Perceived risk likelihood has been considered to be one of the key factors that predict engagement in
health protective behaviors. However, existing studies on perceived risk likelihood and preventive
behaviors have used a mixture of concepts, including perceptions of absolute (i.e., one’s overall
chances of experiencing the risk) versus comparative (i.e., one’s relative chances of experiencing the
risk compared with other people) risk likelihood, and perceptions of conditional (i.e., likelihood
estimated under a condition that specifies future inaction on health behaviors) versus unconditional
(i.e., likelihood that does not specify future behavior engagement) risk likelihood. Therefore, this study
aimed to untangle the mixed conceptualization and past findings by examining the relationshipsbetween each type of perceived risk likelihood and preventive behaviors. We first reviewed the
rationale for the causal relationships between each type of perceived risk likelihood and preventive
behaviors, and critically analyzed the method for verifying the effects of perceived risk likelihood.
Based on the review and critical reasoning, we hypothesized that conditional absolute risk perception
better predicts preventive behaviors in the face of health risks with obvious negative consequences
like infectious diseases, in comparison to other types of risk perceptions, including unconditional
absolute risk perception and comparative risk perception. We further hypothesized that conditional
absolute risk perception is positively associated with preventive behaviors and unconditional absolute
risk perception, while engagement in preventive behaviors has negative effects on unconditional
absolute risk perception when the effect of conditional absolute risk perception is controlled for. A
survey was conducted online in the context of COVID-19 with 445 adults aged 20 to 69 in Korea. The
results supported both hypotheses. First, we found that among the various types of perceived risk
likelihood assessed in the survey, conditional absolute risk perception best predicted COVID-19
preventive behaviors. Conditional absolute risk perception showed the highest correlation with
preventive behaviors compared to unconditional absolute risk perception, direct comparative risk
perception, and indirect comparative risk perception, and had the highest predictive power for
preventive behaviors. Second, we found that conditional risk perception reduced unconditional risk
perception indirectly by increasing preventive behaviors. These findings suggest that conditional
absolute risk perception should be used to accurately measure risk perception and increase predictive
power for preventive behaviors. This study has important theoretical implications for resolving
inconsistencies in findings on the predictive power of different types of perceived risk likelihood on
preventive behaviors. This study also has practical implications for health message design in strategic
communication during infectious disease outbreaks.