Lie detector: Covid-19 and misinformation

Over the past few months, there have been a host of Covid-19 conspiracy theories shared across social media. In some quarters, this has been described as a ‘tsunami’ of misinformation.

In the early phase of the pandemic, wild theories suggested that American business magnate Bill Gates had intentionally caused Covid-19 to implant digital microchips that could track and control people. There have also been conspiracies around vaccines altering human DNA.

Other theories have claimed that technology, such as 5G, is the cause of the virus, while others have argued that the entire pandemic is a hoax and encouraged social media users to film their local hospitals to prove they are empty. Our research focused on the 5G and #FilmYourHospital conspiracy theories, and we published two papers in the Journal of Medical Internet Research in 2020.

It has become really important to have access to tools and methods to rapidly analyse social media data to detect drivers of misinformation. This is because false and misleading information is a serious public health concern. If certain people believe that the Covid-19 pandemic is a hoax, they may ignore lockdown restrictions and/or be sceptical about vaccines.

There is research to support the view that those who might believe in conspiracies may be less likely to follow government recommendations. This could have negative health outcomes, making it important to study the types of misinformation circulating on social media and the drivers of that misinformation.

In the case of the ‘5G and Covid-19’ conspiracy, 5G phone masts were vandalised across the UK, as well as in Europe. For the #FilmYourHospital conspiracy, people attempted to enter hospitals to film inside them.

Using the Microsoft Excel plug-in NodeXL, we were able to retrieve tweets related to a time when there was heightened interest in the two conspiracies. NodeXL provides quick access to data from a number of different social media platforms.

In our first study, data relating to the hashtag #FilmYourHospital was retrieved from Twitter for a seven-day period between 13 and 20 April 2020. In total, 22,785 tweets were captured, sent by 11,333 Twitter users.

For our second study on Covid-19 and the 5G conspiracy theory, we retrieved data from between 27 March and 4 April. In total, 10,140 tweets were retrieved, sent by 6,556 Twitter users.

Social network analysis was used, drawing upon the Clauset-Newman-Moore algorithm. This allowed us to analyse and identify network structures within the Twitter conversation taking place.

For the 5G conspiracy theory, we found that the two largest structures were an ‘isolates’ group and a ‘broadcast’ group. For the #FilmYourHospital conspiracy, the largest clusters identified were broadcast networks.

An isolates group occurs when a large number of users tweet about a topic without mentioning other users or retweeting tweets. These users tend to be outsiders from the core discussion.

A broadcast network structure occurs when a single user, and/or a group of users, is being retweeted in high frequency.

Our research revealed that influential users were distributed across the network, forming their own groups and audiences. We also found that ordinary citizens were among the key drivers of the conspiracies, and users would link to fake news websites and/or videos on YouTube that contained misinformation in order to make their point.

Certain Twitter accounts engaging on the 5G conspiracy theory network may have had ulterior motives, because they could seek to profit from it – including websites claiming to sell products to protect against 5G.

An interesting aspect of both conspiracy networks is that tweets tend to build steadily and then peak, causing a flurry of tweets that then die down. For the ‘#FilmYourHospital’ conspiracy, our research also checked to see whether users were likely to be bots – an automated account that can be controlled by a third party.

We used the Botometer tool to detect bots within the network and found that only a small number of accounts displayed bot behaviour – indicating that, potentially, the network was comprised of ordinary citizens.

Our recommendations were that public health authorities could make use of quick, targeted interventions to reduce the impact of misinformation. Moreover, we suggested that users could use the report feature of a social media platform to flag content that may go against the terms and conditions of Twitter. We also recommended enlisting the assistance of influential users, celebrities, and popular-culture figures to help share factual information, because official government accounts may be viewed with scepticism.

Wasim Ahmed is a lecturer in digital business at Newcastle University Business School.

This article was first published in the April 2021 issue of Impact. 

Ahmed W, Vidal-Alaball J, Downing J, Segui F L. ‘COVID-19 and the 5G Conspiracy Theory: Social Network Analysis of Twitter Data’. Journal of Medical Internet Research 2020, 22( 5 ), e19458.

Ahmed W, López Seguí F, Vidal-Alaball J, Katz S M. ‘COVID-19 and the “Film Your Hospital” Conspiracy Theory: Social Network Analysis of Twitter Data’. Journal of Medical Internet Research 2020, 22( 10 ), e22374.