Trust in government moderates the association between fear of COVID-19 as well as empathic concern and preventive behaviour

The study addresses why individuals support or resist COVID-19 preventive behaviors and how these motivations interact with socio-cultural context, particularly trust in government. Prior work indicates adherence can be driven by self-interest (fear of infection) and other-oriented motives (empathy for vulnerable others). However, it is unclear how these motives depend on contextual conditions such as governmental trust at both individual and country levels. The authors propose that under uncertainty and potential social dilemmas inherent to large-scale cooperation, trust in authorities may shape whether empathy-based or fear-based motives relate more strongly to support for preventive behaviors. They hypothesize that empathic concern will be more strongly associated with support for COVID-19 containment behaviors when trust in government is high (H1a) and that fear of COVID-19 will be more strongly associated when trust is low (H2a). At the country level, they hypothesize similar moderation by general levels of governmental trust: stronger empathy association where trust is high (H1b) and stronger fear association where trust is low (H2b).

Background literature shows mixed evidence on whether prosocial (empathy, altruism) or self-protective (fear) motives dominate COVID-19 preventive compliance across contexts. Studies from Sweden, Germany, USA, and Norway often find other-oriented concerns more influential than fear, while research in Russia, Azerbaijan, and China reported greater importance of self-focused concerns. Trust is vital for cooperation, especially involving large-scale actions and reliance on authorities. While some studies link higher governmental trust to greater adherence, others find weak or inconsistent associations, suggesting a complex relationship. The present work positions governmental trust as a boundary condition moderating links between empathy and fear motives and support for preventive behaviors, considering cross-national socio-cultural variability (e.g., development level, health system capacity, policy stringency, pandemic severity, cultural tightness).

Ethics and preregistration: Approved by the Institutional Review Board of Bahcesehir University (E-8755) and local IRBs where required. Preregistered on OSF (09/09/2021; https://osf.io/k2wjr). Updated on 03/29/2023 to refine hypotheses wording, specify covariates, include generalized trust exploratory analyses, and test robustness with fixed-effects models with cluster-robust SEs. Design and procedure: Cross-national, multi-site online survey as part of the APA Interdivisional Task Force on the Pandemic's IMPACT-C19 project. Convenience sampling via online survey tools (Google Forms/Qualtrics). Translated/adapted via committee approach. Participation voluntary, anonymous, ~25 minutes. Participants: Initial raw N=27,787 through Jan 20, 2022. Exclusions: under 18 or missing age (N=4,951); listwise missing on study variables (N=8,161); sites with <150 complete responses (22 countries; N=1,917). Final dataset: N=12,758 adults (Mage=26.9; 67% women) from 34 countries, collected Feb–Dec 2021. For country-level analyses requiring World Values Survey (WVS) data, N=11,026 from 29 countries (excluding Honduras, India, Israel, El Salvador, and Cuba). Measures: - Trust in government (individual level): 4 items adapted from Kerr et al., 5-point Likert (1–5). Cronbach’s alpha=0.95 overall; 0.85–0.97 by country. - Trust in government (country level): From WVS Wave 7 (2017–2020). Percentage with “a great deal” or “quite a lot” of confidence in government. - Empathic prosocial concern: 3 items from Pfattheicher et al., 7-point Likert (1–7), e.g., concern for those vulnerable to COVID-19. Alpha=0.91 overall; 0.63–0.97 by country. - Fear of COVID-19: 7-item Fear of COVID-19 Scale, 5-point Likert (1–5). Alpha=0.90 overall; 0.83–0.93 by country. - Support for COVID-19 containment behaviors: 13 items (extended from Tepe & Karakulak) rating importance (1–5) of preventive behaviors (masking, staying home, handwashing, etc.). Alpha=0.94 overall; 0.81–0.95 by country. - Covariates: Individual age, gender; country-level HDI, hospital beds per 1000, month of data collection start, government stringency level, 7-day rolling averages of daily new COVID-19 cases and deaths at data collection start. Exploratory: generalized trust (individual: 2 items; r=0.78; country: WVS). Data analysis: Multilevel linear regression using maximum likelihood with individuals nested in countries. Steps: (1) Null model (variance decomposition). (2) Random intercepts with fixed slopes for main effects. (3) Random slopes model for main effects. (4) Add hypothesized two-way interactions. Predictors (individual-level) group-mean centered; country-level governmental trust grand-mean centered. Model comparisons via deviance (−2 log likelihood) and chi-square tests. Robustness checks with fixed-effects OLS and cluster-robust SEs (Supplementary). Analyses conducted in jamovi 2.0.

Sample and variance: N=12,758 individuals from 34 countries. ICC(1)=0.34 for support of containment behaviors, indicating 34% variance between countries. Individual-level (trust in own government as moderator): Main effects all positive and significant in random slopes model: - Trust in government: B=0.05 (SE=0.01), 95% CI [0.03, 0.08], t(34.5)=4.11, p<0.001. - Empathic concern: B=0.14 (SE=0.01), 95% CI [0.12, 0.15], t(30.4)=18.08, p<0.001. - Fear of COVID-19: B=0.30 (SE=0.02), 95% CI [0.26, 0.35], t(33.1)=12.71, p<0.001. Adding interactions improved fit: deviance reduced from 24,864 to 24,789; χ2(3,N=12,758)=75, p<0.001. Significant interactions: - Trust × Empathic concern: B=0.01 (SE=0.005), 95% CI [0.003, 0.02], t=2.56, p=0.011. - Trust × Fear of COVID-19: B=−0.02 (SE=0.01), 95% CI [−0.03, −0.005], t=−2.64, p=0.008. Simple slopes: Empathic concern associated with support at all trust levels, strongest at high trust (β=0.14; 95% CI [0.12, 0.16]) vs low trust (β=0.12; 95% CI [0.10, 0.14]). Fear associated at all trust levels, strongest at low trust (β=0.32; 95% CI [0.27, 0.37]) vs high trust (β=0.29; 95% CI [0.24, 0.34]). Supports H1a and H2a. Country-level (WVS governmental trust as moderator): N=11,026 from 29 countries. Main effects (random slopes): Empathic concern B=0.14 (SE=0.01), p<0.001; Fear B=0.31 (SE=0.03), p<0.001. Country-level trust main effect not significant. Adding interactions improved fit: deviance 21,254 to 21,191; χ2(3,N=11,026)=63, p<0.001. Interactions: - Country trust × Empathic concern: not significant (B=−0.0003, SE=0.0003, p=0.42). No support for H1b. - Country trust × Fear of COVID-19: significant negative (B=−0.002, SE=0.0008, p=0.007). Simple slopes: fear-effect strongest where country trust low (β=0.37; 95% CI [0.30, 0.43]) and weakest where trust high (β=0.26; 95% CI [0.19, 0.32]). Supports H2b. Exploratory analyses: Including generalized trust as covariate did not alter core results; individual-level governmental trust still moderated both empathy (B≈0.013, p=0.008) and fear (B≈−0.02, p=0.015) associations. Using generalized trust (individual or country level) as the moderator yielded no significant interactions with empathy or fear. Overall: Both empathy and fear predict greater support for preventive behaviors. Trust in government moderates these associations: empathy linkage slightly stronger with higher individual trust; fear linkage stronger with lower individual and country-level trust.

Findings indicate that motivations for supporting COVID-19 preventive behaviors are shaped by trust in government. Empathic concern relates positively to support across contexts and is slightly more pronounced among individuals with higher governmental trust, aligning with literature linking empathy to prosocial action. Conversely, fear of COVID-19 is a robust predictor whose association with support strengthens when trust in government is low—both at individual and contextual (country) levels. This pattern suggests under uncertain or low-trust conditions, self-protective motives may drive compliance more strongly, whereas in higher-trust environments, other-oriented motives may more readily translate into support. The results refine understanding of the interplay between individual motives and systemic trust in mobilizing large-scale cooperative actions during crises, emphasizing that both motivational pathways are relevant but differentially accentuated by trust context.

This multinational study demonstrates that trust in government moderates how empathic concern and fear of COVID-19 relate to support for preventive behaviors. Empathy consistently predicts support and is modestly amplified by higher individual trust, while fear’s link is stronger under lower trust at both individual and country levels. The work underscores the unique role of governmental trust (over generalized interpersonal trust) in shaping motivational pathways to collective action. Policy implications include fostering governmental trust to leverage empathy-based cooperation and avoiding reliance on fear appeals due to potential mental health costs. Future research should employ experimental and qualitative designs to establish causal mechanisms, explore generalizability to other crises, and examine demographic groups underrepresented in this survey.

The study is cross-sectional and correlational, limiting causal inference. Convenience sampling and online data collection may reduce representativeness (younger, internet-connected, local populations). Certain high-risk groups (e.g., older adults, migrants) were underrepresented. Fear levels were generally low and empathy high across samples, potentially biasing associations. The dependent measure assessed support for behaviors rather than actual behavior, which may differ due to constraints or obligations. Moderation effects, while statistically significant, were small in magnitude and warrant replication with diverse methods, including experimental and qualitative approaches.