Comment threads can influence climate change attitudes by altering perceived consensus

The comment sections of blogs can influence perceptions of consensus, which can in turn influence beliefs about climate change, according to new research published in the journal Memory & Cognition. The findings suggest that readers “may be nudged towards rejection of climate science if they encounter a stream consisting of contrarian comments.”

“Social media has taken an increasingly prominent role in public discourse, and there is now much concern that it has been exploited by bad-faith actors in support of political operations. We know that ‘trolls’ and ‘bots’ — that is accounts that are in some way fake — have been very active during recent elections and may have affected the outcome of the Brexit referendum and the election of Donald Trump in 2016,” said study author Stephan Lewandowsky (@STWorg), a professor of cognitive science at the University of Bristol.

“I am therefore very interested in how social media can affect people’s attitudes, in particular concerning scientific issues (e.g., here is another article in which I expand on that: https://jspp.psychopen.eu/index.php/jspp/article/view/443).”

“One way in which social media can affect attitudes is by creating the illusion of widespread support for an opinion: we know from much previous research that people are often swayed by what the majority of others think,” Lewandowsky told PsyPost.

“This is a very rational strategy in most circumstances: for example, if most people line up to wait for a bus around the corner, then it’s advisable to join them rather than to stake out one’s own claim somewhere else.”

“However, the strategy is subject to abuse if the majority view is a mirage constructed by trolls and bots. To examine this possibility, we conducted a study in which the comment stream of a blog post was manipulated to create the perception of a majority consensus either in favour of the scientific consensus on climate change or against it,” Lewandowsky explained.

“We wanted to see how the comments would affect people’s views of the post, which in turn was also either endorsing the consensus position or opposed it.”

The researchers started by creating two blog posts: one endorsing the scientific view that humans are causing global warming and another rejecting that view. They then created two threads of ten comments either wholly supporting or wholly refuting man-made climate change.

Nearly 400 participants were then randomly assigned to read one of the blog posts, which included one of the two comment streams at the end.

Lewandowsky and his colleagues found that participants tended to be more supportive of the argument in the blog post when the comments were aligned with it.

In other words, supportive comments boosted perceived consensus and made participants more likely to say they supported the basic argument made in the blog post, while critical comments had the opposite effects — regardless of the content of the post.

The comment stream only appeared to directly influence climate change beliefs among careful readers — who spent between 100s and 1800s processing the thread.

But the comments were found to indirectly influence climate change beliefs by altering the participants’ view of the amount of consensus.

“The take-home message is that people were affected by what they perceived to be the majority view among commenters in the stream. That is, the presumed consensus among other readers was a mediating variable to affected people’s attitudes, as we expected,” Lewandowsky told PsyPost.

It is still unclear, however, how much consensus is necessary to sway opinions.

“In our study, the comment streams were unanimous. We chose to do this because we wanted to start out with the strongest possible manipulation — and unanimity is the strongest version of consensus,” Lewandowsky explained.

“The remaining question is what would happen if there was more diversity of opinion in the comment stream. How many dissenting voices would it take to disrupt the perceived social consensus? It would be intriguing to know this because it might tell us how easy or difficult it is to ‘disrupt’ a consensus—is a single troll among 10 serious commenters sufficient? Or do the trolls have to outnumber the serious commenters?”

The study, “Science by social media: Attitudes towards climate change are mediated by perceived social consensus“, was authored by Stephan Lewandowsky, John Cook, Nicolas Fay, and Gilles E. Gignac.