--- title: "Predicting Convincingness in Political Speech: How Emotional" description: "The study offers an early large-scale test of whether persuasion models trained on benchmarks transfer to… 0.87 precision. 3 references." authors: ["Bhuvanesh Verma", "Mounika Marreddy", "Alexander Mehler"] year: 2026 venue: "The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis (WASSA 2026)" source: "https://aclanthology.org/2026.wassa-1.4/" doi: "10.18653/v1/2026.wassa-1.4" licence: "CC BY 4.0 (ACL Anthology)" licence_url: "https://creativecommons.org/licenses/by/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/verma-2026-predicting-convincingness-political-speech-how/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/verma-2026-predicting-convincingness-political-speech-how/) [home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/39/) / Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength # Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength **Authors:** [Bhuvanesh Verma](https://md.research.elamur.ai/free/en/authors/bhuvanesh-verma/), [Mounika Marreddy](https://md.research.elamur.ai/free/en/authors/mounika-marreddy/) and [Alexander Mehler](https://md.research.elamur.ai/free/en/authors/alexander-mehler/) · **Year:** 2026 · **Venue:** [The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis (WASSA 2026)](https://md.research.elamur.ai/free/en/venues/wassa/) ([Workshop on Computational Approaches to Subjectivity, Sentiment and Social Media Analysis 2026](https://md.research.elamur.ai/free/en/venues/wassa/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.wassa-1.4/) · [PDF](https://aclanthology.org/2026.wassa-1.4.pdf) · [DOI 10.18653/v1/2026.wassa-1.4](https://doi.org/10.18653/v1/2026.wassa-1.4) · [OpenAlex](https://openalex.org/W7140142272) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary The paper asks whether emotional tone is linked to how convincing political arguments appear to computational models, using real campaign rhetoric rather than debate or social-media data. The authors work with the Campaign Speeches Text Data (CSTD), a corpus of 1,056 Democratic and Republican speeches from the 2020 U.S. presidential cycle. Their pipeline has four stages: - find topics with unsupervised topic modeling; - extract argumentative passages with a trained detector; - form cross-party argument pairs per topic and rank them with a convincingness model trained on a pairwise evidence-quality benchmark; - label emotions with a GoEmotions-based classifier. Randomization tests show that both parties argue to a similar extent and with similar overall quality. However, there are clear topic-level asymmetries, and Democrats win significantly more topics than chance would predict. Positive emotions tend to accompany stronger predicted persuasiveness. Negative emotions matter less overall but still play a role in some topics. The study offers an early large-scale test of whether persuasion models trained on benchmarks transfer to campaign speech. It also shows how such models can expose strategic emotional framing. ## Key points - **Resources:** CSTD (Chalkiadakis et al., 2025; Jan 2019–Jan 2021) and ArgDetect, which merges the claim data of Levy et al. (2018) with the evidence data of Ein-Dor et al. (2020). Additional resources are EviConv (Gleize et al., 2019) for pairwise quality ranking and SamLowe/roberta-base-go_emotions (27 emotion labels). - **Models compared:** TF-IDF/SE classical classifiers, bert-base-uncased, roberta-base, deberta-v3-large, and two-shot LLMs with constrained generation (Qwen2.5-Coder-14B, Mistral-8B-Instruct-2410, Llama-3.1-8B-Instruct). - **Performance:** argument detection reaches 0.87 precision, and argument quality ranking reaches 0.78 accuracy. - **Significance tests:** - There is no significant party difference in the share of argumentative text or in overall convincingness (Republicans lead by about 0.5%). - At the topic level, 10 topics significantly favor Democrats and 15 favor Republicans. - Democrats win about 64–65% of topics overall, a significant result (the paper reports 64% in the results and 65% in the conclusion). - **Emotion–topic findings:** - Positive tone is tied to persuasiveness in USMCA jobs/agriculture, Border Control Policies and Progressive Tax Reforms. - Negative tone (fear, sadness) appears in City Crime and Civil Unrest and in the Charlottesville topic. *Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.* 📚 Shelves (3): [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works · [🤔 Philosophy](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/philosophy/) · 284 works 🧭 Topics (1): [😊 Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · 972 works ## Cite Cite as: Bhuvanesh Verma, Mounika Marreddy and Alexander Mehler. 2026. Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength. The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis (WASSA 2026). The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{verma2026predicting, title = {{Predicting Convincingness in Political Speech: How Emotional Tone Shapes Persuasive Strength}}, author = {Bhuvanesh Verma, Mounika Marreddy and Alexander Mehler}, booktitle = {The Proceedings for the 15th Workshop on Computational Approaches to Subjectivity, Sentiment Social Media Analysis (WASSA 2026)}, year = {2026}, doi = {10.18653/v1/2026.wassa-1.4}, url = {https://doi.org/10.18653/v1/2026.wassa-1.4} } ``` - [← previous Predict and Fix: A Unified Model for Translation Quality Estimation and Post-Editing](https://md.research.elamur.ai/free/en/papers/modrzejewski-2026-predict-fix-unified-model-translation/) - [next → Predicting Gaze Location without Camera or Eye-Tracker](https://md.research.elamur.ai/free/en/papers/rezapoor-2026-predicting-gaze-location-without-camera/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/39/) ## References 3 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Segun T Aroyehun, Almog Simchon, Fabio Carrella, Jana Lasser, Stephan Lewandowsky, and David Garcia. 2025. Computational analysis of us congressional speeches reveals a shift from evidence to intuition. Nature Human Behaviour, pages 1–12. [find](https://scholar.google.com/scholar?q=Segun+T+Aroyehun%2C+Almog+Simchon%2C+Fabio+Carrella%2C+Jana+Lasser%2C+Stephan+Lewandowsky%2C+and+David+Garcia%2E+2025%2E+Computational+analysis+of+us+congressional+speeches+reveals+a+shift+from+evidence+to+intuitio) 2. Mohamed S Benlamine, Serena Villata, Ramla Ghali, Claude Frasson, Fabien Gandon, and Elena Cabrio. 2017. Persuasive argumentation and emotions: An empirical evaluation with users. In International Conference on Human-Computer Interaction, pages 659–671. Springer. [find](https://scholar.google.com/scholar?q=Mohamed+S+Benlamine%2C+Serena+Villata%2C+Ramla+Ghali%2C+Claude+Frasson%2C+Fabien+Gandon%2C+and+Elena+Cabrio%2E+2017%2E+Persuasive+argumentation+and+emotions%3A+An+empirical+evaluation+with+users%2E+In+International+Con) 3. Sahbi Benlamine, Maher Chaouachi, Serena Villata, Elena Cabrio, Claude Frasson, and Fabien Gandon. 2015. Emotions in argumentation: an empirical evaluation. In Proceedings of the Twenty-Fourth International Joint Conference on Artificial Intelligence, IJCAI 2015, Buenos Aires, Argentina, July 25-31,… [find](https://scholar.google.com/scholar?q=Sahbi+Benlamine%2C+Maher+Chaouachi%2C+Serena+Villata%2C+Elena+Cabrio%2C+Claude+Frasson%2C+and+Fabien+Gandon%2E+2015%2E+Emotions+in+argumentation%3A+an+empirical+evaluation%2E+In+Proceedings+of+the+Twenty%2DFourth+Interna) *Updated 1 November 2026*