Selected Papers of #AoIR2024: The 25th Annual Conference of the Association of Internet Researchers Sheffield, UK / 30 Oct - 2 Nov 2024 Suggested Citation (APA): Bruns, A., et al. (2024, October). Controversies, Problematic Information, and Polarisation: Case Studies across Six Continents. Panel presented at AoIR2024: The 25th Annual Conference of the Association of Internet Researchers. Sheffield, UK: AoIR. Retrieved from http://spir.aoir.org. CONTROVERSIES, PROBLEMATIC INFORMATION, AND POLARISATION: CASE STUDIES ACROSS SIX CONTINENTS Axel Bruns Digital Media Research Centre, Queensland University of Technology Marco Bastos University College Dublin Otávio Vinhas University College Dublin Raquel Recuero Universidade Federal do Rio Grande do Sul Felipe Soares University of the Arts London Laura Iannelli University of Sassari Giada Marino University of Urbino Danilo Serani University of Salamanca Augusto Valeriani University of Bologna Tariq Choucair Digital Media Research Centre, Queensland University of Technology Sebastian Svegaard Digital Media Research Centre, Queensland University of Technology Laura Vodden Digital Media Research Centre, Queensland University of Technology Daniel Whelan-Shamy Digital Media Research Centre, Queensland University of Technology Alia Azmi Digital Media Research Centre, Queensland University of Technology Jennifer Stromer-Galley Syracuse University Panel Introduction Political trends around the world have drawn further scholarly attention to the study of polarisation, especially also as it is expressed and potentially deepened by public communication on digital and social media platforms. The very concept of polarisation itself, however, remains ill-defined especially in communication research, where it is often used as a mere buzzword without sufficient definition – even in spite of a wide range of conceptual approaches that variously emphasise issue-centric, ideological, affective, interpretive, interactional, or other facets of polarisation (Marino & Ianelli, 2023; Esau et al., 2023). Issue-centric approaches to the study of polarisation often connect it with specific controversies, and therefore align well with controversy mapping and related methodological frameworks. Especially where they study such controversies in digital and social media contexts, they also point to the significant intersections between the circulation of problematic information by and the deepening of polarisation between partisan actors, as well as to the often asymmetrical nature of such contestations (where groups on one side of a given controversy are substantially more likely to use problematic information to support their cause than the groups opposing them; Kreiss & McGregor, 2023). Unfortunately, much of the recent research in this field has continued to focus on a small number of key political contexts, with emphasis especially on the US and UK. This panel reviews evidence on the intersection of controversies, problematic information, and polarisation through a series of case studies from six continents: North and South America, Europe, Asia, Australia, and Oceania. In combination, these studies present a substantially more comprehensive picture of global similarities and local differences. Paper 1 explores the polarising impact of disinformation campaigns in favour of incumbent president Jair Bolsonaro in the 2022 Brasilian presidential election. It reveals a potentially unusual bottom-up disinformation pattern that produces a reverse influence flow from grassroots activists to political leaders and complicates standard distinctions between mis- and disinformation; this also creates new challenges for fact-checking efforts. Paper 2 examines the dynamics of Italian public opinion in response to the introduction of COVID-19 restrictions in early 2020. Drawing on longitudinal survey data, it identifies a range of perspectives from extreme communitarian to extreme libertarian, and connects this to patterns of legacy and social media use, attitudes towards political institutions, and levels of exposure to mis- and disinformation. Paper 3 compares the divergent dynamics of political debates on Indigenous rights in Australia and Aotearoa New Zealand. In the campaigning leading up to Australia’s 2023 referendum on greater Indigenous recognition and representation, it identifies a highly asymmetrical contest that flipped public opinion from strong support for Indigenous recognition to a 60% No vote within less than one year. In the heated political debate about Māori rights in Aotearoa New Zealand since the 2023 election of a new, conservative coalition government, it identifies continuing Māori/non-Māori solidarity in resistance to the reduction of rights stemming from the 1840 Treaty of Waitangi. Paper 4 investigates the debates on Twitter about measures to combat sexual violence in Indonesia that came into effect in 2021 and 2022. Drawing on extensive content and network analysis, the study shows that, diverging from the #metoo-style activism against sexual discrimination, harassment, and violence that is common in Western contexts, in Indonesia this agenda is interpreted predominantly through the lens of an underlying polarisation between secular nationalist and Islamist political groupings in the world’s largest Muslim-majority democracy. Paper 5 compares the online dynamics of the abortion debate in the US before and after the Supreme Court’s 2022 Dobbs decision, focussing especially on political candidates’ social media messaging on abortion rights. Analysis of Democrats’ and Republicans’ posts about the issue, and of broader Twitter and Facebook user engagement with the issue, is expected to point to substantial differences between the parties, timeframes, and platforms. In combination, these five papers cover a rich selection of case studies on the intersections between controversies, problematic information, and polarisation around the world. Extended abstracts for all five papers are included on the following pages. References Esau, K., Choucair, T., Vilkins, S., Svegaard, S., Bruns, A., O’Connor Farfan, K. & Lubicz, C. (2023, 30 May). Destructive Political Polarization in the Context of Digital Communication – A Critical Literature Review and Conceptual Framework. International Communication Association, Toronto. https://eprints.qut.edu.au/238775/ Kreiss, D., & McGregor, S. C. (2023). A Review and Provocation: On Polarization and Platforms. New Media & Society. https://doi.org/10.1177/14614448231161880 Marino, G., & Iannelli, L. (2023). Seven Years of Studying the Associations between Political Polarization and Problematic Information: A Literature Review. Frontiers in Sociology, 8. https://doi.org/10.3389/fsoc.2023.1174161 https://eprints.qut.edu.au/238775/ https://doi.org/10.1177/14614448231161880 https://doi.org/10.3389/fsoc.2023.1174161 Paper 1 REVERSE INFLUENCE: DISINFORMATION AND INSTITUTIONAL DISTRUST IN THE BRAZILIAN ELECTIONS Marco Bastos University College Dublin Otávio Vinhas University College Dublin Raquel Recuero Universidade Federal do Rio Grande do Sul Felipe Soares University of the Arts London Extended Abstract The Brazilian presidential election of 2022 was marked by a slew of fraud allegations and conspiratorial narratives aimed at discrediting the electoral process (Ruediger et al., 2022). This disinformation campaign exploited tropes of a perceived ‘communist threat’ (Burity, 2021), the depiction of Bolsonaro as a ‘Messiah’ tasked with warding off leftists and progressive movements, and unsubstantiated claims of fake electronic ballots and vulnerability in Brazil’s electronic voting systems (Cunha, 2023). The disinformation campaign organized by Bolsonaro supporters intensified towards the end of the campaign trail and slowly merged with calls for military intervention and calls for a coup d’état widely distributed across social media platforms. False claims of election fraud, along with voter suppression attempts, would become a strategy integral to Bolsonaro’s failed bid to the presidency in 2022 (Gonçalves & Abbud, 2022) and would culminate in Bolsonaro supporters attacking the congress, the presidential palace, and the supreme court in Brasilia after Lula da Silva was sworn in on 1 January 2023. The polarized and highly controversial developments leading up to the ballot offered a textbook case of disinformation campaign. In this study, however, we inspect the applicability of categorizations like mis- and disinformation in the context of the Brazilian presidential election of 2022 through three case studies. We first explore visual disinformation distributed on Facebook to probe the extent to which the diffusion of disinformation follows the typical cascade activation process informed by epidemiological models of contagion, or alternatively stems primarily from audiences that effectively set the agenda of the political leadership. Our second case study leverages a database of textual disinformation shared on Twitter to examine whether disinformation is too narrow an analytical category to account for the very diverse and disruptive campaign that targeted the presidential election in Brazil. Our third case study relies on a set of interviews with Brazilian fact-checkers with extensive experience in moderating and checking divisive political content to identify how fact-checkers define and operationalize disinformation in contexts of institutional distrust. Our first case study shows that the distribution of disinformation in the Brazilian election deviates from the prevailing epidemiological model informing disinformation studies whereby messages originate from influential accounts and are then disseminated to the broader population through a process of simple or complex contagion. In contrast to this model, we found that such messages emerge initially from grassroots activists and are only later voiced by the political leadership, a process that the fact-checkers we interviewed refer to as ‘bottom-up disinformation’ and that we refer to as reverse influence. This two-way process where both grassroot activists and popular influencers are simultaneously influencing each other in the effective production of disinformation shows the inherently participatory nature of the spread of these narratives on social media. Our second case study shows that the coding process of tweets associated with the election proved challenging to binary, static definitions of disinformation. We struggled to distinguish between what was a ‘clear call for military intervention’ and what was ‘calling into question democratic institutions.’ The separation between these variables proved porous at best and entirely dependent on the context. At any rate, disinformation could only be identified in reference to the context that grounded these narratives, with the content analysis quickly escalating to a context analysis that speaks to the ‘collective memory’ supporting these narratives (Wodak, De Cillia, & Reisigl, 1999) The relatively lower Krippendorff’s alpha achieved for some categories is indicative of the difficulties in separating what was clearly disinformation from what was at times legitimate, if ill- informed and perhaps ill-intentioned discussion about democratic institutions and the deliberative process. These Krippendorff’s alpha scores were nonetheless only achieved after training the coding team to look into the context of the messages. Our third and last case study incorporates interviews with fact-checkers with extensive experience in the verification of political claims in Brazil. These experts made a forceful case that disinformation in Brazil primarily stems from institutional distrust instead of relating to the supply and consumption of information, such as a lack of reliable news sources or inadequate media literacy among the population (Rossini, Mont'Alverne, & Kalogeropoulos, 2023). This problem invariably pivots along contentious issues such as national history, the weak institutional authority, and the fragile balance between institutional actors—including journalism. The relatively fragile social cohesion across opposing political and religious groups, but also across populations from diverse demographic backgrounds, renders it impractical for fact-checkers and social platforms to counter disinformation through binary truth parameters for content moderation. This was particularly salient in Brazil’s latest election cycle, when Bolsonaro supporters took up a central role in promoting narratives antithetical to democratic principles and that endorsed alternative facts with self-validating participatory disinformation ecosystems crafted via social media and messaging applications like WhatsApp and Telegram (Ozawa et al., 2023). The falsehoods eventually disseminated by politicians transcended mere fabrications; rather, they encapsulate broader issues deeply entrenched in the social fabric of the country conjured up through the expressive sharing of disinformation. This ‘bottom-up disinformation’ is somewhat at odds with the limited effects paradigm (Pickard, 2021) because the central point of diffusion stems from atomized user- generated content that sets the agenda for political elites and the mainstream media. The central insights gleaned from experienced fact-checkers, but also from the limits of text and visual classification of disinformation, speak to the limits of approaching communication from a strictly transmissive perspective that ignores the role of shared interaction online for the maintenance of social groups and the representation of their shared beliefs. Many of the cases reviewed in our study foreground a ritualistic view of communication whereby the information being shared, often objectively containing mis- or disinformation, is of limited import to groups dedicated to drawing people together in fellowship and commonality (Carey, 2009). As such, it should not be surprising that these narratives are dedicated to the maintenance of an ordered, meaningful cultural world that can serve as a control and container for human action; a container that is often repressive, and a cultural world that can be markedly reactionary. These forms of participatory disinformation foreground the reading and partaking in stories that are not consumed as units of information conveyed by authoritative sources, but as a collective narrative effort similar to attending a mass, a situation in which nothing new is learned, and where no information is essentially accurate, but in which a particular view of the contending forces in the world is portrayed and confirmed. In particular, many of the narratives covered in our study evoke beliefs commonly held by those involved in their distribution, beliefs that are frequently rooted in historical revisionism but that have been prevalent within the Brazilian political consciousness over several decades, including notions of the communist menace and scepticism towards democratic institutions. They often rely on symbolisms that cannot be strictly fact-checked, a recurrent issue faced by fact-checkers who deemed most of social media content essentially impossible to fact-check. In the end, the reduction of disinformation to information that is false fails to acknowledge that the social role of disinformation for these groups is not to convey information, but to provide a stage for dramatic forces and actions that invite participation. References Burity, J. (2021). The Brazilian conservative wave, the Bolsonaro administration, and religious actors. Brazilian Political Science Review, 15. Carey, J. (2009). Communication as Culture: Essays on Media and Society (Revised Edition). New York: Routledge. Cunha, M. D. N. (2023). "Brazil Above Everything. God Above Everyone." Political- Religious Fundamentalist Expressions in Digital Media in Times of Ultra-Right Populism in Brazil. International Journal of Communication, 17. Gonçalves, E., & Abbud, B. (2022). Partido de Bolsonaro divulga parecer que questiona sem provas segurança das urnas; TSE diz que documento é 'falso' e 'mentiroso'. O Globo. https://web.archive.org/web/20221002005556/https://oglobo.globo.com/politica/eleico es-2022/noticia/2022/09/pl-partido-de-bolsonaro-divulga-parecer-que-questiona- seguranca-de-urnas-a-quatro-dias-das-eleicoes.ghtml Ozawa, J. V., Woolley, S. C., Straubhaar, J., Riedl, M. J., Joseff, K., & Gursky, J. (2023). How Disinformation on WhatsApp Went from Campaign Weapon to Governmental Propaganda in Brazil. Social Media + Society, 9(1), 20563051231160632. https://web.archive.org/web/20221002005556/https:/oglobo.globo.com/politica/eleicoes-2022/noticia/2022/09/pl-partido-de-bolsonaro-divulga-parecer-que-questiona-seguranca-de-urnas-a-quatro-dias-das-eleicoes.ghtml https://web.archive.org/web/20221002005556/https:/oglobo.globo.com/politica/eleicoes-2022/noticia/2022/09/pl-partido-de-bolsonaro-divulga-parecer-que-questiona-seguranca-de-urnas-a-quatro-dias-das-eleicoes.ghtml https://web.archive.org/web/20221002005556/https:/oglobo.globo.com/politica/eleicoes-2022/noticia/2022/09/pl-partido-de-bolsonaro-divulga-parecer-que-questiona-seguranca-de-urnas-a-quatro-dias-das-eleicoes.ghtml Pickard, V. (2021). Unseeing propaganda: How communication scholars learned to love commercial media. Harvard Kennedy School Misinformation Review. Rossini, P., Mont'Alverne, C., & Kalogeropoulos, A. (2023). Explaining beliefs in electoral misinformation in the 2022 Brazilian election: The role of ideology, political trust, social media, and messaging apps. Harvard Kennedy School Misinformation Review, 4(3). Ruediger, M. A., Grassi, A., Dourado, T., Barboza, P., Piaia, V., & Hubert, D. (2022). Online Disinformation and Questioning of Election Results. Wodak, R., De Cillia, R., & Reisigl, M. (1999). The discursive construction of national identities. Discourse and Society, 10(2), 149-173. Paper 2 PUBLIC HEALTH COMES FIRST? DIVERGENCE AND DIFFERENCES BETWEEN COMMUNITARIANS AND LIBERTARIANS IN ITALY DURING THE COVID-19 PANDEMIC Laura Iannelli University of Sassari Giada Marino University of Urbino Danilo Serani University of Salamanca Augusto Valeriani University of Bologna The Context: Disputes on Pandemic Management between Political Elites The WHO's declaration of the Covid-19 pandemic in January 2020 led Italy to introduce, by March, the first mass lockdown in modern Western democracies (Phase 1 of pandemic management). Phase 2 saw a gradual easing of restrictions, but concerns persisted, leading to alternating lighter and stricter lockdown measures (Phase 3). The government which managed the emergency in Italy was an alliance between the Five Star Movement and left-wing parties. From the earliest phases, the government and right-wing opposition parties struggled with the “lockdown” measures (Bertero & Seddone, 2021). Amidst heated disputes between political elites on Covid-19 restrictive measures, this study delves into the dynamics of Italian public opinion, guided by four research questions. Societal-Level Analysis: Exploring the State of Divergence between “Communitarians” And “Libertarians” First, the study aimed to investigate whether Italian public opinion showed signs of divergence during the initial stages of pandemic management between extreme communitarians and libertarians (who claimed, respectively, the clear-cut priority of public health over citizens’ freedoms and vice versa). Divergence is one of multiple dynamics of polarization that can occur at the societal level when preferences move apart to opposite extremes (Fiorina et al., 2005). Previous research suggested that polarization among political elites could lead to social polarization (e.g., Prior, 2013). At the same time, the literature indicated the possibility of a "rally-around-the-flag" effect during a crisis (Baum & Groeling, 2008), potentially resulting in widespread support for government containment measures. We thus posed this RQ: RQ1 – Did Italian public opinion diverge between extreme "communitarian" and "libertarian" positions during the first two phases of the Covid-19 pandemic, and to what extent? Individual-Level Analysis: Exploring the Media and Political Predictors of “Communitarian” And “Libertarian” Positions Our second goal was to analyze the role of disinformation in forming communitarian and libertarian stances at the individual level (Tucker et al., 2018). One significant challenge involved understanding the role of fake news spread through both “older” and “newer” media sources (Giglietto et al., 2019), being both responsible for the infodemic (WHO, 2020). Another challenge was the rise of "post-truth" (Waisbord, 2018). Citizens' knowledge of different "regimes of truths" about the pandemic could lead to varying perceptions of its seriousness and the necessity of safety measures. Considering these challenges, we posed the following RQs: RQ2 – Did Italian citizens' opposite perceptions of traditional media (television, radio, newspapers), social media, and messaging apps as reliable sources correlate with communitarian and libertarian positions, and if so, how? RQ3 – Were the levels of knowledge regarding Covid-related "official truths" proposed by institutions at the beginning of the emergency associated with communitarian and libertarian positions, and if so, how? Finally, the study investigated the role of the support for the Italian government: RQ4 – Was trust in the government that issued the "stay-at-home" order associated with libertarian or communitarian positions, and if so, how? Data and Method The study is based on the first two waves of a longitudinal CAWI survey administered after Phase 1 (May 2020) and at the end of Phase 2 (September) pandemic management to a representative sample of the Italian adult population with internet access. To address RQ1, we utilized a survey question asking respondents to indicate their position on an 11-point scale regarding two opposing statements: one prioritizing "citizens' freedoms always before public health," and the other prioritizing "public health always before citizens' freedoms." This scale ranged from 0 (extreme communitarian position) to 10 (extreme libertarian position). We measured the divergence in these positions using established methods from polarization studies (Lee, 2016; Lelkes, 2016). Additionally, we compared this divergence in public opinion on public health with divergence on other contentious issues in Italian public discourse, such as immigration and European integration, to provide a benchmark for comparison. To address RQ2, RQ3, and RQ4, we employed a series of cross-sectional and self- regressive logistic regression models. We constructed our dependent variables using the same question described earlier. To identify "the libertarians," we coded individuals who selected one of the three scale points closest to the statement "citizens' freedoms always come before public health" as 1 and all others as 0. Conversely, to identify "the communitarians," we coded individuals who chose one of the three scale points closest to the statement "public health always comes before citizens' freedoms" as 1 and all others as 0. For our first independent variable, we considered responses to three questions recording the frequency of encountering fabricated news on television, radio, print, digital newspapers, social media platforms, and instant messaging apps. In our models, we also included an index measuring respondents’ knowledge of official truths about the Covid-19 pandemic coming from institutional sources. Finally, we used a survey battery on institutional trust and included a set of control variables (frequency of news media use, sociodemographic variables). Findings Contrary to the US (e.g. Allcott et al., 2020), data analysis indicated that Italian citizens' opinions did not significantly diverge into extreme libertarian and communitarian positions. In contrast to divisive issues like European integration and immigration, Italy's dominant stance on public health was overwhelmingly communitarian, even at the end of Phase 2. This trend may be attributed to the "rally-around-the-flag" phenomenon. However, such a widespread acceptance of unprecedented limitations on personal freedoms to protect public health could be read as a fertile ground for a normalized suspension of democratic rights within the emergency frame (Agamben, 2020). The logistic regression findings underscored a pronounced distinction between libertarian and communitarian attitudes, extending beyond specific issues to broader perceptions of the news media, official truths, and government. The analysis demonstrated a strong relation between trust in the government, a trustful attitude toward the legacy news media and institutional sources as reliable sources, a vision of digital platforms as infodemic environments, and communitarian positions. Conversely, libertarians tend to exhibit distrust in the government, skepticism towards traditional news media and official truths, and less concern about the spread of fake news on digital platforms. Considering the typical “ceasefire” of journalists when a country experiences an external threat (Bennett, 1990), individuals skeptical of the government might have perceived traditional media coverage of the government's lockdown measures as exaggerated, leading to a stronger resistance to restrictions on personal freedoms. Conversely, those with greater confidence in the government and traditional media sources might have reinforced their belief in prioritizing public health. The opposite dynamic occurred on digital platforms, which often feature criticism of lockdown measures. These opposite attitudes towards traditional media and the government also aligned with levels of knowledge about official truths. Higher levels of knowledge tended to correlate with communitarian positions. Conclusions This study adopted a specific definition of polarization and polarized opinions that helps to focus on specific manifestations of opinion formation – at the societal and individual levels – which indicate unsustainable conflicts for democratic societies. Moreover, the study delved into the intricate dynamics of fake news within the contemporary hybrid media system, also considering various regimes of truth, thus highlighting the value of going beyond the analysis of a single (social) medium. References Agamben, G. (2020). A che punto siamo? L'epidemia come politica. Quodlibet. Allcott, H., Boxell, L., Conway, J., Gentzkow, M., Thaler, M., & Yang, D. (2020). Polarization and public health: Partisan differences in social distancing during the coronavirus pandemic. Journal of Public Economics, 191, 104254. Baum, M., & Groeling, T. (2008). Crossing the Water’s Edge: Elite Rhetoric, Media Coverage and the Rally-Round-the-Flag Phenomenon. The Journal of Politics, 70(4), 1065–1085. Bennett, L. W. (1990). Toward a Theory of Press-State Relations in the United States. Journal of Communication, 40(2), 103–27. Bertero, A., & Seddone, A. (2021). Italy: Populist in the Mirror, (De)Politicizing the COVID-19 from Government and Opposition. In G. Bobba & N. Hubé (Eds.), Populism and the Politicization of the COVID-19 Crisis in Europe (pp. 45–58). Palgrave Macmillan. Fiorina, M. P., Abrams, S. J., & Pope, J. C. (2005). Culture war? The Myth of a Polarized America. Longman Giglietto, F., Iannelli, L., Valeriani, A., & Rossi, L. (2019). “Fake news” is the invention of a liar: How false information circulates within the hybrid news system. Current Sociology, 67(4), 625–42. Lee, F. L. F. (2016). Impact of social media on opinion polarization in varying times. Communication and the Public, 1(1), 56–71. Lelkes, Y. (2016). Mass Polarization: Manifestations and Measurements. Public Opinion Quarterly, 80(S1), 392–410. Prior, M. (2013). Media and Political Polarization. Annual Review of Political Science, 16(1), 101–27. Tucker, J. A., Guess, A., Barbera, P., Vaccari, C., Siegel, A., Sanovich, S., Stukal, D., & Nyhan, B. (2018) (Eds). Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature. http://eprints.lse.ac.uk/id/eprint/87402 Waisbord, S. (2018). Truth is what happens to news: On journalism, fake news, and post-truth. Journalism studies, 19(13), 1866–78. World Health Organization (2020a, 15 February). WHO General Director’s Speech at Munich Security Conference. https://www.who.int/director- general/speeches/detail/munich-security-conference http://eprints.lse.ac.uk/id/eprint/87402 https://www.who.int/director-general/speeches/detail/munich-security-conference https://www.who.int/director-general/speeches/detail/munich-security-conference Paper 3 REPRESENTATION? TREATY? POLARISATION IN NEWS AND SOCIAL MEDIA DEBATES ABOUT INDIGENOUS RIGHTS IN AUSTRALIA AND AOTEAROA NEW ZEALAND Axel Bruns Digital Media Research Centre, Queensland University of Technology Tariq Choucair Digital Media Research Centre, Queensland University of Technology Sebastian Svegaard Digital Media Research Centre, Queensland University of Technology Laura Vodden Digital Media Research Centre, Queensland University of Technology Daniel Whelan-Shamy Digital Media Research Centre, Queensland University of Technology Introduction and Background Australia and Aotearoa New Zealand continue to struggle with their respective colonial legacies, especially when it comes to the rights, recognition, and representation of Indigenous peoples. In 2023 and 2024, such controversies were once again reignited in both countries, if in very different contexts; this paper presents a comparative analysis of public debates in each case, paying particular attention to the intersections between social media discussions and news media coverage and the extent and structures of political polarisation which such controversies reveal. In Australia, the centrist Labor government under Prime Minister Anthony Albanese, elected in May 2022, honoured its election-night promise to hold a referendum on the constitutional recognition of Australia’s Indigenous peoples (Morse, 2022). On 14 October 2023, Australians voted on a proposal to enshrine an Indigenous Voice to Parliament – a federal body representing Aboriginal and Torres Strait Islander peoples and providing advice to parliament on matters that affect Indigenous peoples, which had been proposed in the 2017 Uluṟu Statement from the Heart (National Constitution Convention, 2017). However, in spite of substantial majority support for the constitutional recognition of Indigenous peoples in the early stages of the referendum campaign, the referendum lost 40% to 60% and failed to carry a majority in even a single of Australia’s six states. This was due to substantial extent to a highly effective and well-organised No campaign, aided by substantial support for No campaigners in Australia’s conservative media. In Aotearoa New Zealand, the 2023 national election, held by sheer coincidence on the same day as the Australian Voice referendum, eventually resulted – after several weeks of negotiations – in a change of government from the Labour Party to a coalition of the conservative National Party with populist minor parties NZ First and ACT (Neilson et al., 2023). Emerging amongst the key policy priorities for this new coalition (and especially for ACT) was the rolling back of existing policies recognising the rights, culture, and language of Aotearoa New Zealand’s Indigenous Māori minority; this manifested initially in part in the removal of bilingual English / te reo Māori nomenclature from government departments and other offices (McConnell, 2023), and subsequently centred especially on the reduction of special provisions for Māori language, culture, and self-governance derived from the 1840 Treaty of Waitangi between British colonial representatives and Māori chiefs. The depth of the division between Māori groups and their supporters on the one hand and the conservative government became especially obvious in the variously frosty and angry reception that the coalition party leaders experienced when they attended and spoke at public ceremonies on the eve of Waitangi Day, Aotearoa New Zealand’s annual national day of commemoration, on 5 Feb. 2024 (McKay, 2024). Data and Methods To investigate and analyse the public discourse surrounding these divisive debates about the recognition of Indigenous peoples in Australia and Aotearoa New Zealand, we gathered data on public debate across a range of mainstream social media platforms, including Facebook and Instagram (through CrowdTangle, which covers activity in public groups and pages), Twitter/X (via NodeXL’s scraping service, and subject to the severe limitations that now apply to data gathering from this platform), and YouTube (from the YouTube Data API). In each case, we gathered posts that contained one of a number of relevant keywords relating to either debate – for the Voice to Parliament referendum, from 1 Jan. 2023 to the referendum date of 14 Oct. 2023, and for the Waitangi debate, from the election date of 14 Oct. 2023 (data collection is ongoing, and for the purposes of this paper we will determine an appropriate end date). In addition, we also extracted from these posts any links to external sources (with particular attention to content from news outlets), and similarly examined whether YouTube content appearing in our datasets originated from news sources. For each of these datasets we performed a mixed-methods analysis including Natural Language Processing (NLP), topic modelling, network mapping, and qualitative close reading steps; we have also explored the potential to use Large Language Models (LLMs) like OpenAI’s ChatGPT to systematically assess the expressed stance of individual social media accounts towards the Voice referendum or Waitangi Treaty, and are currently reviewing the quality and reliability of such assessments. This combination of methods enabled us to identify a number of distinct groups of actors in each case, whose news sourcing and sharing practices we also examined. Preliminary Findings Within the space available in this extended abstract we can only sketch out some overall findings, and focus here on the Voice referendum case in Australia. Our analysis here clearly points to a highly asymmetrical contest: over the course of the campaign, the No side was able to very effectively mobilise uncertainty, doubt, and even fear towards the Voice to Parliament proposal (indeed, a deliberate strategy to foment doubt and fear especially amongst comparatively uninformed voters was revealed by media reporting; Sakkal, 2023). This is exemplified by the campaign’s simplistic but highly effective slogan ‘if you don’t know, vote no’. Such efforts were further aided and amplified by a coalition of conservative and far-right politicians and media outlets, with the arch-conservative Sky News Australia assuming an especially prominent position in the discourse network. Fig. 1: The network of Facebook groups and pages sharing on each other’s content demonstrates the considerably more coherent structure of the No campaign, and the centrality of Sky News Australia as a source of No-aligned content, while the Yes campaign is considerably less organised and has no comparably central campaign or media sources. The Yes campaign, by contrast, was considerably less coordinated: it enrolled a loose network of Indigenous activists, progressive politicians, union leaders, media and sports celebrities, and progressive activists, but this network lacked a core driver and failed to engage centrist or progressive media as supporters and amplifiers; this is due in part to the considerable market power of conservative media in Australia (led by Rupert Murdoch’s News Corporation, which also operates Sky News Australia), but also to the editorial choices of public broadcaster ABC and even the progressive Guardian Australia to favour even-handed fact-checking over active campaigning. Such choices represent a principled journalistic stance that is perfectly defensible, of course, but this mismatch between activist and disinterested journalistic reporting did also mean that public debate in news and social media was able to be dominated by the clearly partisan editorialising of Sky News Australia and other conservative and far-right outlets in support of the No campaign. Further Analysis and Interpretation Given the different status quo in Aotearoa New Zealand (where an existing treaty between Indigenous people and colonial society is being undermined), we expect our second case Sky News Australia Senator Jacinta Price – Don’t Divide Us Vote No The Voice 2023 – Yes or No Pauline Hanson’s Please Explain The Uluru Statement from the Heart Yes 23 ABC News Pensioners, Veterans and Seniors Political Discussion Group We Loathe Dutton & Murdoch Facebook On-Sharing Network Black/green/grey/red: public Facebook pages/groups Yes/No stance: LLM-based coding of post content study to exhibit very different patterns of debate, and intend to analyse and highlight these divergences in the full paper. Overall, we will interpret our observations through the lens of polarisation, distinguishing between the various forms of polarisation (issue-based, ideological, affective, interpretive, interactional, …) that have been identified in the relevant literature (cf. Kubin & von Sikorski, 2021; Lelkes, 2016; Marino & Ianelli, 2023; Yarchi et al., 2021), and assessing in particular whether the patterns of polarised discourse and engagement that are evident in our two cases exhibit any of the symptoms of destructive polarisation that Esau et al. (2023) have identified. Should such symptoms of destructive polarisation be present in our data – and our preliminary analysis of the Voice referendum debate certainly suggests that it exhibits some such symptoms – we will also reflect on how such trends could have been avoided and might yet be reversed. While the Voice to Parliament referendum is lost, the struggle to maintain and safeguard Māori rights continues, and in Australia, too, greater recognition and representation of Aboriginal and Torres Strait Islander peoples in policy- making remains on the political agenda despite this major setback. References Esau, K., Choucair, T., Vilkins, S., Svegaard, S., Bruns, A., O’Connor Farfan, K. & Lubicz, C. (2023, 30 May). Destructive Political Polarization in the Context of Digital Communication – A Critical Literature Review and Conceptual Framework. International Communication Association, Toronto. https://eprints.qut.edu.au/238775/ Kubin, E., & von Sikorski, C. (2021). The Role of (Social) Media in Political Polarization: A Systematic Review. Annals of the International Communication Association, 45(3), 188–206. https://doi.org/10.1080/23808985.2021.1976070 Lelkes, Y. (2016). Mass Polarization: Manifestations and Measurements. Public Opinion Quarterly, 80(S1), 392–410. https://doi.org/10.1093/poq/nfw005 Marino, G., & Iannelli, L. (2023). Seven Years of Studying the Associations between Political Polarization and Problematic Information: a literature review. Frontiers in Sociology, 8. https://doi.org/10.3389/fsoc.2023.1174161 McConnell, G. (2023, 24 Nov.). Incoming Government to Change Branding to English and Repeal Treaty Clauses. Stuff. https://www.stuff.co.nz/national/politics/133346834/incoming-government-to-change- branding-to-english-and-repeal-treaty-clauses McKay, B. (2024, 5 Feb.). Heckles and Boos for NZ Government to ‘Protect Treaty of Waitangi’. Sydney Morning Herald. https://www.smh.com.au/world/oceania/heckles- and-boos-for-nz-government-to-protect-treaty-of-waitangi-20240205-p5f2iw.html Morse, D. (2022, 23 May). Anthony Albanese Promised Action on the Uluru Statement from the Heart. So What Is the Proposed Indigenous Voice to Parliament? ABC News. https://www.abc.net.au/news/2022-05-24/federal-election-anthony-albanese- indigenous-uluru-statement/101092816 National Constitution Convention. (2017). Uluru Statement from the Heart. https://ulurustatemdev.wpengine.com/wp- content/uploads/2022/01/UluruStatementfromtheHeartPLAINTEXT.pdf Neilson, M., Pearse, A., & Cheng, D. (2023, 24 Nov.). Coalition Agreement: National, Act, NZ First and the Deal That Delivers New Government. NZ Herald. https://eprints.qut.edu.au/238775/ https://doi.org/10.1080/23808985.2021.1976070 https://doi.org/10.1093/poq/nfw005 https://doi.org/10.3389/fsoc.2023.1174161 https://www.stuff.co.nz/national/politics/133346834/incoming-government-to-change-branding-to-english-and-repeal-treaty-clauses https://www.stuff.co.nz/national/politics/133346834/incoming-government-to-change-branding-to-english-and-repeal-treaty-clauses https://www.smh.com.au/world/oceania/heckles-and-boos-for-nz-government-to-protect-treaty-of-waitangi-20240205-p5f2iw.html https://www.smh.com.au/world/oceania/heckles-and-boos-for-nz-government-to-protect-treaty-of-waitangi-20240205-p5f2iw.html https://www.abc.net.au/news/2022-05-24/federal-election-anthony-albanese-indigenous-uluru-statement/101092816 https://www.abc.net.au/news/2022-05-24/federal-election-anthony-albanese-indigenous-uluru-statement/101092816 https://ulurustatemdev.wpengine.com/wp-content/uploads/2022/01/UluruStatementfromtheHeartPLAINTEXT.pdf https://ulurustatemdev.wpengine.com/wp-content/uploads/2022/01/UluruStatementfromtheHeartPLAINTEXT.pdf https://www.nzherald.co.nz/nz/politics/coalition-agreement-national-act-nz-first-and- the-deal-that-delivers-new-government/4MEDZQBQ7BDK3J3DLG5K5K3HHY/ Sakkal, P. (2023, 12 Sep.). No Campaign’s ‘Fear, Doubt’ Strategy Revealed. Sydney Morning Herald. https://www.smh.com.au/politics/federal/no-campaign-s-fear-doubt- strategy-revealed-20230910-p5e3fu.html Yarchi, M., Baden, C., & Kligler-Vilenchik, N. (2021). Political Polarization on the Digital Sphere: A Cross-platform, Over-time Analysis of Interactional, Positional, and Affective Polarization on Social Media. Political Communication, 38(1–2), 98–139. https://doi.org/10.1080/10584609.2020.1785067 https://www.nzherald.co.nz/nz/politics/coalition-agreement-national-act-nz-first-and-the-deal-that-delivers-new-government/4MEDZQBQ7BDK3J3DLG5K5K3HHY/ https://www.nzherald.co.nz/nz/politics/coalition-agreement-national-act-nz-first-and-the-deal-that-delivers-new-government/4MEDZQBQ7BDK3J3DLG5K5K3HHY/ https://www.smh.com.au/politics/federal/no-campaign-s-fear-doubt-strategy-revealed-20230910-p5e3fu.html https://www.smh.com.au/politics/federal/no-campaign-s-fear-doubt-strategy-revealed-20230910-p5e3fu.html https://doi.org/10.1080/10584609.2020.1785067 Paper 4 ONLINE POLARISATION AS FRAMEWORK IN RESEARCHING SOCIAL MEDIA ACTIVISM AGAINST SEXUAL VIOLENCE IN INDONESIA Alia Azmi Digital Media Research Centre, Queensland University of Technology Background This study proposes online polarisation as the theoretical framework in researching women’s online activism by focusing on the case of Indonesia, for three reasons: (a) the different characteristics of Indonesia’s digital feminist activism from those in the Global North, where the #metoo online movement started with individuals sharing experiences online and coming together as a shared network in voicing similar interests, (b) Indonesia’s ideological contestation between progressives (secular nationalists) and conservatives (Islamism) that often leads to polarisation based on religious interpretation, particularly on women’s issues; and (c) the fact that similar controversies also happened in other countries with majority Muslim populations, and particularly those that have secular constitutions. Digital feminist activism in the Global North is identified by the common use of the internet to counter the mainstream media’s silencing of women’s experiences of sexual harassment (Salter, 2013; Sills et al., 2016). Social media are considered a safe space for victims to tell of their experiences, with the choice of anonymity and in the hope of being heard, believed, and supported (Andalibi et al., 2018; Gorissen et al., 2023). Social media affordances like hashtags allow users to find others with similar interests (Bruns & Burgess, 2015), and in the case of sexual violence can also frame meanings into terms that expose the prevalence of sexual violence (Mendes et al., 2019). Movements like SlutWalk, The HollaBack!, #YesAllWomen, and #MeToo are considered successes in sharing individual stories and using catchphrases to expose sexual violence in the Global North (Fileborn, 2014; Loney-Howes, 2019; Mendes, 2015). In countries of the Global South, sharing sensitive individual stories, such as harassment experiences, is not common, and frowned upon. The disclosure of individual experiences of sexual violence does not always result in inspiration for others to share stories, as in the #MeToo movement, due to different identity, legal, and sociocultural systems. Objections from family and stigma from the community are the biggest challenges for victims and activists to speak up against sexual violence. In Indonesia, speaking up against harassment and expressing feminist perspectives can lead to judgment based on religious stigma. Public discussions about women’s roles in society and the fight against discrimination and violence have been polarising society, which is visible on social media and in policy-making processes. Controversies about women’s issues in Indonesia are rooted in different interpretations of religious teachings. The feminist movement faces challenges from religious groups who claim that feminism, “imported” from Western culture, is incompatible with local and religious values (Nisa, 2019). Discussions about gender roles and gender policy reforms are often influenced by different perspectives between secular nationalists and Islamists (Afrianty, 2020; Hidayahtulloh, 2023), for example relating to controversies about the role of women in the family and the bylaws on head covers for female students in certain regions of the country. The Elimination of Sexual Violence (ESV) bill, proposed in 2016 to the national legislative, was only passed after six years of deliberation and controversies between conservative and progressive groups. In October 2021, the Minister of Education implemented a decree to prevent and address sexual violence on university campuses, which also sparked controversies similar to the ESV Bill. Both the bill and the decree identify sexual violence as sexual advances without the victim’s consent. The definition is considered to align with democratic cosmopolitan values as the basis for gender equality norms; meanwhile, conservative groups and parties accused the proponents of the bill of a hidden agenda to adopt Western liberal values (Hidayahtulloh, 2023). This polarisation stems from Indonesia’s characteristics as the most populous Muslim nation, but with secular constitution, which means that the constitution does not recognise a specific state religion. The constitution, however, does mention the belief in one God as one of the state’s ideological principles, which is often interpreted differently between conservative (Islamism) and the progressive (secular nationalist) groups (Bourchier, 2019; Hidayahtulloh, 2023). The debate is often linked to the history of the creation of the Indonesian constitution, when secular nationalists rejected a proposal from the Islamist parties and organisations to include Islamic sharia as part of the constitution and the nation’s ideology, the five principles of Pancasila. The drive to combine democracy and Islamic teachings, however, remains alive among conservative Islamists until now (Bourchier, 2019). In discussions from social interactions to government policies, it is quite common to have different opinions between religious conservatives and moderate secular nationalists. Other countries with majority Muslim populations have been facing the same problem when dealing with women’s rights. Surveys indicate that women in Muslim countries are often divided on issues such as wearing a veil, the role of wives, inheritance rights for women, and the implementation of sharia (Pew Research Center, 2013). Iran, for example, has seen women’s movements that align themselves with the international human rights framework (Far, 2023). Women in the Middle East and other regions have improved their knowledge of more moderate interpretations of the Quran to eliminate discrimination against women, therefore creating different paths from the traditional patriarchal interpretation of Islamic teachings (Kharroub, 2015). In these countries, there have been efforts to fights for women’s rights and against discrimination and violence, but there are still challenges from the traditional perspective, making polarisation unavoidable. Research Design This study focused on the controversies on social media about the Indonesian Minister of Education’s Decree about Sexual Violence on Campus and the Elimination of Sexual Violence bill. The ministerial decree came into effect in October 2021, in response to increasing reports of sexual violence cases in universities across Indonesia. At the time of the decree, Indonesia did not have any specific laws about sexual violence, and the Elimination of Sexual Violence bill was still being deliberated in the legislative body, the People’s Representative Council (DPR). It was eventually passed in 2022. This study uses posts about the Ministerial Decree and the ESV bill from Twitter, as one of the most used social media platforms in Indonesia. I collected historical tweets including key terms related to the two topics, using the Twitter Academic API in two different timeframes for both topics: (1) the Ministerial Decree from 1 October 2021, when the news about the decree started, to 31 December 2021; and (2) the ESV bill from 1 April 2016, when it was proposed to the legislative, to 12 April 2022, when the bill was passed. The data were analysed to identify important highlights such as peak activities, prominent actors, and posts that attracted the most interactions. I also examined the network of interactions in the data, to identify different clusters or groups that can be understood as representing the various political groups that are polarised against each other. By analysing the posts, actors, and the interaction network, I identify how polarisation about the issue contributes to or works against activism against sexual violence in Indonesia. References Afrianty, D. (2020). Rising public piety and the status of women in Indonesia two decades after reformasi. TRaNS : Trans-Regional and-National Studies of Southeast Asia, 8(1), 65–80. https://doi.org/10.1017/trn.2019.14 Andalibi, N., Haimson, O. L., Choudhury, M. D., & Forte, A. (2018). Social Support, Reciprocity, and Anonymity in Responses to Sexual Abuse Disclosures on Social Media. ACM Transactions on Computer-Human Interaction, 25(5), 28:1-28:35. https://doi.org/10.1145/3234942 Bourchier, D. M. (2019). Two Decades of Ideological Contestation in Indonesia: From Democratic Cosmopolitanism to Religious Nationalism. Journal of Contemporary Asia, 49(5), 713–733. https://doi.org/10.1080/00472336.2019.1590620 Bruns, A., & Burgess, J. (2015). Twitter hashtags from ad hoc to calculated publics. In N. Rambukkana (Ed.), Hashtag publics: The power and politics of discursive networks (pp. 13–27). Peter Lang. Far, T. S. (2023, June 26). Unveiling Resistance: The Struggle for Women’s Rights in Iran | Human Rights Watch. https://www.hrw.org/news/2023/06/26/unveiling- resistance-struggle-womens-rights-iran Fileborn, B. (2014). Online Activism and Street Harassment: Digital Justice or Shouting Into The Ether? Griffith Journal of Law and Human Dignity, 2(1) (Women & Violence: 2014 Special), 23. Gorissen, M., van den Berg, C. JW., Ruiter, S., & Bijleveld, C. CJH. (2023). Sharing unwanted sexual experiences online: A cross-platform analysis of disclosures before, during and after the #MeToo movement. Computers in Human Behavior, 144, 107724. https://doi.org/10.1016/j.chb.2023.107724 Hidayahtulloh, M. A. (2023). Ideological Contestation and Gender Policy Reforms in Post-Reformasi Indonesia. Journal of Contemporary Asia, 0(0), 1–24. https://doi.org/10.1080/00472336.2023.2177558 https://doi.org/10.1017/trn.2019.14 https://doi.org/10.1145/3234942 https://doi.org/10.1080/00472336.2019.1590620 https://www.hrw.org/news/2023/06/26/unveiling-resistance-struggle-womens-rights-iran https://www.hrw.org/news/2023/06/26/unveiling-resistance-struggle-womens-rights-iran https://doi.org/10.1016/j.chb.2023.107724 https://doi.org/10.1080/00472336.2023.2177558 Kharroub, T. (2015, October 4). Five things you need to know about women in Islam: Implications for advancing women’s rights in the Middle East. Arab Center Washington DC. https://arabcenterdc.org/resource/five-things-you-need-to-know- about-women-in-islam-implications-for-advancing-womens-rights-in-the-middle-east/ Loney-Howes, R. (2019). The Politics of the Personal: The Evolution of Anti-rape Activism From Second-Wave Feminism to #MeToo. In B. Fileborn & R. Loney-Howes (Eds.), #MeToo and the Politics of Social Change (pp. 21–35). Springer International Publishing. https://doi.org/10.1007/978-3-030-15213-0_2 Mendes, K. (2015). SlutWalk: Feminism, Activism and Media. Palgrave Macmillan UK. https://books.google.com.au/books?id=JFEMCgAAQBAJ Mendes, K., Keller, J., & Ringrose, J. (2019). Digitized narratives of sexual violence: Making sexual violence felt and known through digital disclosures. New Media & Society, 21(6), 1290–1310. https://doi.org/10.1177/1461444818820069 Pew Research Center. (2013). The World’s Muslims: Religion, Politics, and Society. https://www.pewresearch.org/religion/2013/04/30/the-worlds-muslims-religion- politics-society-women-in-society/ Salter, M. (2013). Justice and revenge in online counter-publics: Emerging responses to sexual violence in the age of social media. Crime, Media, Culture, 9(3), 225–242. https://doi.org/10.1177/1741659013493918 Sills, S., Pickens, C., Beach, K., Jones, L., Calder-Dawe, O., Benton-Greig, P., & Gavey, N. (2016). Rape culture and social media: Young critics and a feminist counterpublic. Feminist Media Studies, 16(6), 935–951. https://doi.org/10.1080/14680777.2015.1137962 https://arabcenterdc.org/resource/five-things-you-need-to-know-about-women-in-islam-implications-for-advancing-womens-rights-in-the-middle-east/ https://arabcenterdc.org/resource/five-things-you-need-to-know-about-women-in-islam-implications-for-advancing-womens-rights-in-the-middle-east/ https://doi.org/10.1007/978-3-030-15213-0_2 https://books.google.com.au/books?id=JFEMCgAAQBAJ https://doi.org/10.1177/1461444818820069 https://www.pewresearch.org/religion/2013/04/30/the-worlds-muslims-religion-politics-society-women-in-society/ https://www.pewresearch.org/religion/2013/04/30/the-worlds-muslims-religion-politics-society-women-in-society/ https://doi.org/10.1177/1741659013493918 https://doi.org/10.1080/14680777.2015.1137962 Paper 5 UNPACKING THE HIGHLY POLARIZED U.S. ABORTION DEBATE ONLINE: POLITICAL CANDIDATES’ MESSAGING BEFORE AND AFTER THE DOBBS RULING Jennifer Stromer-Galley Syracuse University Social media platforms in the United States, especially Twitter and Facebook, have become central communication channels in the public sphere. Politicians and those running for political office actively use platforms such as Twitter and Facebook to communicate to their supporters and the broader public (Stromer-Galley, 2019). They use these platforms to articulate their policy agendas and attack their opponents’ agendas. There is also concern that the platforms foster communication that contributes to political polarization (Bail, 2021). Research suggests, for, example that people are more likely to share negative messages on Twitter than positive messages (Stromer-Galley, Zhang, Hemsley, & Tanupabrungsun, 2018), which can help boost polarization. Access to abortion is a contested and polarizing policy topic in the United States. Political polarization is a process and a consequence of extreme rhetoric that shapes public attitudes (Sunstein, 2002). Of special concern is affective polarization, which is a negative predisposition that partisans have towards out-groups members (Iyengar et al. 2012). Even in the months and years after the U.S. Supreme Court codified a woman’s right to the medical procedure in 1974, it remained a fraught political topic, with Republican politicians typically aligning with the pro-life movement, and Democratic politicians aligning with the pro-choice movement, and both sides articulating affectively polarizing discourse. Legal protections of abortion access have eroded since 1974, with conservatives vowing to overturn the Supreme Court’s ruling in Roe v. Wade that defined those protections. The election of Republican Donald Trump and his appointment of three Supreme Court justices led to the ruling in Dobbs vs. Jackson’s Women’s Health in 2022 that undid the legal protections defined in Roe. Ideally, elections provide the citizenry the opportunity to learn the policy positions of their potential representatives in government, and then decide on who would best represent their interests (Trent, Friedenberg, & Denton, 2011). In practice, U.S. political campaigns are episodic exercises in political polarization, in which ideology and identity coalesce around the candidates of the two major parties (Hernandez et al., 2021; Lau et al., 2017). In this context, the campaigns use social media to not only articulate their policy positions but do so in a way to court supporters and enable them to amplify their message and evangelize to the broader public. Social media becomes an amplifier of affectively- polarizing speech. One weakness of the literature on polarization is that it focuses more on the attitudes of the public than the causes of those attitudes. Political discourse, especially by people in political power, contribute to the polarizing attitudes that are increasingly pervasive. Given that, my project aims to explore how the discourse by political candidates on social media evolved from 2018 to 2022 – from the lead-up to the Dobbs ruling to the election in the year it took effect. Surprisingly little research has yet explored how political candidates have used their social media platforms to either advocate for or attack abortion rights, and we know little about how the discourse evolved in that time, or how the public engaged with those messages. To understand the discourse in this time period, I ask the following research questions: To what extent did political campaign messages on Twitter and Facebook change between 2018, when abortion was still codified in federal law, and 2022, after Roe v. Wade was overturned? Were Republican candidates more likely to share social media messages about abortion as compared with Democrats, and were there changes over time? Were candidates more likely to attack opponents for their positions on abortion, or to instead advocate for their own positions? To what extent did women candidates emphasize abortion differently than male candidates, and what changes if any are in evidence from before to after the Dobbs ruling? Were there differences in the incivility that attacked opponents in abortion messages by party, by gender, or by political race over time? Are there any differences in the messaging strategies on Twitter as compared with Facebook across the three election periods by race and by party? The data for this analysis comes from a multi-year project collecting Facebook posts and Tweets from U.S. political candidates. The corpus includes messages from nearly all candidates with Facebook accounts and Twitter accounts that ran for U.S. senate or governor during the general elections of 2018, 2020, and 2022.1 To analyze the massive volume of messages, as part of a larger project, we developed algorithms that categorize messages on a variety of different dimensions, including on their policy positions, whether the messages attack opponents or advocate for their own positions (Stromer-Galley and Rossini, 2023), and whether the message tone is civil or uncivil. We also collected meta- data from Twitter and Facebook of engagement metrics, including likes, shares/retweets, and comments. We further gathered meta data regarding the candidate, such as their political party and their gender. I am currently in the process of updating our policy topic classifiers to take advantage of the BERT large language model (Devlin, Chang, Lee, & Toutanova, 2018), and so the data are not yet analyzable. We specifically have a policy classifier focused on women’s issues, which includes abortion. Our process will include initially analyzing the “women’s issues” categorized messages and then building a key-word search within those to identify messages pertaining explicitly to abortion. We will then further build classifiers to identify messages that support or that oppose abortion access. Once those messages are categorized, then we will analyze the data with a series of descriptive statistics to answer the research questions. In addition, we will read the messages across time to also provide a richer qualitative analysis of the changes over time in the public messaging by political candidates, especially around polarizing discourse. 1 I note “nearly” because collection problems surfaced in 2018. We failed to collect a few messages or all of the general election messages of less than 10% of the candidates that year, primarily senatorial candidates. We attempted to collect those missing messages in 2023, but a few candidates deleted their campaign accounts creating a gap in our collection. Our aim is to trace how the public messaging by political candidates on social media evolved over the time period from before to after the Dobbs ruling. Political candidates use their social media accounts to articulate their policy positions, but at the same time we are concerned some communicate in ways that denigrate and dehumanize the opposition, which creates the environment for affective polarization. While members of the public and elites can and do disagree intensely on policy matters, polarization is more likely when the debate turns hostile and dehumanizing. References Bail, C. (2021). Breaking the social media prism: How to make our platforms less polarizing. Princeton University Press. Devlin, J., Chang, M-W., Lee, K., & Toutanova, K. (2018). BERT: Pre-training of deep bidirectional transformers for language understanding. arXiv: https://arxiv.org/abs/1810.04805 Hernandez, E., Anduiza, E., & Rico, G. (2021). Affective polarization and the salience of elections. Electoral Studies, 69. 10.1016/j.electstud.2020.102203 Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology: A social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431. 10.1093/poq/nfs038 Lau, R., Anderson, D. J., Ditonto, T. M., Kleinberg, M. S., & Redlawsk, D. P. (2017). Effect of media environment diversity and advertising tone on information search, selective exposure, and affective polarization. Political Behavior, 39, 231 - 255. 10.1007/s11109-016-9354-8 Sunstein, C. (2002). The law of group polarization. Journal of Political Philosophy, 10 (2), 175-195. 10.1111/1467-9760.00148 Stromer-Galley, J. (2019). Presidential Campaigning in the Internet Age (2nd Ed.). Oxford University Press. Stromer-Galley, J., & Rossini, P. (2023). Categorizing political campaign messages on social media using supervised machine learning. Journal of Information Technology & Politics. 10.1080/19331681.2023.2231436. Stromer-Galley, J., Zhang, F., Hemsley, J., & Tanupabrungsun, S. (2018). Tweeting the attack: Predicting gubernatorial candidate attack messaging and its spread. International Journal of Communication, 12, 3511-3532. 1932-8036/20180005 Trent, J.S., Friedenberg, R. V., & Denton, Jr. R. E. (2011). Political campaign communication: Principles and practices (7th Ed.). New York: Rowman & Littlefield Publishers.