Gendered Meaning at Scale: Text-as-Data Approaches to Representation, Power, and Political Conflict

Abstract

Text-as-data approaches are increasingly used to study political behaviour, representation, and discourse. These methods allow scholars to analyse large volumes of political texts, ranging from parliamentary debates and party manifestos to media coverage and online communication. They offer new ways of examining how gender and power are constructed and contested in political life. Yet within gender and politics, these innovations remain dispersed across subfields and national contexts, and there are few dedicated spaces for exchanging methodological insights or discussing their implications. How gender is articulated, contested, and made politically meaningful is centrally mediated through language. Recent advances in text-as-data approaches allow us to analyse this at scale: not only how gender is talked about, but what work those linguistic choices perform. These methods — from supervised classification to embeddings to multilingual topic modelling — have opened up new ways of examining power, representation, and gendered conflict in parliaments, parties, media, and beyond. This panel brings together scholars developing or applying text-based quantitative, computational, and mixed methods in gender and politics research. Together, these papers demonstrate how text-based methods can illuminate (1) how political actors represent women and other gendered groups, (2) how gender equality gets reframed and politicised across issue domains, and (3) how language encodes power, blame, legitimacy, and status in democratic institutions. This panel features studies using a wide range of approaches, including content and discourse analysis at scale, dictionary or topic models, supervised and unsupervised classification, multilingual corpora, emerging applications of machine learning and LLMs, as well as hybrid designs linking text to survey data or behavioural outcomes. Collectively, this panel demonstrates that text-as-data is not simply a technical toolkit, but a fundamentally theoretical approach: one that makes visible the discursive processes through which gender is constructed, deployed, moralised, and weaponised. We also foreground the methodological challenges of this work — including operationalising gender and intersectionality in multilingual corpora, identifying biases in automated tools, and developing transparent designs that allow us to infer meaning, not merely detect words.

Title	Details
From “girls” to “women”: construction of personal responsibility in gendered public policies	View Paper Details
Gendern Mainstreaming? Analysing the use of gender-conscious language by Austrian MPs	View Paper Details
Why do people run for office? Exploring the motivations of first-time candidates during a real candidate selection process	View Paper Details
Leveraging Crowdcoding and LLMs for the analysis of gender mainstreaming: gender- sensitive recovery policies in the EU	View Paper Details
Claiming (y)EU: Gender, Voice, and Representative Claims in the European Commission	View Paper Details
Intersectionality at the margins: A mixed-methods analysis of IO disaster response evaluations (2008–2024)	View Paper Details