--- title: "How Much Does Persuasion Strategy Matter? LLM-Annotated" description: "This work asks whether specific persuasion tactics relate to whether people donate in charity-solicitation chats. 7 references." authors: ["Tatiana Petrova", "Stanislav Sokol", "Radu State"] year: 2026 venue: "Proceedings of the 1st Workshop on Social Context (SoCon) and the 2nd Workshop on Integrating NLP and Psychology to Study Social Interactions (NLPSI) @ LREC 2026" source: "https://aclanthology.org/2026.socon-1.5/" doi: "10.63317/4hquav3a9b4n" licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)" licence_url: "https://creativecommons.org/licenses/by-nc/4.0/" summary_licence: "Elamur Cosmic License" canonical: "https://research.elamur.ai/free/en/papers/petrova-2026-how-much-does-persuasion-strategy/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/petrova-2026-how-much-does-persuasion-strategy/) [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/36/) / How Much Does Persuasion Strategy Matter? LLM-Annotated Evidence from Charitable Donation Dialogues # How Much Does Persuasion Strategy Matter? LLM-Annotated Evidence from Charitable Donation Dialogues **Authors:** Tatiana Petrova, Stanislav Sokol and [Radu State](https://md.research.elamur.ai/free/en/authors/radu-state/) · **Year:** 2026 · **Venue:** [Proceedings of the 1st Workshop on Social Context (SoCon) and the 2nd Workshop on Integrating NLP and Psychology to Study Social Interactions (NLPSI) @ LREC 2026](https://md.research.elamur.ai/free/en/venues/socon/) ([Workshop on Social Context (SoCon) and the Workshop on Integrating NLP and Psychology to Study Social Interactions (NLPSI) @ LREC 2026](https://md.research.elamur.ai/free/en/venues/socon/2026/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/) **Links:** [Publisher's page](https://aclanthology.org/2026.socon-1.5/) · [PDF](https://aclanthology.org/2026.socon-1.5.pdf) · [DOI 10.63317/4hquav3a9b4n](https://doi.org/10.63317/4hquav3a9b4n) · [OpenAlex](https://openalex.org/W7166572483) **Citations:** 0 citations (OpenAlex, 2026-10-14) ## Summary This work asks whether specific persuasion tactics relate to whether people donate in charity-solicitation chats. Earlier analysis of the PersuasionForGood corpus covered only 252 dialogues, used a 10-label scheme, and did not correct for multiple testing. The authors built a 41-strategy, 11-category taxonomy based on Cialdini and on Marwell and Schmitt, plus 9 conversation-management labels. They labeled every persuader turn in the corpus with three locally run open-weight LLMs, using Qwen3:30b as the primary annotator and Mistral-Small-3.2 and Phi-4 as replication checks. Target turns received sentiment and interest labels. They then ran chi-square tests with Bonferroni and FDR corrections and logistic regression. Strategy categories explain very little of the donation outcome, and no category survives correction. Guilt Induction stands out as linked to fewer donations, and Reciprocity is the steadiest positive signal. Both findings hold across all three models even though the models agree only moderately with each other. The results suggest that simply labeling strategies says little about what makes persuasion work, and that guilt appeals may backfire. The annotated corpus is publicly released. ## Key points - **Corpus coverage:** All 10,600 persuader turns across 1,017 PersuasionForGood dialogues were annotated, plus 10,332 target turns labeled for sentiment and interest with Qwen3:30b. The overall donation rate is 53.6%. - **Weak category effects:** Categories yield pseudo R² ≈ 0.015 for all three annotators, and none of the 11 categories is significant after Bonferroni correction. - **Individual strategies:** Guilt Induction is associated with roughly 23 percentage points lower donation rates. Of all strategies, Reciprocity shows the most consistent positive association with donating. Target sentiment and interest predict whether someone donates but barely relate to how much. - **Annotation quality:** Against Wang et al.'s gold labels, mapped to macro-categories over 300 dialogues, agreement is κ = 0.507 with macro F1 = 0.703. An expert check confirmed 84 of 100 sampled turns. Pairwise inter-model κ at the strategy level ranges from 0.38 to 0.54. - **Release and setup:** Data and code are available at github.com/persuasion-nlp/persuasion-strategies. Annotation used Ollama with a hierarchical prompt that picks the category first and then the strategy, at temperature 0.1. *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 - [University of Luxembourg](https://md.research.elamur.ai/free/en/universities/university-of-luxembourg/) — University, Luxembourg, Luxembourg · 28 works · 337 citations - [Interdisciplinary Centre for Security, Reliability and Trust](https://md.research.elamur.ai/free/en/universities/interdisciplinary-centre-for-security-reliability-and-trust/) — University, Belval, Luxembourg · 3 works · 3 citations 📚 Shelves (6): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 1,954 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 626 works 🧭 Topics (2): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,894 works · [🗨️ Dialogue systems](https://md.research.elamur.ai/free/en/topics/dialogue-systems/) · 654 works ## Cite Cite as: Tatiana Petrova, Stanislav Sokol and Radu State. 2026. How Much Does Persuasion Strategy Matter? LLM-Annotated Evidence from Charitable Donation Dialogues. Proceedings of the 1st Workshop on Social Context (SoCon) and the 2nd Workshop on Integrating NLP and Psychology to Study Social Interactions (NLPSI) @ LREC 2026. The paper itself, not this page. BibTeX from the library's record of the work. ```bibtex @inproceedings{petrova2026how, title = {{How Much Does Persuasion Strategy Matter? LLM-Annotated Evidence from Charitable Donation Dialogues}}, author = {Tatiana Petrova, Stanislav Sokol and Radu State}, booktitle = {Proceedings of the 1st Workshop on Social Context (SoCon) and the 2nd Workshop on Integrating NLP and Psychology to Study Social Interactions (NLPSI) @ LREC 2026}, year = {2026}, doi = {10.63317/4hquav3a9b4n}, url = {https://doi.org/10.63317/4hquav3a9b4n} } ``` - [← previous How Much Data for Stable Formant Values? Pipeline for Convergence Detection Based on Read Speech](https://md.research.elamur.ai/free/en/papers/sward-2026-how-much-data-stable-formant/) - [next → How Much Would a Clinician Edit This Draft? Evaluating LLM Alignment for Patient Message Response Drafting](https://md.research.elamur.ai/free/en/papers/seegmiller-2026-how-much-would-clinician-edit/) - [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/36/) ## References 7 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet. 1. Xuewei Wang, Weiyan Shi, Richard Kim, Yoojung Oh, Sijia Yang, Jingwen Zhang, and Zhou Yu. 2019. [Persuasion for good: Towards a personalized persuasive dialogue system for social good](https://md.research.elamur.ai/free/en/papers/wang-2019-persuasion-good-personalized-persuasive-dialogue/). In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 5635–5649. Associa… 2. Suhaib Abdurahman, Alireza Salkhordeh Ziabari, Alexander K. Moore, Daniel M. Bartels, and Morteza Dehghani. 2025. A primer for evaluating large language models in social-science research. Advances in Methods and Practices in Psychological Science, 8(2). [find](https://scholar.google.com/scholar?q=Suhaib+Abdurahman%2C+Alireza+Salkhordeh+Ziabari%2C+Alexander+K%2E+Moore%2C+Daniel+M%2E+Bartels%2C+and+Morteza+Dehghani%2E+2025%2E+A+primer+for+evaluating+large+language+models+in+social%2Dscience+research%2E+Advances+in+) 3. Jack W. Brehm. 1966. A Theory of Psychological Reactance. Academic Press. 9. [find](https://scholar.google.com/scholar?q=Jack+W%2E+Brehm%2E+1966%2E+A+Theory+of+Psychological+Reactance%2E+Academic+Press%2E+9%2E) 4. Natalie A. Carlson and Vanessa Burbano. 2025. The use of LLMs to annotate data in management research: Foundational guidelines and warnings. Strategic Management Journal. Early view. [find](https://scholar.google.com/scholar?q=Natalie+A%2E+Carlson+and+Vanessa+Burbano%2E+2025%2E+The+use+of+LLMs+to+annotate+data+in+management+research%3A+Foundational+guidelines+and+warnings%2E+Strategic+Management+Journal%2E+Early+view%2E) 5. Language Resource References Wang, Xuewei and Shi, Weiyan and Kim, Richard and Oh, Yoojung and Yang, Sijia and Zhang, Jingwen and Yu, Zhou. 2019. PersuasionFor-Good Corpus: 1,017 Persuasion Dialogues for Charitable Donations. distributed via GitHub. PID [https://gitlab.com/ucdavisnlp/persuasionforgoo…](https://gitlab.com/ucdavisnlp/persuasionforgoo…) [↗](https://gitlab.com/ucdavisnlp/persuasionforgoo…) 6. Wanqin Chen and Diyi Yang. 2021. Weakly supervised persuasion strategy identification in online persuasion. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 12704–12711. [find](https://scholar.google.com/scholar?q=Wanqin+Chen+and+Diyi+Yang%2E+2021%2E+Weakly+supervised+persuasion+strategy+identification+in+online+persuasion%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+volume+35%2C+pages+12704%E2%80%93127) 7. Category Strategy n % Norms / Morality / Values Appeal to Values Moral Appeal Guilt Induction Self-feeling Appeal 543 521 129 138 5.1 4.9 1.2 1.3 Rational / Impact Appeal Rational Appeal Logical Appeal 928 45 8.8 0.4 Framing & Presentation Framing Loss Aversion Appeal Bait-and-switch Pretexting 751… [find](https://scholar.google.com/scholar?q=Category+Strategy+n+%25+Norms+%2F+Morality+%2F+Values+Appeal+to+Values+Moral+Appeal+Guilt+Induction+Self%2Dfeeling+Appeal+543+521+129+138+5%2E1+4%2E9+1%2E2+1%2E3+Rational+%2F+Impact+Appeal+Rational+Appeal+Logical+Appea) *Updated 1 November 2026*