# The Virality of Hate Speech on Social Media

**Authors:** Abdurahman Maarouf, Nicolas Pröllochs, Stefan Feuerriegel  
**arXiv:** [2210.13770](https://arxiv.org/abs/2210.13770)  
**Date:** 2022-10-25  
**Summary page:** https://www.summarizepaper.com/en/paper/2210.13770v1/

## Key points

- Hate speech on social media poses a significant threat to vulnerable groups and society as a whole
- Examples of real-world consequences include the Pittsburgh synagogue shooting in 2018
- The authors collected a dataset of Twitter cascades and classified them as hateful or normal content
- Hateful content authored by verified users is more likely to go viral compared to non-verified users
- Verified users' hateful content had larger cascade size, longer cascade lifetime, and greater structural virality compared to normal content
- This study provides insights into the factors contributing to the spread of hate speech on social media platforms

## Summary

In their paper titled "The Virality of Hate Speech on Social Media," authors Abdurahman Maarouf, Nicolas Pröllochs, and Stefan Feuerriegel explore the phenomenon of hate speech going viral on social media platforms. They highlight that online hate speech poses a significant threat to vulnerable groups and society as a whole, citing examples such as the Pittsburgh synagogue shooting in 2018. To address the limited understanding of what factors contribute to the virality of hate speech, the authors collected a dataset consisting of 25,219 Twitter cascades with 65,946 retweets and classified them as either hateful or normal content. They then employed generalized linear regression techniques to estimate differences in the spread of hateful versus normal content based on various author and content variables. Through their analysis, the authors identified important determinants that explain variations in the spreading of hateful versus normal content. Notably, they found that hateful content authored by verified users is disproportionately more likely to go viral compared to non-verified users. Specifically, hateful content from verified users exhibited a cascade size 3.5 times larger than normal content, a cascade lifetime 3.2 times longer than normal content, and a structural virality 1.2 times larger than normal content. Overall, this study provides novel insights into the virality of hate speech on social media platforms by shedding light on the factors that contribute to its spread. It also contributes to our understanding of how online hate speech can have real-world consequences for vulnerable groups and society at large.