AoIR 2024

Polarisation in Newssharing: Reviewing the Evidence from Facebook and Twitter

Felix Münch, Axel Bruns, and Laura Vodden

• 2 Nov. 2024 – AoIR 2024 conference, Sheffield

Presentation Slides

Abstract

Introduction

Conventional conceptualisations of polarisation in political science frequently focus on interpersonal and intergroup differences of opinion on specific issues, broader ideologies, and overall identities (Lelkes, 2016; Esau et al., 2023). These are often investigated through participant self-assessment, for instance using responses to feeling thermometers on a collection of issues (e.g. Alwin, 1997). But in addition to such self-reported measures, recent political communication research has described several other forms of polarisation that may also be observed independently: these include, for example, interpretive polarisation (where different individuals and groups interpret the same issue or event in widely divergent ways; cf. Kligler-Vilenchik et al., 2020) and interactional polarisation (where individuals and groups preferentially engage only with like-minded others; cf. Yarchi et al., 2021).

These latter forms of polarisation lend themselves to operationalisation through digital trace data, and this paper investigates the evidence for interpretive and interactional polarisation on Facebook and Twitter in Australia. In keeping with the overall theme of this panel, we do so specifically in the context of the news, focussing here on how users of both platforms share the news. Newssharing is a widespread, habitual practice on mainstream social media platforms (Bruns, 2018), and can be part of dedicated personal and collective news curation practices (Thomson & Wells, 2016); news content may be shared as is, or accompanied by framing texts that may be supportive or critical of the news reports being shared.

Data and Methods

We draw for our analysis on a number of large-scale datasets. For Twitter, we utilise an excerpt of the decade-long Australian Twitter News Index (ATNIX; see Bruns, 2017), which captured all tweets linking to one of 35 Australian news sources from 2012 to the discontinuation of the Twitter API under Elon Musk in 2023. For Facebook, we draw on a similar dataset (AFNIX), obtained from CrowdTangle, of all posts from public pages and groups that contained links to the same set of Australian news sources between 2018 and 2023, and further supplement this with data from the Facebook URL Shares dataset (Messing et al., 2018), providing aggregate metrics on the circulation of links to these sources in both public and non-public Facebook spaces but excluding the texts of the posts in which these URLs were shared. For the purposes of this paper, we focus on a year-long extract from these datasets, covering the year 2022. We select this year because it was significant for several political events both domestically (including the Australian federal election in May 2022) and internationally (including the Russian full-scale invasion of Ukraine, and the US mid-term elections), and may therefore feature heightened levels of polarisation.

We process and analyse these datasets in order to assess levels of both interactional and interpretive polarisation. First, although interactional polarisation is often conceptualised as occurring between peers (e.g. social media users), here we operationalise it as describing patterns of interaction between social media users and news outlets: in line with and extending the concept of selective exposure (Stroud, 2010), this assumes that social media users preferentially access – and subsequently share – a selection of news sources with which they have a particular affinity. A simplistic assumption would therefore be that our analysis will produce several groups of accounts that have distinct and divergent newssharing repertoires (predominantly sharing progressive or conservative media, for instance). The strength of such distinctions would indicate the level of interactional polarisation in Australian newssharing.

However, reality is likely to be more complex, and social media users may at times also share news content that does not align with their ideology, for instance in order to critique it. Such critical sharing is likely to exhibit features that differ from supportive sharing: in the former, news headlines and URLs may be accompanied by an oppositional framing, while in the latter, users may simply amplify (on Twitter: retweet; on Facebook: share on) the news outlets’ posts themselves, possibly with some limited additional commentary.

Second, therefore, in the ATNIX dataset (for Twitter) and the AFNIX dataset (for public pages and groups on Facebook) we distinguish these posting practices – both by distinguishing direct retweets and on-shares from posts that contain the news headlines and URLs but also add further commentary, and by evaluating that commentary for its stance towards the issues reported in the article. Through this part of our analysis, then, we are able to investigate interpretive polarisation: in addition to their overall newssharing repertoires, how do users differ in their interpretive approaches to the various news outlets (for instance, do they generally amplify one set of outlets uncritically, but often add their own commentary as they share others)? (The limitations of the Facebook URL Shares dataset prevent us from extending this analysis to the sharing of news URLs outside of public groups and pages, unfortunately.)

We investigate these intersecting patterns of interactional and interpretive polarisation through a mixture of large-scale network analysis and natural language processing and close qualitative interpretation of our findings. In addition to analysing our datasets for the full year 2022, we also examine the presence of any fluctuations in sharing patterns over the course of the year – for instance as the Australian federal election campaign potentially produces a temporary increase in political partisanship, or as Russia’s attack on Ukraine creates a shared sense of outrage that transcends partisan camps. Additionally, of course, any systematic divergences in newssharing polarisation patterns between Facebook and Twitter will also be of interest.

Prospective Contributions

We expect any patterns of both interactional and interpretive polarisation that emerge from this analysis to be determined at least in part by the political alignment of users and news outlets, and will draw in interpreting them on existing assessments of the relative political positioning of Australian news outlets on a spectrum from left to right (Park et al., 2021; Fletcher et al., 2020). However, other aspects – such as the relative quality of individual news outlets, from authoritative to tabloid – may also play a considerable role. Indeed, it is also possible that a small number of widely used news outlets (such as the Australian Broadcasting Corporation, Australia’s major public media organisation) are so dominant with social media users that overall newssharing patterns produce no evidence of widespread polarisation, and instead point only to a lack of diversity in Australian Twitter and Facebook users’ news repertoires. From some perspectives, such a state might even be seen as desirable – it would demonstrate that Australians have a broadly similar informational basis for their societal and political participation, and do not exist in highly divergent information environments.

Even if our study were to document such general homogeneity, however, significantly divergent interactional and interpretive patterns may still exist at the fringes of our datasets, and their presence could point to the existence of smaller polarised hyperpartisan individuals and groups in opposition to the broader societal mainstream. Our focus on (in 2022) fairly mainstream platforms like Facebook and Twitter, where they are perhaps underrepresented, might obscure their true strength; future extensions of the present work, to the extent that they are possible in other platform contexts, could therefore also compare our findings in this paper with alternative platforms such as Telegram or Reddit.

References

Alwin, D. F. (1997). Feeling Thermometers versus 7-Point Scales: Which Are Better? Sociological Methods & Research, 25(3), 318-340.

Bruns, A. (2018). Gatewatching and News Curation: Journalism, Social Media, and the Public Sphere. Peter Lang.

Bruns, A. (2017). Making Audience Engagement Visible: Publics for Journalism on Social Media Platforms. In B. Franklin & S. A. Eldridge II (Eds.), The Routledge Companion to Digital Journalism Studies (pp. 325–334). Routledge.

Esau, K., Choucair, T., Vilkins, S., Svegaard, S., Bruns, A., & Lubicz, C. (2023, May 30). 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/

Fletcher, R., Cornia, A., & Nielsen, R. K. (2020). How Polarized Are Online and Offline News Audiences? A Comparative Analysis of Twelve Countries. The International Journal of Press/Politics, 25(2), 169–195. https://doi.org/10.1177/1940161219892768

Kligler-Vilenchik, N., Baden, C., & Yarchi, M. (2020). Interpretative Polarization across Platforms: How Political Disagreement Develops over Time on Facebook, Twitter, and WhatsApp. Social Media + Society, 6(3). https://doi.org/10.1177/2056305120944393

Lelkes, Y. (2016). Mass Polarization: Manifestations and Measurements. Public Opinion Quarterly, 80(S1), 392–410. https://doi.org/10.1093/poq/nfw005

Messing, S., State, B., Nayak, C., King, G., & Persily, N. (2018). Facebook URL Shares: Codebook. Social Science One. https://socialscience.one/files/partnershipone/files/urls_dataset_for_rfp1.pdf

Park, S., Fisher, C., McGuinness, K., Lee, J. Y., & McCallum, K. (2021). Digital News Report: Australia 2021. News and Media Research Centre. https://doi.org/10.25916/KYGY-S066

Stroud, N. J. (2010). Polarization and Partisan Selective Exposure. Journal of Communication, 60(3), 556–576. https://doi.org/10.1111/j.1460-2466.2010.01497.x

Thorson, K., & Wells, C. (2016). Curated Flows: A Framework for Mapping Media Exposure in the Digital Age. Communication Theory, 26(3), 309–328. https://doi.org/10.1111/comt.12087

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