Elamur Research⌕
☰
home / papers / 2026 / Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance

Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance

Paper · 2026 · Serena Tardelli, Lorenzo Alvisi, Lorenzo Cima et al.

Summary

Emoji reactions, which readers attach to posts as lightweight feedback, have often been used as proxy emotion labels for distant supervision. This assumption is borrowed from research on in-text emojis, which tend to mirror the author's affect. The authors test whether reactions actually signal emotional resonance, using public crypto-related Telegram channels.

They collected English-language channels via TGStat and kept messages from 2024 that received at least one standard reaction. Each message was automatically tagged for sentiment, emotion, persuasion strategy and speech act. The sentiment and emotion of the reactions themselves were inferred with lexicons and LLMs.

Across rhetorical strategies and emotional tones, reactions skew positive even when the message is neutral or negative. Reactions therefore seem to work more as social approval or support than as shared feeling. Neutral and negative content also tends to attract more reactions.

The study cautions against treating reactions as ground-truth sentiment labels or as indicators of public opinion. This concern applies especially to misinformation, political and moderation research.

Key points

  • Dataset: 647,879 messages from 993 channels, filtered from an initial pool of 20,695 English crypto channels with at least 5,000 subscribers. Messages span January–December 2024, and custom emojis were excluded.
  • Annotation tools: Persuasion strategies were labeled with the valhalla/distilbart-mnli-12-1 zero-shot model using the Qachfar and Verma (2023) taxonomy. Other classifiers included BERT/RoBERTa-based models and GPT-4o, none of which were fine-tuned or validated on this data.
  • Main finding: Message sentiment and reaction sentiment are systematically mismatched, with positive reactions dominating. An ablation identifies joy as the emotion that most disrupts alignment.
  • Engagement predictors: Emotion Neutral is the strongest single predictor of reactions. Messages with more intense negativity and weaker positivity tend to receive more reactions, which extends the known negativity bias to emoji interactions.
  • Limitations: The study covers a single domain. Channel admins can restrict the available reactions, often to mostly positive options, which may introduce selection bias.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🤝National Research CouncilRome · Italy15 works · 15,649 citations🎓University of PisaPisa · Italy25 works · 3,842 citations🔬Institute of Informatics and TelematicsPisa · Italy3 works · 251 citations

Also on the byline: IMT School for Advanced Studies Lucca.

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksPsychology · 1,552 worksExperimental and Cognitive Psychology · 870 worksHuman-Computer Interaction · 141 works

🧭 Topics 1Research subjects and languages the work is about.Sentiment & emotion · 972 works

Cite

Cite as Serena Tardelli, Lorenzo Alvisi, Lorenzo Cima, Stefano Cresci and Maurizio Tesconi. 2026. Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance. 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. DOI 10.18653/v1/2026.wassa-1.2

@inproceedings{tardelli2026emoji,
  title = {{Emoji Reactions on Telegram: Unreliable Indicators of Emotional Resonance}},
  author = {Serena Tardelli, Lorenzo Alvisi, Lorenzo Cima, Stefano Cresci and Maurizio Tesconi},
  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-1.2},
  url = {https://doi.org/10.18653/v1/2026.wassa-1.2}
}

References

45 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Margherita Gambini, Serena Tardelli, and Maurizio Tesconi. 2024. The anatomy of conspiracy theorists: Unveiling traits using a comprehensive Twitter dataset. Computer Communications, 217:25–40. find
  2. Lorenzo Alvisi, Victoria Popa, Guglielmo Cola, Serena Tardelli, and Maurizio Tesconi. 2025a. From Toxicity to Conformity: Adaptive user behavior to social norms in Telegram communities. arXiv preprint arXiv:2511.17333. — [arXiv:2511.17333](https://arxiv.org/abs/2511.17333) find
  3. Lisa Graziani, Stefano Melacci, and Marco Gori. 2019. Jointly Learning to Detect Emotions and Predict Facebook Reactions. In International Conference on Artificial Neural Networks, pages 185–197. Springer. find
  4. Lorenzo Alvisi, Serena Tardelli, and Maurizio Tesconi. 2024. Unraveling the Italian and English Telegram Conspiracy Spheres through Message Forwarding. In International Conference on Advances in Social Networks Analysis and Mining, pages 204–213. Springer. find
  5. Md Abrar Jahin, Md Sakib Hossain Shovon, MF Mridha, Md Rashedul Islam, and Yutaka Watanobe. 2024. A hybrid transformer and attention based recurrent neural network for robust and interpretable sentiment analysis of tweets. Scientific Reports, 14(1):24882. find
  6. Lorenzo Alvisi, Serena Tardelli, and Maurizio Tesconi. 2025b. Mapping the Italian Telegram Ecosystem: Communities, Toxicity, and Hate Speech. arXiv preprint arXiv:2504.19594. — [arXiv:2504.19594](https://arxiv.org/abs/2504.19594) find
  7. Olusegun Oladele Jegede. 2024. Speech Acts in the Digital Sphere: A Corpus-Based Exploration of Interactional Dynamics on Social Media Platforms. Corpus-based Studies across Humanities, 2(2):359–383. find
  8. William J Brady, Julian A Wills, John T Jost, Joshua A Tucker, and Jay J Van Bavel. 2017. Emotion shapes the diffusion of moralized content in social networks. Proceedings of the National Academy of Sciences, 114(28):7313–7318. find
  9. Amit Khan, Dipankar Majumdar, and Bikromadittya Mondal. 2025. Sentiment analysis of emoji fused reviews using machine learning and Bert. Scientific Reports, 15(1):7538. find
  10. Leo Breiman. 2001. Random forests. Machine learning, 45(1):5–32. find
  11. Petra Kralj Novak, Jasmina Smailović, Borut Sluban, and Igor Mozetič. 2015. Sentiment of emojis. PloS one, 10(12):e0144296. find
  12. Andrea Calamusa, Serena Tardelli, Marco Avvenuti, Stefano Cresci, Ilaria Federigi, Maurizio Tesconi, Marco Verani, and Annalaura Carducci. 2020. Twitter monitoring evidence of Covid-19 infodemic in Italy. European Journal of Public Health, 30(Supplement_5):ckaa165–066. find
  13. Chuchu Liu, Fan Fang, Xu Lin, Tie Cai, Xu Tan, Jianguo Liu, and Xin Lu. 2021. Improving sentiment analysis accuracy with emoji embedding. Journal of Safety Science and Resilience, 2(4):246–252. find
  14. CardiffNLP. 2021. twitter-roberta-base-sentiment. https://huggingface.co/cardiffnlp/ twitter-roberta-base-sentiment. Accessed: July 22, 2025. ↗
  15. Yinxia Lou, Junxiang Zhou, Jun Zhou, Donghong Ji, and Qing Zhang. 2024. Emoji multimodal microblog sentiment analysis based on mutual attention mechanism. Scientific Reports, 14(1):29314. find
  16. V. D. Chaithra and 1 others. 2019. Hybrid approach: naive bayes and sentiment VADER for analyzing sentiment of mobile unboxing video comments. International Journal of Electrical and Computer Engineering, 9(5):4034–4041. find
  17. Scott M Lundberg and Su-In Lee. 2017. A Unified Approach to Interpreting Model Predictions. Advances in neural information processing systems, 30.
  18. Ahmed Mahrous, Jens Schneider, and Roberto Di Pietro. 2023. The Role of Emojis in Sentiment Analysis of Financial Microblogs. In 2023 Fourth International Conference on Intelligent Data Science Technologies and Applications (IDSTA), pages 76–84. IEEE. find
  19. Emil Chifu, Roxana Hatos, Monica Borda, and Alexandra Terec. 2015. Web harvesting and sentiment analysis of consumer feedback. Acta Technica Napocensis: Electronics and Telecommunications, 56(3):27–32. find
  20. Lorenzo Cima, Benedetta Tessa, Amaury Trujillo, Stefano Cresci, and Marco Avvenuti. 2025. Investigating the heterogeneous effects of a massive content moderation intervention via Difference-in-Differences. Online Social Networks and Media, 48:100320. find
  21. Hannah Miller, Jacob Thebault-Spieker, Shuo Chang, Isaac Johnson, Loren Terveen, and Brent Hecht. 2016. “Blissfully Happy” or “Ready toFight”: Varying Interpretations of Emoji. Proceedings of the International AAAI Conference on Web and Social Media, 10(1):259–268. find
  22. Bjarke Felbo, Alan Mislove, Anders Søgaard, Iyad Rahwan, and Sune Lehmann. 2017. Using millions of emoji occurrences to learn any-domain representations for detecting sentiment, emotion and sarcasm. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 1615…
  23. Esteban Morales, Jaigris Hodson, Anatoliy Gruzd, and Philip Mai. 2025. I [thumbs-up] your Hate: Emojis as Infrastructural Platform Violence on Telegram. In Proceedings of the Annual Hawaii International Conference on System Sciences, pages 2316–2325. University of Hawai’i Press. find
  24. Lynnette Hui Xian Ng and Jia Yuan Loke. 2020. Analyzing Public Opinion and Misinformation in a COVID-19 Telegram Group Chat. IEEE Internet Computing, 25(2):84–91. a Moderation Intervention. Engineering Applications of Artificial Intelligence, 162:112375. find
  25. Arsenii Tretiakov, Sergio D’Antonio-Maceiras, Áurea Anguera de Sojo Hernández, and Alejandro Martín. 2025. Toxic Discourse in the Digital Battlefield: Analysing Telegram Channels During the Russia– Ukraine ‘Conflict’. Expert Systems, 42(7):e70081. find
  26. John C Paolillo. 2023. The awkward semantics of Facebook reactions. First Monday. find
  27. F Plaza Del Arco, S Jiménez-Zafra, Marıa-Teresa Martın-Valdivia, and L Alfonso Urena-López. 2018. Using Facebook Reactions to Recognize Emotion in Political Domain. In Proceedings of CAEPIA. find
  28. Aleksandra Urman and Stefan Katz. 2022. What they do in the shadows: Examining the far-right networks on Telegram. Information, communication & society, 25(7):904–923. find
  29. Robert Plutchik. 1994. The psychology and biology of emotion. HarperCollins College Publishers. find
  30. Valhalla. 2020. distilbart-mnli-12-1. https://huggingface.co/valhalla/ distilbart-mnli-12-1. Accessed: July 22, 2025. ↗
  31. Chris Pool and Malvina Nissim. 2016. Distant supervision for emotion detection using Facebook reactions. In Proceedings of the Workshop on Computational Modeling of People’s Opinions, Personality, and Emotions in Social Media (PEOPLES), pages 30–39. find
  32. Soroush Vosoughi, Deb Roy, and Sinan Aral. 2018. The spread of true and false news online. science, 359(6380):1146–1151. find
  33. Dong Wang, Tao Xiao, Teyon Son, Raula Gaikovina Kula, Takashi Ishio, Yasutaka Kamei, and Kenichi Matsumoto. 2023. More Than React: Investigating The Role of Emoji Reaction in GitHub Pull Requests. Empirical Software Engineering, 28(5):123. find
  34. Fatima Zahra Qachfar and Rakesh Verma. 2023. ReDASPersuasion at SemEval-2023 Task 3: Persuasion Detection using Multilingual Transformers and Language Agnostic Features. In Proceedings of the 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 2124–2132. find
  35. Joe Watson, Sander van der Linden, Michael Watson, and David Stillwell. 2024. Negative online news articles are shared more to social media. Scientific Reports, 14(1):21592. find
  36. Bin Tareaf Raad, Berger Philipp, Hennig Patrick, and Meinel Christoph. 2018. find
  37. ASEDS: Towards Automatic Social Emotion Detection System Using Facebook Reactions. In 2018 IEEE 20th international conference on high performance computing and communications; IEEE 16th international conference on Smart City; IEEE 4th international conference on data science and systems (HPCC/SmartC… find
  38. Yuhang Zhou, Yimin Xiao, Wei Ai, and Ge Gao. 2025. The Hidden Language of Harm: Examining the Role of Emojis in Harmful Online Communication and Content Moderation. arXiv preprint arXiv:2506.00583.
  39. Xiaorui Zuo, Yao-Tsung Chen, and Wolfgang Karl Härdle. 2024. Emoji driven crypto assets market reactions. Management & Marketing, 19(2). find
  40. Tulika Saha, Aditya Prakash Patra, Sriparna Saha, and Pushpak Bhattacharyya. 2020. A Transformer based Approach for Identification of Tweet Acts. In 2020 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE. find
  41. Bhadresh Savani. 2021. bert-base-uncased-emotion. https://huggingface.co/bhadresh-savani/ bert-base-uncased-emotion. Accessed: July 22, 2025. ↗
  42. Jonas P Schöne, David Garcia, Brian Parkinson, and Amit Goldenberg. 2023. Negative expressions are shared more on Twitter for public figures than for ordinary users. PNAS nexus, 2(7):pgad219. find
  43. John R Searle. 1969. Speech acts: An essay in the philosophy of language. Cambridge university press. find
  44. Mohammed Shiha and Serkan Ayvaz. 2017. The Effects of Emoji in Sentiment Analysis. Int. J. Comput. Electr. Eng.(IJCEE.), 9(1):360–369. find
  45. Benedetta Tessa, Lorenzo Cima, Amaury Trujillo, Marco Avvenuti, and Stefano Cresci. 2025. Beyond Trial-and-Error: Predicting User Abandonment After find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/tardelli-2026-emoji-reactions-telegram-unreliable-indicators/