key: cord-316963-2ex4c7tj authors: Dai, Bibing; Fu, Di; Meng, Guangteng; Liu, Bingsheng; Li, Qi; Liu, Xun title: The effects of governmental and individual predictors on COVID‐19 protective behaviors in China: a path analysis model date: 2020-05-19 journal: Public Adm Rev DOI: 10.1111/puar.13236 sha: doc_id: 316963 cord_uid: 2ex4c7tj The outbreak of the COVID‐19 pandemic has plunged the world into a crisis. To contain the crisis, it is essential to build full cooperation between the government and the public. However, it is unclear which governmental and individual factors are the determinants and how they interact on protective behaviors against COVID‐19. To resolve this issue, this study built a multiple mediation model and found government emergency public information as detailed pandemic information and positive risk communication had more important impacts on protective behaviors than rumor refutation and supplies. Moreover, governmental factors could indirectly affect protective behaviors through individual factors such as perceived efficacy, positive emotions, and risk perception. These findings suggest that systematic intervention programs for governmental factors need to be integrated with individual factors to finally achieve effective prevention and control of the COVID‐19 pandemic among the public. This article is protected by copyright. All rights reserved. COVID-19 has plunged the world into a crisis, and its effect on people's physical and mental health, economic development and social stability cannot be underestimated (van Gelder et al. 2020 ). China is not only one of the first countries to experience the outbreak of COVID-19 infection, but also one of the few that have largely contained it. This could not be separated from the strict governmental supervision and people's effective protective behaviors (Li, Chen, and Huang 2020). Therefore, drawing on its experience in pandemic prevention and control can help accelerate the world's progress in defeating the pandemic. The Protective Action Decision Model (PADM) was developed to explore people's actions to natural hazards and disaster events. PADM believes that various sources of information cause people's attention, exploration, and comprehension to generate threats perceptions, protective actions perceptions, and stakeholder perceptions, and finally form decisions about how to take self-protective actions (Lindell and Perry 2012; Lindell 2018) . Based on this framework, the current study proposes an Information-Perception/Consideration-Action mediation model to elucidate protective behaviors during a pandemic. In this model, government emergency public information is considered to be the sources of information, the individual's emotional and cognitive perception and consideration are considered to be the extension of perceptions in the PADM model. Additionally, protective behaviors, including preventive (i.e. wearing masks, disinfectants) (Kim et al. 2015) , avoidant (i.e. stringent quarantine, avoiding public places) (Bayham et al. 2015) , and management of disease behaviors (i.e. seeking professional protection or treatment information, paying for preventive and therapeutic drugs) (Hagan, Maguire, and Bopping 2008) , are considered to be the actions (Bish and Michie 2010) . One important issue that should be explored is how this government emergency public information can persuade the public to adopt recommended protective behaviors to control the spread of the COVID-19 pandemic. Government emergency public information should enhance the public's courage and determination, raise their risk awareness, and adopt effective protections to fight the pandemic (Paek et al. 2008) . The Chinese government implemented several effective emergency public information measures through detailed pandemic information, positive risk communication, and rumor refutation (Chon and Park 2019; Li, Chen, and Huang 2020; Xu et al. 2020) . Detailed pandemic information means the released statistical information, such as confirmed cases, dynamic suspected cases, recovered cases, and deaths both in accumulative numbers and daily updates, as well as tracked information including travel history, and trains or flights taken by specific confirmed or suspected patients. During COVID-19, the detailed pandemic information has become the foundation of the current South Korean policy actions to combat COVID-19 successfully (Moon 2020). Some researchers believe that detailed information can increase people's risk perception and promote protective behaviors (Qazi et al. 2020; French 2011) . Positive risk communication conveying positive educational information can result in more appropriate manners (Fewtrell and Bartram 2001) . According to China's fight against COVID-19 (China Daily 2020), the achievements in the fight against the virus and stories of frontline medical workers, volunteers reported by the mainstream media could inspire people to participate in the pandemic. Rumors increase the uncertainty of public information and trigger conspiracy theories and pseudoscientific claims (Dredze, Broniatowski, and Hilyard 2016; Sharma et al. 2017) . One important challenge to control the Ebola haemorrhagic outbreak was numerous rumors (Lamunu et al., 2004) . Timely refutation of rumors can help the government reduce public confusion, perceived risk and panic, build trust, and promote proper protective behaviors (DiFonzo and Bordia 2007; Greenhill and Oppenheim 2017) . In addition, medical supplies during a pandemic are desperately needed (WHO 2015) . For example, during the 2014 West Africa Ebola epidemic, evidence suggests that earlier supplies modestly reduced mortality (Walker and Whitty 2015) . The efforts to add supplies such as the lifesaving medicines and trained clinicians could increase the public trust and promote people to seek clinical care (WHO Ebola Response Team 2014). During this COVID-19 pandemic, the rapid construction of Huoshenshan Hospital made people concern more about the pandemic and feel they have "warriors" in this battle (BBC News 2020). Perceived efficacy, positive emotions, and risk perception are important individual factors affecting protective behaviors (Prati, Pietrantoni, and Zani 2011) . First, perceived efficacy plays a key role in positively predicting protective behaviors (Balkhy et al. 2010; Rubin et al. 2009; Seale et al. 2009 ). According to the Protection-Motivation Theory (Rippetoe and Rogers 1987) , perceived efficacy is made up of self-efficacy and response efficacy. Self-efficacy refers to individuals' confidence in their abilities to carry out protective behaviors, and response efficacy refers to individuals' belief of the effectiveness of protective behaviors in coping with a health threat. People with higher perceived efficacy are more likely to take precautionary behaviors and seek control in avian influenza pandemic (de Zwart et al. hypotheses, this study tries to test two hypotheses. The first is that government emergency public information would promote people to comply with protective behaviors directly. The second is that government emergency public information would contribute to protective behaviors through increasing people's perceived efficacy, positive emotions, and perception of risk. Participants and Data Collection. This cross-sectional design research was approved by the Institutional Review Board of the Institute of Psychology, Chinese Academy of Sciences, and followed the Declaration of Helsinki. Data collection was conducted from 2020-2-24 to 2020-3-3. All the participants were recruited online from 33 provinces in China. After reading and signing the informed consents, we asked participants to rate government, personal, and behavior factors with 21 items on a 7-point Likert scale, which were displayed in Table 1 preventive behaviors items, one avoidant behaviors item, and two management of disease behaviors items. In the present study, these items were chosen to reflect the main components of these variables in the context of COVID-19 pandemic and most of them had good or accepted reliabilities. A total of 1,131 participants finished the survey. Data of 1,022 participants (90.4%) entered final statistical analyses after deleting the invalid data, where participants gave a wrong response on a question used to detect whether they answer the questionnaire carefully. Participants' demographic information is displayed in Table 2 . Comparison of the sample's demographic characteristics to the corresponding census data suggested that the sample over-represented youth, higher education population and students. This article is protected by copyright. All rights reserved. Data Analysis. Data were analyzed using SPSS Version 20.0, Amos Version 23.0, and Mplus 7.0. T-test and one-way ANOVA were used to explore whether there were gender, age, and education differences in protective behaviors. Descriptive statistics were performed to describe the sample characteristics of each factor. Pearson correlation analyses were performed to examine whether associations between factors conformed to the prerequisites for path analysis. Path analysis was conducted to test the model. The squared multiple regression correlation coefficient was estimated to identify the variance in protective behaviors which was explained by proposed factors. Bootstrap resampling was employed to test the significance of direct and indirect variable effects (MacKinnon, Lockwood, and Williams 2004) . We analyzed how gender, age, and education background impact on protective behaviors. T-test showed that gender had a significant effect on protective behaviors, t (1020) = 5.16, p < .001. Females showed more protective behaviors (M ± SD = 29.69 ± 3.87) than males (M ± SD = 28.23 ± 5.18). One-way ANOVA showed that age had a significant effect on protective behaviors, F (3, 1005) = 5.82, p < .001. Post hoc test indicated that participants from 18 to 25 years (M ± SD = 28.49 ± 4.36) showed significantly fewer protective behaviors than participants from 46 to 61 years (M ± SD = 30.13 ± 4.73), p < .01. No significant differences were found between other age groups, ps > .05. One-way ANOVA showed that education background had a significant effect on protective behaviors, F (3, 1018) = 4.33, p < .01. Post hoc test indicated that participants with high school or lower education background (M ± SD = 30.08 ± 4.46) had significantly more protective behaviors than participants with university bachelor's degree (M ± SD = 28.81 ± 4.69), p < .05. No significant differences were found between other groups, ps > .05. Means and standard deviations for the predictors of protective behaviors, as well as the correlation coefficients between them are displayed in Table 3 . Only the association between risk perception and supplies was not significant (r = .03, p > .05). Associations between other factors and protective behaviors reached significance (ps < .05). Furthermore, all the proposed governmental and individual factors were positively correlated with preventive, avoidant, and management of disease behaviors (ps < .05). These three protective behaviors were also positively correlated with each other significantly (ps < .05). Therefore, path analysis could perform based on the current model. Mediational Model. The model-data fit was evaluated using  2 ,  2 /df, root mean square error of approximation (RMSEA), standardized root mean square residual (SRMR), normed fit index (NFI), comparative fit index (CFI), and goodness of fit index (GFI). The RMSEA and SRMR should be less than .08. Regarding NFI, CFI, and GFI, values no less than .90 indicate a good model fit, whereas values above .95 indicate an excellent fit (Cohen et al. 2013) . Because protective behaviors may be associated with a variety of demographic factors, the hypothesized model was performed adding gender, age, and education as control variables, which was a common statistics method considering the confounding effects of personal (Table 4 ). In addition, to further examine whether the mediating effect was significant, the indirect effects were computed using the bias-corrected bootstrapping method; if the 95% confidence interval (CI) did not include 0, it meant that the mediating effect was significant (MacKinnon, Lockwood, and Williams 2004) . Table 5 The current study has several limits and future directions. First, the sample in the current study is not representative of all demographic categories. A large number of participants were young college students and with a bachelor's degree or higher although age and education had no significant effects on protective behaviors. Thus, the applicability of the findings to other samples needs to be further explored. Second, previous study found that people's perceptions of authorities are different across countries and are correlated with their protective actions to pandemic (Wei et al. 2018 ). All participants of the current study were from China, and a cross-country comparative study is needed to expand the applicability of the current findings. Third, Cronbach's alpha coefficients for the detailed pandemic information, perceived efficacy and protective behaviors have accepted reliabilities rather than good reliabilities in the present study, which may be caused by their limited number of items or omission of important items (e.g., hand washing as an important protective behavior item). Future research should adopt questionnaire with more items or adding important items to improve their reliabilities. In order to combat the COVID-19 pandemic effectively, governments should take effective measures in combination with governmental and individual factors. The suspected numbers, infected numbers, critically ill numbers, and death toll in different regions are officially announced every day. 1 (strongly disagree) to 7 (strongly agree) The confirmed patient's recent movements are officially published as soon as possible. 1 (strongly disagree) to 7 (strongly agree) Positive risk communication A lot of information about medical staff and supplies are brought from other areas to the frontline is officially announced. 1(strongly disagree) to 7 (strongly agree) Rumor refutation Fake news is officially refuted in time. 1(strongly disagree) to 7 (strongly agree) Supplies Medical staff are sufficient in your current country or region. 1 (strongly disagree) to 7 (strongly agree) Medical supplies are sufficient in your current country or region. 1 (strongly disagree) to 7 (strongly agree) Living supplies are sufficient in your current country or region. 1 (strongly disagree) to 7 (strongly agree) Mental health support is sufficient in your current country or region. 1 (strongly disagree) to 7 (strongly agree) Patients are treated on time during the pandemic. 1 (strongly disagree) to 7 (strongly agree) Perceived efficacy I believe the pandemic will be fully controlled in the foreseeable future. 1 (strongly disagree) to 7 (strongly agree) I am confident that the pandemic will be overcome. 1 (strongly disagree) to 7 (strongly agree) To cope with the pandemic, I can discriminate between true information and rumors about COVID-19. (strongly disagree) to 7 (strongly agree) When I return home from outside, I disinfect myself with alcohol spray or sanitizer. 1 (strongly disagree) to 7 (strongly agree) Avoidant I will not go out until the pandemic is over unless I have to. 1 (strongly disagree) to 7 (strongly agree) Management of illness As soon as COVID-19 preventive and treatment medications appear on the market, I will pay for them immediately. 1 (strongly disagree) to 7 (strongly agree) I usually get medical information and prevention measures about COVID-19. 1 (strongly disagree) to 7 (strongly agree) Note. All the estimates provided in the table are standardized estimates. * p < .05, ** p < .01, *** p < .001. Eppm and Willingness to Respond: The Role of Risk and Efficacy Communication in Strengthening Public Health Emergency Response Systems Unresolved Issues in Risk Communication Research: The Case of the H1N1 Pandemic Measured Voluntary Avoidance Behaviour During the 2009 A/H1N1 Coronavirus: Cement mixers become celebrities in China lockdown Demographic and Attitudinal Determinants of Protective Behaviours During a Pandemic: A Review This article is protected by copyright. All rights reserved Report: China's fight against COVID-19 Predicting Public Support for Government Actions in a Public Health Crisis: Testing Fear, Organization-Public Relationship, and Behavioral Intention in the Framework of the Situational Theory of Problem Solving Applied multiple regression/correlation analysis for the behavioral sciences Monitoring of Risk Perceptions and Correlates of Precautionary Behaviour Related to Human Avian Influenza During 2006 -2007 in the Netherlands: Results of Seven Consecutive Surveys Rumors influence: Toward a dynamic social impact theory of rumor Zika Vaccine Misconceptions: A Social Media Analysis Public Support for Government Actions During a Flu Pandemic: Lessons Learned from a Statewide Survey A Social-Cognitive Model of Pandemic Influenza H1N1 Risk Perception and Recommended Behaviors in Italy Analyzing Situational Awareness through Public Opinion to Predict Adoption of Social Distancing Amid Pandemic Covid-19 Stockpiling Supplies for the Next Influenza Pandemic Effects of Components of Protection-Motivation Theory on Adaptive and Maladaptive Coping with a Health Threat Public Perceptions, Anxiety, and Behaviour Change in Relation to the Swine Flu Outbreak: Cross Sectional Telephone Survey The Community's Attitude Towards Swine Flu and Pandemic Influenza Zika Virus Pandemic-Analysis of Facebook as a Social Media Health Information Platform The Role of Public Trust During Pandemics Remote Control: How the Media Sustain Authoritarian Rule in China Covid-19: Reducing the Risk of Infection Might Increase the Risk of Intimate Partner Violence Effective Health Risk Communication About Pandemic Influenza for Vulnerable Populations Tackling Emerging Infections: Clinical and Public Health Lessons from the West African Ebola Virus Disease Outbreak Perceived Stakeholder Characteristics and Protective We thank all the participants for their commitment in the present study. We particularly thank Lux Li for his kindly help with proofreading and thank both reviewers for helping us improve the quality and clarity of our manuscripts. This study was supported by grants from the National Natural Science Note. * p < .05, ** p < .01.