Sentiments in the COVID-19 crisis communication of German authorities and independent experts on Twitter: A sentiment analysis for the first year of the pandemic

Abstract

Background: At the beginning of the COVID‑19 pandemic in Germany, there was great uncertainty among the population and among those responsible for crisis communication. A substantial part of the communication from experts and the responsible authorities took place on social media, especially on Twitter. The positive, negative, and neutral sentiments (emotions) conveyed there during crisis communication have not yet been comparatively studied for Germany. Study aim: Sentiments in Twitter messages from various (health) authorities and independent experts on COVID‑19 will be evaluated for the first pandemic year (1 January 2020 to 15 January 2021) to provide a knowledge base for improving future crisis communication. Material and methods: From n = 39 Twitter actors (21 authorities and 18 experts), n = 8251 tweets were included in the analysis. The sentiment analysis was done using the so-called lexicon approach, a method within the social media analytics framework to detect sentiments. Descriptive statistics were calculated to determine, among other things, the average polarity of sentiments and the frequencies of positive and negative words in the three phases of the pandemic. Results and discussion: The development of emotionality in COVID‑19 tweets and the number of new infections in Germany run roughly parallel. The analysis shows that the polarity of sentiments is negative on average for both groups of actors. Experts tweet significantly more negatively about COVID‑19 than authorities during the study period. Authorities communicate close to the neutrality line in the second phase, that is, neither distinctly positive nor negative.