Validation of Self-Oriented Model of Digital Publics (SOMP)

This repository contains data, code, and analytical procedures for the following paper: Lee, H., Seo, S., Park, S., Tam, L., Kim, S., & Jang, Y. (2026). Theory-driven LLM-assisted analysis in crisis communication: A case study of the self-oriented model of digital publics. Public Relations Review, 52(3), 102709.

Authors

• Dr. Hyelim Lee
• Seongbum Seo
• Somin Park
• Dr. Lisa Tam
• Dr. Soojin Kim
• Dr. Yun Jang

Repository Structure

/data

Survey response data and crisis-related tweet datasets. Contains anonymized worker responses, filtered tweet collections, and language detection results.

/zero_shot

Zero-shot classification experiments using LLMs to evaluate I-, You-, and We-involvement in crisis communication without training examples.

/few_shot

Few-shot classification experiments (k=1, 4, 8, 16) testing LLM performance with varying numbers of demonstration examples.

/finetuned

Fine-tuning experiments with LoRA adapters for tweet evaluation. Includes training data, model checkpoints, and evaluation scripts.

/survey

Survey analysis scripts and outputs, including statistical tests and visualization generation.

/scripts

Data processing and analysis utilities for deduplication, aggregation methods comparison, and accuracy measurement.

/pre-survey

Preliminary survey materials and pilot study data.

Data Anonymization

All personally identifiable information has been anonymized:

• Worker IDs replaced with ANONYMOUS### format
• Email addresses replaced with email###@anonymous.com format

Requirements

See requirements.txt for Python dependencies.

Citation

@article{lee2026theory,
  title = {Theory-driven LLM-assisted analysis in crisis communication: A case study of the self-oriented model of digital publics},
  journal = {Public Relations Review},
  volume = {52},
  number = {3},
  pages = {102709},
  year = {2026},
  issn = {0363-8111},
  doi = {https://doi.org/10.1016/j.pubrev.2026.102709},
  url = {https://www.sciencedirect.com/science/article/pii/S0363811126000445},
  author = {Hyelim Lee and Seongbum Seo and Somin Park and Lisa Tam and Soojin Kim and Yun Jang},
  keywords = {Large Language Models, Human–AI Integration, Public Relations Methods, Crisis Communication, Self-Oriented Model of Digital Publics}
}

License

MIT License - see LICENSE file for details.