Skip to main navigation Skip to search Skip to main content

Explicit Stance Detection in the Political Domain: A New Concept and Associated Dataset

  • Alexander R. Caceres-Wright
  • , Naveen Udhayasankar
  • , Grant Bunn
  • , Stef M. Shuster
  • , Kenneth Joseph
  • SUNY Buffalo
  • North Carolina State University
  • Michigan State University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Stance detection, defined as the task of classifying an individual’s attitude towards a target person or concept, offers the potential to understand political opinions at scale using social media data. However, recent studies have questioned the robustness and accuracy of current stance detection methods, highlighting issues such as generalizability in time and inconsistencies in annotations driven by subtle differences in annotation task design. We argue that central to these challenges is the unresolved question of what constitutes an expression of stance. To address this, the present work introduces a distinction between explicit and implicit stance expressions, and argue that a focus on explicit stance detection addresses many of the existing concerns with modern stance detection methods. To facilitate research on explicit stance detection, we then present a novel (and public) dataset of over 1000 tweets across 13 stance targets for explicit stance detection and evaluate baseline models to establish a foundation for future research in this area.

Original languageEnglish
Title of host publicationSocial, Cultural, and Behavioral Modeling - 17th International Conference, SBP-BRiMS 2024, Proceedings
EditorsRobert Thomson, Aryn Pyke, Aravind Hariharan, Scott Renshaw, Patrick Park, Samer Al-khateeb, Annetta Burger
PublisherSpringer Science and Business Media Deutschland GmbH
Pages3-14
Number of pages12
ISBN (Print)9783031722400
DOIs
StatePublished - 2024
Event17th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2024 - Pittsburgh, United States
Duration: Sep 18 2024Sep 20 2024

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14972 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference17th International Conference on Social Computing, Behavioral-Cultural Modeling and Prediction and Behavior Representation in Modeling and Simulation, SBP-BRiMS 2024
Country/TerritoryUnited States
CityPittsburgh
Period09/18/2409/20/24

Keywords

  • Large Language Models
  • Politics
  • Social Media
  • Stance Detection

Fingerprint

Dive into the research topics of 'Explicit Stance Detection in the Political Domain: A New Concept and Associated Dataset'. Together they form a unique fingerprint.

Cite this