Not-so-straightforward links between believing in COVID-19-related conspiracy theories and engaging in disease-preventive behaviours

The COVID-19 pandemic saw a surge in related conspiracy theories, prompting questions about their influence on disease-preventive behaviors. A common assumption was that belief in these theories would decrease preventive actions. This study aimed to investigate the validity of this assumption and explore the possibility of a reversed link—where certain conspiracy theories might be associated with increased preventive behaviors. The study first defined two main types of COVID-19 conspiracy theories: risk-acceptance conspiracy theories (RA-CTs), which accept the pandemic as a serious threat (e.g., COVID-19 as a man-made bioweapon), and risk-rejection conspiracy theories (RR-CTs), which dismiss the threat (e.g., COVID-19 as no worse than influenza). Previous research indicated conflicting links between these two types of theories and behavior; some studies found that RA-CT beliefs correlated with self-centered prepping behaviors, while others linked RR-CT beliefs with decreased disease-prevention behaviors. This study sought to clarify these conflicting findings by exploring the psychological mechanisms underlying the relationship between these two types of conspiracy theories and disease-preventive behaviors. The researchers hypothesized that RA-CT beliefs would be positively associated with risk perception and negative emotions towards COVID-19, leading to increased preventive behaviors. Conversely, they predicted that RR-CT beliefs would be negatively associated with risk perception and negative emotions, leading to decreased preventive behaviors. Furthermore, they investigated whether the stage of the pandemic outbreak moderated the associations between conspiracy theory beliefs and behaviors, suggesting that these associations might differ between earlier (less severe) and later (more severe) stages of the outbreak. Understanding these complex relationships is crucial for effective public health crisis management.

The study reviewed existing literature on the relationship between conspiracy theories and behavior, particularly concerning COVID-19. Studies by Imhoff and Lamberty (2020) and Sternisko et al. (2020) indicated that belief in COVID-19 as a human-made crisis was linked to more self-centered prepping, while belief in it as a hoax was linked to less disease-preventive behavior. However, inconsistencies existed, with some studies showing that believing in COVID-19 as a human-made crisis correlated with more handwashing but also less social distancing. The authors highlighted that conspiracy theories offer causal explanations for crises, potentially influencing individuals' risk perceptions and negative emotions towards the event. This informed their hypotheses about the relationships between RA-CTs and RR-CTs, risk perception, negative emotions, and disease-preventive behaviors. The existing research suggested the need for a more nuanced understanding of the complex relationship between belief in different types of conspiracy theories and behavior, and how this relationship might change over the course of a pandemic. This study aimed to address this gap by examining the mediating role of risk perception and negative emotions, and the moderating role of the outbreak stage.

The researchers conducted a repeated cross-sectional survey across six time points in the United States, from early February (mild stage) to late March 2020 (severe stage). The survey was administered every ten days, targeting 250 adult US participants per time point using Cloudresearch, recruiting from Amazon Mechanical Turk. 1515 participants completed the survey, with 7 removed for failing attention checks. The survey included measures for: Belief in COVID-19 Conspiracy Theories (five items for RA-CTs and two for RR-CTs, rated on an 11-point scale); Risk Perception (three items about the likelihood of infection, rated on a 7-point scale); Negative Emotions towards COVID-19 (eight items measuring fear, anxiety, etc., rated on a 7-point scale, with positive emotions analyzed separately); Disease-Preventive Behaviors (ten behaviors, rated on a 5-point scale); Institutional Trust (four items measuring trust in institutions, rated on a 7-point scale); Political Ideology (average of liberal and conservative ideology scores); and Demographic Information (gender, age, education, income, socioeconomic status, ethnicity, political party affiliation). Three sets of analyses were conducted using SPSS 25.0: zero-order correlation analysis to examine bivariate relationships; parallel mediation analysis using PROCESS (Hayes, 2013) to examine the simultaneous mediating effects of risk perception and negative emotions on the relationship between conspiracy theory beliefs and disease-preventive behaviors, controlling for covariates; and moderation analysis to examine how the outbreak stage moderated the relationships between conspiracy theory beliefs and outcomes. The outbreak stage was coded as a continuous variable (1 to 6).

The pooled sample correlation analysis showed a positive correlation (r = 0.19, p < 0.001) between belief in RA-CTs and disease-preventive behaviors and a negligible correlation (r = -0.01, p = 0.752) between belief in RR-CTs and disease-preventive behaviors. However, across the six time points, these correlations varied. The parallel mediation analysis revealed significant indirect effects of RA-CT beliefs on disease-preventive behaviors via risk perception and negative emotions. Similarly, RR-CT beliefs showed significant indirect negative effects on behaviors via the same mediators. The moderation analysis revealed that the relationship between RA-CT beliefs and risk perception/negative emotions was stronger during the mild stage of the outbreak, while the relationship between RR-CT beliefs and risk perception was stronger during the severe stage. Specifically, the interaction between belief in RA-CTs and outbreak stage was negatively and marginally significant in predicting risk perception (b = -0.07, SE = 0.04, p = 0.089) and negatively significant in predicting negative emotions (b = -0.12, SE = 0.04, p = 0.004). The interaction between belief in RR-CTs and outbreak stage was negatively significant in predicting risk perception (b = -0.15, SE = 0.04, p < 0.001), but non-significant for negative emotions. Moderated mediation analysis showed that the indirect effect of RA-CTs via negative emotions was significantly higher in the less severe stage, and the indirect effect of RR-CTs via risk perception was significantly lower in the less severe stage. Institutional trust was positively related to disease-preventive behavior but unrelated to belief in the conspiracy theories.

The findings revealed complex relationships between belief in COVID-19 conspiracy theories and disease-preventive behaviors, shaped by the content of the theories and the context (outbreak severity). RA-CT beliefs were associated with higher risk perception and negative emotions, leading to more preventive behaviors, particularly during the early, mild stage. Conversely, RR-CT beliefs were associated with lower risk perception and emotions, leading to fewer preventive behaviors, particularly during the later, severe stage. The study highlights three key contributions: First, it shows that different conspiracy theory contents have different implications for perceptions and behaviors; second, it demonstrates the importance of context in shaping these relationships; and third, it contributes to the understanding of how the outbreak stage moderates the link between conspiracy beliefs and outcomes. The authors acknowledge that RA-CT beliefs, while potentially increasing precautions, might also contribute to prejudice and hostility. The differential impact of the conspiracy theories based on the stage of the pandemic underlines the need to tailor public health messaging and interventions according to the evolving context.

This study demonstrated the nuanced and context-dependent relationships between COVID-19 conspiracy beliefs and disease-preventive behaviors. The findings highlight the importance of considering both the content and the context of conspiracy theories in shaping public health interventions. Future research could explore the causal links between these variables using experimental designs and investigate the relationship between conspiracy beliefs and prejudice.

The correlational nature of the study using a repeated cross-sectional design prevents causal inferences. Differences in participants across time points might also influence the results. The reliance on self-reported behaviors may introduce bias, and the limited scope of conspiracy theories considered could affect the generalizability of the findings. Future studies should employ experimental designs, multiple data collection methods, and a more comprehensive set of conspiracy theories for a more robust understanding.