Social media users are more likely to engage with posts that provoke rather than affirm their political beliefs, according to new research published in Organizational Behavior and Human Decision Processes. This “confrontation effect” is driven by outrage, pushing people to respond to opposing viewpoints. The findings shed light on the emotional dynamics that drive online engagement.
The researchers — Daniel Mochon, the Edward H. Austin Jr. Professor of Business Administration and an associate professor of marketing at Tulane University’s A. B. Freeman School of Business, and Janet Schwartz, the executive director of Duke University’s Center for Advanced Hindsight — were motivated by a puzzling contradiction in existing knowledge about how people process information. Prior studies have shown that individuals tend to avoid content that challenges their beliefs, a behavior known as “confirmation bias.”
However, social media platforms frequently showcase highly polarized and combative interactions between individuals with opposing views. This raises questions about why people, who are generally expected to avoid conflicting information, engage more frequently with it online. Mochon and Schwartz sought to explain this paradox, hypothesizing that the outrage provoked by opposing views might be a key factor driving this unexpected engagement.
“We wanted to better understand how people interact with ideologically charged content online,” Mochon told PsyPost. “We commonly observed online settings where negative sentiments posted by users with ideology-inconsistent views dominated the threads, and we wanted to reconcile this pattern with the well-established finding that people tend to avoid information inconsistent with their beliefs.”
To investigate this, the researchers conducted a series of studies using a combination of real-world data from social media platforms and controlled online experiments.
In three initial field studies, the researchers aimed to test whether people engage more with content that opposes their political views (ideology-inconsistent) rather than content that aligns with their beliefs (ideology-consistent). They used Facebook’s advertising platform to target U.S. users with different political views (liberal and conservative).
The researchers created political posts related to three topics: gun control (Study 1A), Obamacare (Study 1B), and President Trump (Study 1C). Each post was designed to support either a liberal or conservative stance. For example, one post might advocate for gun control, while another opposes it. These posts appeared in users’ newsfeeds as paid advertisements.
The results showed that people were more likely to click on and comment on posts that opposed their political views than those that supported them. For example, liberals were more likely to engage with posts supporting President Trump than those criticizing him. This pattern, observed across all three studies, provided evidence for the confrontation effect.
“We were surprised by the size of the confrontation effect in the field studies,” Mochon said. “We found situations where users were four times more likely to engage with content they disagreed with than content they agreed with.”
Next, Mochon and Schwartz sought to replicate the findings of the Facebook field studies in a more controlled setting, allowing for a clearer understanding of the confrontation effect. The researchers also extended their investigation beyond politics to another emotionally charged issue: vegetarianism.
The study recruited 1,001 participants from Amazon Mechanical Turk (MTurk), an online platform commonly used for academic research. Participants were categorized based on their dietary practices (vegetarian/vegan or non-vegetarian) and were shown two tweets: one supporting vegetarianism and another opposing it. They were then given the option to comment on each tweet, and the researchers recorded whether participants chose to engage.
The findings mirrored those of the Facebook studies, with participants more likely to comment on tweets that conflicted with their dietary practices. For instance, non-vegetarians were more likely to respond to pro-vegetarian tweets, often expressing negative emotions such as anger or disgust.
The researchers then explored whether the type of engagement response moderated the confrontation effect. They hypothesized that when people are asked to comment on ideology-inconsistent content, they are more likely to engage than when asked if they want to follow or see more content from the same source.
Participants (802 individuals recruited from the research platform Prolific) were assigned to one of two conditions: they were either asked to comment on an ideology-inconsistent tweet or decide whether they wanted to see more content from the same source. The tweets addressed a highly charged topic: COVID-19 vaccines. Participants’ emotional reactions were measured, and the researchers recorded their engagement behavior in both conditions.
The results showed a stark difference depending on the type of engagement. Participants were more likely to comment on tweets that contradicted their views on vaccines (in line with the confrontation effect). However, when given the option to follow the user who posted the tweet, participants were less likely to do so if the content conflicted with their beliefs. This suggested that while people are drawn to confront opposing views, they avoid prolonged exposure to such content.
Mochon and Schwartz also tested whether the way a message is framed affects the likelihood of engaging with ideology-inconsistent content. Their fourth study was conducted using Facebook again, targeting users with liberal or conservative posts about gun control. The posts were framed either as information-based (low threat) or action-based (high threat). For example, a low-threat post might provide facts about gun control, while a high-threat post might call for immediate action to change gun laws. The researchers measured the engagement rates, including clicks and comments, for each type of post.
The findings revealed that posts framed as high-threat (fighting for a cause) generated more engagement from users with opposing views than low-threat (informational) posts. This suggested that the emotional intensity of the message plays a key role in driving the confrontation effect, with more provocative content leading to greater engagement.
In Study 5, the researchers investigated whether the importance of the topic being discussed affects the confrontation effect. The hypothesis was that people would be more likely to engage with ideology-inconsistent content on important topics, like COVID-19 vaccines, than on less personally relevant topics, such as smartphone brands.
Participants (803 individuals from Prolific) were randomly assigned to read tweets about either COVID-19 vaccines (a high-importance topic) or smartphone brands (a low-importance topic). As in previous studies, participants were asked to comment on ideology-consistent or inconsistent tweets, and their emotional reactions were measured.
The results showed that participants were more likely to engage with opposing views on important topics like vaccines, but not on less important topics like smartphones. This finding indicated that the confrontation effect is stronger when the issue at hand is personally or socially significant, providing a boundary condition for the effect.
Finally, in their fifth study, Mochon and Schwartz examined whether confronting ideology-inconsistent content (by commenting on it) could reduce the emotional distress or outrage people experience. Prior research suggests that people might feel better after expressing their negative emotions, so the researchers tested whether commenting on an opposing viewpoint would lower the outrage participants felt.
In this study, 400 participants from Prolific were exposed to political tweets from either Joe Biden or Donald Trump. Some participants were given the opportunity to comment on the tweet, while others were not. The researchers measured their levels of outrage and satisfaction both before and after exposure to the tweet.
The findings showed that participants who were able to comment on the ideology-inconsistent tweet experienced a reduction in their outrage levels, compared to those who were not given the chance to respond. This suggests that part of the reason people engage with opposing views online is to relieve their emotional discomfort.
“Outrage is a powerful driver of online engagement, often leading users to interact with content they would prefer to avoid,” Mochon told PsyPost. “Users should be more mindful of their reactions to online content and whether it is in their best interest to engage with it.”
Despite its important insights, the study also had limitations. One of the main challenges was the complexity of measuring engagement across different types of online platforms. Each social media platform has unique features that may influence how users engage with content, and the study focused on specific actions like clicking or commenting.
Future research could explore other forms of engagement, such as sharing or following, to gain a fuller understanding of how people interact with ideology-inconsistent content. Additionally, the study largely focused on political content, leaving room for further exploration of how the confrontation effect operates in non-political contexts.
“Social media is a complex and constantly changing environment,” Mochon said. “While outrage is a strong driver of engagement, the specific behaviors it leads to online may depend on the platform and social content.”
Future research could also explore whether this effect varies across cultures or social groups, as well as how it might evolve with changing social media dynamics. There may be other emotions besides outrage, such as fear or sadness, that drive engagement with opposing viewpoints. Understanding these nuances could help researchers develop more comprehensive theories about how people interact with conflicting information online.
“We would like to understand how to help people engage more constructively and reduce the amount of toxic content online,” Mochon said.
The study, “The confrontation effect: When users engage more with ideology-inconsistent content online,” was published October 5, 2024.
