While polarisation is on the rise in Western Europe, there is little descriptive evidence about which issues divide voters and how. There are even fewer studies that seek to explain how party salience strategies respond to polarising issues. This study addresses these gaps, underscoring a conceptual distinction that is often missed in the literature on issue competition — namely, that issues can divide the electorate at large (‘general issue polarisation’), or according to the party they support (‘party issue polarisation’). Survey data about voter preferences in six Western European countries shows that the correlation between these two types of issue polarisation is weak. For example, while some cultural issues like soft drug policy and same-sex marriage are highly polarising in general, they do not divide supporters of opposing parties so much. Using party Twitter activity as the dependent variable, it is shown that parties put less emphasis on issues the more extreme voter preferences are, running contrary to predictions based on prior work. By contrast, issues higher in party polarisation received more attention. This study highlights the importance of taking a multidimensional approach to issue polarisation, and calls for more refined theories of how parties deal with polarising issues.

The COVID-19 pandemic triggered an unprecedented surge in Twitter activity, providing a rich source of public opinions and emotions. This study proposes a fusion model combining a bidirectional GRU (BiGRU) and a DistilBERT transformer, with their learned features concatenated and fed into an XGBoost meta-classifier for final sentiment prediction. We evaluate our approach on over one million COVID-19-related English-language tweets collected from eight countries between January and April 2020. The fusion model (BiGRU + DistilBERT + XGBoost) achieves a classification accuracy of ~ 85.8%, outperforming individual models (for example, DistilBERT alone at 85.5% accuracy) in sentiment detection. Key results indicate that public sentiment evolved with pandemic phases: negative sentiments peaked during surges in cases and deaths, while positive sentiments rose during recovery periods. We also observe notable regional differences - for instance, sentiment trends vary across countries and between English- vs. non-English-speaking regions. These findings demonstrate the effectiveness of our context-infused approach, offering valuable insights for policymakers on social media sentiment during health crises.

The price of controversy: CEO Twitter activity and stock market performance

Using a novel dataset on Twitter activity as well as a novel corpus of law journal publications, this paper examines the impact of social media activity on the scholarly success of U.S. law professors. We find that joining Twitter increases citation counts by an average of 22% per year and improves article placements by up to 10 ranks for law professors, relative to a synthetic control group. These positive returns apply across nearly all classes of scholars and are magnified for those who post frequently about their own work. The identified citation boost would be even larger than 22% if it were not partially offset by a decline in citations to articles published pre-Twitter. Overall, our results suggest that social media participation yields concrete benefits in the legal academy—indeed, benefits outstripping those that prior studies have identified in other disciplines—along with a number of potential downsides.

ABSTRACT This article examines inter-individual differences in the communication patterns of German political journalists by analyzing a snapshot of the ongoing Russian-Ukrainian war discourse on Twitter. Journalists exhibit significant variation in Twitter activity and posting of individual contributions. We use aggregated individual data from content analysis as the foundation for network comparisons. By analyzing war communication from an aggregated individual perspective, the study addresses understudied aspects of journalists’ social media use. Specifically, we analyze how tweets correspond to war/peace journalism framing and how journalistic accounts within this network relate to this distribution. To achieve this, we identify indicators of war/peace journalism framing over a five-month research period starting in September 2022. The content analysis reveals that peace journalism framing predominates over war journalism framing at the aggregated actor level. However, journalists inclined toward war journalism framing occupy significantly higher ranks in network centrality. Additionally, we observe substantial inter-individual differences among journalists in terms of opinion expression in tweets, with a positive association between a journalist’s tendency to express opinions and their centrality rank. The article illustrates how network-based attributes and insights from journalistic framing analysis can be integrated to explore inter- and intra-individual dynamics among journalistic actors.

ABSTRACT This study examines how multimodal resources in Kenyan Twitter activism against police brutality construct representations of social actors and conveys ideological positions. Drawing on 496 purposively sampled tweets under the #JusticeForKianjokomaBrothers, the study employs a multimodal critical discourse analysis approach. It examines how semiotic resources frame victims, state actors, activists, and family while exposing embedded ideologies. Linguistic resources such as naming and familial nouns, humanize victims as innocent and vulnerable, reinforced by emotive visual cues. In contrast, police are delegitimized through collectivized nouns, negative lexical items and symbolic imagery. Activists are portrayed as a unified collective using plural nouns and solidarity symbols, while maternal grief and collective suffering frame the family emphasizing the social impact. The findings present police brutality as systemic, stress the moral obligation to resist and position digital activism as a powerful tool for social change. This research contributes to digital protest discourse by showing how multimodal elements mobilize public sentiment and challenge entrenched power structures.

This study focused on the influence of sports brands on social media, and selected a number of representative activities on the official Twitter accounts of Nike and Adidas Football as the research object. By collecting the relevant tweet data during 2022-2023, including the independent variable event type, the number of followers of the protagonist, and the release time; Based on variables such as comments, likes, retweets and page views, this paper applies the quantitative analysis of social platform influence, multi-factor analysis and descriptive statistics methods to deeply explore the key factors affecting the influence of sports brands on social media, as well as the influence of different activity types on user interaction behavior. The results show that different types of activities have significant differences on brand influence, among which "star micro film shooting activities" score the highest, but the stability is poor; The performance of "champion reporting activities" was more stable. The study also suggests optimizing sports brands' social media strategies, aiming to increase brands' influence and user engagement on social media.

This study aims to explore the experiences of adolescents who are active Twitter users in building virtual communities as safe spaces. Using a literature review method, this study examines various scholarly sources that discuss adolescent engagement on social media, identity formation, and the dynamics of online communities. The findings indicate that Twitter is used by adolescents not only for social interaction but also as a means of identity exploration and emotional support. The virtual communities formed provide a sense of safety, solidarity, and a space for expression that is often unavailable in offline social environments. However, challenges such as digital toxicity, cyberbullying, and exclusion are also part of the dynamics within these communities. Therefore, a deeper understanding of the role of social media in adolescent life is necessary, along with efforts to strengthen ethical and empathetic digital literacy.

Firearm injuries are a leading cause of death in the United States, surpassing fatalities from motor vehicle crashes. Despite this significant public health risk, Americans continue to purchase firearms in large quantities. Commonly cited drivers of firearm acquisition include fear of violent crime, fear of mass shootings, and panic-buying. Additionally, advocacy groups’ activity on social media may capitalize on emotions like fear and influence firearm acquisition. The simultaneous effects of these variables have not been explored in a causal framework. In this study, we aim to elucidate the causal roles of media coverage of firearm laws and regulations, media coverage of mass shootings, media coverage of violent crimes, and the Twitter activity of anti- and proregulation advocacy groups in short-term firearm acquisition in the United States. We collect daily time series for these variables from 2012 to 2020 and employ the PCMCI+ framework to investigate the causal structures among them simultaneously. Our results indicate that the Twitter activity of antiregulation advocacy groups directly drives firearm acquisitions. We also find that media coverage of firearm laws and regulations and media coverage of violent crimes influence firearm acquisition. Although media coverage of mass shootings and online activity of proregulation organizations are potential drivers of firearm acquisition, in the short term, only the lobbying efforts of antiregulation organizations on social media and specific media coverage appear to influence individuals’ decisions to purchase firearms.

A prominent approach in remote sensing poverty prediction uses deep learning satellite imagery with daytime and nighttime light (NTL). While results yield acceptable average predictions, certain areas consistently perform poorly. For high-stakes aid allocation decisions, however, achieving near-perfect predictions across all regions is crucial. To address this, we apply explainability techniques to evaluate vision models, enhance predictions with multimodal architectures, and investigate challenging cases. Our diagnostic analysis shows that fine-tuning with NTL tends to fail in areas where NTL poorly reflects wealth. Integrated Gradients and Guided Grad-CAM techniques also reveal that VGG, ResNet, and ViT architectures tend to focus on brightly lit features, such as roads and buildings. To complement vision models, a multimodal model integrates non-visual features. Metrics such as X (formerly known as Twitter) activity, distance from residential roads, and internet speed are especially predictive. Among 10 tested architectures, the Stack Ensemble Late Stage model, combining pre-trained weights, multimodal data, and NTL transfer learning, achieves the best performance with high accuracy and low error variance. Qualitative analysis of difficult-to-predict regions highlights the challenges of modeling wealth due to inherent aleatoric uncertainty. Two core issues, spatial inequality in small areas and the presence of rare features, underscore the need for principled cluster segmentation.

ABSTRACT When Prince Philip died on 9th April 2021 aged 99, the media coverage was extensive. Not only did the BBC dedicate wall-to-wall coverage to his passing (and received nearly 110,000 complaints from viewers for it), but social media, too, participated in what had quickly become a major media event. Social media remembrance has become an expected part of the death of a celebrity or well-known figure. However, what stood out in the Twitter activity, in particular, was not only sombre remembrance but also many instances of what can be considered bad-taste humour. Particularly prominent were comparisons of Philip to zombies, often including images of him. We explore the ways in which Twitter reactions to Prince Philip's death may highlight the monarchy as an outdated, presumably useless (yet undying) institution. The use of the zombie metaphor complicates the reading of these Tweets are merely offensive, as they are used to articulate resistance to royal power. We thus argue that the use of zombie metaphors in instances of bad taste, while considered offensive, may at the same time also offer a critique of the royal family as an institution at a particular moment in timeng and offer reflections on the ethics of researching controversial social media content.

Abstract Political communication on social media is increasingly prevalent; however, the factors influencing message content and profile characteristics that contribute to its reach remain underexplored. This study analyzes these factors through a case study of a significant global political event: the Colombian peace process. We monitored the Twitter activity of 27 key accounts, comprising politicians, public figures, and organizations actively engaged in the peace process. Employing a mixed-methods research approach, we combined qualitative thematic and content analysis of the tweets with quantitative analyses to identify predictors of reach. The findings indicate that negative emotion, controversial topics related to the peace process, and opposition stance are strong predictors of retweet behavior. Additionally, user profile characteristics—such as a higher follower count and clear opposition or support stance on the peace process—also influence retweet diffusion. This study aims to elucidate the behavior associated with tweets by public figures, considering both message content and profile attributes. Understanding retweet behavior is essential, as it sheds light on the processes and variables involved in the dissemination of opinions and ideas, particularly within the fields of political marketing and communication. The originality of this article lies in its holistic approach, which integrates emotional and topical dimensions alongside profile characteristics within the context of a political event in a developing country. This research addresses a gap in the existing literature by providing comprehensive insights into the impact of social media on political communication.

ABSTRACTThe existing literature on human–computer interactions is rich in studies on how consumers interact with brands on social media. However, there is a gap in the research on consumers' responses to the language style of social media branded messages. Therefore, this study aims to analyze e‐retailers' Twitter (now X) posts through text analysis to identify the content attributes that are most effective in generating higher numbers of retweets. R software was used for the data extraction and analysis of 28,737 tweets posted by e‐retailers in India. We used a variety of text analysis approaches, including retweet analysis, hashtag analysis, word cloud, network analysis, and sentiment analysis to analyze the collected tweets. We observed that tweets that included questions, product names, and promotional activities attracted better retweets, and that hashtags coupled with campaigns, products, and events were dominant. On average, positively charged tweets (specifically commanded by trust) were three times more popular than negative tweets. The four most prominent themes emerging in our network analysis are help and support, contests, discounts and offers, and query handling and resolution, which induce positive intentions among online shoppers towards e‐retailers. Our findings offer insights into how e‐retailers can improve their Twitter (X) activities to engage their audiences.

This study examines the impact of COVID-19 on the interaction between citizens and local governments through Twitter and conducts sentiment analysis within the accountability process guided by dialogical theory. An open-source intelligence tool extracted 1,224,725 tweets from 126 Hispanic–American municipalities during the first wave. Statistical methods identified tweet publication factors, and the NRC Emotion Lexicon was used for natural language processing. The results demonstrate that Twitter activity increased during the crisis, both reflecting changes in accountability dynamics and heightened dialogue between citizens and local governments and highlighting negative tones and emotions such as trust and anticipation.

PurposeThis study aims to understand luxury brands’ communication strategies on social media and to evaluate the effectiveness of marketer-generated content (MGC) on Twitter.Design/methodology/approachMGC’s characteristics in terms of media richness, content orientation, media type and post-timing were examined. The data from luxury handbag brands posted in 2022 were collected using Vicinitas. The final data set includes a total of 9,977 tweets from 18 brands. Tweets were analyzed using Latent Dirichlet Allocation and multiple regression with robust standard errors. To facilitate interpretations, word clouds were also used.FindingsEmpirical results demonstrated that context-specific topics exist. These topics together with information richness, content orientation, media type and timing factors significantly affected consumer engagement. Frequent keywords used in tweets in terms of influential context-specific topic and hashtags were shown.Research limitations/implicationsThis study extends theories on engagement on SM, provides deep insights for luxury brand literature, offers recommendations for luxury handbag brand managers to evaluate and further revise their strategies.Practical implicationsLuxury handbag brands should talk more about their new collections, use more hashtags, mentions and graphics interchange format media types, avoid using only texts and long message lengths, choose photos over videos and post in the afternoon rather than in the evening to increase engagement.Originality/valueThis research determines effective types of content which should be used by luxury brands and presents new insights regarding social media engagement specifically for luxury products. Despite the fact that past research mainly focused on Facebook, this paper examines activities in Twitter and MGC, which are arguably faster growing platforms.

This study explores the factors and motivations driving Twitter usage by firms and their CEOs, offering insights into the strategic and personal dynamics of social media adoption. Using hand-collected data from CEO and firm Twitter accounts, the findings reveal that firms adopt Twitter at a significantly higher rate and two years earlier than CEOs. Both firm and CEO Twitter usage are positively associated with size and valuation. However, CEO age and tenure are associated only with CEO Twitter activity, suggesting that CEO Twitter usage is discretionary and influenced by personal factors. Firms in environments with greater information asymmetry are less likely to use Twitter, suggesting a desire to minimize scrutiny or misinterpretation. However, CEO Twitter activity is unaffected by the firm’s information environment, emphasizing its personal nature. These findings highlight the distinct motivations for Twitter adoption by firms and CEOs, underscoring its role in modern corporate communication and leadership practices.

X (formerly known as Twitter) is a social media platform gaining popularity in neurosurgery. Other disciplines have demonstrated a positive correlation between Twitter activity and traditional citation metrics. This study aims to determine if X activity is a greater predictor of citation rates than a journal's 5-year impact factor (IF) among major neurosurgical journals. Using a mixed linear model, we compared the predictive value between alternative metrics (such as mentions on X and Altmetric attention score, a weighted aggregate of the attention an article receives on various platforms) and traditional citation metrics (5-year journal IF) on the number of citations an article received by analyzing 7592 articles published from January 2022 to December 2023 in 18 neurosurgical journals. It was necessary to also account for the confounding variable time since publication in the model to determine the true effect of altmetrics. The relative importance (RI) of each predictor variable was determined through permutation testing. X mentions, time since publication, 5-year journal IF, and Altmetric attention score all significantly predict citation rates (P < .001). RI of X mentions on citation rates indicate that X (RI = 0.13) is approximately 8.7x times greater of a predictor of citations than 5-year IF (RI = 0.015) and 5.4x times greater of a predictor than the Altmetric attention score (RI = 0.024). Time of publication remains the strongest predictor (RI = 0.83). Our study shows that in neurosurgical research, while social media mentions (X mentions) are significant, they are weaker predictors of citation rates than time since publication. Traditional journal IF and Altmetric attention scores have weaker predictive value. These findings indicate that altmetrics, especially X mentions, can complement traditional citation metrics.

Exploring Nonprofit Organizations’ Twitter Activism during the #StopAAPIHate Movement

ABSTRACT This study examined the use of the hashtags #NorthEastMatters and #AChapterForNE by northeast Indians to protest curriculum marginalization and seek racial justice on Twitter. Drawing on the lens of TribalCrit theory, hashtag activism, and racial microaffirmation, a qualitative thematic analysis of the Twitter tweets and semi-structured interview data with key stakeholders is conducted. Twitter’s advanced search function is used to retrieve English-language posts on June 4, 2021, using the two hashtags, and then manually reviewed using Microsoft Word and Excel for data familiarization and coding purposes. A data set of 123 English textual tweets was selected for thematic analysis. Tweets reveal seven main themes: collective memory, equal citizenship, collective voice, collective rights, curriculum inclusivity, racial detoxification, and Twitter activism. Hashtags are used for collective storytelling to expose the inactivity of the authorities and their failure to resolve them, suggest solutions, and put pressure on the government and officials for social change. The in-depth interview addresses unfair representation, tackling racism, and protest participation. This study advances understandings of Twitter as a digital counter-space, hashtag activism platform, and site of resistance for marginalized groups.

This study analyses the participation on Twitter (X) during the 15 days following the deaths of Veronica Forque (suicide) and Alex Casademunt (traffic accident). The primary objective of the study is to compare the reactions generated on Twitter (X) to the suicide versus the traffic accident of a famous person. To achieve this, the Application Programming Interface Tweet Downloader was used to download batches of tweets (posts) corresponding to each period. Pearson's chi-square test was utilised to identify differences in the messages. Additionally, a content analysis was conducted on a sample of the retrieved tweets (n=621), in which various variables were measured. The results indicate that despite both victims being similarly well-known to the public, Twitter (X) activity was higher in the case of Veronica Forque than in the case of Alex Casademunt. Furthermore, the topic of suicide prompted discussions about guilt (internal and external), preventive measures, and their perceived usefulness compared to the traffic accident, which elicited more expressions of emotions and mentions of the cause of death.