The writer is associate professor of behavioural science at Warwick Business School
To my mind, Denis Villeneuve’s epic sci-fi remake of Dune is a critique of the idea of a messiah as a form of societal control. After logging on to X, however, I saw that some out there do not share this view: a few far-right pundits were lauding the new film for upholding conservative gender roles and a “pro-life” stance. I wanted to object to this take, but what could I do?
Social media platforms such as X, Facebook and Instagram make it easy to express approval: I can share, like, or “heart” a post. But showing instant disapproval? There isn’t an option for that.
I could leave a negative comment, of course — but it is time-consuming both for me to write and the post’s author to read. The critical feedback drowns in the sea of noise.
What does this mean for the content itself? In a collaboration between Warwick Business School and Pompeu Fabra University, our team hypothesised that such imbalanced feedback led to more extreme opinions being expressed online.
First, our experiments verified several basic assumptions of user behaviour on social media: we found that people were more likely to post opinions that had received positive feedback in the past, and less likely to post those that were received negatively. Plus, they were more likely to provide positive feedback to those they agreed with, and less likely to do so for posts they did not support.
We then simulated a mathematical model that implemented these assumptions and showed that when users could “like” or “dislike” a post, they were less likely to express extreme opinions compared with when only positive feedback was available.
Our findings suggest that because the author of an extreme post can easily see only the number of likes but not the silent dislikes, they are more likely to post something similar again. In other words, they confuse the lack of balance for wholehearted endorsement.
This design feature has profound implications.
For one, it facilitates a toxic environment online. Authors of hateful messages feel validated because their opinions mainly receive support, while the targets of the hate find themselves isolated, given their allies don’t have any easy way of expressing solidarity. This is especially problematic for women and minority ethnic groups, who receive a disproportionate level of abuse.
This feature can also play a cruel joke on companies that rely on social media to taste test new products or concepts. The imbalance of online feedback might create a false perception of how the public will react to a marketing campaign.
Is the solution to introduce a dislike button? If so, why haven’t social media companies done so yet?
Well, it’s complicated. We assume that the social media platforms are not fans of this option as they are afraid it would discourage users and, therefore, affect advertising profits.
In addition, social media users themselves are not necessarily in favour of a dislike button. Participants in our studies were more likely to respond to posts with a like than a dislike, and said they left positive comments more often than negative ones. Out of 400 prolific participants, 42 per cent said that they wouldn’t want a dislike button, citing the possibility of it being used for harassment.
Feedback on social media is an under-researched topic. All I have discussed here is its asymmetry. There are still many questions left. Does this imbalance skew our worldview given that the reactions of others signal what society thinks about an issue? What form of dislike button would be the most beneficial? And, when the dislike count rises, would those with extreme opinions reconsider their views? These are the dilemmas we should be puzzling over, as societal polarisation grows.