Misinformation thrives on outrage, study finds

Misinformation travels far and fast online, and according to new research published in Science, outrage is the fuel behind its spread. By examining millions of social media interactions, researchers found that misinformation sources consistently provoke more outrage than credible news, making it more likely to be shared. Alarmingly, this emotional response encourages users to share misinformation impulsively, often without verifying its content.

“Understanding the psychology of misinformation consumption is crucial, as platforms continue to grapple with limiting its spread (and it might be even more important now that platforms such as Meta are giving up on professional fact checkers),” said study author William J. Brady, an assistant professor at the Kellogg School of Management at Northwestern University.

“Much of the existing research assumes that people prioritize sharing accurate information—what we term ‘epistemic motivations’ for sharing. However, social media behavior is also driven by non-epistemic motivations, such as the desire to share information just because it aligns with your group’s beliefs (see here for example some of our prior work).”

“Expressions of moral outrage, in particular, are frequently tied to such behaviors. Building on this reasoning, we hypothesized that misinformation exploits our inclination to share moral outrage online. While previous studies have examined the role of emotion in misinformation consumption, we theorized that moral outrage plays a particularly critical role in shaping this psychology.”

The researchers analyzed data from Facebook and Twitter, two of the most widely used social media platforms. On Facebook, the dataset included over one million shared links from the URL Shares dataset. This dataset provides aggregated metrics on public interactions with links shared on the platform, including the number of reactions, such as “Anger,” “Love,” and “Wow.” Outrage was operationalized by counting the number of “Anger” reactions a link received, as this reaction closely aligns with moral outrage, a mix of anger and disgust.

On Twitter, the researchers examined over 44,000 tweets, focusing on original posts that linked to either misinformation or trustworthy sources. They also collected over 24,000 responses to these tweets, such as replies and quote tweets. To measure outrage in these responses, the team used a machine-learning-based classifier called the Digital Outrage Classifier. This algorithm was specifically trained to detect expressions of moral outrage in political tweets, offering a robust and scalable way to quantify outrage on the platform.

Both platforms’ data spanned multiple years (2017 and 2020–2021), allowing the researchers to assess the consistency of their findings over time. To ensure reliability, they classified links as “misinformation” or “trustworthy” based on the quality of their source, as determined by professional organizations and fact-checking databases.

The researchers found that misinformation sources tended to provoke stronger feelings of outrage than trustworthy news sources. On Facebook, links from misinformation sources received significantly more “Anger” reactions compared to links from credible sources. Similarly, on Twitter, tweets containing misinformation links elicited a higher proportion of responses expressing moral outrage.

On Facebook, links that received more “Anger” reactions were also shared more frequently, and this relationship was stronger for misinformation than for trustworthy content. Similarly, on Twitter, tweets with outrage-filled responses were more likely to be reshared, further amplifying their reach. In addition, links on Facebook that evoked more outrage were more likely to be shared without being read first, particularly when the links came from misinformation sources.

“The key findings revealed that misinformation tends to elicit more outrage than trustworthy news, which in turn increases its likelihood of being shared. Additionally, when individuals feel outraged, their sharing decisions are more likely to be driven by non-epistemic motivations—such as sharing information simply because it aligns with their group’s beliefs. For example, we find that outrageous misinformation is more likely to be shared even before people have read the actual news article.”

In addition to observational analyses, the researchers conducted two controlled experiments to establish causal relationships. A total of 1,475 participants were recruited for these studies. Participants were shown news headlines in a simulated social media environment and were asked to evaluate these headlines on various criteria.

In the first experiment, participants were shown 20 headlines, half of which were from trustworthy sources and half from misinformation sources. The headlines were further divided based on their outrage-evoking potential (high or low). Participants rated how likely they were to share each headline on a four-point scale, ranging from “not at all likely” to “very likely.” This design allowed researchers to test whether outrage influenced sharing behavior differently for misinformation and trustworthy news.

Headlines evoking high levels of outrage were significantly more likely to be shared than those with low outrage potential. Interestingly, the researchers found no significant difference in the likelihood of sharing between misinformation and trustworthy headlines. This suggests that outrage influences sharing behavior uniformly, regardless of whether the content is true or false.

The second experiment followed a similar design but focused on accuracy discernment. Participants evaluated the perceived accuracy of the same set of headlines, using a four-point scale from “not at all accurate” to “very accurate.” This experiment aimed to determine whether outrage affected participants’ ability to distinguish true from false headlines, providing insights into the epistemic motives for sharing.

Participants were generally able to discern trustworthy headlines from misinformation, rating the former as more accurate overall. However, the level of outrage evoked by a headline did not significantly affect participants’ accuracy judgments. High-outrage headlines were neither more nor less likely to be judged accurately compared to low-outrage headlines. This finding suggests that while outrage drives sharing intentions, it does not directly impair or enhance participants’ ability to evaluate the accuracy of information.

“An important implication of our findings is that that policy interventions focused on getting users to pay more attention to ‘facts’ might not be effective for the most outrageous misinformation,” Brady told PsyPost. “However, further research is needed to identify the most effective interventions for curbing the consumption of outrage-inducing misinformation. At a minimum, our findings underscore the importance of focusing on misinformation that is most likely to evoke outrage, as this type of content is more prone to spreading widely and potentially having significant impact.”

The study, “Misinformation exploits outrage to spread online,” was authored by Killian L. McLoughlin, William J. Brady, Aden Goolsbee, Ben Kaiser, Kate Klonick, and M. J. Crockett.