The Self-Fulfilling Prophecy of Internet Outrage

The UCLA researchers noted that the most accessible way for most people to engage in political discussion is through social media (especially in a country like the U.S., where political divisions often follow geographic lines). Ideally, social media would help decrease polarization by exposing people to a diversity of viewpoints. But a growing body of research suggests that this may not be the case. 

Instead, online platforms may actually be giving us inaccurate ideas about how other people think and feel. In particular, recent studies are finding that social media causes us to overestimate polarization and hostility. 

One such study, published earlier this year, found that participants were likely to over-exaggerate the amount of outrage a person was expressing in politically charged posts.³ Researchers asked participants to read a series of tweets about political content, and then rate how happy or outraged they thought the author felt. When evaluating negative posts, observers consistently imagined that the original poster had been angrier than they actually were. 

Even more telling: the more a participant relied on social media as their source for political information, the stronger their overperception of outrage. 

Why does this happen? It's important to remember that while social media itself is a modern invention, its design activates a range of cognitive biases that humans have evolved over millennia. 

Researchers observed that not only did participants incorrectly measure the level of outrage in individual social media posts, but also relied heavily on the most outrage-filled posts when evaluating collective levels of outrage. In other words, the most negative posts had a disproportionate influence on their beliefs. This may be caused by negativity bias, our tendency to give more weight to negative events than we do to positive ones. 

While there are several theories of why negativity bias exists, one possibility is that our brains spend more cognitive resources processing negative events, in order to help us take action to minimize potential undesired consequences. As a result, these events become more experientially intense and therefore memorable.

We can see evidence of this bias in the fact that participants did not over-exaggerate the level of happiness in positive posts, nor did positive posts have a disproportionate influence on their beliefs about collective outrage. 

In a previous article for TDL, Paridhi Kothari outlined evidence that social media posts judging opposing political groups are more likely to grab our attention. She explores how this may occur because of our evolutionary interest in group identification — our hardwired need to distinguish ingroups from outgroups. This likely derives from evolved survival instincts, encouraging us to seek safety by aligning ourselves with a larger group. 

Due to this often subconscious preoccupation with retaining group membership, we’re likely to treat outgroup members more harshly than those with whom we identify (known as ingroup bias). Therefore, our attitudes about others and our behaviors towards them can actually shift based on how we assess the group dynamics in a situation. 

For an example of this, let’s return to the UCLA Zoom study. Most of the experiments involved a private one-on-one discussion between just two people with opposing views on a political topic. However, a subset of the discussions also included an additional participant from each side of the political spectrum who played the role of a silent onlooker during the conversation. The researchers were interested in discovering whether the known presence of an ingroup member would change the way that people interacted.

The answer: conversations that included these “ingroup audiences” had higher levels of conflict than those without. The researchers noted increased levels of what they called “unproductive” types of conflict, such as “frustration and bad-faith arguing.” This may have been an indication that the presence of an ingroup member created more pressure for participants to boldly defend their positions in order to avoid being perceived as disingenuous to their ingroup. Meanwhile, in conversations without audiences, discussion partners may have been more motivated and able to find common ground. 

If the presence of an audience member noticeably altered civility in a low-stakes and relatively private experiment, we can speculate how strongly these social pressures influence the highly public domain of social media. As the UCLA researchers put it: on social media, “rather than simply posting their honest opinion or thoughts about something, [users] must navigate a complex landscape of engaging with the outgroup while signaling loyalty with their ingroup." 

Findings like these can help us understand why the way people communicate and respond online may not be an accurate way to gauge how divided we really are, or how capable we are of getting along.  

We know that algorithms, targeted political advertisements, and other external forces influence what we see and believe online. But even without these intentional attempts to skew public opinion, social media may still contribute to an overperception of collective moral outrage and ideological extremity. 

Our inaccurate beliefs may increase our own levels of hostility, as we come to believe that expressing outrage is a social norm. Our fear and mistrust of others may also contribute to discussion avoidance, creating barriers to engaging in the types of restorative cross-ideological conversations that can correct our misperceptions about one another. 

Scientists and nonprofit organizations are fighting an uphill battle to outweigh the negative impacts of the highly prevalent and well-funded social media industry. Thankfully, the latest research provides hope that if we create the right opportunities and incentives for people to engage, we are still capable of respectful discourse and understanding. 

Anyone interested in supporting these efforts can keep these insights in mind:

1. Negativity is overrepresented. The salience of negative content (and the fact that algorithms tend to boost it), may skew our evaluation of the collective. But even if these negative voices dominate our minds and social media feeds, in truth, they’re just one small slice of the demographic pie. 
2. Face-to-face discussions have different dynamics. The harmful effects of group dynamics are vastly reduced when we talk to someone privately and directly, rather than in a public online forum. Video conferencing can assist us in bridging geographical distance without compromising psychological safety. 
3. It’s helpful to reframe. Instead of approaching cross-ideological conversations as if they were political “debates” to be won or lost, framing discussions as “opportunities to find common ground” can help create social norms of civility and positively capitalize on our desire to belong.