Online comments can shape how social media content about politics is perceived, even when people's opinions are hard to change, a new study shows.
New research suggests that while attitudes may be stable, the way people interpret political messages can still shift depending on the surrounding conversation online. Engaging with problematic content can make a difference.
The survey shows how counter arguments and comments can shape whether people recognise sexist or misleading communication.
Their findings highlight that political communication on social media is rarely understood or interpreted in isolation. Instead, replies, reactions and comments provide signals about how messages can be interpreted.
The findings come from two large research projects involving more than 11,000 respondents across five European countries. Across both studies, researchers found that comments appearing underneath social media posts can influence how audiences interpret the original message.
"In polarized debates, persuasion may be difficult," says Professor Florian Stoeckel, from the University of Exeter, who led the projects. "But how people interpret a specific message remains more flexible than we often assume."
The research examined two different types of contested online communication: social media posts challenging gender equality and false information.
Participants in survey experiments were shown social media posts together with different types of comments appearing underneath them. This allowed researchers to examine how user responses influence the interpretation of the original message. The posts used in the experiments were based on real-world examples of sexist communication and misinformation circulating online.
One set of experiments examined sexist social media posts that challenge gender equality and the role of user comments responding to them. Recognising sexist communication is not always straightforward. Messages can be framed ambiguously or embedded in broader political arguments. As a result, comments that explain why and how a specific post is sexist can help audience recognize sexism.
The study also examined whether certain types of counter-arguments are more effective than others, for example responses appealing to fairness or empathy or using humour. The results did not identify a single strategy that consistently outperformed others. Effects varied across topics and countries, suggesting that there is no guaranteed formula for effective responses. It seems more important that an attempt is made at all at calling out sexism or correcting misinformation.
"People sometimes assume that sexism or misinformation might be obvious," Professor Stoeckel said. "But in practice, this isn't necessarily the case. Comments that highlight problematic aspects of a message can therefore influence how others understand it."
The researchers also explored the potential role of generative artificial intelligence in this process. Tools such as ChatGPT were used to draft potential responses to sexist posts. These comments proved capable of highlighting problematic aspects of the messages and helping users recognise sexist communication.
Similar dynamics appeared in the misinformation research. When participants encountered corrective comments responding to misleading claims, they were more likely to distinguish between factual and inaccurate information.
"Our findings show that the surrounding conversation matters," said Professor Stoeckel. "Even in a polarized environment where people hold strong views, the context in which a message appears can influence how it is understood."
