This page describes the team and the research project behind our participation in SemEval 2023 Task 3, “Detecting the Category, the Framing, and the Persuasion Techniques in Online News in a Multilingual Setup”.
Team
- Boris Gorelik, data scientist. Senior Lecturer, Azrieli College of Engineering Jerusalem (JCE). Contact: borisge@jce.ac.il
- Shirley Druker Shitrit, communication studies. Department of Communication and Research Authority, Sapir Academic College.
- Ella Ben Atar, communication studies. Head of the Department of Communication, Sapir Academic College.
Topic
We study how news framing varies across languages and outlets. Our current project is a comparative framing analysis of press coverage of the September 2024 Lebanon pager operation, in Hebrew, Arabic and English, following Entman’s framing framework and critical discourse analysis.
We are building a coding instrument for article length news text that measures actor salience, source attribution, quotation type, entity linked descriptor valence, positional salience and macro frame prevalence. A second strand applies the same instrument to news text generated by large language models prompted about the same event in different languages, so that model output can be compared against the real press of each language rather than only against other models.
The SemEval 2023 Task 3 corpus interests us as a validated multilingual anchor for the framing layer of that instrument. It gives us gold framing labels across nine languages against which our own coding can be calibrated and its cross language behaviour checked.