SSOAR Logo
    • Deutsch
    • English
  • Deutsch 
    • Deutsch
    • English
  • Einloggen
SSOAR ▼
  • Home
  • Über SSOAR
  • Leitlinien
  • Veröffentlichen auf SSOAR
  • Kooperieren mit SSOAR
    • Kooperationsmodelle
    • Ablieferungswege und Formate
    • Projekte
  • Kooperationspartner
    • Informationen zu Kooperationspartnern
  • Informationen
    • Möglichkeiten für den Grünen Weg
    • Vergabe von Nutzungslizenzen
    • Informationsmaterial zum Download
  • Betriebskonzept
Browsen und suchen Dokument hinzufügen OAI-PMH-Schnittstelle
JavaScript is disabled for your browser. Some features of this site may not work without it.

Download PDF
Volltext herunterladen

(externe Quelle)

Zitationshinweis

Bitte beziehen Sie sich beim Zitieren dieses Dokumentes immer auf folgenden Persistent Identifier (PID):
https://doi.org/10.17645/mac.9677

Export für Ihre Literaturverwaltung

Bibtex-Export
Endnote-Export

Statistiken anzeigen
Weiterempfehlen
  • Share via E-Mail E-Mail
  • Share via Facebook Facebook
  • Share via Bluesky Bluesky
  • Share via Reddit reddit
  • Share via Linkedin LinkedIn
  • Share via XING XING

Ideology and Policy Preferences in Synthetic Data: The Potential of LLMs for Public Opinion Analysis

[Zeitschriftenartikel]

Lee, Keyeun
Park, Jaehyuk
Choi, Suh-hee
Lee, Changkeun

Abstract

This study investigates whether large language models (LLMs) can meaningfully extend or generate synthetic public opinion survey data on labor policy issues in South Korea. Unlike prior work conducted on people's general sociocultural values or specific political topics such as voting intentions, ou... mehr

This study investigates whether large language models (LLMs) can meaningfully extend or generate synthetic public opinion survey data on labor policy issues in South Korea. Unlike prior work conducted on people's general sociocultural values or specific political topics such as voting intentions, our research examines policy preferences on tangible social and economic topics, offering deeper insights for news media and data analysts. In two key applications, we first explore whether LLMs can predict public sentiment on emerging or rapidly evolving issues using existing survey data. We then assess how LLMs generate synthetic datasets resembling real-world survey distributions. Our findings reveal that while LLMs capture demographic and ideological traits with reasonable accuracy, they tend to overemphasize ideological orientation for politically charged topics - a bias that is more pronounced in fully synthetic data, raising concerns about perpetuating societal stereotypes. Despite these challenges, LLMs hold promise for enhancing data-driven journalism and policy research, particularly in polarized societies. We call for further study into how LLM-based predictions align with human responses in diverse sociopolitical settings, alongside improved tools and guidelines to mitigate embedded biases.... weniger

Thesaurusschlagwörter
öffentliche Meinung; künstliche Intelligenz; politische Einstellung; Umfrageforschung; Datenverarbeitung

Klassifikation
Allgemeines, spezielle Theorien und Schulen, Methoden, Entwicklung und Geschichte der Kommunikationswissenschaften

Freie Schlagwörter
AI-generated text; ChatGPT; large language models; news media; policy preferences

Sprache Dokument
Englisch

Publikationsjahr
2025

Zeitschriftentitel
Media and Communication, 13 (2025)

Heftthema
AI, Media, and People: The Changing Landscape of User Experiences and Behaviors

ISSN
2183-2439

Status
Veröffentlichungsversion; begutachtet (peer reviewed)

Lizenz
Creative Commons - Namensnennung 4.0


GESIS LogoDFG LogoOpen Access Logo
Home  |  Impressum  |  Betriebskonzept  |  Datenschutzerklärung
© 2007 - 2025 Social Science Open Access Repository (SSOAR).
Based on DSpace, Copyright (c) 2002-2022, DuraSpace. All rights reserved.
 

 


GESIS LogoDFG LogoOpen Access Logo
Home  |  Impressum  |  Betriebskonzept  |  Datenschutzerklärung
© 2007 - 2025 Social Science Open Access Repository (SSOAR).
Based on DSpace, Copyright (c) 2002-2022, DuraSpace. All rights reserved.