Categorising Anti-Muslim Discourse on Twitter: Hate Speech and Toxicity in the Spanish Context

Tuesday, 8 July 2025: 09:30
Location: FSE036 (Faculty of Education Sciences (FSE))
Oral Presentation
William GONZÁLEZ BAQUERO, Universidad de Salamanca, Spain
Javier JIMENEZ AMORES, Universidad de Salamanca, Spain
Carlos ARCILA, University of Salamanca, Spain
In 2023, anti-Muslim discourse on social media has increased to 25% from the 20% recorded in 2022, making it the third most prevalent form of hate speech in Spain according to OBERAXE. This underscores the imperative for an extensive examination of online Islamophobic discourse. The study focuses on messages related to Islam and the Spanish-speaking Muslim community on Twitter in Spain from January 1, 2010, to December 31, 2022, utilizing a substantial sample of 21,326 tweets obtained through Twitter's (now X) API v.2. Geographical coordinates were collected to pinpoint regions where Islamophobic messages are most concentrated. Furthermore, an analysis of bigoted anti-Muslim language in tweets was conducted using a codebook based on established instruments. The analysis also incorporated the Perspective API to evaluate toxicity levels in Islamophobic and anti-Muslim discourse. The results revealed that 28.2% of messages were identified as Islamophobic, with 2,475 displaying explicit and irrational hatred. Additionally, rhetoric associating Muslims with invasion, terrorism, and crime was prevalent. The Perspective API demonstrated limited effectiveness in detecting Islamophobic hate speech, highlighting the necessity for a specialized tool tailored to this purpose.