Towards Psychological Herd Immunity: Cross-cultural Evidence for Two Prebunking Interventions against COVID-19 Misinformation

"...findings add to the emerging insight that interventions informed by behavioural science are a crucial tool to help mitigate the spread of misinformation."

Misinformation about COVID-19 has been found to lead to a variety of negative outcomes, such as reducing willingness to comply with health guidelines and affecting vaccine uptake intentions around the world. Preemptively debunking ("prebunking") misinformation may be a promising strategy for building attitudinal resistance against misinformation. Prebunking is a key component of inoculation theory, which is based on a biological analogy of the immunisation process. This 2-part study tests two technique-based prebunking interventions aimed at improving people's ability to spot misinformation about COVID-19.

The first intervention is Go Viral!, a freely available five-minute choice-based browser game that exposes three manipulation techniques commonly used in COVID-19 misinformation: fearmongering, using fake experts, and spreading conspiracy theories. Available in three languages (English, French, and German), Go Viral! functions as an active inoculation against future manipulation attempts by preemptively warning and exposing people to weakened doses of COVID-19 misinformation and letting them generate their own psychological "antibodies".

In Study 1, the researchers implemented a voluntary pre-post survey within the Go Viral! game. Consenting participants (n-1,771) were shown three misinformation and three real news social media posts (tweets) relating to COVID-19, and they were asked to rate the manipulativeness of each post on a 1-7 Likert scale (1 being "not at all" and 7 being "very"). After completing the game, players were again asked to rate the manipulativeness of the same social media posts they saw in the pre-test, and they were presented with a series of demographic questions.

Study 1 revealed that Go Viral! players found misinformation about COVID-19 to be significantly more manipulative after the game than before, whereas their assessment of real news did not change in a meaningful sense. This finding held irrespective of their demographic background, aside from political ideology (identifying as left-wing is associated with a higher post-pre inoculation effect in terms of veracity discernment than people who identify as right-wing). The absence of a randomised control group allows for limited causal inference, and the survey was run in one language (English); these issues were addressed through Study 2.

The second prebunking intervention consists of a series of infographics about COVID-19 misinformation. As part of its #ThinkBeforeSharing prebunking campaign, the United Nations Educational, Scientific and Cultural Organization (UNESCO), with input from inoculation researchers, created a social media package of images that explain how COVID-19 misinformation is created and spreads.

For Study 2, a preregistered randomised controlled trial, the active (inoculation) condition involved participants playing Go Viral! The passive (infographics) condition involved reading through the UNESCO infographics. The control condition involved attentively playing the puzzle game Tetris for a minimum of five minutes, approximately the same amount of time it takes to complete Go Viral! To begin, participants (n=1,777) performed an item-rating task, where they were randomly shown nine real news posts and nine misinformation posts (three per manipulation technique: fearmongering, fake experts and conspiracy). For each post, participants rated the following statements on a 1-7 scale: (i) this post is manipulative; (ii) I am confident in my assessment of this post's manipulativeness (attitudinal certainty); (iii) I would share this post with people in my network. After completing this task, participants were randomly assigned to one of the treatment conditions (active inoculation or infographics) or the control condition (1:1:1). Next, participants were given additional tasks and questions, as outlined in the paper.

Study 2 found that both prebunking interventions (inoculation and infographics) significantly increase the perceived manipulativeness of misinformation about COVID-19, compared to the control group. This result is in line with Study 1 and remained valid in a randomised controlled setting and across three different languages. Go Viral! participants rated misinformation about COVID-19 as significantly more manipulative one week after the intervention. They were also significantly more confident in their judgments and experienced more motivational threat to defend their attitudes. Unlike Study 1, playing Go Viral! increased the perceived manipulativeness of real news immediately after the intervention, whereas this effect is not observed for the infographics. However, this scepticism of real news among Go Viral! participants dissipates entirely after one week, unlike for misinformation.

Furthermore, people were significantly more willing to share the Go Viral! game with others in their social media network than the infographics, which points towards a potential relative benefit of active versus passive prebunking interventions. Though playing Go Viral! significantly reduced willingness to share misinformation about the virus, this finding is not significant at the country level (although directionally similar). Furthermore, this effect was no longer significant after one week, and there was no difference in sharing willingness for the UNESCO infographics.

In short, Go Viral! was found across the studies to have (a) increased the perceived manipulativeness of misinformation about COVID-19, (b) improved people's attitudinal certainty (confidence) in their ability to spot misinformation, and (c) reduced self-reported willingness to share misinformation with others. The first two effects remain significant for at least one week after gameplay. Also, reading real-world infographics from UNESCO improves people's ability and confidence in spotting COVID-19 misinformation (albeit with descriptively smaller effect sizes than the game).

Editor's notes:

• The researchers' Open Science Framework (OSF) page includes information for those seeking to replicate their findings and methods, including datasets, Qualtrics surveys, the full list of items (social media posts), preregistrations, supplementary tables, figures and analyses, and analysis and visualisation scripts.
• This article is a part of special theme issue on Studying the COVID-19 Infodemic at Scale. To see a full list of all articles, please click here.