ZeMKI goes ICA

This year’s annual conference of the International Communications Association (ICA) will take place in Goald Cost, Australia, from June 20 to 24. The ZeMKI will be represented with numerous contributions and ZeMKI members will present their research.

Saturday, June 22, 2024

10:30 AM

Christian Katzenbach; Daria Dergacheva; Vasilisa Kuznetsova; Adrian Kopps: How Have Social Media Platforms Governed ‚Misinformation‘ From Their Inception Until Today: A Longitudinal Study

This research is the first study that analyzes all policies of five major social media platforms (Facebook, Instagram, YouTube, X (formerly Twitter) and TikTok starting from the beginning of their existence until the present day in connection to the concept of misinformation. Given that platforms are constantly changing their policies (Katzenbach, 2021; Barrett & Kreiss, 2019), the novelty of this paper lies in a systematic coverage of these changes over time and across platforms.
While scholarship has focused strongly on specific harms of social media platforms, such as hate speech and misinformation, there has been little systematic research on how they dealt with these harms over time. So to understand the role of platforms in their approach to misinformation, we need to systematically study how platforms have approached it since the time the concept had first appeared, in the context of societal transformations and technological developments (Katzenbach, 2021).
Methodologically, this research uses a mixed-method approach. We apply computational text analysis, including words in context and topic modeling (Chen et al., 2021) to understand the overarching themes and topics in policy documents and to narrow down the scope for qualitative analysis. The study then employs critical discourse analysis and qualitative content analysis to trace the changes around the concept of misinformation through time and platforms under study (Jacobs & Tschötschel, 2019).

3:30 PM

Yuru Li; Jing Zeng; Cornelius Puschmann; Andreas Breiter: Computational Approaches to Visual Framing in Communication Studies

In recent years, computational visual analysis has rapidly developed. Nonetheless, the exploration of visual framing remains a formidable challenge within the domain of communication research. This study introduces a systematic solution for the extraction of visual frames utilizing computational methods, especially visual-language pre-training models. The dataset utilized in this investigation comprises TikTok videos addressing climate change, spanning from 2018 to 2023. To retrieve the visual frames, four distinct approaches were evaluated: 1) an inductive approach of image captioning coupled with topic modeling, 2) a deductive approach of zero-shot classification, 3) a hybrid approach involving image captioning, zero-shot classification, and topic modeling and 4) a multimodal approach based on the hybrid approach in which text data from posts are included in the topic modeling model. After visualization, we did a manual validation of these four approaches from the dimensions of semantic coherence and accuracy, and the results indicate that the hybrid approach yields the most favorable performance. In contrast, the multimodal approach is helpful for finding more abstract visual frames.

Sunday, June 23, 2024

1:30 PM

Felix V. Münch; Philipp Kessling; Gregor Wiedemann: Digging Deep and Wide: Explorable Hierarchical Topic Modeling in Large-Scale Cross-Platform Online Discourses Based on Temporal Semantic Similarity Networks and Community Detection

In times of a so-called poly-crisis, i.e., the current culmination of several global crisis developments at once, such as climate change, pandemics, and conflicts affecting the post-cold-war balance of powers, it seems a daunting task to gain an overview of the relevant parts of the greater public discourse. Within this poly-crisis, superficially unrelated discourses influence and overlap each other on multiple levels of scopes, from hyperlocal to global, and on various timescales, from days to decades. Therefore, when investigating discourses related to large-scope and long-term topics, such as climate change or the Russian invasion of Ukraine, this endeavour necessitates approaches that can simplify the inherent complexity of these topics at a top-level while allowing data analysts to choose where to dig deeper. For this endeavor, not only the interactive explorability of topic relations and sub-topic structures is key, but ever-changing topics and crises necessitate the introduction of time as a topic-defining parameter. To our knowledge, both are not provided either by standard methods like Latent Dirichlet Allocation or by state-of-the-art topic detection methods such as BERTopic. Therefore, we propose an alternative combination of methods, leaning on the embedding-clustering sequence of BERTopic, but based on semantic similarity networks, augmented with a time-component, and tried-and-tested network-based community detection methods that allow for hierarchical clustering and exploration of these networks. In this contribution we describe our method and its validation across two large-scale test cases: 1) the discourse during autumn and winter 2022 around gas deliveries from Russia on German-language Twitter (now called X) after the Russian invasion of Ukraine; and 2) the German-language cross-platform discourse around climate change on Facebook, Instagram, Twitter, and Telegram from 2019 to 2023.

Monday, June 24, 2024

12:00 PM

Cornelius Puschmann; Sebastian Stier; Patrick Zerrer; Helena S. Rauxloh: Politicized and Paranoid? How Conspiracy Ideation Predicts Alternative News Consumption 

The study of right-wing alternative online news sources has moved to the center of scholarly attention in recent years. Such sources cater to news consumers who are frequently characterized by their extreme political views and mistrust towards established mainstream sources. Right-wing alternative news outlets have also been linked to the spread of conspiracy myths in areas such as immigration and climate change. However, research into the predictors of alternative news consumption is still comparatively scarce. We approach these gaps by combining online tracking and survey data from 2,098 German panel participants collected over a period of 1.5 years. In doing so, we are able to capture news consumption both as a binary variable (dividing subjects into alternative media users and non-users) and in terms of usage frequency (subdividing the former into occasional and frequent users). We then measure the tendency of participants towards conspiratorial thinking through an automated content analysis of open-ended survey items on four controversial topics. We find conspiratorial thinking to be a reliable predictor of alternative news consumption, both in absolute terms (users who manifest conspiratorial thinking are more likely to use alternative news than those who do not) and in relative terms (the more conspiratorial thinking a user manifests the more alternative media they consume). Our findings contribute towards the development of a deeper understanding of alternative news consumers.

Stephanie Geise, Katharina V. Hajek, Katharina Maubach: Look – Don’t Look! Psychological Reactance in Multimodal Framing

Media framing plays a crucial role in shaping viewers’ perception of reality, particularly within the growing dominance of multimodal news environments. There are divergent findings as to whether images enhance or hinder successful framing effects in multimodal framing. This study is the first to explore the role of psychological reactance within the process of multimodal framing. We conducted a 2×3 between- and within-subjects experimental study (N = 336) that combined pre- and post-test measures varying modality (text vs. text-visual) and the presented topic of the news article (Covid-19, climate change, and economic crisis). We find a significant impact on modality on perceived manipulation and perceived manipulation on reactant behavior, yet no significant impact of the modality on anger (H3a) or reactant behavior (H1). However, an interesting link emerged between modality, perceived manipulation, and reactant behavior. We discuss our results as an indication of an increased media literacy of the recipients in dealing with images.

3:00 PM

Uwe Hasebrink, Andreas Hepp, Wiebke Loosen: The Refiguration of Public Communication: A Relational and Process-oriented Perspective

In the field of communication and media research, the idea of the “public sphere” and its “structural change” has long been a central focus, particularly in the context of how commu-nication evolves in relation to society and politics. The problem with this concept is its imprecise normative profile and vague empirical scope. This paper aims to introduce a novel theoretical approach characterized by a more rigorous emphasis on relational and process-oriented thinking. Specifically, we propose to conceptualize the transformation of public communication as a process of refiguration. To this end, we deal first with media-environmental change from the perspective of mediatization theory. Next, we delve into how this shift in public communication can be comprehended as refiguration, emphasizing that public communication isn’t confined to a singular public sphere but rather pertains to the transformation of the relational dynamics among at least three distinct types of publics. Finally, our paper addresses how exactly such a relational and process-oriented perspective can be realized analytically. While our argument is based on much empirical research of our own, we see this paper as a theoretical contribution.

Niklas Venema; Erik Koenen; Christian Schwarzenegger; Jördis Krey: Adaptation, Expansion, or Misuse of a Theoretical Concept? A Systematic Literature Review of the Research on Counterpublics

In recent years, there has been a great deal of research on both negatively and positively connoted phenomena of digital communications, which have likewise been characterized as counterpublics. Therefore, the same public arenas, practices, and communication strategies, once idealized as sentinels for democracy, speech, and political participation, are increasingly met with suspicion regarding their contribution to societal polarization, spreading conspiracy myths, and undermining democracy through manipulation. Against this background, the paper addresses the transformation of the concept of counterpublics. A systematic literature review is used to analyze the differences and continuities in the theoretical foundation and empirical operationalization of this idea. For this purpose, a corpus of 313 English-language scientific journal articles issued from 1999 to 2023 was collected. The study contributes to contextualizing the phenomena and makes an important contribution to clarifying the controversial theoretical question of what can or should be understood as counterpublics.