Propaganda by Other Means: Affective Politics on Facebook

• In Event: Emerging Challenges to Digital Campaigning

Abstract

As democratic backsliding takes hold across the globe in established democracies, this study turns attention to the global south context of Bangladesh and to a context that affords the exploration of the political role of Facebook in a South-Asian authoritarian nation. Bangladesh has long been defined by increasing authoritarian rule and democratic decline despite hopeful assumptions that the country was on the path to a political, liberal democracy. To date, there is little research on Bangladeshi politics due to dominance in communication scholarship on the global North. Regarding social media technologies developed by Silicon Valley, communication scholarship has also remained biased towards its usage and deployment within American political contexts, which has created gaps in our understanding of how technologies like Facebook are leveraged and deployed by ruling and oppositional political parties to further their political messaging within international contexts during heightened political events and crisis. To address these academic gaps and build research in this area, this study examines how the two largest Bangladeshi political parties, Bangladesh Nationalist Party-BNP and Bangladesh Awami League-AL, used Facebook to sway public support during a pre-election political crisis that occurred on October 28, 2023. On that day, there was significant political tensions and unrest, defined by a grand rally by the Bangladesh Nationalist Party (BNP) in Dhaka that demanded both the current government’s Awami League (AL) resignation and free and fair elections under an interim non-partisan government. In response, the AL organized a counter-rally, and these dueling rallies resulted in the death of four people including a policeman. To examine this day of political crisis within the context of Facebook, this study advanced two primary research questions: RQ1- What issues gained prominence in the AL’s and BNP’s Facebook pages on this day? and RQ2- How did these parties apply affective and propaganda-based strategies to sway the public in their Facebook content on this day? To explore these questions, this study analyzed a total of 150 posts scraped from both parties’ Facebook pages, with these posts all generated on the single day of October 28, 2023, as the political crisis unfolded in real time.
This study utilized two main theoretical frameworks to decipher the role that Facebook played as a platform and channel for both parties to advance their countervailing positions. Primarily, this study advanced affective politics as a framework for understanding how platforms like Facebook impact and alter the emotional intensity of political communication. Defined through both the positive and negative tone of sentiment in content and the emotional expressions that define these sentiments, this study analyzed the content extracted from Facebook on October 28, 2023, to determine how affect drove and shaped political messaging within both parties. The second theoretical framework, computational propaganda, acknowledged that political manipulation is now increasingly driven by the affordances of social platforms like Facebook through platform architecture and algorithmic infrastructure. Politicians can now use platforms to sidestep mass media influence to establish deeper control of issues presented to the public in efforts to “manufacture consensus” via automated means. Propaganda can now deepen in scope and reach through algorithmic acceleration on platforms like Facebook. Methodologically, this study adopted a mixed-methods approach, combining quantitative network analysis with deeper qualitative framing analysis, utilizing a range of methods that included manual scraping for data collection, semantic network analysis and visualization (to answer RQ1), and in-depth qualitative framing analysis (to answer RQ2). These methods were utilized for uncovering manifest and latent messaging of sentiments and their emotional expressions.
To date, preliminary results for RQ1 via semantic network analysis revealed that both parties emphasized the political opposition, current political issues, and elections. Both parties also utilized contesting discourses primarily through negative sentiments in issues. Preliminary analysis for RQ2 revealed through framing analysis that both parties used propaganda structures to 1, establish themselves as more eligible, 2. dismiss the validity of the other party, and 3. back their claims by presenting evidence. Both parties leveraged the affordances of Facebook’s architecture through hashtags, pictures, videos, and news links, to present evidence of the other party’s wrongdoings. Both parties also built agendas through selective attention to different historical events to construct one-sided agendas. The affective architecture of Facebook allowed both parties to engage in computational propaganda achieved through negative affective sentiments and emotional expressions.

Authors

• Priyanka Kundu, University of Illinois at Chicago

• Sharon Meraz, University of Illinois at Chicago