Thread by @AnaSKozyreva

How to resolve content moderation dilemmas between free speech and harmful misinformation? Thrilled to share our new article just out in PNAS
With @JasonReifler @stefanmherzog @lorenz_spreen @STWorg @MLeiser and Ralph Hertwig @arc_mpib @mpib_berlin 1/11
pnas.org/doi/10.1073/pn…

Content moderation of online speech is a moral minefield, especially when 2 key values come into conflict: upholding freedom of expression and preventing harm caused by misinformation. In our study, we examined how the U.S. public would approach such difficult trade-offs. 2/11

In a conjoint survey experiment, U.S. respondents indicated whether they would remove problematic social media posts on 4 misinformation topics and whether they would take punitive action against the accounts. 3/11

Our hypothetical scenarios included misinformation topics where active polices have already been implemented by social media platforms: politics (“election denial”), health (“anti-vaccination”), history (“Holocaust denial”), and the environment (“climate change denial”). 4/11

The conjoint design allowed us to systematically vary factors that could influence moral judgments. These factors included: characteristics of the account and of the shared content; whether this was a repeated offense; and the consequences of sharing the misinformation. 5/11

The majority of respondents chose to take some action to prevent the spread of misinformation. On average, 66% said they would remove the offending posts, and 78% would take some action against the account (of which 33% opted to “issue a warning”). 6/11

Content moderation of harmful misinformation was a partisan issue: Across all four scenarios, Republicans were consistently less willing than Democrats or independents to remove posts or penalize the accounts that posted them. 7/11

What were key factors that affected people’s decisions to quash harmful misinformation? The topic of misinformation, the severity of the consequences, and whether it was a repeated offense had the strongest impact on decisions to remove posts and suspend accounts. 8/11

Characteristics of the account itself—the person behind the account, their partisanship, and number of followers—had little to no effect on respondents’ decisions. Respondents did not penalize political out-group accounts more than in-group accounts. 9/11

Our results also show that, despite partisan differences, there is ground for agreement in approaching these difficult dilemmas. We hope that our study can contribute to the process of establishing transparent and consistent rules for content moderation. 10/11

The paper is published with #openaccess. Many thanks to my wonderful co-authors, insightful reviewers and the editor, our scientific editor Deb Ain, RA Spela Vrtovec, and our PR team🙏. For my earlier thread on this study see here 11/11 End🧵

https://twitter.com/AnaSKozyreva/status/1538460120170078209