IAMCR 2024

‘If you don’t know, vote no’: Symptoms of Destructive Polarisation in the 2023 Voice to Parliament Referendum in Australia

Axel Bruns, Tariq Choucair, Sebastian Svegaard, Samantha Vilkins, Katharina Esau, and Laura Vodden

• 1 July 2024 – Paper presented at the IAMCR 2024 conference, Christchurch

Presentation Slides

Abstract

Much of the research on polarisation stops short of sufficiently defining and conceptualising the concept. This can lead to the conflation of different forms of polarisation in the design and findings of empirical studies; the over-diagnosis of problematic and pernicious forms of polarisation instead of mere disagreement and antagonism; and the unquestioned adoption of technologically determinist perspectives in the search for scapegoats and solutions.

Building on a systematic, crossdisciplinary review of the different forms of polarisation that have been proposed and identified in recent studies, Esau et al. (2023) introduced destructive polarisation as a particularly pernicious form of polarisation that is distinguished from more ordinary and less problematic forms of polarisation by a number of distinct symptomatic features; these include (a) breakdown of communication; (b) discrediting and dismissing of information; (c) erasure of complexities; (d) exacerbated attention and space for extreme voices; and (e) exclusion through emotions.

Building on this conceptual work, this paper addresses the challenge of translating the definitions of these symptoms into a methodological approach for identifying them in contemporary political debates, and for the empirical assessment of their severity – and therefore, for diagnosing the presence and extent of destructive dysfunction. In doing so, it also considers whether there are additional symptoms of destructive polarising dynamics that may need to be added to the initial list presented in Esau et al.

We develop and test this approach by applying it to a cross-platform dataset of social media debates related to the 14 October 2023 constitutional referendum on the establishment of an Indigenous Voice to Parliament in Australia. Although opinion polling early in 2023 indicated that nearly two thirds of Australians were in favour of better constitutional recognition for Australia’s Indigenous peoples, the referendum eventually failed, with some 60% of voters choosing to vote No. This starkly illustrates continued polarisation in the Australian electorate about greater Indigenous recognition – and we suggest that the referendum campaigns and associated public debate, especially also on social media, exhibits several of the proposed symptoms of destructive polarisation.

This is perhaps obvious for the No campaign’s official slogan ‘if you don’t know, vote no’, which clearly seeks to erase the complexities of the decision voters are asked (or indeed, under Australia’s compulsory electoral system, required) to make, and the widespread repetition of the slogan both by referendum campaigners and members of the general public demonstrates its resonance with some voters. However, the presence of other symptoms of destructive polarisation is less easily assessed, and requires more complex approaches to the operationalisation of the symptom definitions for analysis and evaluation.

In this paper, we therefore translate the symptom definitions into qualitative as well as quantifiable criteria, which we systematically apply to large-scale datasets of public debate on the Voice to Parliament referendum that were drawn from Facebook, Instagram, Twitter, and YouTube for the period from 1 January to 14 October 2023. Using a mixed-methods approach that involves manual and computational content analysis, network analysis, and AI-supported coding and categorisation, we document the presence of destructive polarisation in the Voice debate and assess its severity; in doing so, we convert the symptom definitions by Esau et al. into an analytical toolkit that can be applied to polarised debates on a wide range of topics.