Influence of trust on two different risk perceptions as an affective and cognitive dimension during Middle East respiratory syndrome coronavirus (MERS-CoV) outbreak in South Korea: serial cross-sectional surveys

Jang, Won Mo; Kim, Un-Na; Jang, Deok Hyun; Jung, Hyemin; Cho, Sanghyun; Eun, Sang Jun; Lee, Jin Yong

Univariate analyses using chi-square test were performed in the four consecutive surveys, entirely and respectively, to identify the relationships between risk perception and each demographic variable. We used multivariable logistic regression analyses to explore factors influencing risk perceptions (affective and cognitive) in the four surveys, entirely and respectively. Multivariable logistic regression model was adjusted for gender, age, educational attainment, occupation, self-reported household economic status, affected area, residential area, presidential job approval, and party identification. The self-reported household economic status was excluded in survey 4 model with cognitive risk perception. These exclusions were because there was small sample size of those who perceived cognitive risk in the upper economic level in survey 4. Missing values of any variable were ≤2.7%. Using logistic regression analysis for each affective and cognitive risk perception, "y = 1" was used respectively when "worried" in affective and when "spread" in cognitive, otherwise "y = 0" was used.