COVID-19 (Mis)Information Exposure and Messaging Effects in the United Kingdom

Lead Research Organisation: University of Exeter
Department Name: Politics

Abstract

Fighting the COVID-19 pandemic requires understanding what information people have about the disease, which misperceptions might be prevalent, and how officials can improve public knowledge and encourage behaviours that will protect public health. First, this study will measure beliefs and attitudes about COVID-19, providing a thorough map of accurate knowledge, but also of misperceptions, particularly those driven by conspiratorial thinking. Second, the study will catalogue the online sources from which our respondents get COVID-19-related information. This will allow us to gauge which accurate information sources reach a wide audience and which sources of online misinformation are systematically distorting people's views. Third, the study will test whether public health messages that seek to correct misinformation are actually effective in changing people's beliefs.

Behavioural and survey data will be collected in a multi-wave nationally representative survey that measures both prevalence of false and accurate beliefs about COVID-19 (including beliefs in misperceptions and conspiracy theories) and support for recommendations from public health authorities. To evaluate responses to information from public health officials, the second survey wave will include a randomized experiment evaluating the effects of messaging from health and medical authorities. Finally, the study will measure the quality of the information people consume online about the pandemic by analysing the behavioural data provided by respondents using a combination of human-coded and machine learning approaches. These data will make it possible to identify which groups are most frequently exposed to inaccurate or untrustworthy information about COVID-19, which should aid in the design of effective interventions.
 
Description To date via this award we have achieved several key milestones, and there are three accepted papers (Nature Human Behaviour, PNAS Nexus, Journal of Experimental Political Science) with additional papers in progress.

1) We find that fact-checks are effective in addressing COVID-19 related misperceptions. However, these effects do not last and they do not accumulate. These results can be found in our paper published in Nature Human Behaviour "The ephemeral effects of fact-checks on COVID-19 misperceptions in the US, Great Britain, and Canada" (https://doi.org/10.1038/s41562-021-01278-3).

Paper abstract: Widespread misperceptions about COVID-19 and the novel coronavirus threaten to exacerbate the severity of the pandemic. We conducted preregistered survey experiments in the United States, Great Britain and Canada examining the effectiveness of fact-checks that seek to correct these false or unsupported beliefs. Across three countries with differing levels of political conflict over the pandemic response, we demonstrate that fact-checks reduce targeted misperceptions, especially among the groups who are most vulnerable to these claims, and have minimal spillover effects on the accuracy of related beliefs. However, these reductions in COVID-19 misperception beliefs do not persist over time in panel data even after repeated exposure. These results suggest that fact-checks can successfully change the COVID-19 beliefs of the people who would benefit from them most but that their effects are ephemeral.

2) We find in our paper published in the Journal of Experimental Political Science that descriptive norms messages (how people actually behave) can have positive effects on pro-social pandemic behaviors. In general, people underestimate the number of people who engage in pro-social behavior (in this case mask-wearing). Giving people information about the extent of mask-wearing can (but does not always) increase survey measures of mask wearing intentions.

Abstract: Public health officials have faced resistance in their efforts to promote mask-wearing to counter the spread of COVID-19. One approach to promoting behavior change is to alert people to the fact that a behavior is common (a descriptive norm). However, partisan differences in pandemic mitigation behavior mean that Americans may be especially (in)sensitive to information about behavioral norms depending on the party affiliation of the group in question. In July--August 2020, we tested the effects of providing information to respondents about how many Americans, co-partisans, or out-partisans report wearing masks regularly on both mask-wearing intentions and on the perceived effectiveness of masks. Learning that a majority of Americans report wearing masks regularly increases mask-wearing intentions and perceived effectiveness, though the effects of this information are not distinguishable from other treatments.

3) In a paper accepted for publication at PNAS Nexus, we find that injunctive norm information (telling people that most people support getting vaccinated) does not increase vaccination intent.

Abstract: Does information about how other people feel about COVID-19 vaccination affect immunization intentions? We conducted pre-registered survey experiments in Great Britain (5,456 respondents across three survey waves from September 2020 - February 2021), Canada (1,315 respondents in February 2021), and the state of New Hampshire in the United States (1,315 respondents in January 2021). The experiments examine the effects of providing accurate public opinion information to people about either public support for COVID-19 vaccination (an injunctive norm) or public beliefs that the issue is contentious. Across all three countries, exposure to this information had minimal effects on vaccination intentions even among people who previously held inaccurate beliefs about support for COVID-19 vaccination or its perceived contentiousness. These results suggest that providing information on public opinion about COVID vaccination has limited additional effect on people's behavioral intentions when public discussion of vaccine uptake and intentions is highly salient.

There are additional potential papers in various stages (from planning to draft) that examine the following topics:
1) How the COVID vaccination campaign has affected general vaccine sentiment (using a crossly design from the panel study)
2) Whether or not to use "don't know" in vaccine attitude questions
3) Consumption of COVID misinformation online
4) Public reactions to the DARS-CoV-2 Omicron variant
1) The paper and i
Exploitation Route We believe that all of these outcomes have potential relevance to other users.

As for the basic surveillance of accurate and inaccurate coronavirus beliefs, we believe that this can be of consequence to political, health, and science communicators so they know the basic lay of the land for effective communications regarding the coronavirus. For example, the COVID-5G conspiracy belief has attracted a lot of attention by press or circulating on social media, but very very few people actually believe this. Given limited bandwidth to communicate to the public, we believe it is not worth the opportunity costs of addressing these beliefs.

Following on from this, our findings also show the crucial importance of repeatedly correcting the most worrisome misperceptions. Fact-checking works, but the effects fade. Corrective information probably needs to be repeated (perhaps not unlike booster shots).

Finally, our results show that social norm messaging in general has limited effects. This is particularly important for public health communicators because early during the pandemic our view is that this was seen as an important and persuasive way to move public opinion. Our results are far more equivocal. (However, an important possibility is that these efforts were meaningful at motivating behavior, and thus our messages delivered in a survey context could not effectively *further* move public opinion.)
Sectors Communities and Social Services/Policy,Education,Healthcare,Government, Democracy and Justice,Other