Since the outbreak of COVID-19, the flow of related news items is inescapable. Whether you watch the news, read the paper or open Social Media, there will always be some new information about the global pandemic we are in. The stage of social media becomes more prominent for both reader and writer. In times of a highly uncertain global pandemic, accompanied with the fact that everybody has the agency to spread information, fact-checking has never been more important. It can diminish the fear around our day-to-day lives and the distrust in governmental institutions. With this article, we take the first step and look at whether we can find a possible solution for this issue.
The focus of our investigation is on visual content because we know humans process visual information faster, and most people remember 80% of what they see against 20% of what they read. Also, visual content is more likely to be shared, as it is eye-catching and has a direct effect on human emotions (Manic, 2015). These properties made it urgent to focus on tweets containing a visual element. By manually fact-checking a sample of tweets from a large database, we tried to formulate an answer on the following research question: “How reliable are the most retweeted COVID-19-related tweets containing a visual element?”
To answer this research question, 59 tweets were checked for their truthfulness. These tweets were selected from a database with most retweeted tweets containing a visual element by Pointer. Initially, this database consisted of more than five thousand tweets. To draw fair conclusions about the reliability of the whole database, a random sample was taken from it. The truthfulness of this selection was checked using different tools and sources to investigate the reliability of accounts, media and statements (see report for exact method). Eventually, results showed that 68% of these fact-checked tweets were considered as (mostly) true.
The truthfulness of prestigious sources
While interpreting the results, the majority of the tweets came from prestigious sources, of which some accounts had the verification mark. Examples of these are politicians, scientists, and the National Institute for Public Health and Environment. Comfortingly, almost all of these sources’ tweets were found to be factually correct. Often, media, such as a picture, is used to lend force to the statement in the text. However, the source of the picture is, in most cases, not clearly stated. This makes it difficult for Twitter users to decide if what they see is true. For example, Fleur Agema, member of the Dutch political party PVV, states in the tweet below that 11% of the institutes for disabled have been dealing with corona and finds this worrisome. To lend force to her statement, she shared a picture of highlighted sentences from a governmental document, stating these numbers as well. After digging into political reports, we can confirm that this picture is indeed a paragraph from a governmental document. However, the readers of her tweet could have thought that she made up these numbers herself. Therefore, we could stress that adding a source will reduce some suspicion or distrust.
Furthermore, we found that non-verified accounts spread most false information. We, therefore, investigated whether the account below has a humanlike activity pattern or rather shows some suspicious activities. This analysis showed that this account could be a so-called ‘Astroturf’ bot (see report for detailed explanation). However, we can reject this concern since the activity patterns are humanlike. This means that we expect the account not to be a bot, but an account created to provoke discussion and fear of political matters. The latter becomes more obvious when looking at some of the last tweets and most used hashtags. Furthermore, we did not find any evidence of this tweet’s statement that citizens have to pay for the damage COVID-19 costs to our society. Posted by this account, it seems just another fear-provoking tweet.
Provoking fear
Provoking fear was seen as a thread throughout the tweets we found to be (to some degree) false. This picture of Bill Gates and Prime minister Rutte was posted, by an account stating that 5G causes COVID-19 symptoms. Moreover, it claimed Bill Gates to be the ‘inventor of crime against humanity’, of which no evidence was found, and even counter-evidence can be found. This insinuates that Mark Rutte and Bill Gates have been meeting on this topic and are together in this ‘crime’. A Google reverse image search shows us the truth: this photo was taken on January 26, 2016. On this day, Rutte and Gates talked about the importance of sustainable development goals and not about the subject shown in the tweet (Volkskrant, 2016). Together with other provoking statements, this tweet can be seen as an incitement to fear.
Another example of a fear-provoking tweet is the one below, containing a video in which a claim is made about the COVID-19 software. The person in the video claims that, COVID-19 software has been installed automatically in our phones, and that from now on, it will be easy for the government to install the tracking app on our phones. A lot of panicking comments of people concerned about their privacy were elicited. The official website of Apple explains that the software is only activated when you manually install an official COVID-19 app. Also, the user decides for himself whether to turn tracking on. It is clear that the goal of this video is to elicit fear and collective distrust in the government by creating false panic about COVID-19 software.
The importance of data visualisation
The tweet below shows us that we need to stay critical because even (verified) accounts that are expected to be reliable can spread fake news. A Twitter account managed by the Dutch national broadcaster “NPO 1” posted a video of mathematician Bert Slagter. In this video, he explained that the data visualisation of deaths caused by COVID-19 in several countries is misleading. The tweet stated that The Netherlands, in comparison to other countries, are doing ‘not too badly’ and is confirmed by virologist Jaap van Dissel. This statement was a motivation for Slagter to express his criticism on how the numbers of Corona cases are distributed in the media. However, the explanation of Bert Slagter was not following reality. First of all, we found that Bert Slagter is not a mathematician. A search on his LinkedIn profile told us that he started a study in computer science at Utrecht University, but never completed this, or any related study. Furthermore, Slagter states that the x- and y-axes are not important in the graph, which is not true, because axes allow readers to interpret a graph objectively (Yang et al., 2020). Therefore, his statement that The Netherlands is one of the countries with the highest death rates’ is a wrong interpretation, as the axes showed that rate was per capita. In short, this example visualises that even prestigious sources can be part of spreading fake news.
The urge to be critical before disseminating tweets
Finally, after analysing 59 tweets, we can conclude that visual tweets related to COVID-19 show some division to reliability. These tweets give us insight into which sources disseminate either fake or genuine news. We found that regular accounts spread fake news, while the institutions and journalists usually shared messages based on the truth. This trend can be traced back to where we started. As we explained, people gain less confidence in institutions such as the government and science (Lewandowsky, Ecker & Cook, 2017). As a result, people express their fear in tweets, creating a vicious circle in which these tweets, in turn, contribute to increased distrust in intuitive parties.
This effect, defined as Elite Discourse, explains how communication shared by the elite about fake news affects people’s ability to accurately assess whether the news is true or fake (Van Duyn & Collier, 2019). This means that even when news institutions and journalists disseminate real news, many people will distrust this information. The lack of good judgment is alarming because more people are spreading misinformation on Twitter (see report for the statistics).
A possible solution to this problem is to give people tools with which they can control information. As our research has shown, institutions are not transparent in their resources on Twitter. It is necessary to list resources systematically, even on a transient platform such as Twitter. Hopefully, this will result in a more critical view of the institutions themselves, which may lead to a more in-depth fact check before distributing news and thus, reduce the unintentional dissemination of fake news from their side. With this article, we show the urge to take steps from the transmitter’s side to regain confidence in the news and the institutions of a fundamental part of the population.
References
- Van Duyn, E., & Collier, J. (2019). Priming and fake news: The effects of elite discourse on evaluations of news media. Mass Communication and Society, 22(1), 29-48.
- Lewandowsky, S., Ecker, U. K., & Cook, J. (2017). Beyond misinformation: Understanding and coping with the “post-truth” era. Journal of applied research in memory and cognition, 6(4), 353-369.
- Manic, M. (2015). Marketing engagement through visual content. Bulletin of the Transilvania University of Brasov. Economic Sciences. Series V, 8(2), 89.
- Yang, B., Vargas-Restrepo, C., Stanley, M., & Marsh, E. (2020). Truncating bar graphs persistently misleads viewers. Journal of Applied Memory and Cognition, in press.