Bayesian probability revision and infection prevention behavior in Japan: A quantitative analysis of the first wave of COVID-19

The relationship between cognitive biases and infection prevention behavior remains unexplored in the existing literature. This study uses data from a questionnaire survey conducted in Japan on the first wave of Coronavirus Disease 2019 (COVID-19) from February to May 2020 to empirically investigate the impact of Bayesian probability inference, the influence of cognitive biases of PCR test results on infection prevention behavior, and the discrepancy between infection prevention intentions and behaviors. We used a bivariate ordinal probit model when considering the correlation between behaviors. The results showed that the higher probability responses, implying pessimistic biases, were more likely to indicate that declaring a state of emergency was necessary and effective, and were more health-oriented in ensuring infection prevention behavior even at the expense of the economy. However, the study found that although they wanted to reduce the frequency of their outings and the number of people they met, they did not reduce them in terms of actual behavior change. It also found that those with pessimistic biases had a higher WTP for the vaccine.