276 | nº 38, pp. 275-292 |January-June of 2024Analysis of sentiment in the European Central Bank’s social media activity during the Covid-19 pandemic and Ukraine War...ISSN: 1696-019X / e-ISSN: 2386-3978doxa.comunicaciónReceived: 10/08/2023 - Accepted: 17/11/2023 - Early access: 04/12/2023 - Published: 01/01/2024Recibido: 10/08/2023 - Aceptado: 17/11/2023 - En edición: 04/12/2023 - Publicado: 01/01/2024Abstract:To assess and comprehend the communication strategy used by the European Central Bank (ECB) throughout the Covid-19 pandemic and war in Ukraine, this study considers the crisis communication tactics of the ECB. e study analyses 11,781 tweets from the ECB's ocial Twitter account using sentiment analysis to look at how the Bank interacts with the public and how it uses adaptive communication strategies in these challenging times. is research innovatively employs the AFINN and NRC sentiment analysis methods to evaluate the ECB's crisis communication. NRC's use of a comprehensive lexicon for emotional identication complements the pre-computed sentiment ratings of AFINN. e ECB's use of positive, clear communication built public trust and underlines sentiment analysis's utility in strategic crisis communication. e study's detailed investigation oers insights for improving crisis communication plans for better public engagement and economic stability.Keywords: European Central Bank; crisis communication; sentiment analysis; social media engagement; public trust.Resumen:Para evaluar y comprender la estrategia de comunicación utilizada por el Banco Central Europeo (BCE) a lo largo de la pandemia del Covid-19 y la guerra en Ucrania, este estudio examina las tácticas de comunica-ción de crisis del BCE. El estudio analiza 11.781 tweets de la cuenta o-cial de Twitter del BCE utilizando el análisis de sentimientos para obser-var cómo interactúa el Banco con el público y cómo utiliza estrategias de comunicación adaptativas en estos tiempos difíciles. Esta investigación emplea de forma innovadora los métodos de análisis de sentimientos AFINN y NRC para evaluar la comunicación de crisis del BCE. El uso por parte de NRC de un léxico exhaustivo para la identicación emo-cional complementa las calicaciones de sentimiento precalculadas de AFINN. El uso por el BCE de una comunicación positiva y clara generó conanza en el público y subraya la utilidad del análisis de sentimientos en la comunicación estratégica de crisis. La investigación detallada del estudio ofrece ideas para mejorar los planes de comunicación de crisis con el n de mejorar la participación pública y la estabilidad econó-mica.Palabras clave: Banco Central Europeo; comunicación de crisis; análisis de sentimientos; participación en redes sociales; conanza pública.Tariq Tawfeeq Yousif Alabdullah. Associate professor at University of Basrah, from Iraq and he has PhD in accounting. He got his PhD from Malaysia from Universiti Sains Malaysia (USM) from School of Management. Based on the Global AD Scientic Index, and within his major of accounting, nance, management and economics, He has got No 1 at the level of Iraqi universities , No. 6 in Asia and No 57 in the world, the link below shows this ranking: https://www.adscienticindex.com/scientist/tariq-tawfeeq-yousif-alabdullah/291502 He participated and presented 30 international conferences in Europe and in the USA. Some of the conferences were in Harvard University, Oxford University, University of Texas, and University of Rome. He is a member in the editorial boards in more than 70 international journals and conferences in Europe and USA.University of Basrah, Iraq[email protected]ORCID: 0000-0003-1494-69831. IntroductionAs we live in the digital age, social media has become an integral part of our lives, drastically reshaping communication dynamics worldwide. Its inuence is particularly felt in public communication, where institutions have had to adapt to this fast-paced, interactive communication landscape, particularly in crisis times (Bjola & Zaiotti, 2020). is paper seeks to examine the interplay between sentiment analysis in social media and the public communication of public institutions during crises.
doxa.comunicación | nº 38, pp. 275-292 January-June of 2024Tănase Tasente, Maria Alina Carataș and Tariq Tawfeeq Yousif Alabdullah ISSN: 1696-019X / e-ISSN: 2386-3978| 277Sentiment analysis, or opinion mining, utilizes computational linguistics, text analysis, and natural language processing to discern and extract subjective information from source materials (Lyu, et al., 2021). When applied to social media, sentiment analysis allows us to capture the pulse of public opinion instantaneously, giving us unprecedented insights into the concerns, fears, and hopes of people, especially during crises (Mee, et al., 2021). Public institutions have begun to realize the importance of eective communication during crises. Traditionally, these institutions would employ “one-way” communication strategies, where they would disseminate information without much scope for public feedback. However, social media has changed the game. e rise of digital diplomacy, characterized by open, dialogic communication with the public, is evident, enabling a more participative public sphere (Bjola & Zaiotti, 2020). is transformation has been particularly crucial in crisis communication, where public sentiment can signicantly aect the course of crisis management eorts. Crisis communication refers to the exchange of information between the institution facing a crisis and the stakeholders who can aect or be aected by the crisis outcome (Baba & Nastase-Anysz, 2020). In this regard, social media has emerged as an eective tool for public institutions to manage crises. rough platforms like Twitter and Facebook, these institutions can not only disseminate vital information quickly but also gauge public sentiment towards the crisis and their management eorts (Arroyo Barrigüete, et al., 2022). Such feedback, which can be captured and analysed through sentiment analysis, can be invaluable in ne-tuning the institution’s crisis management strategies. Several scholars have explored dierent facets of this complex interplay between sentiment analysis, social media, and public communication during crises. However, there are still many aspects of this relationship that remain underexplored. is paper seeks to provide a comprehensive review of the existing literature in this domain, identify gaps in the current understanding, and propose directions for future research. It is essential to underscore that the digital landscape is rapidly evolving. erefore, while this review strives to be comprehensive, it also recognizes the transient nature of its ndings. As new platforms emerge and public institutions nd innovative ways to communicate during crises, future research will need to adapt and evolve accordingly. is paper aims to deepen our understanding of how public institutions use social media for crisis communication, the role of sentiment analysis in capturing public opinion during crises, and how these insights can inform and improve crisis management strategies. e ongoing interplay between these elements in our increasingly connected world makes this an exciting and dynamic eld of study.Based on the above explanation and argument, there are still many undiscovered facets of the complex interaction between sentiment analysis, social media, and public communication during crises, even though multiple experts have looked at various aspects of it. Although there is an important volume of material discussing the topic, there is still a critical need to discover and ll in knowledge gaps. It is essential to perform an exhaustive evaluation and analysis of the available literature to increase our knowledge in this area. By doing this, we may identify the areas that need more research, illuminate previously unrecognized factors, and eventually advance our comprehension of the way sentiment analysis aects social media involvement and public discourse during crises. is analysis is crucial to developing ecient crisis communication strategies and ensuring prompt and suitable reactions to lessen the negative eects of crises on public perception and attitude. 278 | nº 38, pp. 275-292 |January-June of 2024Analysis of sentiment in the European Central Bank’s social media activity during the Covid-19 pandemic and Ukraine War...ISSN: 1696-019X / e-ISSN: 2386-3978doxa.comunicación2. Literature reviewe digital age has revolutionised crisis communication, fundamentally changing the way information is disseminated and consumed. Social media platforms are now essential channels for government agencies to communicate their crisis management strategies and engage with the public. With immediate access and wide reach, these platforms facilitate real-time updates and foster a two-way dialogue between authorities and citizens. e use of digital communication tools facilitates transparency, speeds up the distribution of information and helps manage public perception during emergency situations.Extending this digital narrative, the study by Drescher et al. (2023) examines Twitter communication during the early stages of the COVID-19 pandemic. e study shows that while German authorities maintained a more even-handed tone, independent experts favoured a more pessimistic tone, particularly during the more testing periods of the pandemic. is variation in tone illustrates the multiple perspectives and conicting ideas in the midst of a crisis.e literature shows the signicant involvement of the European Union (EU) and its agencies in social media, which is a fundamental channel for public communication during crises. Müller (2022) highlights the importance of EU agencies in attracting public attention on social media. is is inuenced by several factors, including the importance of the news, the longevity of the account and the level of interaction in the communication.During the pandemic, European leaders showed emotion in their communication, but according to Arroyo Barrigüete et al. (2022), contrary to expectations, there was no signicant dierence in empathy shown by leaders from countries dierently aected by COVID-19. Bunea & Nørbech (2022) analyse the European Commission’s strategic communication, which focused on reinforcing its reputation as a responsible policymaker rather than responding to public demand. Following this line of inquiry, Carvache-Franco et al. (2022) examine the language used on Twitter to discuss crisis and risk. eir ndings show that these discussions often focus on key areas such as the economy, health and government, suggesting that social media plays a crucial role in shaping public perceptions and response strategies in times of crisis.Furthermore, Nisch (2023) oers an analysis of President Zelensky’s tweets during the invasion of Ukraine. His ndings reveal a consistent positive sentiment and a call for unity and support, demonstrating the power of social media to foster solidarity in the face of adversity.Hancu-Budui et al (2020) conducted a study on the use of social media by European audit institutions and found that their presence on these platforms eectively promotes public services. However, they also found a lack of strategy for online engagement, particularly in relation to environmental and sustainability messaging.Building on this theme of digital engagement, MacKay et al. (2022) examined the use of Instagram by Canadian inuencers during the pandemic. eir research suggests that while inuencers did not make extensive use of Health Belief Model constructs, their posts still managed to generate varying levels of public interaction, demonstrating the nuances of inuencer impact during health crises.Sentiment analysis has been widely used to study public discourse during crises. Lyu et al. (2021) and Monselise et al. (2021) conducted sentiment analysis on discussions related to COVID-19 vaccination, highlighting the public’s concerns and changing attitudes towards vaccination. e results showed that there was a general increase in positive sentiment over time, doxa.comunicación | nº 38, pp. 275-292 January-June of 2024Tănase Tasente, Maria Alina Carataș and Tariq Tawfeeq Yousif Alabdullah ISSN: 1696-019X / e-ISSN: 2386-3978| 279with trust being the primary emotion. Mee et al. (2021) conducted a sentiment analysis of Brexit discourse among UK MPs, revealing discreet dierences in language use across the political spectrum. Kovács et al. (2021) used sentiment analysis to examine public reactions on Twitter to the murder of Ján Kuciak, demonstrating how social media can reveal patterns in public sentiment that correspond to political events and narratives. Vrana et al. (2023) conducted an analysis of EU citizens’ Twitter discussions during the energy crisis of 2022-23. ey found that fear and sadness predominated, indicating mostly negative sentiments that remained consistent across languages.By way of comparison, Maghyereh and Abdoh (2022) examined how sentiment aected asset volatility during the nancial crisis and the COVID-19 pandemic. eir ndings highlight the inuential role of public sentiment in nancial markets, aecting dierent types of assets in dierent ways during times of crisis.Similarly, Vrana et al. (2023) analysed sentiment in Twitter conversations among EU citizens during the 2022-23 energy crisis. eir work shows a dominant tone of fear and sadness, reecting widespread public concern about energy prices and political developments, regardless of language barriers.Comparative studies such as Heidenreich et al. (2020) provide insights into migration discourse on social media, demonstrating the frequent and negative expression of migration issues by political factions on both the extreme left and right. Traber et al. (2020) examine the speeches of prime ministers during economic crises and nd an increased shift of blame to external actors when domestic unemployment rises. Tasente (2023) examines the online communication methods used by world leaders during the Ukraine crisis, highlighting the importance of a wide range of communication strategies in eectively engaging global audiences.e online reputation of global organisations, particularly the EU during the refugee crisis, is analysed by Bjola and Zaiotti (2020), who suggest the need for proactive online reputation management in crisis situations. Yoon and Chung (2020) conducted an evaluation of the European Union’s public diplomacy campaigns in Asia by analysing sentiment and semantic networks. eir study found that the EU’s social media content was ineective in generating public engagement, suggesting the need for a more targeted narrative development strategy.Nastase-Anysz and Baba’s (2020) research highlights the response of the public sector during crises and promotes the implementation of relationship marketing principles to strengthen crisis response and communication. In the eld of public health, Ali (2021) explores the function of sentiment analysis, particularly in the context of online learning during COVID-19, and proposes methodologies that can help public health ocials manage crises.ere is a clear gap in the current literature regarding the implementation of relationship marketing principles for crisis management and communication in the public sector. While there are studies, such as Nastase-Anysz & Baba (2020), that touch on this topic, a more thorough investigation is needed to understand how these principles can be used eectively during crises.erefore, social media sentiment analysis is an important asset for understanding and managing public discussions in critical situations. e literature conrms that social media play an important role in the crisis communication of public institutions. However, there is a need for strategic approaches that incorporate relationship marketing principles to improve the eectiveness of crisis communication. 280 | nº 38, pp. 275-292 |January-June of 2024Analysis of sentiment in the European Central Bank’s social media activity during the Covid-19 pandemic and Ukraine War...ISSN: 1696-019X / e-ISSN: 2386-3978doxa.comunicación3. Research methodTo conduct our research, we employed a data centralization approach using the platform Fanpagekarma.com. We extracted a total of 11,781 posts from the ocial Twitter page of the European Central Bank, spanning from October 28, 2013, to June 2, 2023. Additionally, to identify potential dierences in communication strategies, we divided the dataset into three periods:–Period 1: October 28, 2013, to March 10, 2020: is period can be characterized as one without major geopolitical incidents.–Period 2: March 11, 2020, to February 23, 2022: is period was marked by the crisis caused by the Covid-19 pandemic.–Period 3: February 24, 2022, to June 6, 2023: is period was characterized by the crisis generated by the invasion of the Rus-sian army in Ukraine and the ongoing war between the two belligerent parties.For the analysis of sentiment in online discourse, we utilized RStudio along with the following packages: tm, SentimentAnalysisand syuzhet. We conducted a discourse analysis using two distinct methods: the AFINN method and the NRC method.When assessing the attitude expressed in the European Central Bank’s (ECB) announcement, the usage of the AFINN method and the NRC approach has various advantages and is extremely helpful. First o, the AFINN approach oers a quantitative evaluation of sentiment through its pre-computed sentiment scores, enabling a quick and simple study of the ECB’s conversation (Lennox et al., 2020). With the use of this technique, researchers can recognize and follow changes in mood over time, giving them important insights into how well the ECB’s communication tactics performed throughout the Covid-19 pandemic and Ukraine war. e NRC technique, on the other hand, enables a more detailed analysis by collecting particular sentiments and emotions expressed in the ECB’s tweets thanks to its extensive lexicon (Ali, 2021). R