README.md

About

This repository contains materials for the NAACL 2021 paper Modeling Framing in Immigration Discourse on Social Media.

• dataset.zip contains the full set of tweet IDs used for analysis. Human-annotated data for training frame detection models is located in the annotated_data folder, and machine-predicted frame labels are located in the predicted_data folder.

• codebook.pdf contains guidelines for frame annotation. It includes detailed descriptions of issue-generic policy, immigration-specific, and episodic/thematic frames.

• code/ contains all code for data collection, assessing annotations, and building and evaluating models

• notebooks/ contain Jupyter notebooks for framing analyses, including regressions and plots

Frame Detection Models

Multilabel RoBERTa classification models for identifying frames, fine-tuned on our full set of immigration-related tweets, can be found here:

• https://huggingface.co/juliamendelsohn/framing_issue_generic
• https://huggingface.co/juliamendelsohn/framing_immigration_specific
• https://huggingface.co/juliamendelsohn/framing_narrative

Please see this Colab notebook for how to use the frame classification models.

Citation

  @inproceedings{mendelsohn2021modeling,
  title={Modeling Framing in Immigration Discourse on Social Media},
  author={Mendelsohn, Julia and Budak, Ceren and Jurgens, David},
  booktitle={Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
  pages={2219--2263},
  year={2021}
}

Contact

Please email Julia Mendelsohn (juliame@umich.edu) with any issues, questions, or comments.