Previous research finds that public sentiment on immigration is influenced to a large extent not by changes in immigrant stocks and flows, but by specific highly visible events (e.g., a terrorist attack) and media coverage. However, public opinion data is available only at infrequent intervals, making it difficult to establish a clearer relationship between specific news events or changes in the news environments and migration sentiment. In this paper, we explore the utility of natural language processing in analyzing the U.S. migration discourse in the first two years of the Trump presidency, connecting national news coverage to public sentiment through the use of Twitter data. We suggest that this connection, and the polarized sentiment it engenders, could have significant implications for migration policy in the near future.
Presented in Session 189. Flash Session: New and Pressing Immigration Issues