A harmonious oasis of nuanced debate and reason… is sadly not an accurate description of most social media platforms. Let’s face it, things can quickly get very nasty in those comment sections – but why is it like that? A new study explores what the authors call the “confrontation effect”, in which users become highly engaged with content that challenges their political ideology. In other words, it makes them mad. 

We often speak about echo chambers on social media, and there is evidence to suggest that people seek out information that’s consistent with their own beliefs. That’s true in general, whether it be gravitating towards particular newspapers, news networks, or YouTube channels that align with a prior worldview. 

“However,” write the authors of the new paper, “a mere glance at social media posts on highly charged topics (e.g., immigration, gun control, climate change) shows a more complex pattern that challenges such traditional notions of congeniality bias.”

After all, if everyone was only reading content they agreed with, the arguments in the comments sections wouldn’t get so heated. Rarely is this more clearly visible than during an election period, when feeds are flooded with partisan messaging and commentary on hot-button issues. 

In a series of field experiments, the study team showed political Facebook posts to more than 500,000 Americans in two camps: those for and against then-President and current candidate Donald Trump. The content covered the topics of gun control, Obamacare, and Trump himself, and they took a similar approach to previous studies looking at targeted Facebook ads. 

The confrontation effect was observed across all three experiments. Engagement was higher on posts that challenged the user’s own ideology; for example, liberals were more likely to interact with posts praising Trump and conservatives were more likely to interact with pro-gun control content. 

“The research helps explain the large amount of toxic discourse we observe online. Our results reveal that individuals are strongly driven to voice their outrage toward those with whom they disagree,” said Dr Daniel Mochon of Tulane University, who co-authored the work with Dr Janet Schwartz of Duke University, in a statement. “While previous studies show that people avoid content inconsistent with their beliefs, we found that counter-ideological content actually drives higher engagement.”

The findings from the field tests were then replicated in a lab environment, using pro- and anti-vegetarianism messaging rather than political content. The team then delved deeper, looking at some of the factors that could be driving the confrontation effect. They found that the more important a topic was deemed, the more likely users were to be outraged enough to comment. The effect could also be heightened if the information was presented in a more ideology-threatening manner. 

Essentially, content that appears to seriously threaten someone’s personal core values is more likely to elicit sufficient outrage for them to click or comment. Importantly, none of this poses a particular issue for the social media sites themselves: “Platforms benefit from keeping users active, regardless of whether the interaction is positive or negative,” said Mochon.

One of the limitations of the study that the authors flag is the complexity of the real-world dynamics that could impact the findings. For example, people may deliberately share content to their own network that they know will elicit a confrontation effect – the concept is something that content creators are already getting good at exploiting for so-called “rage bait”, which is literally geared towards irritating people. 

Overall, the study underscores the fact that greater engagement with content does not necessarily equal greater agreement and shows that a lot of the reaction to divisive political content comes from dissenters, not supporters. 

Mochon said, “We hope our findings provide a more balanced perspective on the interplay between ideology and online engagement.”

The study is published in the journal Organizational Behavior and Human Decision Processes.