A Case Study on Australian Political Debates About Transgender Rights
Led by DMRC Research Fellow Dr. Tariq Choucair in collaboration with the GenAI Lab, this project develops methods for computationally identifying references to social and identity groups, and considers the application of these method to Australian political debates on transgender rights.
Identifying social and identity groups in text data – including, for instance, news stories, political speeches, and social media posts – is an important task for Media and Communication studies. This identification allows researchers to investigate, for example, the discursive representation of these groups, sentiments towards these groups, and how these groups are framed. Automating this task allows for larger corpora to be investigated, making possible more complex, broader, and inclusive analysis, including comparisons across times, arenas, and platforms. Past research has performed the identification of mentions to social and identity groups mainly through manual identification and dictionaries. But the identification of other specific references in text has increasingly been done with computational methods, specially Named Entity Recognition (NER).
The project offers two key contributions: first, the development of several sophisticated Named Entity Recognition (NER) models capable of detecting these groups (e.g., “trans women”, “indigenous high school graduates”, “children living with a disability”), with a granularity that enables detailed analysis and differentiations between specific sub-groups or identities; and second, this model’ application to produce empirical insights on which and how social and identity groups are referenced or appealed to in debates about transgender rights in Australia.
