Majority in Favor of Action against Online Hate Speech, but Further-Reaching Consequences Are Controversial

How to deal with hate speech online? An international study shows strong public support for regulation, but what is perceived as hate speech often varies depending on the individual point of view.

Press release from 19 Februar 2025
Print version (PDF)

Although a large majority is in favor of measures against hate speech online under certain conditions, consequences that go beyond online sanctions find limited acceptance.This is the result of an international study by Mannheim political scientist Professor Dr. Richard Traunmüller (University of Mannheim) and his colleagues Professor Dr. Simon Munzert (Hertie School), Dr. Pablo Barberá (University of Southern California), Professor Dr. Andrew Guess (Princeton University), and Dr. JungHwan Yang (University of Illinois).

“Against the backdrop of stricter EU regulations such as the Digital Services Act and criminal prosecution in Germany, the question arises: What should be deleted or punished on the Internet? The answer is difficult, as the question of what is considered hate speech and how its destructive effect is to be assessed is answered very differently depending on cultural background and political convictions,” explains Richard Traunmüller, Professor of Empirical Democracy Research at the University of Mannheim and project manager at the Mannheim Center for European Social Research (MZES).

Respondents rated realistic social media posts
The research team conducted studies with over 2,500 respondents from Germany and the US. The study participants were each confronted with eight dialogs from social media and each exchange began with a message and a subsequent response that varied in intensity from harmless to strongly hateful. The respondents were then asked to find an appropriate response to the answer, whether through measures taken by the platforms (e.g. deleting the message) or through consequences for non-digital everyday life (e.g., paying a fine).

The results show: The stronger the perceived hate content, the more likely people are to be in favor of regulation. More overt forms of hate speech are rejected much more strongly than more subtle forms such as discriminatory language. In addition, a large proportion of respondents in both countries (more than 70 percent in Germany and more than 60 percent in the US) are in favor of restricting the freedom of expression when it comes to extreme insults or calls for violence.

However, a significant proportion of respondents reject consequences for non-digital everyday life, such as fines or job loss, even in the case of particularly extreme forms of hate speech. Specifically: One third of respondents in Germany decided against and two thirds decided in favor of far-reaching sanctions for extreme forms of hate speech. More than 95 percent of respondents, however, support a restriction at platform level in the case of extreme hate speech. In the US, the ratio is around 50–50, when it comes to far-reaching consequences for extreme hate speech. “Our study shows that the Internet platforms have a duty to act. But relatively large sections of the population do not support further-reaching consequences beyond the digital space,” says Richard Traunmüller.

What explains the resistance to restricting hate speech?
In the public debate, regulating hate speech is often seen as a violation to the freedom of expression. The study reveals an additional psychological mechanism: The perception of hate speech is influenced by one's own group affiliation. Specifically, the study results suggest that people are more tolerant of hate speech coming from their own ideological group and judge hate speech from another group more harshly.

“We call this in-group bias. It means that regulation is primarily met with resistance when it affects one's own group and is more likely to be supported when it affects the other side,” explains Professor Dr. Simon Munzert, head of the study at the Hertie School. Particularly with regard to automated or AI-assisted moderation and the question of appropriate government regulation of hate speech, there are significant differences in perceptions of what is problematic. “There is no social consensus that makes it possible to make generally accepted decisions,” continues Munzert.

The study with the title Citizen preferences for online hate speech regulation was recently published in the journal PNAS Nexus.

Further information and contact:
Prof. Dr. Richard Traunmüller
MZES project leader and Professor of Empirical Democracy Research
University of Mannheim
Phone: +49 621181 – 2288
E-mail: traunmuellermail-uni-mannheim.de
https://www.richardtraunmueller.com/

Professor Dr. Simon Munzert
MZES project leader and Professor of Data Science and Public Policy
Hertie School
Phone: +49 30 259 219-450
E-mail: munzertmail-hertie-school.org
https://simonmunzert.com/

Nikolaus Hollermeier
Public Relations
Mannheim Centre for European Social Research (MZES)
University of Mannheim
Phone: +49 621 181-2859
E-mail: dillhoefermail-mzes.uni-mannheim.de
www.mzes.uni-mannheim.de