You are not alone. Misunderstandings abound in digital communication, where “there is a limited wealth of information compared to face-to-face interaction,” he explains William J. Bradyassistant professor of management and organizations at the Kellogg School.
New research from Brady and co-authors shows that the same basic principle applies to interpreting the emotions of people who post on social media. Specifically, the study finds that people perceive more moral outrage in political posts than the authors felt when they wrote them. This is especially true for people who regularly use social media to learn about politics.
This skewed outrage detection—a kind of large-scale, “wait, they’re mad at…everyone?”—has important downstream consequences, the researchers found. Over-perceiving outrage in their feeds makes people over-perceive anger on the social media platform as a whole and increases people’s belief in the ideological extremism and political polarization of other platform users. In other words, when we see outrage that was not intended and may not exist, the misunderstanding colors our sense of how much outrage on a platform is normal.
“There’s a lot of discussion around how to improve debate and political discourse on social media and make it less irreverent and less toxic,” Brady explains. “One of the things this work strongly suggests is that it’s not just about targeting specific users. It’s also about people’s social perceptions—because even if we could tone it down, people might pick up on it.”
Overrated outrage on Twitter
Brady worked with Molly Crockett and Killian McLoughlin of Princeton University and Maria GedronKara Luo, and Mark Torres of Yale University, for conducting the research. They started by looking at real Twitter users and their posts.
They used a machine learning classifier to identify people tweeting about American politics with what the classifier judged to be either high or low levels of outrage. The researchers then contacted the tweet authors via direct message and asked them to report, within 15 minutes of posting, how happy or outraged they felt when they wrote a particular tweet.
Then, in several studies, the researchers asked a total of 650 participants to look at the tweets and rate how happy and how angry they thought the author felt when they posted. Participants also reported how often they used social media to learn about American politics.
Analysis of this data revealed a consistent mismatch: observers perceived more outrage in tweets than writers reported feeling when they wrote them. (Interestingly, participants did not overperceive the authors’ happiness, suggesting that positive emotions are less readily perceived than negative ones.) And overperception of resentment was stronger among participants who used social media to learn more. for politics.
“We think the association is explained by the fact that if you spend a lot of time on social media, you’re going to create expectations about how people express themselves and how they feel,” explains Brady. “And once you form those expectations, it will affect your next round of judgments when you see something that looks like outrage”—triggering a vicious cycle of anticipating outrage and finding out again and again.
When resentment seems ordinary
The researchers then wanted to understand whether people might over-perceive anger on a social media platform as a whole, not just individual posts or their personal news feeds.
They used the tweets from the first experiment to create two simulated news streams. The two streams were equally matched in the levels of outrage reported by the authors of the tweets, but differed in how they were perceived by previous participants, leading the researchers to call high overperception news and low overperception news.
The researchers recruited 600 young participants and randomly assigned them to view either high or low news hyperperception. Participants were then asked to rate on a scale of one to seven how angry they thought members of the entire social media platform—not just that feed—were on average.
Participants who viewed the high hyperperception stream rated outrage on the overall platform higher than those who viewed the low hyperperception stream (an average of 5.82 out of seven compared to 3.53). Additional statistical analyzes revealed that participants who saw the high hyperperception stream relied heavily on the most outraged tweets when making ratings of collective outrage—suggesting that particularly inflammatory posts can have a large influence on how users perceive the entire platform.
Another study with 1,200 young participants reached a similar conclusion. After viewing one of the same two simulated news feeds from the previous experiment, participants were then presented with ten new tweets: five in which the authors intended to convey high levels of outrage and five intended to be more neutral. Participants were asked to rate how socially appropriate these tweets would be for the platform. They also reported how politically polarized and ideologically extreme they perceived users of the platform to be.
Participants who viewed the high hyperperception news feed rated posts expressing high levels of outrage as more appropriate than those who viewed the low hyperperception news feed. Participants high in overperception also saw platform users as more polarized and more ideologically extreme, the researchers found.
“People see the social network as more polarized than it actually is. They think that outrage is normative, and that’s really important, because we know that norms are a strong predictor of group behavior,” Brady explains. “This suggests that when outrage is over-perceived or even over-represented by algorithms, people may be more likely to conform to an outrage expression rule” — even though outrage may not actually be the rule.
Breaking the cycle of resentment
The long-term solution to excessive outrage—real and perceived—on social media may lie in changing the way the platforms are designed and the posts they reward. Brady also believes it’s important for everyday users to understand how these platforms work, as well as the biases and misconceptions they bring to social media.
“It’s important to raise awareness that the things you see in your feed aren’t necessarily representative of your social network online and especially offline,” he says. In some cases, people may not be as crazy as you think.