Objective  To construct a theoretical framework of information-emotion-perception-behavior chain model (abbreviated as ‘risk IEPB’ model) for the public′s response to major public health emergencies and to examine relationships among components of the model.

  Methods  A mechanism framework for the ‘risk IEPB’ model was constructed first with literature studies. Then an online survey on the variables in the constructed model was conducted among 19 132 internet media users in 31 provincial-level administration regions with self-designed questionnaires during 3 – 13 of August, 2020. AMOS24.0 statistical software was used to test the model and bootstrap method was used to test the mediating effect of the variables in the model.

  Results  The constructed model is of a relatively high goodness-of-fit, with the values of 46.498 for chi-square minimum/degree of freedom (CMIN/DF), 0.049 for root mean square error of approximation (RMSEA), 0.946 for goodness-of-fit index (GFI), 0.934 for adjusted goodness-of-fit index (AGFI), 0.933 for normed fit index (NFI), 0.923 for relative fit index (RFI), 0.934 for incremental fit index (IFI), 0.819 for parsimony normed fit index (PNFI), and 0.772 for parsimony goodness-of-fit index (PGFI), respectively. The mediating effect analysis indicated that direct effect on decision-making for behaviors about major public health emergency risk is significant for following variables: trust in government media (effect value = 0.525), trust in non-government media (– 0.037), positive emotion (0.136), negative emotion (– 0.071), and risk perception (0.027); trust in government media, trust in non-government media, and negative emotion affect risk perception directly, with the effect values of 0.069, – 0.020, and 0.535, but the direct effect of positive emotion on risk perception is not statistically significant (P ≥ 0.05); trust in government and non-government media exert direct effect on both positive emotion (effect value = 0.233 and 0.043) and negative emotion (effect value = – 0.169 and 0.229). Emotion plays a partial mediating role in the interaction path between trust in media and risky behavior-related decision-making. In the interaction path between trust in media and risky behavior-related decision-making, risk perception plays a partial mediating role. In the pathway from trust in media to risk perception, negative emotion plays a partial mediating role; in the path of negative emotion to risky behavior-related decision-making, risk perception plays a partial mediating role. The direct and indirect effect of trust in government and non-government media on risk perception and of negative emotion on risky behavior-related decision-making are opposite, which belongs to ‘masking effect’.

  Conclusion  The constructed ‘risk IEPB’ model is of good theoretical and empirical significance, which can reflect the mechanism framework of decision-making for risky behaviors related to public health emergencies and interactive relationships among relevant variables, suggesting that the model could provide a theoretical support for interventions on public health emergency-related risky behaviors in the public.