Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Which Identities Are Mobilized: Towards an automated detection of social group appeals in political texts

Felicia Riethmüller, Julian Dehne, Denise Al-Gaddooa

TL;DR

Riethmüller et al. tackle how political actors mobilize social identities and how to detect these cues at scale in European manifestos. They introduce a hybrid detection pipeline—dictionary seeds, BERT-based expansion, and a Mistral-7B with Embedding Space-based Filtering—to identify social group mentions across the Manifesto Corpus in 15 countries from 1980 to 2021. The LLM-ESF approach provides the strongest coverage and a scalable workflow, enabling cross-national mapping of party group images and comparisons between Radical Right and mainstream parties. Their empirical tests find no robust contagion: Radical Right strength or mainstream vote losses do not consistently drive convergence in group appeals, although patterns vary by country; the method offers a transferable tool for studying group-based rhetoric in political texts and related CSS contexts.

Abstract

This paper proposes a computational text classification strategy to identify references to social groups in European party manifestos and beyond. Our methodology uses machine learning techniques, including BERT and large language models, to capture group-based appeals in texts. We propose to combine automated identification of social groups using the Mistral-7B-v0.1 Large Language Model with Embedding Space-based filtering to extend a sample of core social groups to all social groups mentioned in party manifestos. By applying this approach to RRP's and mainstream parties' group images in manifestos, we explore whether electoral dynamics explain similarities in group appeals and potential convergence or divergence in party strategies. Contrary to expectations, increasing RRP support or mainstream parties' vote loss does not necessarily lead to convergence in group appeals. Nonetheless, our methodology enables mapping similarities in group appeals across time and space in 15 European countries from 1980 to 2021 and can be transferred to other use cases as well.

Which Identities Are Mobilized: Towards an automated detection of social group appeals in political texts

TL;DR

Riethmüller et al. tackle how political actors mobilize social identities and how to detect these cues at scale in European manifestos. They introduce a hybrid detection pipeline—dictionary seeds, BERT-based expansion, and a Mistral-7B with Embedding Space-based Filtering—to identify social group mentions across the Manifesto Corpus in 15 countries from 1980 to 2021. The LLM-ESF approach provides the strongest coverage and a scalable workflow, enabling cross-national mapping of party group images and comparisons between Radical Right and mainstream parties. Their empirical tests find no robust contagion: Radical Right strength or mainstream vote losses do not consistently drive convergence in group appeals, although patterns vary by country; the method offers a transferable tool for studying group-based rhetoric in political texts and related CSS contexts.

Abstract

This paper proposes a computational text classification strategy to identify references to social groups in European party manifestos and beyond. Our methodology uses machine learning techniques, including BERT and large language models, to capture group-based appeals in texts. We propose to combine automated identification of social groups using the Mistral-7B-v0.1 Large Language Model with Embedding Space-based filtering to extend a sample of core social groups to all social groups mentioned in party manifestos. By applying this approach to RRP's and mainstream parties' group images in manifestos, we explore whether electoral dynamics explain similarities in group appeals and potential convergence or divergence in party strategies. Contrary to expectations, increasing RRP support or mainstream parties' vote loss does not necessarily lead to convergence in group appeals. Nonetheless, our methodology enables mapping similarities in group appeals across time and space in 15 European countries from 1980 to 2021 and can be transferred to other use cases as well.
Paper Structure (17 sections, 4 equations, 3 figures, 5 tables)

This paper contains 17 sections, 4 equations, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Radical right parties' contested group image
  3. A group perspective on mainstreaming niche and challenger parties
  4. Research Design
  5. Case selection
  6. Computation of Social Group Labels
  7. Dictionary Approach
  8. BERT-based Classification of Social Group Labels
  9. Combined approach Mistral-LLM(LLM) and Embedding Space-based Filtering (ESF)
  10. Social Group Detection Approaches
  11. Operationalization
  12. Dependent variable:
  13. Independent variables
  14. Modelling strategy
  15. Descriptive results
  16. ...and 2 more sections

Figures (3)

  • Figure 1: Visualization of the Embedding Space-based Filtering (ESF)
  • Figure 2: Similarity (0-100) to the Radical Right's group appeal per election. The calculation is based on electoral manifestos (Lehmann et al. 2023a).
  • Figure 3: Groups mentioned with highest relative frequency by Radical Right compared to Centre Left Parties. Translation for a, from above: foreigners, guest workers, criminals, Muslims, asylum seekers, immigrants, nursing assistants, civil servants, apprentices, (manual) workers, Christians, older people, rural areas, those in need of care, people with migration backgrounds, start-up founders, the rich, women, young people, self-employed. Translation for b, from above: foreigners, criminals, asylum seekers, guest workers, Muslims, farmers, immigrants, soldiers, apprentices, Christians, older people, gay people, young women, lesbian people, young people, the rich, fathers, men, tenants, women