Presence of cognitive bias cues in COVID-19 fake news on social media and its role in individuals' vulnerability to believing them

Abstract

The COVID-19 global pandemic has had a severe impact on nations, causing numerous deaths, severe economic losses, and disruptions in everyday life. One concern is the rising amount of COVID-19 related fake news that has penetrated the digital space, causing confusion and uncertainty during these times. Examining this issue from a psychological and communications sciences lens is important in combating and deterring such fake news. However, there is little research into how fake news may contain cues that result in the use of cognitive biases when consuming them. This study examined the role of the authority bias, negativity bias, and confirmation bias, in individuals’ consumption of COVID-19 related fake news in comparison to real news. This was achieved using a text mining approach to extract cues of the three biases, from a dataset of 500 fake news and 500 real news collected from social media. ANOVA results found a stronger presence of authority and confirmation bias cues in COVID-19 related fake news than real news. Further, logistic regression found several cognitive bias cues, which include message length, word length, use of negation words, first-person pronouns, numerics, and emotions of hate and torment, to be significant in the detection of fake news. These findings have implications for areas of fake news interventions of improving fake news corrections, media literacy, and automated fact-checking techniques.