Analyzing the Association between Latinx Mental Health and Anti-Latinx Online Hate Speech

• In Event: 40861 - Section on Sociology of Population Roundtables
  In Refereed Roundtable (60 minute): Table 3: Racial and Ethnic Inequality and Health

Abstract

This study examines the relationship between online racial discrimination (ORD) against the Latinx population and mental health outcomes. Using a novel approach that measures the volatility in hate speech rather than raw scores, we analyze Twitter data from 2014-2019 containing Latinx-related terms alongside mental health data from the Behavioral Risk Factor Surveillance System (BRFSS). We employ the Measuring Hate Speech Model, a neural network classifier, to quantify hate speech on a continuous scale, and introduce "Hate Speech Volatility"—the standard deviation of hate speech scores over three-month periods—to account for changing baseline thresholds of online hate. Our time series regression analysis reveals that increased volatility in anti-Latinx hate speech significantly correlates with worsened mental health among internet-using Latinx individuals, who reported an additional 10.27 days of poor mental health. Importantly, this relationship was not observed among non-Hispanic White internet users or Latinx individuals who do not use the internet, suggesting a unique vulnerability among Latinx internet users to online hostility targeting their group. These findings contribute to the literature by: (1) extending research on ORD effects beyond youth to the broader Latinx adult population, (2) utilizing an advanced neural network classifier to measure hate speech on a nuanced scale, and (3) introducing volatility as a key metric for understanding the impact of online hate speech on mental health. The results underscore how digital environments can contribute to racial health disparities and highlight the tangible psychological harms of online discrimination against marginalized communities, particularly during periods of heightened anti-Latinx political rhetoric.

Authors

• Angelita Repetto, University of California-Davis

• Pratik S. Sachdeva, University of California, Berkeley