Understanding policy, media, and academic narratives around cycles of disadvantage in Australia
The way society thinks about an issue determines how it responds. This project seeks to understand public discourse around cycles of disadvantage in Australia by analysing public texts across media and parliamentary speeches. This will enable us to better understand what the public thinks about disadvantage and the factors that shape it. It will also help us understand what forces in turn shape the discourse.
Project background and aims
We have built an interactive interface which will be featured on large wall panels in the PRF exhibition space at Darlinghurst and on the web (https://prfviz.org
). It allows visitors to explore the changing narratives around disadvantage in the Australian media and parliament over multiple decades. The data is compiled from the text of thousands of articles and speeches and so is complex and multifaceted. The interactive visualisation we have created uses a novel blend of visual data storytelling to deliver important findings from our statistical analysis of the data as well as free-form exploration to allow users to test their own hypotheses and form new insights.
How do narratives change over time, and what can we learn about who is and isn't included or heard in these public forums? When discussing complex issues of multidimensional disadvantage and wellbeing, which topics are framed in relation to each other, and what does that tell us about our policy focus? How can narrative analysis inform our advocacy as communities and the for-purpose sector?
The Paul Ramsay Foundation (PRF) has a bold ambition to disrupt disadvantage in Australia through pioneering innovations. This is a vision to break cycles of disadvantage and bring about lasting, systemic change. But achieving this ambition requires deep, evidence-based understanding of the narratives that surround disadvantage in this country. This vital knowledge will be the foundation for relevant and meaningful long-term change.
Using a transdisciplinary perspective and cutting-edge algorithmic discourse analysis methods, this project will provide a path-breaking tool to identify and track changes in public narratives about disadvantage. This tool will provide a structured way of understanding public attitudes to disadvantage – and ultimately inform their own planning, reporting, impact assessments, and activities aiming to influence public narratives about disadvantage as they emerge. The large-scale, longitudinal data generated by this project will provide the key to answering a broad range of questions about how narratives emerge, and how they can influence, reinforce and break cycles of disadvantage. It will open the door to a broad array of future projects that link this dataset to others (e.g. statistics on employment, crime, health, social capital, etc.) to understand how narratives influence employment, crime, health, and well-being outcomes; test the effect of campaigns that aim to alter public narratives; and identify events that disrupt intergenerational and intersectional cycles of disadvantage.
Our discourse of disadvantage is multi-dimensional and intersectional: News & Opinion dataset.
Radial dendrogram and clustering presentation of co-occurrence analysis of issues in the News & Opinion dataset. Issue pairs that more often co-occur in articles are linked by smaller bridges (e.g. ChildProtection and Violence), whilst those which are less likely to co-occur are far apart (e.g. Shelter and ProtectingHeritage). Issues placed together via optimised clustering are coloured accordingly. Broadly, the analysis identifies three main branches of intersectionality: Wellbeing (A), Economy (B), and Autonomy (C).
Disadvantage is a complex and multidimensional issue. It has many intersecting elements that can be viewed through a multitude of lenses. Recognising this complexity, our proposal captures content matter expertise through an expert panel of leading scholars and practitioners working on issues of disadvantage (e.g. economics, social inclusion, youth, criminal justice, education). This panel will provide essential insights, feedback and decisions at key points of the project. We plan to combine this transdisciplinary content matter expertise with state-of-the-art technical capabilities (e.g. evidence review, artificial intelligence, data science). Using cutting-edge social science methodology and artificial intelligence methods, we will conduct reviews and case studies analysing changes in public narratives. In parallel, we will investigate the potential to develop new technologies that can automate some or all of these analyses; develop new technologies that can improve on them; and provide proof-of-concept demonstrations of their potential.
See our interactive web tool here
- O'Neill, Lachlan, Anantharama, Nandini, Buntine, Wray and Angus, Simon D, (2021), Quantitative Discourse Analysis at Scale - AI, NLP and the Transformer Revolution, No 2021-12, SoDa Laboratories Working Paper Series, Monash University, SoDa Laboratories.
- Measuring Social Inclusion: The Inclusive Australia Social Inclusion Index. Faulkner, N., Borg, K., Zhao, K. & Smith, L., Sep 2020, Australia: Inclusive Australia. 34 p. (report)