The psychology of fake news: how disinformation spreads online

6 in 10 people will share an article online after having read only its title. 23% of Americans admit to having posted a fake story online. Yet, most of us consider ourselves immune to online disinformation – but if that’s so, why has fake news become so pervasive? Here is what psychology says about the new aspects of the good old propaganda –  – and how the problem is growing in the connected age.

Fake news runs faster than the truth

A recent study done at MIT analyzed the dissemination of 126,000 stories on Twitter between 2006 and 2017. Around 3m Twitter users spread those stories with 4.5m tweets. The analysis showed that false stories reach 100x  bigger audiences, and spread much faster and broader than true stories  (the top 1% of the analyzed false news reached between 1000 and 100,000 users, while the fact-based news reached on average only 1000 people).

Interestingly, the analysis also shows that bots have less to do with sharing fake news on Twitter than you likely imagine – the stories spread so quickly because they were shared by real peopleas opposed to automatic accounts.

What makes people fall for fake news? MIT’s researchers provided two explanations.

  • First, the element of surprise is a strong driving force for sharing content online. False stories are written with the aim to arouse interest and, thus, they have a higher level of novelty than the true stories. According to the study, sharing such content may correspond to people’s preference to share breaking news and novel information with their peers.
  • The second reason lies in the emotional reactions false stories evoke – such content is associated with strong emotions, like fear, disgust, and surprise, while true stories inspire emotions like sadness, joy, or trust. Although we know that emotional content gets shared more often on social media than neutral content (for example, this study analyzed 165,000 tweets and found a clear link between the number of retweets of emotionally charged tweets compared to neutral tweets), it is the intensity of emotions that plays a more important role here.
Content evoking high-arousal emotions such as anger or awe gets more shares than content with low-arousal emotions such as sadness or content

In their article “What makes online content viral?” researchers Johan Berger and Katherine L. Milkman discover that content evoking high-arousal emotions such as anger or awe gets more shares than content with low-arousal emotions.

High-arousal emotions can be both positive and negative, but while Bergman and Milkman found that positive content tends to get shared more often online – it is not the valence, but the intensity of emotions, that makes fake news go viral.

You will know them by their language

A characteristic of fake news is the sensationalist language they use.

Psycholinguistics can help in detecting false information by pinpointing concrete linguistic cues related to deception. For example, this study done at the University of Washington compared the language used in real news, satire, hoaxes, and propaganda, and found that words used to exaggerate – subjectives, superlatives, and modal adverbs – are used more in fake news. By contrast, truthful news contains more words and expressions with concrete figures – comparatives, money, and numbers.

The researchers also found that first-person and second-person pronouns are used more in less reliable news types, while trustworthy news is more likely to avoid language that seems too personal.

Fake news reports also commonly contain two clear signs of deceptive language – more vague expressions in general, and more hedge words, or words used to soften or lessen the impact of the statements that follow  (e. g. “a little bit”, “somewhat”, “maybe”, “apparently”, “sort of”).

Fake news use pompous language and words to exaggerate – subjectives, superlatives, and modal adverbs

Another paper analyzes the linguistic traits of fake news, and concluded that these types of stories differ substantially from the style of real news, despite the common conception that false stories are written to resemble trustworthy stories as closely as possible.

Fake news is closer to satire than to real news, concluded the authors – their titles are longer, use few stop words (short function words such as “the”, “which”, “on”, “is”), and fewer nouns, but more proper nouns. Fake news packs the main claim of the article in the long title, which is often a claim about a person or an entity, while the body of the article remains repetitive, short, and less informative.

Blame the brain? But not only!

We have access to billions of resources on the Internet – how come we fall prey to fake news, which we can easily double-check with a click?

Sadly, people tend to lounge in their echo-chambers rather than actively search for new information online, and social media makes it frighteningly easy to do so.

According to this study, which analyzed the online behavior of 326m Facebook users over a period of 6 years, for most people news consumption online is limited to a very small number of Facebook pages. The phenomenon is called selective exposure, and it determines the formation of tight communities around particular Facebook pages, which leads to further segregation. According to the researchers, the growing polarization of users on specific narratives drives the rapid diffusion of disinformation online.

The way we evaluate incoming information is another factor for our susceptibility to fake news. In a recent paper, Belgian researchers found that people’s cognitive ability determines how well they can adjust their attitudes after false information is being corrected.

The study first assessed people’s judgment about a fictitious character after receiving negative information about her, but later corrected that information and assessed whether changes in the subjects’ original judgments occurred. In turns out, only participants with high cognitive ability adjusted their opinions after being presented with the contrasting information. Participants with low cognitive ability persevered with their initial opinion, despite the new facts.

Our cognitive ability determines how we process and remember fake news

Our cognitive ability is related to our working memory, analyzes this article at Scientific American. Some people are less able to discard no longer relevant information from their working memory, which makes them an easy victim to fake news. As we grow older, we become less adept at self-correcting false information, because our working memory naturally declines with age (which can explain why elderly people are more likely to fall for fake news).

But things are even more complex than that – the propagation of fake news is a multi-factor phenomenon that depends on both cognitive and environmental factors.

This study published in Nature Human Behaviour posits that both information overflow and users’ limited attention span contribute to the virality of low-quality content on social media. The researchers find a weak correlation between quality and popularity of content, and conclude that in a network oversaturated with information, low-quality content is just as likely to go viral as high-quality content.

In fact, seeing too much information (for example, in the form of a constant flow of posts in the News feed), affects cognition: confirmation bias, or our tendency to share information that supports our beliefs, is exacerbated on social networks, because we have no time to properly analyze all incoming posts, and tend to pay attention only to the ones we like.

Retweet a lie often enough…

…and yes, it will become the truth. One reason why fake news is so hard to debunk is that it is largely shared on social media, which, combined with the fact that more and more people are getting their news primarily from social networking sites, creates the perfect environment for the so called “illusory truth effect”. This effect postulates that we are more likely to believe a statement is true if we’ve read it somewhere before. That’s because repetition enhances cognitive fluency, or speeds up our ability to process information, which makes us (mistakenly) imply that the information we already know must be correct.

We are more likely to believe a statement is true if we’ve read it before

To prove this effect on social media, the authors of a study conducted by MIT and Yale University showed people fake and real news headlines in the format of Facebook posts, and found that fake news headlines that were shown in the first part of the experiment were more likely to be perceived as accurate in the second part – even if the headlines were flagged as “disputed by 3rd party fact-checkers,” were opposite to the participant’s political orientation, or were not even explicitly remembered by the participant!

Difficult to debunk

One thing that is true about fake news is that they are extremely hard to debunk. The reason for this is, again, our emotions – it turns out, once we are hooked on something, we become extra alert, and as a result, we are more likely to remember it.

Psychologist Johanna Kaakinen from the University of Turku explores our reactions to fake news at the physiological level. According to her research on eye movements, when we read a story that provokes emotional reactions, and is relevant to us, we get a sort of tunnel vision and our attentiveness increases manifold, which makes us more likely to remember the story, and more disinclined to re-evaluate the information later.

In 2017, Facebook admitted that their technique of flagging articles as “fake news” was not working – the mere flags turned users’ attention towards the content, and reinforced inherent false beliefs. As noted – a false statement receives legitimacy when it gets visibility.

But how do we debunk a myth without mentioning it? In their Debunking Handbook psychologist Stephan Lewandowsky and media professor John Cook come up with some advice on how to dispel false information:

  • First of all, to focus on the facts rather than on the myth, and mention the myth only after the facts have been explained;
  • Second, don’t bury the audience in facts, but rather use simple language and just a few facts. Lewandowsky and Cook talk about the gap in people’s knowledge after a myth has been debunked – in order to be effective, debunking needs to fill that gap with alternative information.

Whether it is realistic to expect debunking to happen at the first attempt is another issue. A meta-analysis of 65 studies on correcting misinformation shows that some myths are harder to correct than others – especially misinformation in the context of politics, compared to topics such as health or crime.

What makes debunking even harder is that most people have difficulty believing they themselves could be a victim of online disinformation. This interesting survey explores the so-called Third-person effect, or our tendency to believe that others are more vulnerable to the effects of mass media, in the context of political communication on social media. The results show that people are more likely to believe that their political opponents are influenced by false narratives online, while they overestimate their own capacity to detect misleading information.

No one-size-fits-all solution

The explosion of fake news in the past couple of years showed that, sadly, we have not been made smarter by our smartphones.

Propaganda has always existed, but the difference now is that it has become easier to manipulate public opinion using humanity’s psychological shortcomings such as low attention spam and cognitive biases – all of which thrive in a dynamic online environment.

We can try to protect ourselves from disinformation by keeping an open mind and actively selecting our sources, however, one cannot transcend one’s limitations. Fake news is a social problem, which won’t be solved in a few quick steps.


Related posts