Why do people believe in fake news?

Synthesis produced by the Fondation Descartes of the following research papers:

Pennycook, G., Cannon, T. D., & Rand, D. G. (2018). Prior exposure increases perceived accuracy of fake news. Journal of experimental psychology: general, 147(12), 1865.
Pennycook, G., & Rand, D. G. (2019). Lazy, not biased: Susceptibility to partisan fake news is better explained by lack of reasoning than by motivated reasoning. Cognition, 188, 39-50.
Bronstein, M. V., Pennycook, G., Bear, A., Rand, D. G., & Cannon, T. D. (2019). Belief in fake news is associated with delusionality, dogmatism, religious fundamentalism, and reduced analytic thinking. Journal of Applied Research in Memory and Cognition, 8(1), 108-117.

Here, we present a series of three studies by a team of psychology researchers. Published in 2018 and 2019 respectively, these articles study how the human brain works when it comes to believing fake news. They identify two cognitive processes that may impact belief in fake news - analytic thought, which makes people more sceptical, and intuitive thought, which makes them more gullible. The first article shows that repeated exposure to fake news increases our credulity, even when there are explicit warnings to alert the reader; the second article analyses the influence of intuitive reasoning on our gullibility; and the last article confirms the role of intuition in believing fake news.

Article 1: Prior Exposure Increases Perceived Accuracy Of Fake News. (2018)

Gordon Pennycook, Tyrone D. Cannon, and David G. Rand

Summary

This article shows that repeated exposure to fake news increases its perceived accuracy. It should be noted that this study was carried out in English, and participants had to judge the "accuracy" of the news stories presented to them. It is interesting to note that, for the authors of this study, believing in something is equivalent to deeming it accurate.

Introduction

Many psychological studies have shown that repeating information causes individuals to judge it more accurate. The phenomenon, named the “illusory truth effect”, could be explained by the fact that repeating a piece of information makes it easier to access in your memory. We then take this ease of access as proof that the statement is true and are therefore more likely to believe it. In other words, the easier it is to remember a statement, the more likely we are to believe it. Unfortunately, this effect can apply for fake news - we are more likely to believe headlines that appear multiple times in our social media news feeds. The authors tested this hypothesis with three studies.

Study 1

Method:

The first study in this article tests the effect of repeating false and implausible statements, such as “the earth is a perfect square".

The 409 participants in the final sample had to rate the interestingness of these false and implausible statements on a scale of 1 to 6 (participants were not told that the statements were false). The authors also presented them with false and true statements that the participants were unlikely to know (for example, it is true that Billy the Kid's last name is Bonney but very few people know). Then, after asking them some personal questions (age, sex, etc.), the authors showed them these statements again and asked them to rate their accuracy.

To check the effect of repeating the statements, the authors also presented statements that the participants had not seen at the start of the study. The authors aimed to determine whether statements shown earlier would be deemed more accurate than statements shown one time only.

Results:

Whether presented once or twice, false and implausible statements were, on average, judged highly implausible by participants. On the other hand, repetition had a positive effect on true or false statements that participants were unlikely to know - on average, repeated statements were judged more accurate than statements that were only shown once.

Conclusion :

• The illusory truth effect does not seem to work on highly improbable statements.
• However, it does work on false statements that are hard to determine as accurate or not at first glance.

Study 2

The authors then investigated fake news. Could repetition make individuals more likely to believe fake news? Could this effect persist even with warning labels?

Method:

There were 949 participants in the study. The procedure was essentially the same as in Study 1.

Here, four types of statements were presented to participants, in a format imitating news headlines on social media:

1. Fake news headlines
2. Fake news headlines with “disputed” labels (i.e. that indicates on the photo that the headline has been disputed by a fact-checking organisation)
3. Real news headlines
4. Real news headlines with “disputed” labels.

The experiment was run in the same way as the first study. When participants had to assess the accuracy of statements, new statements (from all four categories) were mixed in with the others.

Results:

• Participants fell victim to the illusory truth effect when faced with fake news - they were more likely to believe false headlines that were presented twice. The authors found that this was also the case for “disputed” fake news - the “disputed” label did not prevent participants from being more likely to believe the headlines with repeated exposure.

Fake news only needed to be shown once again for it to increase in perceived accuracy for participants.

Study 3

This time, the authors aimed to determine if the illusory truth effect would work if fake news was presented a third time, one week after the experiment.  

Method:

The method was the same as for the second study, only, one week later, participants were invited to give their judgements again. They were therefore exposed three times to the same fake news.

Results:

• Once again, the more that fake news is repeated, the more likely people are to believe it. This effect also works with fake news with "disputed” labels. However, it should be noted that, on average, the label makes fake news less likely to be believed.

General conclusion

• Labelling has an immediate effect on the perception of fake news - it allows people to be more wary. But this tool, used by social media organisations to flag news that has been fact-checked and disputed, is not enough to contain the problem. The more people are exposed to fake news, the more likely they are to believe it, even if they are explicitly warned. The actual effect could be even bigger than that measured in the study, as on social media, a headline may be seen many more times than three.
• This shows that it is necessary to raise awareness of the illusory truth phenomenon, to try to make concrete efforts against fake news.

Thanks to this first article, we know that repeated exposure to fake news increases its perceived accuracy. However, we still do not know what process is responsible for people believing in fake news. The second article gives us greater insight on that point.

Article 2: Lazy, Not Biased: Susceptibility to Partisan Fake News Is Better Explained by Lack of Reasoning Than by Motivated Reasoning (2019)

Gordon Pennycook and David G. Rand

Introduction

The title of this article leaves little to the imagination. In the previous article, the authors also observed that belief in fake news was stronger when the headlines aligned with the participants’ political ideology. In other words, pro-Democrat fake news will be more likely to convince a Democrat and vice versa. Though this may seem obvious, it is important to be able to explain this result.

Two explanations were proposed and tested in this article: classical reasoning theory and motivated reasoning theory.

Classical reasoning theory uses a model that identifies two distinct processes in human cognition. The first, “analytic” process works slowly, is deliberate, and requires cognitive effort; the second, “intuitive” process works quickly, is rushed, and requires few cognitive resources. This theory is very widespread in psychology and has notably been championed by famous psychologists Tversky and Kahneman (Nobel Prize in Economics). In short, it states that, depending on events, human thought can react very quickly, at the risk of making mistakes, or take more time, at the risk of reacting late but making fewer mistakes. According to this theory, belief in fake news could be caused by a lack of attention and cognitive effort - intuition leads us to make mistakes while reason pushes us to reject fake news.  

The second theory, called motivated reasoning theory, is also based on this distinction between two types of cognitive processes. It stipulates that the more people use their reasoning ability, the more likely it is that they will strengthen their preconceived beliefs. This theory may be counter-intuitive - you might think that the more we reason, the more likely we are to drop our certainties and beliefs. However, many studies lean towards the opposite conclusion - engaging in intense cognitive effort can reinforce our initial ideas. We use this cognitive effort to defend and reinforce our already-held beliefs. This theory is championed by Dan Kahan. For example, he showed that on the subject of climate change, individuals that are more inclined to reason are those who adopt the most polarised (or the most extreme) positions on the question. In the case of fake news, we believe it because it serves our cause!

The authors of the article aimed to find out if participants’ belief in fake news that aligns with their ideology would be further strengthened after cognitive effort, as predicted by motivated reasoning theory.

Method:

The researchers undertook two experimental studies. We will only be looking at the second study, which replicated the first on a larger scale.

1463 participants were recruited for the study, which took place in the summer of 2017. They were sorted politically by a question where they had to choose whether they support Hillary Clinton or Donald Trump (no alternative was given, to force participants to respond). 55.6% of participants chose Clinton.

The participants were presented with 24 headlines in Facebook post format (photo, headline, by-line). Half were fake news taken from a fact-checking website (Snopes.com), the other half were verified headlines. For each type of headline, the authors made sure that half were pro-Clinton and the other half pro-Trump (by first asking around 100 people).

To test the type of reasoning (intuitive or analytic) used by participants, the authors employed a widely used test - the Cognitive Reflection Test (CRT).

It requires participants to solve three problems that suggest an immediate response, one that seems obvious, but is wrong. As an example, here is one of the problems:

“A bat and ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost?”

The obvious answer seems to be $0.10, which is the answer given by 65% of participants in a previous study by Pennycook. However, the correct answer is $0.05. According to dual-process theory, this second answer (the analytic answer) requires you to ignore the first answer that jumps to mind (the intuitive answer) and analyse the problem more carefully.

The authors argue the following:

• If motivated reasoning theory is correct, participants that make more cognitive effort (and give the correct answers in the CRT) would also be those who give more weight to fake news that matches their political leaning - the theory states that the motivation to reflect strengthens our beliefs and therefore justifies the news (even fake news) that aligns with our preferences.
• But if classical reasoning theory is correct, participants who make more cognitive effort will also be less likely to believe fake news that panders to their opinion.

For each headline presented, the participants had to judge its accuracy (the same question as in the team's first article). At the end of the questionnaire, they completed the CRT.

Results

The results confirmed the predictions of classical reasoning theory - the participants with the best scores on the CRT did not give more weight to fake news that aligns with their political positions than participants who did worse on the test. Rather, the more intuitive participants of the group judged fake news to be more accurate than the more analytic participants. However, the authors noted that, on the whole, Trump supporters were more likely to believe fake news headlines (regardless of their content). This result was not interpreted by the researchers, as they had not formed any prior hypothesis relating to it (in experimental research, recommended practice is to not try to explain an unanticipated result after the fact).

Conclusion

• Given that people who use analytic reasoning are better at distinguishing fake news, this kind of reasoning could be promoted and developed to fight the phenomenon of disinformation online.
• However, the authors raise a subtle yet fundamental point. Certain studies (in particular, those by Pennycook) show that there is a difference between individuals in their ability to use analytic reasoning. In general, we can say that some people have an analytic disposition and others do not. So, the question is whether analytic reasoning abilities, once learned, could be deliberately used by all individuals, no matter their disposition.

Article 3: Belief in Fake News Is Associated with Delusionality, Dogmatism, Religious Fundamentalism, and Reduced Analytic Thinking. (2019)

M. Bronstein, G. Pennycook, A. Bear, D. G. Rand, and T. Cannon

Here, we present the last article in this series that confirms the link between analytic thinking and fake news.

Once again, the methodology is close to that in the previous studies.

Methodology:

948 participants were recruited to assess the accuracy of 24 headlines (12 fake and 12 verified). To obtain a psychological profile of the participants, the authors gave them a series of tests, including the CRT presented in the previous study.  

In this study, the authors hypothesised a possible link between believing fake news and three psychological dimensions: delusionality, dogmatism, and religious fundamentalism. They used a range of tests to measure the level of each of these psychological dimensions in participants.

To test the effect of the different dimensions on belief in fake news, the authors performed mediation analysis. This statistical method allowed them to evaluate the overall effect of a dimension on believing fake news. It was also able to measure the influence of a second dimension on this relationship (we will illustrate this point in the Results section).

Results:

• There is a statistical link between believing fake news and the three dimensions in question: people who scored highly on the delusionality, dogmatism and religious fundamentalism tests tended to have a higher score on the fake news belief test.
• Mediation allowed the authors to identify the key role of intuitive thinking in believing fake news. This result supports the previous study. Intuitive thinking explains why individuals who scored higher for the three aforementioned psychological dimensions (delusionality, dogmatism and religious fundamentalism) believe more strongly in fake news, on average. In other words, intuitive thinking is used more by individuals who have a higher level of these psychological dimensions, which would explain their susceptibility to believing fake news.

• There is a statistical link between believing fake news and the three dimensions in question: people who scored highly on the delusionality, dogmatism and religious fundamentalism tests tended to have a higher score on the fake news belief test.
• Mediation allowed the authors to identify the key role of intuitive thinking in believing fake news. This result supports the previous study. Intuitive thinking explains why individuals who scored higher for the three aforementioned psychological dimensions (delusionality, dogmatism and religious fundamentalism) believe more strongly in fake news, on average. In other words, intuitive thinking is used more by individuals who have a higher level of these psychological dimensions, which would explain their susceptibility to believing fake news.

For more information, readers can also directly consult two of these articles that are free to access:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6279465/

https://www.df.cl/noticias/site/artic/20181116/asocfile/20181116172919/2018_pennycook_rand_cognition.pdf

The third article can be accessed in the Fondation Descartes documentation centre.