CHAPTER 10

Misinformation and Conspiracies in COVID Times

Aaron Clark-Ginsberg, Luke J. Matthews, Carlos Villegas, and Joie Acosta

Introduction

In the early stages of the COVID-19 pandemic the World Health Organization raised concerns about the growing “infodemic”: an overabundance of information, both accurate and inaccurate, that can make finding trustworthy sources incredibly difficult.1 Indeed, misinformation and conspiracies abounded during the pandemic, covering everything from masking, the origins of the virus, rationales for interventions, and other rumors. While the level of harm has varied by topic and population, collectively misinformation and conspiracy beliefs contribute to the growing mistrust of official government response agencies.

The purpose of this chapter is to understand how social networking sites (SNSs) contributed to the infodemic during the pandemic and provide paths forward for combating online disaster-related disinformation and misinformation. Online platforms such as Facebook, Twitter (now X), and Instagram that people use to build social networks are critical sites of information sharing, including the sharing of misinformation. Today over 70 percent of Americans use social media, with most users visiting at least one SNS each day.2 Globally, over 50 percent of the population use SNSs.3

SNSs have many benefits in the context of risk and disaster. Households are turning to these sites to organize and support each other during crisis, using them to communicate, maintain situational awareness, and share resources,4 a digital extension of the well-documented phenomena of emergent, democratized responses to disaster.5 Indeed, SNSs can be a key tool for sharing and amplifying messages widely to enhance knowledge on public health topics and motivate health-promoting behaviors or practices.6 The utility and influential role of SNSs has been especially critical during public health emergencies, where the large-scale and timely dissemination of information can be lifesaving.7 As these platforms easily connect many individuals together, they can also be used to support interactions between households and broader response agencies, both governmental and nongovernmental.8

However, at the same time, SNSs increasingly are being used to spread misinformation and disinformation about public health emergencies,9 which can consequently erode the trustworthiness of public health science, officials, and institutions. Of concern, vulnerable populations are disproportionately targeted by misinformation about public health emergencies, which may worsen health inequities.10 Vulnerable populations may also have less access to SNS information, as there is a very real digital divide, with some people on SNSs and others not.11 These divides extend to emergencies, situations in which locations with uneven SNS connectivity, and certain groups such as the elderly and information-poor who are less likely to use social media in disasters.12 As a result, vulnerable groups may be less able to leverage the resources that SNSs provide during disasters. Problems are further compounded by uneven use by response agencies,13 in part because effective utilization requires insights into both the communication preferences of different groups and norms and practices for using SNSs that many responders may not have.14

To meaningfully analyze these information dynamics, it is important to be clear about what is meant by disinformation and misinformation. Most studies of misinformation characterize it as false information believed to be true by its main disseminators, while disinformation is defined as false information known to be false by the disseminator, who spreads it for ulterior motives.15 For most disinformers, whether they are state propagandists or seeking monetary profit, disinformation ultimately has to go viral online for them to achieve their goals; that is, the original disinformation must convert into misinformation spread by believers.16 Analyzing the underlying beliefs and motives of online actors is a laborious task, so for the purposes of this chapter we will use the term “misinformation” as a catch-all for content that likely is misinformation but potentially could be disinformation.

Studies of misinformation must grapple with the reality that some misinformation comprises belief in ideas that are inherently difficult to verify as false or true. For example, the evolutionary origins of particular diseases are inherently difficult to study scientifically because they constitute unique and particular events.17 Even after decades of research to uncover the likely zoonotic origins of HIV, the origins are still debated today.18

We sought to understand the spread of misinformation via SNSs during the pandemic through social network analysis of online discussions on Twitter about one of the most rampant topics of misinformation: the use of hydroxychloroquine to treat COVID-19. Social network analysis is useful for mapping out and assessing connections (or “edges”) between groups or individuals (“nodes”),19 making it well suited as a technique for understanding the spread of misinformation online, and has also been used widely in the study of disasters to better understand an array of topics.20

Approach and Empirical Findings

The use of hydroxychloroquine to treat COVID-19 emerged as a therapeutic suggestion near the very start of the pandemic.21 Originating from studies in China and then Europe, the suggested use of hydroxychloroquine shot to prominence in the United States partly due to its early endorsement by President Donald Trump. The Food and Drug Administration initiated an emergency use authorization for hydroxychloroquine in March 2020,22 which it ended by June 2020 when scientific studies conclusively refuted its efficacy as a treatment.23 Hydroxychloroquine lived on as a therapy long after the emergency use authorization was revoked, however, in large part due to the continued championing of the treatment by conservative media personalities and by some physicians with public followings. This made it a highly pertinent focus for our research.

We sampled hydroxychloroquine-related messages posted on Twitter for the entirety of 2020. This period was critical, as it included the time when hydroxychloroquine gained prominence and after it was revoked as a valid COVID-19 treatment. We used the following query formula to identify relevant tweets, performing a manual thematic coding process to determine relevance and establish our network:

(plaquenil OR quineprox OR hydroxy OR hydroxychloroquine OR chloroquine OR quinoline OR quinine OR HCQ)

We then followed fairly standard processes for social network analysis of SNSs,24 with the exception that we focused on nodes that each had at least one incoming and one outgoing connection to another node in the dataset via a tweet at (@) or retweet. Our rationale for this was that any conversation involves give-and-take, meaning those involved in the conversation on Twitter both mentioned another and were mentioned by others (not necessarily the same others, however).

Retweets and tweets@ are the most common forms of mention of one Twitter user by another and thus served to establish incoming and outgoing connections for our study. A core discussion network of 4,621 users reflected the single-largest connected subnetwork in which each user can be reached by some number of social steps between them (Figure 10.1). This alone turned out to be a key insight of our study, because it showed that hydroxychloroquine misinformation flowed among only a few thousand users (4,621 in total). Despite hundreds of thousands of users (534,096) tweeting about it, information primarily flowed through a handful of well-connected actors at the center (Table 10.1). The most central actors all belonged to a single network community (Community 2), which are densely connected regions of the overall network. The central actors are predominantly representatives of the media. This community also included representation of other groups: medical doctors, politicians, and some members of the public.

These findings suggest that contrary to broadly held perceptions, social media has done little to democratize the flow of information as compared to previous journalism models (e.g., newspapers, broadcast news). Instead, for any given topic, social media has simply replaced a few thousand people working at vetted publications such as the New York Times and the Wall Street Journal with a few thousand people mostly lacking both professional credentials and proper journalistic standards who also usually are not typical members of the public. To our knowledge, this is a novel insight from this study.

We also aimed to understand who could function as a bridge between different communities using betweenness centrality, a measure of how a node is situated in the structure of the overall network (Table 10.2). High betweenness nodes sit at positions in the network that connect otherwise disconnected groups. In our analysis, we found that users with high betweenness represented more diverse types of actors compared to the best-connected actors in the core discussion network (e.g., activists, social media influencers).

Discussion and Conclusions

In the social structure and hierarchy of the Twitter ecosystem, nonexpert accounts occupied central positions in the hydroxychloroquine network. For instance, @seanhannity functions like an expert in the core discussion network even though he possesses no particular expertise on medical issues. This is at least in part because @seanhannity had tremendous connectivity and likely influence over this population’s perceptions of hydroxychloroquine.

The network does not function as an egalitarian society of citizen-researchers all working to uncover the truth about hydroxychloroquine for themselves. The network also is not a tool for the democratization of research but is merely the replacement of credentialed experts with earned expertise (e.g., career scientists at the Centers for Disease Control and Prevention [CDC]) with opinion-led individuals (such as media personalities). This underscores that what needs to occur is the renewed building of trusted relationships between the audience and experts with the relevant expertise for a given topic.

To combat public health emergency-related misinformation, public health officials, emergency managers, and other responders must find a way to change these discussions and ultimately elevate credible and consistent voices who also have skill in risk communication. This could be done through direct outreach from various public health departments and related agencies mandated to address public health crises through both medical and social action. However, our analysis found that these responders were largely absent from our lists of the most influential users. Given the decades of underinvestment in both public health and emergency management across the United States,25 lack of influence might be attributed to limitations in staff with the time and expertise to directly engage in SNS messaging. It may also require a social network–savvy strategy. Instead of engaging directly in SNS messaging to attempt to break into an influential position within a network, messages might better be relayed through trusted intermediaries. Given the critical role of intermediaries in SNS communication (including national media and experts and social media influencers), this could be more effective.

Combating misinformation during an emergency will likely require multiple strategies. One strategy will work to communicate trusted information to local populations that have self-identified as being interested in and ready to act upon this information. Different communication and trust-building approaches will be needed to reach populations who may not implicitly trust public health officials. Someone already convinced that vaccines are safe and effective and that hydroxychloroquine is not likely needs only simple messages from public health authorities as to where and when they can receive a newly available vaccination. Others might be on the fence about receiving a vaccine because of specific concerns about its safety or the potential impacts on, for example, fertility. These populations may require more tailored strategies based on addressing specific concerns and building trust through, for instance, local community-based organizations and leaders who are already trusted by these populations.

Some individuals or groups might remain consistently resistant to certain types of messages. They might push back against any vaccinations and thus are active in their resistance to a new vaccine, such as for COVID-19. Changing their views may require a more intensive and long-term approach to trust-building. Since public officials dealing with crises and disasters often have extremely limited resources, there is an argument to be made for focusing on self-identified supporters and those who are on the fence about whether to take public health action rather than trying to change the hearts and minds of those with more resistant or entrenched views. However, these populations can also undermine public health’s trustworthiness and impede effective public health response to and recovery from an emergency, as our analysis suggests. For example, people continue to decline the COVID-19 vaccines, making it more likely that the virus will continue to cause potentially avoidable morbidity and mortality. Therefore, to change behaviors, public health officials must build their capacity to address misinformation efforts that are actively impeding public health emergency response efforts.

It is important to emphasize as well that the three attitude groups we just described are a heuristic device. Antivaccine beliefs were in fact distributed on a continuum among a nationally representative sample of US parents prior to the COVID-19 pandemic,26 in spite of our collective perception that there are two camps of belief.

Relatedly, what is truthful information versus misinformation is not always clear, and this is especially the case for topics that are inherently difficult to study scientifically, such as disease origins.27 Across the continuum of antivaccine belief, beliefs in true and false side effects and true and false conspiracies were correlated;28 that is, individuals tended to believe or disbelieve in side effects/conspiracies irrespective of the evidence for each particular side effect or conspiracy. While this can be interpreted to indicate that individuals believe in side effects and conspiracies because they are, a priori, motivated to distrust medical science, it can equally be interpreted to mean that individuals disbelieve misinformation only because they are equally motivated to trust medical science rather than engage in any evaluation of evidence because they also disbelieve actual side effects and conspiracies.

Our analysis documents that social media networks, contrary to intuition and to the gestalt messages of executives at Meta and Twitter (now X), have not democratized global conversations. Instead, these networks have allowed nonexpert opinion–led individuals to dominate. Fortunately, the structure of online communications may also make it possible for experts to reassert themselves with the right strategies. The trust rebuilding we are recommending is achievable because messages from groups such as the CDC are not automatically drowned out by the multitude. If anything, social media may have contracted global conversations. Using a very simple and intuitive definition of “conversation”—that a user both mentioned other users and is mentioned themselves—our analysis shows that only a few thousand users controlled the entirety of the hydroxychloroquine conversation on Twitter. Given the prominence of the topic during the period sampled, we suspect that our findings will generalize to many other instances of misinformation online. A few thousand individuals are comparable to the total number of employees at a large traditional media agency; for example, the New York Times employs about five thousand individuals globally.29

It is for this reason that we emphasize the potential role for novel kinds of brokers (e.g., conservative national media) with whom public health officials may have had limited partnerships in the past. The empirical analyses summarized in this chapter highlight the specific kinds of online actors who could at least be broached as change agents to create a more effective public health communication for future emergencies.

Funding Statement

Research reported in this book chapter has been funded by the US Centers for Disease Control and Prevention, an agency of the Department of Health and Human Services, under CDC contract 75D30122C14256. The findings and conclusions in this book chapter are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Notes

1. 1.  World Health Organization, Novel Coronavirus(2019-nCoV) Situation Report—13, February 2, 2020, https://www.who.int/docs/default-source/coronaviruse/situation-reports/20200202-sitrep-13-ncov-v3.pdf.
2. 2.  “Social Media Fact Sheet,” Pew Research Center, June 12, 2019, https://www.pewresearch.org/internet/fact-sheet/social-media/.
3. 3.  Jacob Poushter, Caldwell Bishop, and Hanyu Chwe, Social Media Use Continues to Rise in Developing Countries but Plateaus Across Developed Ones, Pew Research Center, June 19, 2018, https://assets.pewresearch.org/wp-content/uploads/sites/2/2018/06/15135408/Pew-Research-Center_Global-Tech-Social-Media-Use_2018.06.19.pdf.
4. 4.  DeeDee M. Bennett, “How Do Emergency Managers Use Social Media Platforms?,” Journal of Emergency Management 12, no. 3 (2014): 251–56, https://doi.org/10.5055/jem.2014.0178; Chao Fan, Yucheng Jiang, and Ali Mostafavi, “Emergent Social Cohesion for Coping with Community Disruptions in Disasters,” Journal of the Royal Society Interface 17, no. 164 (2020), https://doi.org/10.1098/rsif.2019.0778; Mark E. Keim and Eric Noji, “Emergent Use of Social Media: A New Age of Opportunity for Disaster Resilience,” American Journal of Disaster Medicine 6, no. 1 (2011): 47–54, https://doi.org/10.5055/ajdm.2011.0044; Carmen Mei Ling Leong, Shan L. Pan, Peter Ractham, and Laddawan Kaewkitipong, “ICT-Enabled Community Empowerment in Crisis Response: Social Media in Thailand Flooding 2011,” Journal of the Association for Information Systems 16, no. 3 (2015), https://doi.org/10.17705/1jais.00390; Jooho Kim and Makarand Hastak, “Social Network Analysis: Characteristics of Online Social Networks After a Disaster,” International Journal of Information Management 38, no. 1 (2018): 86–96, https://doi.org/10.1016/j.ijinfomgt.2017.08.003; Carlos A. Villegas and Matthew J. Martinez, “Lessons from Harvey: Improving Traditional Damage Estimates with Social Media Sourced Damage Estimates,” Cities 121 (February 2022): 103500, https://doi.org/10.1016/j.cities.2021.103500; and Zheye Wang, Xinyue Ye, and Ming-Hsiang Tsou, “Spatial, Temporal, and Content Analysis of Twitter for Wildfire Hazards,” Natural Hazards 83, no. 1 (2016): 523–40, https://doi.org/10.1007/s11069-016-2329-6.
5. 5.  John Twigg and Irina Mosel, “Emergent Groups and Spontaneous Volunteers in Urban Disaster Response,” Environment and Urbanization 29, no. 2 (2017): 443–58, https://doi.org/10.1177/0956247817721413; Thomas E. Drabek and David A. McEntire, “Emergent Phenomena and Multiorganizational Coordination in Disasters: Lessons from the Research Literature,” International Journal of Mass Emergencies and Disasters 20, no. 2 (2002): 197–224, https://doi.org/10.1177/028072700202000208; Thomas E. Drabek and David A. McEntire, “Emergent Phenomena and the Sociology of Disaster: Lessons, Trends and Opportunities from the Research Literature,” Disaster Prevention and Management 12, no. 2 (2003): 97–112, https://doi.org/10.1108/09653560310474214; and Sarah Lee Saunders and Gary A. Kreps, “The Life History of the Emergent Organization in Times of Disaster,” Journal of Applied Behavioral Science 23, no. 4 (1987): 443–62, https://doi.org/10.1177/002188638702300402.
6. 6.  S. Anne Moorhead, Diane E. Hazlett, Laura Harrison, Jennifer K. Carroll, Anthea Irwin, and Ciska Hoving, “A New Dimension of Health Care: Systematic Review of the Uses, Benefits, and Limitations of Social Media for Health Communication,” Journal of Medical Internet Research 15, no. 4 (2013): e85, https://doi.org/10.2196/jmir.1933.
7. 7.  Yan Wang, Haiyan Hao, and Lisa Sundahl Platt, “Examining Risk and Crisis Communications of Government Agencies and Stakeholders During Early-Stages of COVID-19 on Twitter,” Computers in Human Behavior 114 (January 2021): 106568, https://doi.org/10.1016/j.chb.2020.106568; Melissa MacKay, Taylor Colangeli, Abhinand Thaivalappil, Ariana Del Bianco, Jennifer McWhirter, and Andrew Papadopoulos, “A Review and Analysis of the Literature on Public Health Emergency Communication Practices,” Journal of Community Health 47 (February 2022): 150–62, https://doi.org/10.1007/s10900-021-01032-w; Stine Eckert et al., “Health-Related Disaster Communication and Social Media: Mixed-Method Systematic Review,” Health Communication 33, no. 12 (2018): 1389–1400, https://doi.org/10.1080/10410236.2017.1351278.
8. 8.  Yu Xiao, Qunying Huang, and Kai Wu, “Understanding Social Media Data for Disaster Management,” Natural Hazards 79 (2015): 1663–79, https://doi.org/10.1007/s11069-015-1918-0; Tomer Simon, Avishay Goldberg, and Bruria Adini, “Socializing in Emergencies: A Review of the Use of Social Media in Emergency Situations,” International Journal of Information Management 35, no. 5 (2015): 609–19, https://doi.org/10.1016/j.ijinfomgt.2015.07.001; and Yury Kryvasheyeu et al., “Rapid Assessment of Disaster Damage Using Social Media Activity,” Science Advances 2, no. 3 (2016): e1500779, https://doi.org/10.1126/sciadv.1500779.
9. 9.  Nashit Chowdhury, Ayisha Khalid, and Tanvir C. Turin, “Understanding Misinformation Infodemic During Public Health Emergencies Due to Large-Scale Disease Outbreaks: A Rapid Review,” Journal of Public Health 31, no. 4 (2023): 553–73, https://doi.org/10.1007/s10389-021-01565-3.
10. 10.  Brian G. Southwell, Jessica Otero Machuca, Sabrina T. Cherry, Melissa Burnside, and Nadine J. Barrett, “Health Misinformation Exposure and Health Disparities: Observations and Opportunities,” Annual Review of Public Health 44 (2023): 113–30, https://doi.org/10.1146/annurev-publhealth-071321-031118.
11. 11.  Poushter, Bishop, and Chwe, Social Media Use Continues to Rise in Developing Countries but Plateaus across Developed Ones.
12. 12.  Xiao et al., “Understanding Social Media Data for Disaster Management.”
13. 13.  Bennett, “How Do Emergency Managers Use Social Media Platforms?”
14. 14.  Simon et al., “Socializing in Emergencies.”
15. 15.  See Michael J. Mazarr, Abigail Casey, Alyssa Demus, Scott W. Harold, Luke J. Matthews, Nathan Beauchamp-Mustafaga, and James Sladden, Hostile Social Manipulation: Present Realities and Emerging Trends (RAND Corporation, 2019), https://www.rand.org/pubs/research_reports/RR2713.html/.
16. 16.  For a fuller discussion, see Luke J. Matthews and Paul Robertson, Theorizing the Anthropology of Belief: Magic, Conspiracies, and Misinformation (Routledge, 2024).
17. 17.  Charles L. Nunn, The Comparative Approach in Evolutionary Anthropology and Biology (University of Chicago Press, 2011).
18. 18.  Lutz G. Gürtler and Josef Eberle, “Aspects on the History of Transmission and Favor of Distribution of Viruses by Iatrogenic Action: Perhaps an Example of a Paradigm of the Worldwide Spread of HIV,” Medical Microbiology and Immunology 206 (2017): 287–93, https://doi.org/10.1007/s00430-017-0505-2; and William Carlsen, “Did Modern Medicine Spread an Epidemic?,” San Francisco Chronicle, January 15, 2001, http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2001/01/15/MN162301.DTL&type=science.
19. 19.  Julie K. Maldonado, “The Practical and Policy Relevance of Social Network Analysis for Disaster Response, Recovery, and Adaptation,” in Social Network Analysis of Disaster Response, Recovery, and Adaptation, ed. Eric C. Jones and A. J. Faas (Elsevier, 2017); and Stanley Wasserman and Katherine Faust, Social Network Analysis: Methods and Applications (Cambridge University Press, 1994).
20. 20.  Aaron Clark-Ginsberg, “Disaster Risk Reduction Is Not ‘Everyone’s Business’: Evidence from Three Countries,” International Journal of Disaster Risk Reduction 43 (February 2020): 101375, https://doi.org/10.1016/j.ijdrr.2019.101375; Citra S. Ongkowijoyo and Hemanta Doloi, “Understanding of Impact and Propagation of Risk Based on Social Network Analysis,” Procedia Engineering 212: 1123–30, https://doi.org/10.1016/j.proeng.2018.01.145; Branda Nowell, Toddi Steelman, Anne-Lise K. Velez, and Zheng Yang, “The Structure of Effective Governance of Disaster Response Networks: Insights from the Field,” American Review of Public Administration 48, no. 7 (2018): 699–715, https://doi.org/10.1177/0275074017724225; Kim and Hastak, “Social Network Analysis”; Eric C. Jones and A. J. Faas, “An Introduction to Social Network Analysis in Disaster Contexts,” in Social Network Analysis of Disaster Response, Recovery, and Adaptation, ed. Eric C. Jones and A. J. Faas (Elsevier, 2017); Maldonado, “The Practical and Policy Relevance of Social Network Analysis,” 255–67; Danielle M. Varda, “Strategies for Researching Social Networks in Disaster Response, Recovery, and Mitigation,” in Social Network Analysis of Disaster Response, Recovery, and Adaptation, ed. Eric C. Jones and A. J. Faas (Elsevier, 2017); and Uuf Brajawidagda and Akemi Chatfield, “Twitter Tsunami Early Warning Network: A Social Network Analysis of Twitter Information Flows,” ACIS 2012 Proceedings 56 (2012), https://aisel.aisnet.org/acis2012/56.
21. 21.  Ilan S. Schwartz, David R. Boulware, and Todd C. Lee, “Hydroxychloroquine for COVID19: The Curtains Close on a Comedy of Errors,” The Lancet Regional Health–Americas 11 (July 2022): 100268, https://doi.org/10.1016/j.lana.2022.100268; and Erisa Alia and Jane M. Grant-Kels, “Does Hydroxychloroquine Combat COVID-19? A Timeline of Evidence,” Journal of the American Academy of Dermatology 83, no. 1 (2020): e33–e34, https://doi.org/10.1016/j.jaad.2020.04.031.
22. 22.  Denise M. Hinton, “Request for Emergency Use Authorization for Use of Chloroquine Phosphate or Hydroxychloroquine Sulfate Supplied from the Strategic National Stockpile for Treatment of 2019 Coronavirus Disease,” letter to Rick Bright, Food and Drug Administration, March 28, 2020, https://www.fda.gov/media/136534/download.
23. 23.  US Food and Drug Administration, “Coronavirus (COVID-19) Update: FDA Revokes Emergency Use Authorization for Chloroquine and Hydroxychloroquine,” news release, June 15, 2020, https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-revokes-emergency-use-authorization-chloroquine-and; and US Food and Drug Administration, “FDA Cautions against Use of Hydroxychloroquine or Chloroquine for COVID-19 Outside of the Hospital Setting or a Clinical Trial Due to Risk of Heart Rhythm Problems,” last modified July 1, 2020, https://www.fda.gov/drugs/drug-safety-and-availability/fda-cautions-against-use-hydroxychloroquine-or-chloroquine-covid-19-outside-hospital-setting-or.
24. 24.  We processed the Twitter data into networks and calculated all network measures using R Version 4.2.3 with the igraph package. We visualized the networks using Gephi, an open-source software designed for social network analysis visualization and analysis, using the ForceAtlas algorithm, a force-directed layout for network visualization. Wasserman and Faust, Social Network Analysis; Varda, “Strategies for Researching Social Networks”; and Mathieu Jacomy, Tommaso Venturini, Sebastien Heymann, Mathieu Bastian, “ForceAtlas2, a Continuous Graph Layout Algorithm for Handy Network Visualization Designed for the Gephi Software,” PLOS One 9, no. 6 (2014): e98679, https://doi.org/10.1371/journal.pone.0098679.
25. 25.  Nason Maani and Sandro Galea, “COVID-19 and Underinvestment in the Public Health Infrastructure of the United States,” Milbank Quarterly 98, no. 2 (2020), 250–59, https://doi.org/10.1111/1468-0009.12463; Sandro Galea and Roger Vaughan, “Preparing the Public Health Workforce for the Post–COVID-19 Era,” American Journal of Public Health 111, no. 3 (2021): 350–52, https://ajph.aphapublications.org/doi/abs/10.2105/AJPH.2020.306110; and Daniel H. Xu and Rashmita Basu, “How the United States Flunked the COVID-19 Test: Some Observations and Several Lessons,” American Review of Public Administration 50, nos. 6–7 (2020): 568–76, https://doi.org/10.1177/0275074020941701.
26. 26.  Luke J. Matthews and Paul Robertson, Theorizing the Anthropology of Belief: Magic, Conspiracies, and Misinformation (Routledge, 2024).
27. 27.  See the introduction to this chapter and also Matthews and Robertson, Theorizing the Anthropology of Belief.
28. 28.  Luke J. Matthews et al., “Belief Correlations with Parental Vaccine Hesitancy: Results from a National Survey,” American Anthropologist 124, no. 2 (2022): 291–306, https://doi.org/10.1111/aman.13714.
29. 29.  “About,” New York Times, accessed October 13, 2023, https://www.nytco.com.