). Twitter and LIWC have been used in a number of research contexts, from elections (e.g., Tumasjan, Sprenger, Sandner, & Welpe, 2011) to the European refugee crisis (Pope & Griffith, 2016).

For this study, the public Twitter API will be used to download all tweets published by the 100 focus universities during the four months surrounding the March 11, 2020 pandemic declaration. Tweets will be categorized as either pre-pandemic (January 11- March 10) or during pandemic (March 11- May 10), based on the creation date, and as either pandemic-related or other-related, based on the existence of COVID-19-related words in the tweet content. Individual universities will be the unit of analysis.

During a pilot test of 30 universities, we analyzed: 7,199 pre-pandemic and 8,571 during-pandemic tweets; 2,938 pandemic-related and 12,832 other-related tweets; and, 10,185 U.S. and 5,585 German tweets (n=15,770 for the pilot study). Tweets of each university were grouped into documents based on period and topic. Using LIWC, each document was analyzed for sentiment (positive and negative emotion), emotional tone, analytical thinking, clout, authenticity, time orientation, and personal concerns (Pennebaker et al., 2015a). In addition, public engagement with tweets was measured (i.e., Twitter likes and retweets), giving an indication of the effectiveness of different communication approaches. The data were then statistically-analyzed to determine differences in Twitter use and their reactions between the U.S. and German universities.

Limitation and significance of the study
Though the study examines every tweet within the given time period, it only includes 50 universities from each country, based on their rankings. CATA scores are approximations of actual constructs (e.g. positive emotion) so it is recommended future research examines the tweets using additional methods, such as qualitative thematic analysis. Public perception and a uniquely diverse demographic (Hare, 2009) pose additional limitations. This research nevertheless provides insight into an extensive ongoing public health crisis and how universities communicate with their stakeholders.
The comparison of linguistic data gained through LIWC allows for the description and analysis of organizational behavior on the social media platform during the exceptional global COVID-19 crisis in HE. The results can inform future organizational managerial practices and decisions on how HEIs successfully use social media as a tool in crisis intervention. This study will shed light on the imperative of having policies in place for organizational communication via social media prior to a crisis to improve risk and crisis communication efforts. Social media interaction and the organizations’ desire for information control need to be balanced in order to most effectively communicate with stakeholders and the public. This study serves as a starting point for understanding HEIs’ conversations in the early stages of COVID-19 in the U.S. and Germany in elite universities and may foreshadow future risk perceptions and behaviors associated with COVID-19. We suggest that similar studies be conducted for other types of HEIs and organizations.

Authors

• Antigoni Papadimitriou, Western Kentucky University

• Kevin P Taylor, STETSON UNIVERSITY

• Sarah Schiffecker, Texas Tech University