AI & Society
Posting hateful content on X is linked to higher anger, anxiety, and third-person language, study finds
A study of 6,002 Twitter/X users found that people who posted hateful content expressed more anger, anxiety, and negative emotion than a matched group of users who did not post hateful content. Researchers from Toronto Metropolitan University (Canada) and the University of Guelph (Canada) used propensity score matching, a method that pairs similar users so the comparison is less affected by other differences. Hateful posts also used more third-person pronouns, words such as "they" and "them" that the authors interpret as a sign of social detachment. The journal Information Processing & Management published the paper on February 5, 2025.
What did the study compare?
The researchers compared users who posted hateful content with a matched group of non-hateful users. The matching method paired similar users before measuring differences in their emotional and linguistic expression.
What changed in emotion and language?
Users who posted hateful content showed significantly higher levels of anger, anxiety, and negative emotions. They also used third-person pronouns more often. Moral outrage and profanity peaked during hateful posts and declined over time, but they stayed above the levels seen in non-hateful posts.
What did topic modeling show?
Topic modeling, a computational way to find themes across many posts, showed that hateful posts were more connected to one another and covered more diverse topics. Hateful posts were also more similar to each other. Within individual posts, themes were less cohesive, meaning the ideas in one post hung together less tightly. Across posts, the hateful messages formed a more cohesive network.
FAQ
How many users did the researchers study?
The researchers studied 6,002 Twitter/X users.
Did posting hateful content cause these emotional and language changes?
The authors used causal inference methods, including propensity score matching, to compare similar users. The abstract reports these differences as causal findings, but it does not say that users were randomly assigned to post hateful content.
What does third-person pronoun use mean in this study?
The authors interpret increased third-person pronoun use as a sign of greater social detachment.
Limits and what's next
The abstract does not state limitations or planned next steps. The paper is peer-reviewed. The abstract does not say whether a preprint exists.
Source: Information Processing & Management, 2025-02-05