Two Voices, One War: NLP and LDA for Sociological Analysis of Political Communication on Twitter vis-a-vis the Ukrainian War

Two Voices, One War: NLP and LDA for Sociological Analysis of Political Communication on Twitter vis-a-vis the Ukrainian War | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Case Report Two Voices, One War: NLP and LDA for Sociological Analysis of Political Communication on Twitter vis-a-vis the Ukrainian War Andrzej Meler This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6771056/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 14 Dec, 2025 Read the published version in Social Network Analysis and Mining → Version 1 posted 12 You are reading this latest preprint version Abstract In moments of pivotal geopolitical importance, such as Russia’s aggression against Ukraine, sentiment analysis of political communication offers a unique lens through which to examine how national interests are articulated and emotional tones are conveyed in public discourse. This study explores how Members of Parliament (MPs) in Poland and the United Kingdom addressed the war on the X platform (formerly Twitter) during its first year, treating parliamentary communication as an indicator of collective political positioning. While both countries have consistently supported Ukraine, the analysis revealed notable differences: British MPs expressed significantly more positive sentiment overall, whereas Polish MPs, despite addressing the Ukrainian topic more frequently, communicated in a more emotionally restrained and neutral tone. The study thus sheds light on the contrasting communicative styles and underlying political orientations that emerge in response to the same international crisis. Contemporary natural language processing tools—such as language models and thematic analysis—were employed to systematically compare messages across two distinct political and linguistic contexts. political communication sentiment analysis computational sociology BERT LDA Figures Figure 1 Figure 2 Introduction Contemporary democracies, despite idealistic beliefs in citizen equality, may be perceived as imagined constructs that serve efficient governance by a minority over the majority (Rancière, 2006, p. 52). This statement is not an assessment but rather an attempt to describe reality, suggesting that despite a certain degree of cynicism, democracy may still be the most beneficial system for individuals. However, for such governance to be effective, citizens must believe in the legitimacy of the actions of political elites. In this process, emotions play an essential role, exerting a strong influence on shaping political messages. This analysis aims to examine the political communication of British and Polish parliamentarians using machine learning methods such as the Bidirectional encoder representations from transformers (BERT) model for sentiment analysis and Latent Dirichlet Allocation (LDA) for content analysis. The study focuses on posts on the social media platform X (formerly Twitter) concerning the war in Ukraine—a significant topic in political debates in both countries during the analyzed period. The key question I aim to answer is how machine learning tools can support the analysis of political communication, what similarities and differences in expressed emotions appear in the British and Polish parliamentarians’ statements regarding the war in Ukraine, and what factors underlie these differences. Analyzing political communication on X is not a new issue (Vergeer, 2015 ), nor is the use of sentiment analysis to study such communication (Mueller & Saeltzer, 2020 ). However, to my knowledge, comparative analyses in this area are rare. Through this analysis, I hope to deepen the understanding of politicians’ use of emotions in political messaging and the various cultural and political contexts shaping the reception of this communication. While modern democracies are often idealized as systems of equal citizen participation, they can also be understood—as Rancière (2006, p. 52) suggests—as imagined constructs that enable the efficient rule of a few over the many. This is not meant as a critique but as a reflection of political reality. Even if this view carries a hint of cynicism, democracy may still offer the most advantageous framework for individual well-being. Yet for such a system to function, citizens must trust in the legitimacy of political elites—a trust that is deeply shaped by emotion. Emotions are not peripheral; they are central to how political messages are framed, communicated, and received. This study explores how British and Polish members of parliament use emotional language when communicating about the war in Ukraine on the social media platform X (formerly Twitter)—a topic that has played a central role in both countries' political discourse during the period analyzed. Drawing on machine learning tools—including Bidirectional Encoder Representations from Transformers (BERT) for sentiment analysis and Latent Dirichlet Allocation (LDA) for content modeling—this research asks: how can such tools help us understand political communication more effectively? What emotional patterns emerge across British and Polish discourse? And what contextual factors may explain the similarities and differences? Analyzing political messaging on social media is not new (Vergeer, 2015 ), nor is the use of sentiment analysis in this context (Mueller & Saeltzer, 2020 ). However, comparative approaches remain relatively rare. By examining these two cases side by side, this study aims to shed light on how emotions function in political communication and how national and cultural contexts shape their use and impact. Political communication Political communication is a key element of public discourse, which forms the foundation for shaping social norms (Bleiker, 2003 ; Hutchison & Bleiker, 2020 ). The ideas and values that dominate the public sphere determine the distinction between acceptable and desirable norms and norms that face criticism, ostracism, or sanctions. In this context, politicians play a particular role as primary actors in the political sphere, possessing greater agency than the average citizen. They can translate the ideals and values they hold into formal social rules, contributing to their broad legitimization. Interests, as one of the fundamental analytical categories, are a crucial tool in shaping political communication. Weber ( 2019 ) emphasized the role of interests as the primary driver of social actions, where purposeful-rational actions take precedence, with emotions playing a secondary role. Simultaneously, political interests are often shielded by cultural symbolism, with dominant groups imposing their interests as universally binding (Bourdieu, 1984 ). Researchers such as Milliken ( 1999 ), Hansen (2006), Epstein (2008) and Zybertowicz ( 1995 ) observe the strong link between language and power, emphasizing that language functions as symbolic violence, through which certain norms and attitudes are promoted. Contemporary political communication occurs primarily through the media, making public discourse almost entirely a media discourse. In this context, social media platforms, especially X, play a fundamental role due to their global reach and significant opinion-shaping power. Research indicates that discourse on X can influence actual political outcomes, such as election victories (Kartsounidou et al., 2023 ; Kruikemeier, 2014 ; Rotesi, 2019 ). Politicians shape their messages per party lines by using professional communication management, thereby effectively influencing public opinion (Mueller & Saeltzer, 2020 ). Sentiment The study of emotions in politics emerged in the social sciences as a counterpoint to the rational approach to social actions (Demertzis, 2020 , pp. 4–5; Hutchison & Bleiker, 2020 ). Today, we know that emotions not only complement rationality in political communication but often play a central role there (Koschut, 2020 ). Politicians manage emotions to evoke collective social reactions (Hutchison & Bleiker, 2020 , p. 37). In pursuing their interests, while politicians seek to gain support and approval for their actions, they must also act as a team, which often results in the building of sentiment (Meler, 2022 ). Early studies on emotions relied on an ontological approach, viewing emotions as an evolutionary legacy that preserves reactions conducive to survival and reproduction. For example, fear triggers defensive reactions, disgust protects against poisoning, and anger enables competition for resources or social standing. Happiness, in turn, motivates actions that encourage a return to this state of happiness (Plutchik, 1962 , pp. VI–VIII). However, these initial, mainly psychological approaches to emotions focus on individual experiences. Meanwhile, politics is a collective phenomenon (Berezin, 2002 ). Therefore, when focusing on emotions in political communication, a cognitive-constructivist approach is more appropriate, suggesting that we learn emotions culturally, both in experiencing and expressing them (Koschut, 2020 , pp. 3–12). The ontological nature of emotions remains a subject of debate among researchers (Barrett, 2018 ), but is it always necessary to determine their ontological status? If we consider, for instance, the distinction between cognitive and constructivist approaches, it primarily concerns the ontological dimension. The cognitive approach attributes biological-evolutionary foundations to emotions but postulates that the communication of emotions is strongly shaped by culture. Constructivism goes further and assumes, even ontologically, that humans learn emotions through socialization. The ontology of emotions is irrelevant to my analysis since it focuses on the cultural nature of the communication of emotions. Politicians on platforms such as X appeal to socially rooted emotional figures, attempting to evoke emotions that translate into collective actions among the audience. Political efforts to build emotional energy, either supporting or opposing policies, are well captured by the sociological concept of communities of feeling (Berezin, 2002 , p. 38). To complete the conceptual framework regarding emotions in politics, it is essential to mention the distinction between emotions and affects (Hutchison & Bleiker, 2020 , p. 5). Affects are sometimes triggered in politics by unforeseen events, but it is primarily emotions that play a dominant role in political communication as more enduring representations of moral judgments. In political communication, emotions are a tool that facilitates the exercise of symbolic violence, as people follow messages that elicit specific emotions. Often, provoking anger, fear, or general negativity in the audience can yield a quicker and more lasting promotional effect for a particular political idea than rational arguments. Materials and Methods Sentiment Analysis Emotions are a very important component of political communication, and it is impossible to determine its impact without analyzing them. Nowadays, data science provides useful tools that allow for relatively objective analysis of emotions even in large text corpora. Sentiment analysis (SA) is the automated process of detecting and understanding the emotions conveyed through written text (Gunasekaran, 2019 ). It is used by researchers across various fields to determine attitudes toward something or someone or to analyze moods. This methodology serves to predict stock market trends (Bollen et al., 2011 ; Turner et al., 2021 ) in economics and is applied to disease recognition through social media analysis (Yadav et al., 2019 ) in the medical field. In pedagogy, it has been used to study youth subcultures (Rudzionis et al., 2023 ). Given the significant role of emotions in politics and political communication, SA has been applied in the current study as well. In SA, we can identify two primary methodological approaches: the lexicon-based approach and the machine-learning approach (Gunasekaran, 2019 ). My analysis was based on the latter, using the BERT transformer, where sentiment measurement is performed through machine-recognized attention (Devlin et al., 2018 ). This analysis results in identifying the level of a specific sentiment for a given text, with the same text potentially containing a certain level of negative, neutral, or positive sentiment. Variants of different models may help specify the level of extremely negative (or positive) and moderately negative (or moderately positive) sentiments. Transformers such as BERT are more advanced SA tools because they can “understand” context and even recognize the sender’s intentions (Hardalov et al., 2023 ). In this analysis, BERT models were chosen to ensure compatibility for English and Polish languages and include variables of negative, neutral, and positive sentiment. To identify and understand differences in political communication, a simple test of the BERT tool was performed by comparing the sentiment of identical sentences in both languages. Data The research material for this analysis includes 127,000 tweets by Polish MPs and 210,000 tweets by British MPs posted during the first year of the Russia-Ukraine war, from February 2022 to January 2023. The tweets were collected from the accounts of 790 British MPs and 329 Polish MPs, which accounts for 55% of all MPs in both countries. In the following text, I use the terms MPs, parliamentarians, and the elite interchangeably. The latter is used rhetorically, fully aware that the set of all parliamentarians is not identical to the set of 'the elite', but with the conviction that MPs are an important part of the country's political elite. For each national dataset, keyword-based subsets of posts dedicated to the war in Ukraine were extracted. The datasets were then subjected to automatic SA using BERT models trained for English and Polish. Models were selected to provide the same sentiment measurement scale: negative, neutral, and positive. The output of the BERT transformer analysis was to determine for each post the proportion of negative, neutral, and positive sentiment as a decimal fraction, which summed up to 1. Then, based on the assumption that the mood of media communication emerges only after aggregating individual statements into larger datasets (Meler, 2024 ), sentiment for both countries and both thematic subsets was aggregated by monthly periods. This provided averaged SA results for British parliamentarians’ communications on Ukraine and other topics as well as analogous results for Polish parliamentarians’ communications on the same topics. The aggregation method is depicted in Tables 3 and 4. Topics analysis Latent Dirichlet Allocation (LDA) is a probabilistic generative model used for topic modeling in large corpora of text. The model assumes that documents are mixtures of topics, where each topic is a distribution over words, and each document is represented as a distribution over topics. LDA operates under the assumption that each word in a document is generated from one of the document's topics, which in turn is drawn from a fixed set of topics shared across the corpus (Blei et al., 2003 ). By applying a Dirichlet prior distribution to both the topic distributions for each document and the word distributions for each topic, LDA allows for the discovery of latent structures within a collection of texts. This methodology has been successfully applied to analyze the behavior of politicians on X (Kushwaha et al., 2020 ). The shaping of social attitudes does not come out of nowhere, and by examining political communication, we can trace their emergence. Results Significance of Russia-Ukraine War in Overall Political Communication on X The significance of the Russia-Ukraine War in political communication, understood as the share of tweets on this subject in the overall pool of X posts by Polish and British parliamentarians, experienced a dynamic pattern. They showed high shares in the first two months and a drop in the next two months, subsequently stabilizing at a relatively low level by the end of the analyzed period. Although the trend lines run similarly, there is a difference in the share levels: tweets related to the conflict in Ukraine constituted between 10% and 56% of the Polish dataset, while for British MPs, it ranged from 4% to 39%. Table 1. The Topic of Ukraine vs. Overall X Communication. Period (year-month) UK PL Ukraine Other Topics N Ukraine Other Topics N 2022–02 29% 71% 24,458 44% 56% 15,057 2022–03 39% 61% 27,642 56% 44% 18,773 2022–04 16% 84% 18,776 33% 67% 10,875 2022–05 9% 91% 18,870 22% 78% 9,319 2022–06 8% 92% 17,131 19% 81% 9,816 2022–07 4% 96% 18,801 11% 89% 9,777 2022–08 7% 93% 12,838 10% 90% 8,361 2022–09 6% 94% 13,121 15% 85% 9,973 2022–10 5% 95% 19,524 13% 87% 9,804 2022–11 4% 96% 16,016 14% 86% 9,279 2022–12 5% 95% 12,957 12% 88% 8,214 2023–01 6% 94% 10,668 21% 79% 8,619 Total 14% 86% 210,802 26% 74% 127,867 Comparison of Sentiment on the Topic of Ukraine Against Other Topics Compared to tweets on other topics, significant differences can be observed in the political communication of British and Polish parliamentarians about Ukraine. Positive and negative sentiments concerning Ukraine varied across periods in both countries. Meanwhile, the aggregated sentiment of other tweets showed no significant differences between periods or between countries (the standard deviation was twice as high for the Ukraine topic compared to others). A ratio, commonly used in decision trees, was applied to determine the scale of differences between individual sentiments, representing the value ratio for the first group in a given period relative to the second group. A 100% indicator would denote identical sentiment for both groups. Table 2. Average Sentiment Loadings in MPs’ Tweets Over Periods: Ukraine vs. Other Topics. Period (year-month) UK PL Ukraine Other Topics Ukraine Other Topics Neg. Neu. Pos. Neg. Neu. Pos. Neg. Neu. Pos. Neg. Neu. Pos. 2022–02 0.32 0.48 0.21 0.28 0.37 0.34 0.30 0.46 0.24 0.29 0.46 0.24 2022–03 0.26 0.43 0.31 0.27 0.36 0.37 0.29 0.46 0.26 0.29 0.46 0.25 2022–04 0.29 0.40 0.30 0.28 0.36 0.37 0.33 0.47 0.19 0.29 0.46 0.25 2022–05 0.20 0.37 0.43 0.26 0.36 0.38 0.29 0.48 0.23 0.26 0.46 0.29 2022–06 0.23 0.37 0.41 0.25 0.34 0.40 0.22 0.49 0.29 0.25 0.46 0.29 2022–07 0.22 0.35 0.43 0.27 0.37 0.37 0.27 0.49 0.24 0.27 0.45 0.28 2022–08 0.15 0.34 0.51 0.27 0.37 0.37 0.26 0.45 0.29 0.30 0.46 0.24 2022–09 0.28 0.39 0.33 0.26 0.40 0.35 0.29 0.50 0.20 0.28 0.47 0.26 2022–10 0.25 0.40 0.35 0.32 0.36 0.32 0.31 0.49 0.20 0.27 0.47 0.27 2022–11 0.22 0.42 0.36 0.28 0.36 0.36 0.26 0.53 0.21 0.24 0.50 0.27 2022–12 0.23 0.38 0.39 0.26 0.35 0.39 0.26 0.48 0.26 0.26 0.44 0.30 2023–01 0.22 0.41 0.37 0.29 0.37 0.35 0.28 0.49 0.23 0.28 0.47 0.25 Total mean 0.24 0.39 0.37 0.27 0.36 0.36 0.28 0.48 0.24 0.27 0.46 0.26 Std. dev. 0.04 0.04 0.08 0.02 0.01 0.02 0.03 0.02 0.03 0.02 0.01 0.02 When we compare the sentiment regarding Ukraine (first group) to that of other topics (second group), we see that British parliamentarians’ communication about Ukraine was more positive overall throughout the year than on other topics. This indicator also showed greater variability over time—from 112% in February 2022 to 57% negative sentiment in August 2022. In contrast, Polish parliamentarians’ tweets about Ukraine were more negative, compared to other topics. An analysis of British MPs’ tweets reveals that in eight out of 12 months, the share of positive sentiment in the Ukraine topic was higher than for other topics. Negative sentiment was lower in nine out of 12 months. Meanwhile, in Polish parliamentarians’ tweets about Ukraine, there was mainly a decline in positive sentiment in most of the analyzed period (eight out of 12 months). Comparison of the Sentiment of the Topic of Ukraine Between Countries When comparing the sentiment of tweets about Ukraine between British and Polish parliamentarians, the main difference was found in the share of positive sentiment. In British MPs’ tweets, this share was higher (an average of 36%) than in Polish MPs’ tweets (an average of 26%). Notably, this was a general difference, applying to both Ukraine-related and other topics. The indicator for the share of negative sentiment tweets among British (first group) and Polish parliamentarians (second group) in the context of Ukraine ranged from 105% in February 2022 to 57% in August 2022. Only in February and June 2022 were British MPs’ tweets more negative than those of their Polish counterparts, though the differences were small. The sentiment trend regarding Ukraine, represented by dashed lines (chart 1), shows that in the first two months of the analyzed period, British parliamentarians expressed significantly more negative opinions about Ukraine than about other topics. Later, their communication about Ukraine became more positive vis-a-vis other topics, peaking in August 2022; after September 2022, it no longer differed significantly from the rest of the political communication. For Polish parliamentarians, at the beginning of the war, the sentiment regarding Ukraine did not deviate significantly from general political communication, but in subsequent months, it became noticeably less positive compared to other topics. The largest sentiment discrepancies between British and Polish parliamentarians occurred in July and August 2022 and January 2023. Topic Analysis Using LDA To understand the reasons behind the convergence of positive sentiment levels in Polish and British communication (generally divergent in other periods) in February 2022, as well as the divergence of negative sentiment levels (previously quite similar), we conducted content analysis using LDA (see Appendix). In February 2022, British parliamentarians’ tweets began to feature words like “war,” “invasion,” “attack,” and “sanction,” which explain the lowered sentiment. In Polish communications, words like “war” (pol.: “wojna”) and "aggression" (pol.: “agresja”) also appeared, but terms like “help” and “support” (pol.: “pomoc”, “wsparcie”) were equally prominent. Further, in February, the topic of “refugee” (pol.: “uchodźcy”) emerged in Polish tweets, becoming a dominant theme in their communications. However, the term was much less present in British discourse. The Polish discourse began to refer to events in Ukraine as “war” more frequently only in April, which influenced a more negative tone in content. In February, however, both the Polish and British discourses centered around terms like “war,” “invasion,” and “attack.” In Polish tweets, there was an additional topic related to “support,” referring to the assistance for Ukrainian refugees who were warmly received at both government and social levels. Most refugees found shelter in private homes, eliminating the need for camps, even as their numbers exceeded one million within the first week of the war (Zalewski, 2022). The topic of support evoked a positive sentiment in Polish discourse, while British messaging maintained more neutral or negative tones, thus bringing positive sentiment levels closer between the two countries. In March, British tweets began to prominently feature terms like “support,” “people,” “refugees,” “children,” and “family,” which significantly increased the positive tone of the messaging. Interestingly, in Polish tweets, “refugees,” “support,” and “help” were as dominant as in February, keeping Polish sentiment almost unchanged. In August, at the height of the differences between messages, British MPs wrote mainly about “support,” “people,” “humanitarian,” and “help,” which reinforced the positive sentiment. Meanwhile, in Polish discourse, terms like “campuspolska” or “campus” appeared, referring to a cyclical party event of the then-opposition, where Ukraine was also a significant topic. The frequency of posts about “refugees” and “support” declined, shifting the focus of Polish messaging. Sentiment Comparison by Political Affiliation In Polish communication, especially in the latter half of the analyzed period, the topic was defined by the names of prominent Polish politicians, raising the question of whether the communication sentiment was shaped by affiliation with the ruling or opposition camp. Visual analysis in Chart 2 already suggests that the sentiment of communication by the ruling and opposition camps differed significantly more for British parliamentarians. A T-test indicates a statistically significant difference in both negative and positive sentiments between the ruling and opposition camps in both countries, but the eta squared coefficient shows that political affiliation had a much stronger impact on sentiment in the UK than in Poland. In the UK, 81% of the variation in negative sentiment in individual periods can be explained by political affiliation; in contrast, the percentage in Poland was 60%. For positive sentiment, this relationship was even more divergent: 66% in UK communication and 47% in Poland. Table 3. Value of the Eta Squared in Sentiment Comparison Between Governing and Opposition Country Negative Neutral Positive UK 0.81* 0.01 0.66* PL 0.60* 0.31 0.47* * T-test significance (p-value <0.05); sentiment as dependent Discussion Social sciences have long emphasized the importance of the so-called “humanistic coefficient,” underscoring the need to consider both individual and collective interpretations of reality in research (Znaniecki, 1934 ). An analysis of behaviors on social media platforms can be an effective tool in achieving this goal, where sentiment in online communication represents one of the key dimensions of communication. The first important conclusion from the analysis concerns methodology. Automated discourse analysis methods not only enable the examination of larger datasets but also contribute to the objectivity of results. When a researcher interprets text individually, as is standard in social sciences (van Dijk, 1998 ), there is a greater risk of personal beliefs affecting the outcomes. Automated methods such as transformer models or LDA significantly reduce this risk. For example, the BERT model, trained on a very large corpus of texts, mitigates some individual biases. Nevertheless, a methodological test shows that analyzing politically charged sentences may require more specific model training. Turning to the SA results of British and Polish MPs’ political communication, British communication was significantly more positive than Polish communication. We also observe that the topic of Ukraine held more significance for the Polish political class than for the British, when comparing it to their entire scope of political communication. Polish MPs dedicated relatively more posts to this topic, with their communications on the Ukraine war reflecting a negative tone than on other issues. Thus, Polish politics has been significantly more engaged in Ukrainian affairs than British politics. Interestingly, British MPs referred to the Ukraine war in a less negative tone compared to other topics. The peak of enthusiasm in British posts about Ukraine occurred in August 2022, largely due to the British Prime Minister’s visit to Kyiv during Ukraine’s Independence Day celebrations. In Polish communication, visits by the Polish Prime Minister and President, which occurred much more frequently than visits by other high-ranking officials, did not have such a marked effect. The war has not directly impacted British or Polish societies. However, the political elites of both countries adopted a strategy to deter Russia through multi-level support for Ukraine that caused significant financial strains on both countries. Politicians had to gain public support for decisions that ultimately meant burdens on the population. SA indicates that the British political class quickly began building support based on negative messaging. Negative communication can be a mobilizing factor (McNoir, 2018 ; van Dijk, 1998 ), but its prolonged use may discourage voters (Ansolabehere & Iyengar, 1995 ; Perloff, 2022 ). British communication followed the textbook approach; it increased negative messaging in the first two months, then diminished negative emotions and shifted to a relatively positive tone before eventually returning to an average sentiment level. One may question why the sentiment management pattern in Polish political communication was different. Initially, the tone was not more negative than in the rest of their communication, but later, it became negative and remained so for longer than that of their British counterparts. Content analysis using LDA and contextual knowledge suggested that Polish politicians in February and March 2022 first needed to effectively alleviate the migration crisis. Managing this crisis required a positive attitude within Polish society, reflected in maintaining a more positive tone (than the British). Differences in sentiment between British and Polish MPs are better understood through the interpretive framework of the security state and community of feelings (Berezin, 2002 , p. 38). We can assume that the more negative sentiment in Polish communication are manifestations of fear about losing security-state status, which Polish society has only recently started to take for granted after a turbulent unsecurity state period that concluded with the systemic transformation in the early 1990s. The UK has, for centuries, represented an archetype of a secure state due to its political structure and safe, insular location. The more positive tone of British MPs indicates their ability to build a community of feeling as during times of somber but non-war-related events, such as the death of Princess Diana (Berezin, 2002 , p. 40). Appropriate communication supports the mobilization of the community for Ukraine. Ukraine finds itself attacked by a great power that other powers eagerly supported, similar to Belgium at the start of World War I. In Polish communication, the creation of a community of feeling manifested in diverse ways but was dominated by a strong sense of threat to their own country. The more negative sentiment in Polish tweets reflect fear, as concerns about security are highly current. Russia’s aggressive actions revive historical fears for Poland. The prospect of Ukraine falling into Russia’s sphere of influence evokes memories of dependence on Russia, still alive in many Poles’ minds and familiar to all from history lessons. For the UK, Russia might appear to be yet another international problem, especially UK has previously been victorious (Crimean War) and the countries have been allies on several occasions. Consequently, Poles writing about Russian aggression culturally and instinctively limit positive tones, as a Ukrainian defeat would signal a return to the “Russian world,” remembered by generations of Poles as economic, social, and cultural degradation. Unsurprisingly, Polish politicians find it difficult to maintain higher proportion of positive sentiment, which seems significantly easier for their British counterparts. Here, we can speak of a stronger “flag effect” in Poland. Noticeably, the topic of Ukraine was critical only in the first two months and then lost prominence. As this topic receded, sentiment levels for both groups of MPs began to fluctuate within narrower limits. This suggests that in political communication (and perhaps more broadly in politics), there is little room for affects; only emotions remain relevant according to the earlier distinction (Hutchison & Bleiker, 2020 ). Declarations Funding details: This work was supported by the Nicolaus Copernicus University in Toruń under Grant IDUB/Debiuty_6_Andrzej Meler. Disclosure statement: The author report there are no competing interests to declare. Biographical note: A Nicolaus Copernicus University’s Institute of Sociology graduate. He received the Polish Sociological Association award for his master’s thesis, which focused on the analysis of media discourse on the Polish judiciary system. He worked as a web analyst for Polska Press Group for ten years. Currently his academic research focuses on political discourse in media. Data availability statement: Data availability after contact by ResearchGate: https://www.researchgate.net/profile/Andrzej-Meler References Ansolabehere, S., & Iyengar, S. (1995). Going Negative: How Political Advertisements Shrink and Polarize the Electorate . Free Press. Barrett, L. F. (2018). How Emotions Are Made: The Secret Life of the Brain . Pan MacMillan. Berezin, M. (2002). Secure states: Towards a political sociology of emotion. Sociological Review , 50 (S2), 33–52. https://doi.org/10.1111/j.1467-954X.2002.tb03590.x Blei, D. M., Ng, A. Y., & Jordan, M. I. (2003). Latent Dirichlet Allocation. Journal of Machine Learning Research , 3 , 993–1022. https://doi.org/10.1016/B978-0-12-411519-4.00006-9 Bleiker, R. (2003). Discourse and Human Agency. Contemporary Political Theory , 2 (March 2003), 25–47. https://doi.org/10.1057/palgrave.cpt.9300073 Bollen, J., Mao, H., & Zeng, X. (2011). Twitter mood predicts the stock market. Journal of Computational Science , 2 (1), 1–8. https://doi.org/10.1016/j.jocs.2010.12.007 Bourdieu, P. (1984). Distinction: A Social Critique of the Judgement of Taste. (R. Nice (ed.)). Harvard University Press. Demertzis, N. (2020). The political sociology of emotions . Routledge. Devlin, J., Chang, M.-W., Lee, K., Google, K. T., & Language, A. I. (2018). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Naacl-Hlt 2019 , Mlm , 4171–4186. https://aclanthology.org/N19-1423.pdf Gunasekaran, K. P. (2019). Review on Techniques in Natural Language Processing. International Journal of Scientific Research in Computer Science and Engineering , 7 (5), 1–4. https://doi.org/10.26438/ijsrcse/v7i5.14 Hardalov, M., Koychev, I., & Nakov, P. (2023). Enriched Pre-trained Transformers for Joint Slot Filling and Intent Detection. International Conference Recent Advances in Natural Language Processing, RANLP , 480–493. https://doi.org/10.26615/978-954-452-092-2_054 Hutchison, E., & Bleiker, R. (2020). Emotion, agency and power in word politics. In S. Koschut (Ed.), The Power of Emotions in World Politics . Routledge. Kartsounidou, E., Papaxanthi, D., & Andreadis, I. (2023). Measuring the impact of candidates’ tweets on their electoral results. Journal of Information Technology & Politics , 20 (2), 154–168. https://doi.org/10.1080/19331681.2022.2069181 Koschut, S. (2020). Emotion, discourse, and power in world politics. In S. Koschut (Ed.), The Power of Emotions in World Politics (Issue 112, p. 165). Routledge. Kruikemeier, S. (2014). How political candidates use Twitter and the impact on votes. Computers in Human Behavior , 34 , 131–139. https://doi.org/10.1016/j.chb.2014.01.025 Kushwaha, A. K., Mandal, S., Pharswan, R., Kar, A. K., & Ilavarasan, P. V. (2020). Studying Online Political Behaviours as Rituals: A Study of Social Media Behaviour Regarding the CAA. IFIP Advances in Information and Communication Technology , 618 , 315–326. https://doi.org/10.1007/978-3-030-64861-9_28 McNoir, B. (2018). An Introduction to Political Communication (6th ed.). Routledge. Meler, A. (2022). Parlamentarne ćwierkanie o pandemii. Analiza sentymentu tweetów parlamentarzystów publikowanych podczas pierwszych 12 miesięcy pandemii koronawirusa w Polsce. Studia Socjologiczne , 2022 (2), 113–136. https://doi.org/10.24425/sts.2022.141425 Meler, A. (2024). In The Beginning, Let There Be The Word: Challenges and Insights in Applying Sentiment Analysis to Social Research. In C. Tat_seng (Ed.), WWW ’24: Proceedings of the ACM Web Conference 2024 (pp. 1214–1217). Association for Computing Machinery. https://doi.org/10.1145/3589335.3651264 Milliken, J. (1999). The study of discourse in international relations: A critique of research and methods. European Journal of International Relations , 5 (2), 225–254. https://doi.org/10.1177/1354066199005002003 Mueller, S. D., & Saeltzer, M. (2020). Twitter made me do it! Twitter’s tonal platform incentive and its effect on online campaigning. Information Communication and Society , 0 (0), 1–26. https://doi.org/10.1080/1369118X.2020.1850841 Perloff, R. M. (2022). The Dynamics of Political Communication Media and Politics in a Digital Age (3rd ed.). Routledge. Plutchik, R. (1962). The Emotions. Facts, Theories, and a New Model . Random House. Ranciere, J. (2006). Hatred of Democracy. In S. Corcoran (Ed.), What Is Democracy and How Do We Study It? Verso. https://doi.org/10.1080/09614520801899259 Rotesi, T. (2019). The impact of Twitter on political participation. Bocconi University, Milan, Italy Working Paper . Rudzionis, V., Ramanuskaite, E., & Kairaityte-Uzuper, A. (2023). Sentiment Analysis of Lithuanian Youth Subcultures Zines Using Automatic Machine Translation. In A. Lopata, R. Butkiene, & D. Gudoniene (Eds.), Information and Software Technologies (pp. 201–221). Springer. https://doi.org/10.1007/978-3-031-48981-5 Turner, Z., Labille, K., & Gauch, S. (2021). Lexicon-based sentiment analysis for stock movement prediction. Journal of Construction Materials , 2 (3), 149–154. https://doi.org/10.36756/jcm.v2.3.5 van Dijk, T. A. (1998). Oppionions and Ideologies in the Press. In A. Bell & P. Garrett (Eds.), Approaches to Media Discourse (pp. 21–63). Blackwell. Vergeer, M. (2015). Twitter and Political Campaigning. Sociology Compass , 9 , 745–760. https://doi.org/10.1111/soc4.12294 Weber, M. (2019). Economy and Society (K. Tribe (ed.)). Harvard University Press. Yadav, S., Ekbal, A., Saha, S., & Bhattacharyya, P. (2019). Medical sentiment analysis using social media: Towards building a patient assisted system. LREC 2018 - 11th International Conference on Language Resources and Evaluation , 2790–2797. Zalewski, P. (2022). Działania administracji państwowej w Polsce wobec uchodźców z Ukrainy w pierwszych tygodniach wojny w Ukrainie 2022 roku. Aspekty prawne i securitologiczne. Studia Politicae Universitatis Silesiensis , 34 , 101–124. https://doi.org/10.31261/spus.13934 Znaniecki, F. (1934). The Method of Sociology . Rinehart & Company. Zybertowicz, A. (1995). Przemoc i poznanie (1st ed.). Wydawnictwo Uniwersytetu Mikołaja Kopernika. Additional Declarations No competing interests reported. Supplementary Files Apendixs.docx Cite Share Download PDF Status: Published Journal Publication published 14 Dec, 2025 Read the published version in Social Network Analysis and Mining → Version 1 posted Editorial decision: Revision requested 15 Jul, 2025 Reviews received at journal 14 Jul, 2025 Reviews received at journal 10 Jul, 2025 Reviews received at journal 05 Jul, 2025 Reviewers agreed at journal 16 Jun, 2025 Reviewers agreed at journal 15 Jun, 2025 Reviewers agreed at journal 13 Jun, 2025 Reviewers agreed at journal 13 Jun, 2025 Reviewers invited by journal 01 Jun, 2025 Editor assigned by journal 01 Jun, 2025 Submission checks completed at journal 29 May, 2025 First submitted to journal 28 May, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6771056","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Case Report","associatedPublications":[],"authors":[{"id":465312416,"identity":"ea329ace-b03e-439d-9313-169a6ab8c668","order_by":0,"name":"Andrzej Meler","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYHACAyBmA2LmBgaGCgYGCQYGxgNEamEEajkD1sJAjBYGiBbGNiK08M9u3vbxRwUfg257Y+PjynmHoyUbeA/g1SJx51jxDIkzbAxmZw42G57ddjh3NgNfAn6H3cgxZjBsA2q5kdgm2QjUMo+BxwCvFnmQlsR/QC33H7b/bJxDhBYDkJaDDSBbGNsYGxtADiOgxfBGWjFjwzE2HrMzic2SDcfSc2c2E/CL3I3kzYw/ao7JmR0/fPBjQ4117ozjvQcf4NMCBcd4EGxmHtzqkEANMoc4LaNgFIyCUTByAABZgU9UwgMmtgAAAABJRU5ErkJggg==","orcid":"","institution":"Nicolaus Copernicus University","correspondingAuthor":true,"prefix":"","firstName":"Andrzej","middleName":"","lastName":"Meler","suffix":""}],"badges":[],"createdAt":"2025-05-28 21:08:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6771056/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6771056/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s13278-025-01538-4","type":"published","date":"2025-12-14T15:59:15+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":83905829,"identity":"c57ff2e3-c71d-41bf-930e-4d9f4c826ad8","added_by":"auto","created_at":"2025-06-04 10:17:38","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":169548,"visible":true,"origin":"","legend":"\u003cp\u003eSentiment of Topic of Ukraine Compared to Sentiment of Other Topics\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6771056/v1/a5b6738b479b47131e3c68b0.png"},{"id":83906566,"identity":"15fcb70f-d500-4c27-8954-cf19abcc139e","added_by":"auto","created_at":"2025-06-04 10:25:38","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":181843,"visible":true,"origin":"","legend":"\u003cp\u003eSentiment on Ukraine by Representatives of the Ruling and Opposition Camps\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6771056/v1/6ddc745ebc1a9008bb4fbe96.png"},{"id":98246049,"identity":"720b1222-4fae-4e59-a7ea-891f85370acc","added_by":"auto","created_at":"2025-12-15 16:18:46","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1210168,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6771056/v1/07a97d9e-627c-4464-b237-c94b1c2a3ec7.pdf"},{"id":83905831,"identity":"a81305a9-2942-4403-bdd8-dbcd011ca8c2","added_by":"auto","created_at":"2025-06-04 10:17:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2071494,"visible":true,"origin":"","legend":"","description":"","filename":"Apendixs.docx","url":"https://assets-eu.researchsquare.com/files/rs-6771056/v1/f2a2d2191d1636ac09cabee5.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Two Voices, One War: NLP and LDA for Sociological Analysis of Political Communication on Twitter vis-a-vis the Ukrainian War","fulltext":[{"header":"Introduction","content":"\u003cp\u003eContemporary democracies, despite idealistic beliefs in citizen equality, may be perceived as imagined constructs that serve efficient governance by a minority over the majority (Ranci\u0026egrave;re, 2006, p. 52). This statement is not an assessment but rather an attempt to describe reality, suggesting that despite a certain degree of cynicism, democracy may still be the most beneficial system for individuals. However, for such governance to be effective, citizens must believe in the legitimacy of the actions of political elites. In this process, emotions play an essential role, exerting a strong influence on shaping political messages.\u003c/p\u003e \u003cp\u003eThis analysis aims to examine the political communication of British and Polish parliamentarians using machine learning methods such as the Bidirectional encoder representations from transformers (BERT) model for sentiment analysis and Latent Dirichlet Allocation (LDA) for content analysis. The study focuses on posts on the social media platform X (formerly Twitter) concerning the war in Ukraine\u0026mdash;a significant topic in political debates in both countries during the analyzed period. The key question I aim to answer is how machine learning tools can support the analysis of political communication, what similarities and differences in expressed emotions appear in the British and Polish parliamentarians\u0026rsquo; statements regarding the war in Ukraine, and what factors underlie these differences.\u003c/p\u003e \u003cp\u003eAnalyzing political communication on X is not a new issue (Vergeer, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), nor is the use of sentiment analysis to study such communication (Mueller \u0026amp; Saeltzer, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, to my knowledge, comparative analyses in this area are rare. Through this analysis, I hope to deepen the understanding of politicians\u0026rsquo; use of emotions in political messaging and the various cultural and political contexts shaping the reception of this communication.\u003c/p\u003e \u003cp\u003eWhile modern democracies are often idealized as systems of equal citizen participation, they can also be understood\u0026mdash;as Ranci\u0026egrave;re (2006, p. 52) suggests\u0026mdash;as imagined constructs that enable the efficient rule of a few over the many. This is not meant as a critique but as a reflection of political reality. Even if this view carries a hint of cynicism, democracy may still offer the most advantageous framework for individual well-being. Yet for such a system to function, citizens must trust in the legitimacy of political elites\u0026mdash;a trust that is deeply shaped by emotion. Emotions are not peripheral; they are central to how political messages are framed, communicated, and received.\u003c/p\u003e \u003cp\u003eThis study explores how British and Polish members of parliament use emotional language when communicating about the war in Ukraine on the social media platform X (formerly Twitter)\u0026mdash;a topic that has played a central role in both countries' political discourse during the period analyzed. Drawing on machine learning tools\u0026mdash;including Bidirectional Encoder Representations from Transformers (BERT) for sentiment analysis and Latent Dirichlet Allocation (LDA) for content modeling\u0026mdash;this research asks: how can such tools help us understand political communication more effectively? What emotional patterns emerge across British and Polish discourse? And what contextual factors may explain the similarities and differences?\u003c/p\u003e \u003cp\u003eAnalyzing political messaging on social media is not new (Vergeer, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2015\u003c/span\u003e), nor is the use of sentiment analysis in this context (Mueller \u0026amp; Saeltzer, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, comparative approaches remain relatively rare. By examining these two cases side by side, this study aims to shed light on how emotions function in political communication and how national and cultural contexts shape their use and impact.\u003c/p\u003e\n\u003ch3\u003ePolitical communication\u003c/h3\u003e\n\u003cp\u003ePolitical communication is a key element of public discourse, which forms the foundation for shaping social norms (Bleiker, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2003\u003c/span\u003e; Hutchison \u0026amp; Bleiker, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The ideas and values that dominate the public sphere determine the distinction between acceptable and desirable norms and norms that face criticism, ostracism, or sanctions. In this context, politicians play a particular role as primary actors in the political sphere, possessing greater agency than the average citizen. They can translate the ideals and values they hold into formal social rules, contributing to their broad legitimization. Interests, as one of the fundamental analytical categories, are a crucial tool in shaping political communication.\u003c/p\u003e \u003cp\u003eWeber (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) emphasized the role of interests as the primary driver of social actions, where purposeful-rational actions take precedence, with emotions playing a secondary role. Simultaneously, political interests are often shielded by cultural symbolism, with dominant groups imposing their interests as universally binding (Bourdieu, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e1984\u003c/span\u003e). Researchers such as Milliken (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e1999\u003c/span\u003e), Hansen (2006), Epstein (2008) and Zybertowicz (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e1995\u003c/span\u003e) observe the strong link between language and power, emphasizing that language functions as symbolic violence, through which certain norms and attitudes are promoted.\u003c/p\u003e \u003cp\u003eContemporary political communication occurs primarily through the media, making public discourse almost entirely a media discourse. In this context, social media platforms, especially X, play a fundamental role due to their global reach and significant opinion-shaping power. Research indicates that discourse on X can influence actual political outcomes, such as election victories (Kartsounidou et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Kruikemeier, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Rotesi, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). Politicians shape their messages per party lines by using professional communication management, thereby effectively influencing public opinion (Mueller \u0026amp; Saeltzer, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eSentiment\u003c/h2\u003e \u003cp\u003eThe study of emotions in politics emerged in the social sciences as a counterpoint to the rational approach to social actions (Demertzis, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, pp. 4\u0026ndash;5; Hutchison \u0026amp; Bleiker, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Today, we know that emotions not only complement rationality in political communication but often play a central role there (Koschut, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Politicians manage emotions to evoke collective social reactions (Hutchison \u0026amp; Bleiker, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, p. 37). In pursuing their interests, while politicians seek to gain support and approval for their actions, they must also act as a team, which often results in the building of sentiment (Meler, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Early studies on emotions relied on an ontological approach, viewing emotions as an evolutionary legacy that preserves reactions conducive to survival and reproduction. For example, fear triggers defensive reactions, disgust protects against poisoning, and anger enables competition for resources or social standing. Happiness, in turn, motivates actions that encourage a return to this state of happiness (Plutchik, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e1962\u003c/span\u003e, pp. VI\u0026ndash;VIII). However, these initial, mainly psychological approaches to emotions focus on individual experiences. Meanwhile, politics is a collective phenomenon (Berezin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2002\u003c/span\u003e). Therefore, when focusing on emotions in political communication, a cognitive-constructivist approach is more appropriate, suggesting that we learn emotions culturally, both in experiencing and expressing them (Koschut, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, pp. 3\u0026ndash;12).\u003c/p\u003e \u003cp\u003eThe ontological nature of emotions remains a subject of debate among researchers (Barrett, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2018\u003c/span\u003e), but is it always necessary to determine their ontological status? If we consider, for instance, the distinction between cognitive and constructivist approaches, it primarily concerns the ontological dimension. The cognitive approach attributes biological-evolutionary foundations to emotions but postulates that the communication of emotions is strongly shaped by culture. Constructivism goes further and assumes, even ontologically, that humans learn emotions through socialization. The ontology of emotions is irrelevant to my analysis since it focuses on the cultural nature of the communication of emotions. Politicians on platforms such as X appeal to socially rooted emotional figures, attempting to evoke emotions that translate into collective actions among the audience. Political efforts to build emotional energy, either supporting or opposing policies, are well captured by the sociological concept of communities of feeling (Berezin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, p. 38).\u003c/p\u003e \u003cp\u003eTo complete the conceptual framework regarding emotions in politics, it is essential to mention the distinction between emotions and affects (Hutchison \u0026amp; Bleiker, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, p. 5). Affects are sometimes triggered in politics by unforeseen events, but it is primarily emotions that play a dominant role in political communication as more enduring representations of moral judgments. In political communication, emotions are a tool that facilitates the exercise of symbolic violence, as people follow messages that elicit specific emotions. Often, provoking anger, fear, or general negativity in the audience can yield a quicker and more lasting promotional effect for a particular political idea than rational arguments.\u003c/p\u003e \u003c/div\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eSentiment Analysis\u003c/h2\u003e \u003cp\u003eEmotions are a very important component of political communication, and it is impossible to determine its impact without analyzing them. Nowadays, data science provides useful tools that allow for relatively objective analysis of emotions even in large text corpora. Sentiment analysis (SA) is the automated process of detecting and understanding the emotions conveyed through written text (Gunasekaran, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). It is used by researchers across various fields to determine attitudes toward something or someone or to analyze moods. This methodology serves to predict stock market trends (Bollen et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Turner et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) in economics and is applied to disease recognition through social media analysis (Yadav et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) in the medical field. In pedagogy, it has been used to study youth subcultures (Rudzionis et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Given the significant role of emotions in politics and political communication, SA has been applied in the current study as well. In SA, we can identify two primary methodological approaches: the lexicon-based approach and the machine-learning approach (Gunasekaran, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMy analysis was based on the latter, using the BERT transformer, where sentiment measurement is performed through machine-recognized attention (Devlin et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). This analysis results in identifying the level of a specific sentiment for a given text, with the same text potentially containing a certain level of negative, neutral, or positive sentiment. Variants of different models may help specify the level of extremely negative (or positive) and moderately negative (or moderately positive) sentiments. Transformers such as BERT are more advanced SA tools because they can \u0026ldquo;understand\u0026rdquo; context and even recognize the sender\u0026rsquo;s intentions (Hardalov et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In this analysis, BERT models were chosen to ensure compatibility for English and Polish languages and include variables of negative, neutral, and positive sentiment.\u003c/p\u003e \u003cp\u003eTo identify and understand differences in political communication, a simple test of the BERT tool was performed by comparing the sentiment of identical sentences in both languages.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eData\u003c/h3\u003e\n\u003cp\u003eThe research material for this analysis includes 127,000 tweets by Polish MPs and 210,000 tweets by British MPs posted during the first year of the Russia-Ukraine war, from February 2022 to January 2023. The tweets were collected from the accounts of 790 British MPs and 329 Polish MPs, which accounts for 55% of all MPs in both countries. In the following text, I use the terms MPs, parliamentarians, and the elite interchangeably. The latter is used rhetorically, fully aware that the set of all parliamentarians is not identical to the set of 'the elite', but with the conviction that MPs are an important part of the country's political elite. For each national dataset, keyword-based subsets of posts dedicated to the war in Ukraine were extracted. The datasets were then subjected to automatic SA using BERT models trained for English and Polish. Models were selected to provide the same sentiment measurement scale: negative, neutral, and positive. The output of the BERT transformer analysis was to determine for each post the proportion of negative, neutral, and positive sentiment as a decimal fraction, which summed up to 1.\u003c/p\u003e \u003cp\u003eThen, based on the assumption that the mood of media communication emerges only after aggregating individual statements into larger datasets (Meler, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), sentiment for both countries and both thematic subsets was aggregated by monthly periods. This provided averaged SA results for British parliamentarians\u0026rsquo; communications on Ukraine and other topics as well as analogous results for Polish parliamentarians\u0026rsquo; communications on the same topics. The aggregation method is depicted in Tables\u0026nbsp;3 and 4.\u003c/p\u003e\n\u003ch3\u003eTopics analysis\u003c/h3\u003e\n\u003cp\u003eLatent Dirichlet Allocation (LDA) is a probabilistic generative model used for topic modeling in large corpora of text. The model assumes that documents are mixtures of topics, where each topic is a distribution over words, and each document is represented as a distribution over topics. LDA operates under the assumption that each word in a document is generated from one of the document's topics, which in turn is drawn from a fixed set of topics shared across the corpus (Blei et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). By applying a Dirichlet prior distribution to both the topic distributions for each document and the word distributions for each topic, LDA allows for the discovery of latent structures within a collection of texts. This methodology has been successfully applied to analyze the behavior of politicians on X (Kushwaha et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe shaping of social attitudes does not come out of nowhere, and by examining political communication, we can trace their emergence.\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003eSignificance of Russia-Ukraine War in Overall Political Communication on X\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eThe significance of the Russia-Ukraine War in political communication, understood as the share of tweets on this subject in the overall pool of X posts by Polish and British parliamentarians, experienced a dynamic pattern. They showed high shares in the first two months and a drop in the next two months, subsequently stabilizing at a relatively low level by the end of the analyzed period. Although the trend lines run similarly, there is a difference in the share levels: tweets related to the conflict in Ukraine constituted between 10% and 56% of the Polish dataset, while for British MPs, it ranged from 4% to 39%.\u003c/p\u003e\n\u003cp\u003eTable 1.\u0026nbsp;The Topic of Ukraine vs. Overall X Communication.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePeriod\u003cbr\u003e\u003c/strong\u003e\u003cstrong\u003e(year-month)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 41px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUK\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 41px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePL\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003eUkraine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003eOther Topics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003eUkraine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003eOther Topics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e29%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e71%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e24,458\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e44%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e56%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e15,057\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;03\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e39%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e61%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e27,642\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e56%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e44%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e18,773\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;04\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e16%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e84%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e18,776\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e33%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e67%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e10,875\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;05\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e18,870\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e22%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e78%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,319\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;06\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e8%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e92%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e17,131\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e19%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e81%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,816\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;07\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e18,801\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e11%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e89%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,777\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;08\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e93%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e12,838\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e10%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e8,361\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;09\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e94%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e13,121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e15%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e85%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,973\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;10\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e19,524\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e13%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e87%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,804\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e16,016\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e14%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e86%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e9,279\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;12\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e12,957\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e12%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e88%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e8,214\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2023\u0026ndash;01\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e94%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e10,668\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e21%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e79%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e8,619\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTotal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e14%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e86%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e210,802\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e26%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 17px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e74%\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 11px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e127,867\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eComparison of Sentiment on the Topic of Ukraine Against Other Topics\u0026nbsp;\u003c/h2\u003e\n\u003cp\u003eCompared to tweets on other topics, significant differences can be observed in the political communication of British and Polish parliamentarians about Ukraine. Positive and negative sentiments concerning Ukraine varied across periods in both countries. Meanwhile, the aggregated sentiment of other tweets showed no significant differences between periods or between countries (the standard deviation was twice as high for the Ukraine topic compared to others). A ratio, commonly used in decision trees, was applied to determine the scale of differences between individual sentiments, representing the value ratio for the first group in a given period relative to the second group. A 100% indicator would denote identical sentiment for both groups.\u003c/p\u003e\n\u003cp\u003eTable 2.\u0026nbsp;Average Sentiment Loadings in MPs\u0026rsquo; Tweets Over Periods: Ukraine vs. Other Topics.\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePeriod\u003cbr\u003e\u003c/strong\u003e\u003cstrong\u003e(year-month)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 38px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUK\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"6\" valign=\"top\" style=\"width: 48px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePL\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 19px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUkraine\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 19px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOther Topics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 20px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUkraine\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\" valign=\"top\" style=\"width: 28px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eOther Topics\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeg.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeu.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePos.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeg.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeu.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePos.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeg.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeu.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePos.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeg.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeu.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePos.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;03\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;04\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;05\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;06\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;07\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;08\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;09\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;10\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;11\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.24\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2022\u0026ndash;12\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.39\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e2023\u0026ndash;01\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTotal mean\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.39\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.37\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.36\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.28\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.48\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.24\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.27\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.46\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.26\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 13px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eStd. dev.\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.04\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.04\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.08\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.01\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 7px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.03\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 6px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.03\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 10px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.01\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 9px;\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.02\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eWhen we compare the sentiment regarding Ukraine (first group) to that of other topics (second group), we see that British parliamentarians\u0026rsquo; communication about Ukraine was more positive overall throughout the year than on other topics. This indicator also showed greater variability over time\u0026mdash;from 112% in February 2022 to 57% negative sentiment in August 2022. In contrast, Polish parliamentarians\u0026rsquo; tweets about Ukraine were more negative, compared to other topics. An analysis of British MPs\u0026rsquo; tweets reveals that in eight out of 12 months, the share of positive sentiment in the Ukraine topic was higher than for other topics. Negative sentiment was lower in nine out of 12 months. Meanwhile, in Polish parliamentarians\u0026rsquo; tweets about Ukraine, there was mainly a decline in positive sentiment in most of the analyzed period (eight out of 12 months).\u003c/p\u003e\n\u003ch2\u003eComparison of the Sentiment of the Topic of Ukraine Between Countries\u003c/h2\u003e\n\u003cp\u003eWhen comparing the sentiment of tweets about Ukraine between British and Polish parliamentarians, the main difference was found in the share of positive sentiment. In British MPs\u0026rsquo; tweets, this share was higher (an average of 36%) than in Polish MPs\u0026rsquo; tweets (an average of 26%). Notably, this was a general difference, applying to both Ukraine-related and other topics. The indicator for the share of negative sentiment tweets among British (first group) and Polish parliamentarians (second group) in the context of Ukraine ranged from 105% in February 2022 to 57% in August 2022. Only in February and June 2022 were British MPs\u0026rsquo; tweets more negative than those of their Polish counterparts, though the differences were small.\u003c/p\u003e\n\u003cp\u003eThe sentiment trend regarding Ukraine, represented by dashed lines (chart 1), shows that in the first two months of the analyzed period, British parliamentarians expressed significantly more negative opinions about Ukraine than about other topics. Later, their communication about Ukraine became more positive vis-a-vis other topics, peaking in August 2022; after September 2022, it no longer differed significantly from the rest of the political communication. For Polish parliamentarians, at the beginning of the war, the sentiment regarding Ukraine did not deviate significantly from general political communication, but in subsequent months, it became noticeably less positive compared to other topics. The largest sentiment discrepancies between British and Polish parliamentarians occurred in July and August 2022 and January 2023.\u003c/p\u003e\n\u003ch2\u003eTopic Analysis Using LDA\u003c/h2\u003e\n\u003cp\u003eTo understand the reasons behind the convergence of positive sentiment levels in Polish and British communication (generally divergent in other periods) in February 2022, as well as the divergence of negative sentiment levels (previously quite similar), we conducted content analysis using LDA (see Appendix).\u003c/p\u003e\n\u003cp\u003eIn February 2022, British parliamentarians\u0026rsquo; tweets began to feature words like \u0026ldquo;war,\u0026rdquo; \u0026ldquo;invasion,\u0026rdquo; \u0026ldquo;attack,\u0026rdquo; and \u0026ldquo;sanction,\u0026rdquo; which explain the lowered sentiment. In Polish communications, words like \u0026ldquo;war\u0026rdquo; (pol.: \u0026ldquo;wojna\u0026rdquo;) and \u0026quot;aggression\u0026quot; (pol.: \u0026ldquo;agresja\u0026rdquo;) also appeared, but terms like \u0026ldquo;help\u0026rdquo; and \u0026ldquo;support\u0026rdquo; (pol.: \u0026ldquo;pomoc\u0026rdquo;, \u0026ldquo;wsparcie\u0026rdquo;) were equally prominent. Further, in February, the topic of \u0026ldquo;refugee\u0026rdquo; (pol.: \u0026ldquo;uchodźcy\u0026rdquo;) emerged in Polish tweets, becoming a dominant theme in their communications. However, the term was much less present in British discourse.\u003c/p\u003e\n\u003cp\u003eThe Polish discourse began to refer to events in Ukraine as \u0026ldquo;war\u0026rdquo; more frequently only in April, which influenced a more negative tone in content. In February, however, both the Polish and British discourses centered around terms like \u0026ldquo;war,\u0026rdquo; \u0026ldquo;invasion,\u0026rdquo; and \u0026ldquo;attack.\u0026rdquo; In Polish tweets, there was an additional topic related to \u0026ldquo;support,\u0026rdquo; referring to the assistance for Ukrainian refugees who were warmly received at both government and social levels. Most refugees found shelter in private homes, eliminating the need for camps, even as their numbers exceeded one million within the first week of the war (Zalewski, 2022). The topic of support evoked a positive sentiment in Polish discourse, while British messaging maintained more neutral or negative tones, thus bringing positive sentiment levels closer between the two countries.\u003c/p\u003e\n\u003cp\u003eIn March, British tweets began to prominently feature terms like \u0026ldquo;support,\u0026rdquo; \u0026ldquo;people,\u0026rdquo; \u0026ldquo;refugees,\u0026rdquo; \u0026ldquo;children,\u0026rdquo; and \u0026ldquo;family,\u0026rdquo; which significantly increased the positive tone of the messaging. Interestingly, in Polish tweets, \u0026ldquo;refugees,\u0026rdquo; \u0026ldquo;support,\u0026rdquo; and \u0026ldquo;help\u0026rdquo; were as dominant as in February, keeping Polish sentiment almost unchanged.\u003c/p\u003e\n\u003cp\u003eIn August, at the height of the differences between messages, British MPs wrote mainly about \u0026ldquo;support,\u0026rdquo; \u0026ldquo;people,\u0026rdquo; \u0026ldquo;humanitarian,\u0026rdquo; and \u0026ldquo;help,\u0026rdquo; which reinforced the positive sentiment. Meanwhile, in Polish discourse, terms like \u0026ldquo;campuspolska\u0026rdquo; or \u0026ldquo;campus\u0026rdquo; appeared, referring to a cyclical party event of the then-opposition, where Ukraine was also a significant topic. The frequency of posts about \u0026ldquo;refugees\u0026rdquo; and \u0026ldquo;support\u0026rdquo; declined, shifting the focus of Polish messaging.\u003c/p\u003e\n\u003ch2\u003eSentiment Comparison by Political Affiliation\u003c/h2\u003e\n\u003cp\u003eIn Polish communication, especially in the latter half of the analyzed period, the topic was defined by the names of prominent Polish politicians, raising the question of whether the communication sentiment was shaped by affiliation with the ruling or opposition camp.\u003c/p\u003e\n\u003cp\u003eVisual analysis in Chart 2 already suggests that the sentiment of communication by the ruling and opposition camps differed significantly more for British parliamentarians. A T-test indicates a statistically significant difference in both negative and positive sentiments between the ruling and opposition camps in both countries, but the \u003cem\u003eeta squared\u0026nbsp;\u003c/em\u003ecoefficient shows that political affiliation had a much stronger impact on sentiment in the UK than in Poland. In the UK, 81% of the variation in negative sentiment in individual periods can be explained by political affiliation; in contrast, the percentage in Poland was 60%. For positive sentiment, this relationship was even more divergent: 66% in UK communication and 47% in Poland.\u003c/p\u003e\n\u003cp\u003eTable 3.\u0026nbsp;Value of the \u003cem\u003eEta Squared\u003c/em\u003e in Sentiment Comparison Between Governing and Opposition\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"100%\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 35px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCountry\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNegative\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eNeutral\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePositive\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 35px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUK\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003e0.81*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e0.66*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 35px;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePL\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 27px;\"\u003e\n \u003cp\u003e0.60*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 18px;\"\u003e\n \u003cp\u003e0.47*\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e* T-test significance (p-value \u0026lt;0.05); sentiment as dependent\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eSocial sciences have long emphasized the importance of the so-called \u0026ldquo;humanistic coefficient,\u0026rdquo; underscoring the need to consider both individual and collective interpretations of reality in research (Znaniecki, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e1934\u003c/span\u003e). An analysis of behaviors on social media platforms can be an effective tool in achieving this goal, where sentiment in online communication represents one of the key dimensions of communication.\u003c/p\u003e \u003cp\u003eThe first important conclusion from the analysis concerns methodology. Automated discourse analysis methods not only enable the examination of larger datasets but also contribute to the objectivity of results. When a researcher interprets text individually, as is standard in social sciences (van Dijk, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1998\u003c/span\u003e), there is a greater risk of personal beliefs affecting the outcomes. Automated methods such as transformer models or LDA significantly reduce this risk. For example, the BERT model, trained on a very large corpus of texts, mitigates some individual biases. Nevertheless, a methodological test shows that analyzing politically charged sentences may require more specific model training.\u003c/p\u003e \u003cp\u003eTurning to the SA results of British and Polish MPs\u0026rsquo; political communication, British communication was significantly more positive than Polish communication. We also observe that the topic of Ukraine held more significance for the Polish political class than for the British, when comparing it to their entire scope of political communication. Polish MPs dedicated relatively more posts to this topic, with their communications on the Ukraine war reflecting a negative tone than on other issues. Thus, Polish politics has been significantly more engaged in Ukrainian affairs than British politics. Interestingly, British MPs referred to the Ukraine war in a less negative tone compared to other topics. The peak of enthusiasm in British posts about Ukraine occurred in August 2022, largely due to the British Prime Minister\u0026rsquo;s visit to Kyiv during Ukraine\u0026rsquo;s Independence Day celebrations. In Polish communication, visits by the Polish Prime Minister and President, which occurred much more frequently than visits by other high-ranking officials, did not have such a marked effect.\u003c/p\u003e \u003cp\u003eThe war has not directly impacted British or Polish societies. However, the political elites of both countries adopted a strategy to deter Russia through multi-level support for Ukraine that caused significant financial strains on both countries. Politicians had to gain public support for decisions that ultimately meant burdens on the population. SA indicates that the British political class quickly began building support based on negative messaging. Negative communication can be a mobilizing factor (McNoir, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; van Dijk, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e1998\u003c/span\u003e), but its prolonged use may discourage voters (Ansolabehere \u0026amp; Iyengar, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Perloff, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). British communication followed the textbook approach; it increased negative messaging in the first two months, then diminished negative emotions and shifted to a relatively positive tone before eventually returning to an average sentiment level.\u003c/p\u003e \u003cp\u003eOne may question why the sentiment management pattern in Polish political communication was different. Initially, the tone was not more negative than in the rest of their communication, but later, it became negative and remained so for longer than that of their British counterparts. Content analysis using LDA and contextual knowledge suggested that Polish politicians in February and March 2022 first needed to effectively alleviate the migration crisis. Managing this crisis required a positive attitude within Polish society, reflected in maintaining a more positive tone (than the British). Differences in sentiment between British and Polish MPs are better understood through the interpretive framework of the \u003cem\u003esecurity state\u003c/em\u003e and \u003cem\u003ecommunity of feelings\u003c/em\u003e (Berezin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, p. 38). We can assume that the more negative sentiment in Polish communication are manifestations of fear about losing security-state status, which Polish society has only recently started to take for granted after a turbulent unsecurity state period that concluded with the systemic transformation in the early 1990s. The UK has, for centuries, represented an archetype of a secure state due to its political structure and safe, insular location. The more positive tone of British MPs indicates their ability to build a \u003cem\u003ecommunity of feeling\u003c/em\u003e as during times of somber but non-war-related events, such as the death of Princess Diana (Berezin, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2002\u003c/span\u003e, p. 40).\u003c/p\u003e \u003cp\u003eAppropriate communication supports the mobilization of the community for Ukraine. Ukraine finds itself attacked by a great power that other powers eagerly supported, similar to Belgium at the start of World War I. In Polish communication, the creation of a community of feeling manifested in diverse ways but was dominated by a strong sense of threat to their own country. The more negative sentiment in Polish tweets reflect fear, as concerns about security are highly current. Russia\u0026rsquo;s aggressive actions revive historical fears for Poland. The prospect of Ukraine falling into Russia\u0026rsquo;s sphere of influence evokes memories of dependence on Russia, still alive in many Poles\u0026rsquo; minds and familiar to all from history lessons. For the UK, Russia might appear to be yet another international problem, especially UK has previously been victorious (Crimean War) and the countries have been allies on several occasions. Consequently, Poles writing about Russian aggression culturally and instinctively limit positive tones, as a Ukrainian defeat would signal a return to the \u0026ldquo;Russian world,\u0026rdquo; remembered by generations of Poles as economic, social, and cultural degradation. Unsurprisingly, Polish politicians find it difficult to maintain higher proportion of positive sentiment, which seems significantly easier for their British counterparts.\u003c/p\u003e \u003cp\u003eHere, we can speak of a stronger \u0026ldquo;flag effect\u0026rdquo; in Poland. Noticeably, the topic of Ukraine was critical only in the first two months and then lost prominence. As this topic receded, sentiment levels for both groups of MPs began to fluctuate within narrower limits. This suggests that in political communication (and perhaps more broadly in politics), there is little room for affects; only emotions remain relevant according to the earlier distinction (Hutchison \u0026amp; Bleiker, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding details:\u003c/strong\u003e This work was supported by the Nicolaus Copernicus University in Toruń under Grant IDUB/Debiuty_6_Andrzej Meler.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure statement:\u003c/strong\u003e The author report there are no competing interests to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBiographical note:\u003c/strong\u003e A Nicolaus Copernicus University\u0026rsquo;s Institute of Sociology graduate. He received the Polish Sociological Association award for his master\u0026rsquo;s thesis, which focused on the analysis of media discourse on the Polish judiciary system. He worked as a web analyst for Polska Press Group for ten years. Currently his academic research focuses on political discourse in media.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability statement:\u003c/strong\u003e Data availability after contact by ResearchGate: https://www.researchgate.net/profile/Andrzej-Meler\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAnsolabehere, S., \u0026amp; Iyengar, S. (1995). \u003cem\u003eGoing Negative: How Political Advertisements Shrink and Polarize the Electorate\u003c/em\u003e. Free Press.\u003c/li\u003e\n\u003cli\u003eBarrett, L. F. (2018). \u003cem\u003eHow Emotions Are Made: The Secret Life of the Brain\u003c/em\u003e. Pan MacMillan.\u003c/li\u003e\n\u003cli\u003eBerezin, M. (2002). Secure states: Towards a political sociology of emotion. \u003cem\u003eSociological Review\u003c/em\u003e, \u003cem\u003e50\u003c/em\u003e(S2), 33\u0026ndash;52. https://doi.org/10.1111/j.1467-954X.2002.tb03590.x\u003c/li\u003e\n\u003cli\u003eBlei, D. M., Ng, A. Y., \u0026amp; Jordan, M. I. (2003). Latent Dirichlet Allocation. \u003cem\u003eJournal of Machine Learning Research\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e, 993\u0026ndash;1022. https://doi.org/10.1016/B978-0-12-411519-4.00006-9\u003c/li\u003e\n\u003cli\u003eBleiker, R. (2003). Discourse and Human Agency. \u003cem\u003eContemporary Political Theory\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(March 2003), 25\u0026ndash;47. https://doi.org/10.1057/palgrave.cpt.9300073\u003c/li\u003e\n\u003cli\u003eBollen, J., Mao, H., \u0026amp; Zeng, X. (2011). Twitter mood predicts the stock market. \u003cem\u003eJournal of Computational Science\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(1), 1\u0026ndash;8. https://doi.org/10.1016/j.jocs.2010.12.007\u003c/li\u003e\n\u003cli\u003eBourdieu, P. (1984). \u003cem\u003eDistinction: A Social Critique of the Judgement of Taste.\u003c/em\u003e (R. Nice (ed.)). Harvard University Press.\u003c/li\u003e\n\u003cli\u003eDemertzis, N. (2020). \u003cem\u003eThe political sociology of emotions\u003c/em\u003e. Routledge.\u003c/li\u003e\n\u003cli\u003eDevlin, J., Chang, M.-W., Lee, K., Google, K. T., \u0026amp; Language, A. I. (2018). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. \u003cem\u003eNaacl-Hlt 2019\u003c/em\u003e, \u003cem\u003eMlm\u003c/em\u003e, 4171\u0026ndash;4186. https://aclanthology.org/N19-1423.pdf\u003c/li\u003e\n\u003cli\u003eGunasekaran, K. P. (2019). Review on Techniques in Natural Language Processing. \u003cem\u003eInternational Journal of Scientific Research in Computer Science and Engineering\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(5), 1\u0026ndash;4. https://doi.org/10.26438/ijsrcse/v7i5.14\u003c/li\u003e\n\u003cli\u003eHardalov, M., Koychev, I., \u0026amp; Nakov, P. (2023). Enriched Pre-trained Transformers for Joint Slot Filling and Intent Detection. \u003cem\u003eInternational Conference Recent Advances in Natural Language Processing, RANLP\u003c/em\u003e, 480\u0026ndash;493. https://doi.org/10.26615/978-954-452-092-2_054\u003c/li\u003e\n\u003cli\u003eHutchison, E., \u0026amp; Bleiker, R. (2020). Emotion, agency and power in word politics. In S. Koschut (Ed.), \u003cem\u003eThe Power of Emotions in World Politics\u003c/em\u003e. Routledge.\u003c/li\u003e\n\u003cli\u003eKartsounidou, E., Papaxanthi, D., \u0026amp; Andreadis, I. (2023). Measuring the impact of candidates\u0026rsquo; tweets on their electoral results. \u003cem\u003eJournal of Information Technology \u0026amp; Politics\u003c/em\u003e, \u003cem\u003e20\u003c/em\u003e(2), 154\u0026ndash;168. https://doi.org/10.1080/19331681.2022.2069181\u003c/li\u003e\n\u003cli\u003eKoschut, S. (2020). Emotion, discourse, and power in world politics. In S. Koschut (Ed.), \u003cem\u003eThe Power of Emotions in World Politics\u003c/em\u003e (Issue 112, p. 165). Routledge.\u003c/li\u003e\n\u003cli\u003eKruikemeier, S. (2014). How political candidates use Twitter and the impact on votes. \u003cem\u003eComputers in Human Behavior\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e, 131\u0026ndash;139. https://doi.org/10.1016/j.chb.2014.01.025\u003c/li\u003e\n\u003cli\u003eKushwaha, A. K., Mandal, S., Pharswan, R., Kar, A. K., \u0026amp; Ilavarasan, P. V. (2020). Studying Online Political Behaviours as Rituals: A Study of Social Media Behaviour Regarding the CAA. \u003cem\u003eIFIP Advances in Information and Communication Technology\u003c/em\u003e, \u003cem\u003e618\u003c/em\u003e, 315\u0026ndash;326. https://doi.org/10.1007/978-3-030-64861-9_28\u003c/li\u003e\n\u003cli\u003eMcNoir, B. (2018). \u003cem\u003eAn Introduction to Political Communication\u003c/em\u003e (6th ed.). Routledge.\u003c/li\u003e\n\u003cli\u003eMeler, A. (2022). Parlamentarne ćwierkanie o pandemii. Analiza sentymentu tweet\u0026oacute;w parlamentarzyst\u0026oacute;w publikowanych podczas pierwszych 12 miesięcy pandemii koronawirusa w Polsce. \u003cem\u003eStudia Socjologiczne\u003c/em\u003e, \u003cem\u003e2022\u003c/em\u003e(2), 113\u0026ndash;136. https://doi.org/10.24425/sts.2022.141425\u003c/li\u003e\n\u003cli\u003eMeler, A. (2024). In The Beginning, Let There Be The Word: Challenges and Insights in Applying Sentiment Analysis to Social Research. In C. Tat_seng (Ed.), \u003cem\u003eWWW \u0026rsquo;24: Proceedings of the ACM Web Conference 2024\u003c/em\u003e (pp. 1214\u0026ndash;1217). Association for Computing Machinery. https://doi.org/10.1145/3589335.3651264\u003c/li\u003e\n\u003cli\u003eMilliken, J. (1999). The study of discourse in international relations: A critique of research and methods. \u003cem\u003eEuropean Journal of International Relations\u003c/em\u003e, \u003cem\u003e5\u003c/em\u003e(2), 225\u0026ndash;254. https://doi.org/10.1177/1354066199005002003\u003c/li\u003e\n\u003cli\u003eMueller, S. D., \u0026amp; Saeltzer, M. (2020). Twitter made me do it! Twitter\u0026rsquo;s tonal platform incentive and its effect on online campaigning. \u003cem\u003eInformation Communication and Society\u003c/em\u003e, \u003cem\u003e0\u003c/em\u003e(0), 1\u0026ndash;26. https://doi.org/10.1080/1369118X.2020.1850841\u003c/li\u003e\n\u003cli\u003ePerloff, R. M. (2022). \u003cem\u003eThe Dynamics of Political Communication Media and Politics in a Digital Age\u003c/em\u003e (3rd ed.). Routledge.\u003c/li\u003e\n\u003cli\u003ePlutchik, R. (1962). \u003cem\u003eThe Emotions. Facts, Theories, and a New Model\u003c/em\u003e. Random House.\u003c/li\u003e\n\u003cli\u003eRanciere, J. (2006). Hatred of Democracy. In S. Corcoran (Ed.), \u003cem\u003eWhat Is Democracy and How Do We Study It?\u003c/em\u003e Verso. https://doi.org/10.1080/09614520801899259\u003c/li\u003e\n\u003cli\u003eRotesi, T. (2019). The impact of Twitter on political participation. \u003cem\u003eBocconi University, Milan, Italy Working Paper\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eRudzionis, V., Ramanuskaite, E., \u0026amp; Kairaityte-Uzuper, A. (2023). Sentiment Analysis of Lithuanian Youth Subcultures Zines Using Automatic Machine Translation. In A. Lopata, R. Butkiene, \u0026amp; D. Gudoniene (Eds.), \u003cem\u003eInformation and Software Technologies\u003c/em\u003e (pp. 201\u0026ndash;221). Springer. https://doi.org/10.1007/978-3-031-48981-5\u003c/li\u003e\n\u003cli\u003eTurner, Z., Labille, K., \u0026amp; Gauch, S. (2021). Lexicon-based sentiment analysis for stock movement prediction. \u003cem\u003eJournal of Construction Materials\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(3), 149\u0026ndash;154. https://doi.org/10.36756/jcm.v2.3.5\u003c/li\u003e\n\u003cli\u003evan Dijk, T. A. (1998). Oppionions and Ideologies in the Press. In A. Bell \u0026amp; P. Garrett (Eds.), \u003cem\u003eApproaches to Media Discourse\u003c/em\u003e (pp. 21\u0026ndash;63). Blackwell.\u003c/li\u003e\n\u003cli\u003eVergeer, M. (2015). Twitter and Political Campaigning. \u003cem\u003eSociology Compass\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e, 745\u0026ndash;760. https://doi.org/10.1111/soc4.12294\u003c/li\u003e\n\u003cli\u003eWeber, M. (2019). \u003cem\u003eEconomy and Society\u003c/em\u003e (K. Tribe (ed.)). Harvard University Press.\u003c/li\u003e\n\u003cli\u003eYadav, S., Ekbal, A., Saha, S., \u0026amp; Bhattacharyya, P. (2019). Medical sentiment analysis using social media: Towards building a patient assisted system. \u003cem\u003eLREC 2018 - 11th International Conference on Language Resources and Evaluation\u003c/em\u003e, 2790\u0026ndash;2797.\u003c/li\u003e\n\u003cli\u003eZalewski, P. (2022). Działania administracji państwowej w Polsce wobec uchodźc\u0026oacute;w z Ukrainy w pierwszych tygodniach wojny w Ukrainie 2022 roku. Aspekty prawne i securitologiczne. \u003cem\u003eStudia Politicae Universitatis Silesiensis\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e, 101\u0026ndash;124. https://doi.org/10.31261/spus.13934\u003c/li\u003e\n\u003cli\u003eZnaniecki, F. (1934). \u003cem\u003eThe Method of Sociology\u003c/em\u003e. Rinehart \u0026amp; Company.\u003c/li\u003e\n\u003cli\u003eZybertowicz, A. (1995). \u003cem\u003ePrzemoc i poznanie\u003c/em\u003e (1st ed.). Wydawnictwo Uniwersytetu Mikołaja Kopernika.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"social-network-analysis-and-mining","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"snam","sideBox":"Learn more about [Social Network Analysis and Mining](http://link.springer.com/journal/13278)","snPcode":"13278","submissionUrl":"https://submission.nature.com/new-submission/13278/3","title":"Social Network Analysis and Mining","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"political communication, sentiment analysis, computational sociology, BERT, LDA","lastPublishedDoi":"10.21203/rs.3.rs-6771056/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6771056/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"In moments of pivotal geopolitical importance, such as Russia’s aggression against Ukraine, sentiment analysis of political communication offers a unique lens through which to examine how national interests are articulated and emotional tones are conveyed in public discourse. This study explores how Members of Parliament (MPs) in Poland and the United Kingdom addressed the war on the X platform (formerly Twitter) during its first year, treating parliamentary communication as an indicator of collective political positioning. While both countries have consistently supported Ukraine, the analysis revealed notable differences: British MPs expressed significantly more positive sentiment overall, whereas Polish MPs, despite addressing the Ukrainian topic more frequently, communicated in a more emotionally restrained and neutral tone. The study thus sheds light on the contrasting communicative styles and underlying political orientations that emerge in response to the same international crisis. Contemporary natural language processing tools—such as language models and thematic analysis—were employed to systematically compare messages across two distinct political and linguistic contexts.","manuscriptTitle":"Two Voices, One War: NLP and LDA for Sociological Analysis of Political Communication on Twitter vis-a-vis the Ukrainian War","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-04 10:17:33","doi":"10.21203/rs.3.rs-6771056/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-07-15T18:37:11+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-15T01:08:35+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-10T15:06:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-05T18:16:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"62280611611259198074893442228179330562","date":"2025-06-16T15:27:00+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"168627858768618397412626085904047615661","date":"2025-06-15T11:57:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"278285606545913361106584078937620752345","date":"2025-06-13T11:05:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"323572805314314963771266534698493055020","date":"2025-06-13T09:34:28+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-06-01T22:33:01+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-01T22:31:03+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-29T14:02:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"Social Network Analysis and Mining","date":"2025-05-28T20:56:18+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"social-network-analysis-and-mining","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"snam","sideBox":"Learn more about [Social Network Analysis and Mining](http://link.springer.com/journal/13278)","snPcode":"13278","submissionUrl":"https://submission.nature.com/new-submission/13278/3","title":"Social Network Analysis and Mining","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"2d61678d-2520-4e0d-acbe-1e62d0524195","owner":[],"postedDate":"June 4th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-15T16:16:58+00:00","versionOfRecord":{"articleIdentity":"rs-6771056","link":"https://doi.org/10.1007/s13278-025-01538-4","journal":{"identity":"social-network-analysis-and-mining","isVorOnly":false,"title":"Social Network Analysis and Mining"},"publishedOn":"2025-12-14 15:59:15","publishedOnDateReadable":"December 14th, 2025"},"versionCreatedAt":"2025-06-04 10:17:33","video":"","vorDoi":"10.1007/s13278-025-01538-4","vorDoiUrl":"https://doi.org/10.1007/s13278-025-01538-4","workflowStages":[]},"version":"v1","identity":"rs-6771056","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6771056","identity":"rs-6771056","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}