Emotionalisation of political discourse: Sentiment and topic analysis of public comments on Austrian political parties

Erokhin, D. ORCID: https://orcid.org/0000-0002-5191-0579 (2026). Emotionalisation of political discourse: Sentiment and topic analysis of public comments on Austrian political parties. In: Day of Parliamentary Research, 16 June 2026, Vienna, Austria.

[thumbnail of Parlament 2026.pdf]
Preview
Text
Parlament 2026.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial.

Download (2MB) | Preview

Abstract

This study examines the extent to which emotions dominate public discourse in comments on YouTube videos that appear as most relevant results in searches for major Austrian political parties (ÖVP, NEOS, FPÖ, SPÖ, GRÜNE). It explores how emotional tone varies across parties and whether this contributes to societal polarization through digital media, particularly by prioritizing emotions over facts in politically charged debates.
The research retrieves most relevant YouTube videos as of 16th of April 2026 from searches using each party acronym. In total, 436 ÖVP-, 540 FPÖ-, 284 NEOS-, 501 SPÖ-, and 271 GRÜNE-related videos are analyzed. They constitute a mix of content, including news reports, interviews, debates, and commentary on topics such as coalition negotiations, government performance, integration policies, and controversial postings. Comments to these videos are collected via YouTube API and processed using the OpenAI API model gpt 5.4-minifor multilingual sentiment analysis detecting positive, negative, and neutral tones and specific emotions such as anger, fear, and hope, as well as topic modeling. With 67,415 comments, most frequently, users commented on FPÖ-related videos. In comparison, GRÜNE-related videos received 9,085 comments.
The analysis reveals similarities and differences across the party-related discources. For example, comments on ÖVP- and FPÖ-related videos were close in terms of comment language used, whereas GRÜNEand NEOS-related videos had a close level of engagement. Emotions across comments varied from anger and frustration to joy depending on the topics discussed. The same was true for the level of toxicity, polarization, and average sentiment. Comments under GRÜNE-related videos were more polarizing, comments under NEOS-related videos had a higher level of emotional intensity, and comments under FPÖ-related videos had a higher sentiment. As expected, discussions were mostly emotional rather than factual, but it again varied across parties and topics. For instance, comments on climate, energy, and transport were more factual under ÖVPand SPÖ-related videos, whereas discussions under FPÖ-related videos were more factual on economy, costs, and taxes.
The findings provide parliaments with empirical evidence on how digital media shapes public perceptions of parties and democratic processes. They support the development of strategies to promote evidence-based discussions in emotionally charged contexts, mitigate polarization, and improve monitoring of emotional dynamics in political communication. This type of research can serve as a complement to traditional surveys and polls, though it should also acknowledge different limitations and ethical aspects.

Item Type: Conference or Workshop Item (Poster)
Research Programs: Advancing Systems Analysis (ASA)
Advancing Systems Analysis (ASA) > Cooperation and Transformative Governance (CAT)
Depositing User: Luke Kirwan
Date Deposited: 17 Jun 2026 11:57
Last Modified: 23 Jun 2026 07:35
URI: https://pure.iiasa.ac.at/21659

Actions (login required)

View Item View Item