Using LLM to Trace Public Sentiment and Policy Change Toward Canadian International Students

• In Event: AERA Roundtable Session Wednesday 1:45 pm - Gold 1
  In Roundtable Session: Virtual Communication in Higher Education

Abstract

Recent Canadian higher education policies, such as the federal international student cap, reflect a growing nationalist-populist orientation. To understand how such policy directions emerge and gain traction, our study moves upstream from policy content to public discourse. We conduct a thematic analysis of social media posts (on Twitter/X and YouTube) using a Large Language Model (LLM), comparing pre-pandemic (2015-2020) and post-pandemic (2020-2025) discourse. Our findings reveal a post-pandemic surge in negative sentiment, increasingly framing international students as competitors for housing, jobs and public resources. This shift suggests that emerging public sentiment not only reflects individual dissatisfaction, but may also contribute towards driving restrictive policy measures.

Authors

• Ya Xuan Wang, York University

• Qiang Zha, York University