Risk-Likelihood Perceptions and Preventive Behaviors Against Infectious Diseases: Testing Different Types of Risk-Likelihood Perceptions in the Context of MERS and COVID-19

ABSTRACT

Effective public health communication should take into the consideration how communicated risk perceptions might influence the public. Although many health behavior theories claim that high perceived risk likelihood motivates preventive health behaviors, this general prediction is complicated by different ways of estimating risk likelihood. Our review of 14 previous studies on perceived risk of infectious diseases revealed different measures of risk likelihood and inconsistent findings regarding the relationship between perceived risk likelihood and behavioral intention. Drawn from the review, we examined the predictive power of different types of risk likelihood on intention to prevent infectious diseases. Using two separate data sets, collected during the 2015 outbreak of the Middle East respiratory syndrome (N = 557) and the COVID-19 pandemic of 2021 (N = 1,001), we demonstrated that perceived risk likelihood that specifies future inaction on health behaviors (i.e., conditional absolute risk perception) better predicted behavioral intention than did other types of risk perceptions. When the effect of conditional absolute risk perception was controlled, the behavioral intention and unconditional absolute risk perception showed a negative relationship. The findings have theoretical and practical implications that can inform strategic communication during future outbreaks.

Acknowledgments

The authors would like to thank Minsun Shim for her assistance with data collection and for her feedback on an earlier version of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Similarly, in studies on the perceived impact of harmful media messages on attitude toward censorship, it was not comparative (i.e., the discrepancy between perceived harmful impact on others and that on oneself) but absolute risk perception that better predicted censorship attitude (Chung & Moon, 2016).

2. Throughout the entire procedure, Study 1 followed the scientific and ethical practices to the best of researchers’ ability who were properly trained on responsible and ethical conduct of research. However, back in 2015, the Institutional Review Board’s ethical approval was not a required nor widely preformed practice in South Korea. We expected that there would be none to minimal risk associated with the two studies reported in this research, but having recognized the importance of ensuring the rights, safety, and welfare of human research participants, we sought for formal IRB approval for Study 2.

3. Cohen and Cohen (1983) highlighted the problem of using a difference score in multivariate analyses. They explained that a correlation between a difference score (e.g., A – B) and a covariate C can be confounded with the correlations between A and C or between B and C. Cohen and Cohen suggested analysis of partial variance to test the relationship between the difference score and the covariate.

4. A series of independent-samples t tests was conducted to compare the responses of those who were asked about absolute risk perception prior to comparative risk perceptions (Group 1; n = 512) and those who received the questionnaire that alternated the order of the two types of risk perceptions (Group 2; n = 489). The results of a series of independent-samples t tests are as follows. Conditional absolute risk: Group 1 (M = 46.44, SD = 21.92) and Group 2 (M = 48.19, SD = 20.46), t(999) = > −1.31, p = .191; Unconditional absolute risk: Group 1 (M = 45.06, SD = 21.56) and Group 2 (M = 43.94, SD = 20.55), t(999) = .84, p = .400; Indirect comparative risk: Group 1 (M = 9.69, SD = 14.93) and Group 2 (M = 8.22, SD = 14.66), t(999) = 1.57, p = .117; Direct comparative risk: Group 1 (M = > −.27, SD = 1.29) and Group 2 (M = > −.19, SD = 1.16), t(999) = > −.95, p = .344; Vaccination intention Group 1 (M = 7.20, SD = 2.63) and Group 2 (M = 7.29, SD = 2.60), t(999) = > −.55, p = .582.

Additional information

Funding

This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea [2021S1A5A2A01065446].