The price of prosociality in pandemic times

The study investigates how individual differences in prosociality-related factors—specifically empathy, theory of mind (ToM), and moral judgments—predict key pandemic behaviors and perceptions: COVID-19 risk perception, impact estimation, and acceptance of quarantine. The SARS-CoV-2 pandemic drastically altered daily life and imposed social restrictions, with amplified challenges in countries with high inequality and limited state capacity, such as Colombia. Prior theoretical and empirical work suggests that social cognition (empathy, ToM) and moral cognition (utilitarian vs. deontological tendencies regarding instrumental harm and impartial beneficence) can shape prosocial behaviors and compliance with public health measures. The purpose is to quantify the contribution of these prosociality-linked traits to pandemic-related perceptions and behaviors, thereby informing behavioral interventions to improve public health compliance.

Prosociality encompasses cognitive and affective processes that benefit others, including helping, sharing, comforting, and even self-sacrifice. It can be decomposed into social cognition (empathy and ToM) and moral cognition (judgments in instrumental harm (IH) and impartial beneficence (IB) contexts). Prior research shows that empathy (affective concern and personal distress) and ToM relate to coping with uncertainty and adherence to warnings; empathy can heighten sensitivity to social information and personal distress under stress. Moral orientations influence regulation-following in stressful contexts, with IH (normally forbidden actions tolerated for greater good) and IB (personal sacrifice for common good) modulating responses. Risk perception and acceptance of restrictions have social roots and are promoted by prosocial behavior. However, the influence of moral tendencies (utilitarian vs. deontological) on pandemic responses remained unclear, motivating this study.

Design: Cross-sectional online survey conducted in Colombia during pandemic restrictions. Sample: N = 413 (314 women, 99 men); mean age 23.46 years (SD = 9.76); mean education 15.29 years (SD = 3.28); SES mean 4.19 on 1–8 scale (SD = 1.40). No participants reported prior SARS-CoV-2 infection. Recruitment was voluntary via internet invitation. Measures: - Social cognition (Empathy): Short Interpersonal Reactivity Index (IRI) assessing dispositional empathy with subcomponents including personal distress, empathic concern, perspective taking, and fantasy (Davis, 1983; short version validated by Davis et al., 2017). - Theory of Mind (ToM): Reading the Mind in the Eyes Test (RMET), abridged 17-item version; score is percent correct in inferring emotions from eye-region photographs. - Moral cognition: • Instrumental Harm (IH): Four COVID-19-related utilitarian moral dilemmas. For each, participants rated moral valence (−5 extremely bad to 5 extremely good), likelihood of behaving like the protagonist (1–10), and perceived intention to harm (1–10). • Impartial Beneficence (IB): Four COVID-19-related scenarios emphasizing common good at personal cost; same questions as IH. Dependent variables: - Risk perception: Estimated personal probability (1–100%) of contracting COVID-19 over the next year. - Impact estimation: Six items (three medical, three social) indicating perceived consequences; summed positive responses (0–6). - Acceptance of quarantine: Single item rating agreement with quarantine requirements (1–6; higher indicates greater acceptance). Covariates: Age, gender, years of education, SES, depressive symptoms (PHQ-9), anxiety symptoms (GAD-7/9 as reported). Analysis: Multiple linear regressions for each dependent variable with stepwise/hierarchical models. Separate models for social cognition predictors and for moral measures (moral judgment, harm detection, behavior acceptance). All models controlled for demographics and mood covariates. Effect sizes reported as R² and η²; post hoc sensitivity indicated power = 0.95 to detect medium effect size (η² ≈ 0.13; critical t = 1.98). Analyses performed in JASP v0.14.1.

- Risk perception: • Social cognition model significant: F(4,412)=8.77, p<0.001, R²≈0.19–0.20. Higher personal distress (β=0.14, t=2.84, p<0.01, η²=0.18), higher empathic concern (β=0.23, t=4.41, p<0.001, η²=0.18), and older age (β=0.20, t=4.07, p<0.001, η²=0.08) predicted higher risk perception. ToM showed positive association in hierarchical models. Gender, education, depression, and anxiety were not significant in these models. • Moral measures model significant: F=5.96, p<0.001, R²≈0.16. Higher moral condemnation of IH and greater harm detection in IH related to higher risk perception. - Impact estimation: • Social cognition model significant: F(9,412)=9.35, p<0.0001, η²≈0.20. Higher personal distress (β=0.14, t=2.97, p<0.01, η²=0.20), higher empathic concern (β=0.25, t=4.73, p<0.001, η²=0.22), and older age (β=0.21, t=4.18, p<0.001, η²=0.18) predicted greater perceived impact. Other covariates non-significant. • Moral measures-only model significant: F(1,412)=3.39, p<0.01, R²=0.06. Stronger moral judgment against IH (β≈0.27, t=2.44, p<0.01, η²=0.14) and higher IH harm assessment (β≈0.14, t=1.99, p<0.05, η²=0.10) predicted greater impact estimation. Hierarchical models with demographics and mood remained significant. - Acceptance of quarantine: • Social cognition model significant: F(9,412)=3.18, p<0.01, R²=0.09. Acceptance was associated with affective empathy (especially empathic concern; see Table 4) and age effects were observed across models. • Moral measures: Greater acceptance of IB scenarios and lower acceptance of IH scenarios predicted higher acceptance of quarantine (as per Fig. 2C and Table 4). - Demographics and mood: • Older age consistently predicted higher risk and impact perceptions and was associated with acceptance of quarantine in moral models. • Female gender predicted higher risk perception and impact estimation in models including moral judgments. • Depressive and anxiety symptoms did not show robust associations with the dependent measures.

Findings indicate that prosociality-related dispositions, particularly affective empathy (personal distress and empathic concern), are key predictors of how individuals perceive COVID-19 risk and impact, and how willing they are to accept quarantine measures. Moral cognition also contributes: stronger deontological-leaning responses to instrumental harm (greater harm detection and moral disapproval) relate to heightened perceived risk and impact, while support for impartial beneficence and rejection of instrumental harm relate to greater acceptance of quarantine. Age and gender further shape these outcomes, with older individuals showing higher perceived risk/impact and evidence of greater quarantine acceptance, and women displaying higher risk and impact perceptions in moral-judgment models. Notably, depression and anxiety were not significant determinants after accounting for prosociality-related traits. Overall, affective empathy exerted more robust and consistent effects than cognitive components or moral judgments, suggesting that relatively stable empathic traits may more directly guide pandemic-related behaviors. These insights can inform interventions (e.g., empathy-based messaging) to increase compliance with public health measures, particularly in settings with heightened inequality and vulnerability.

The study demonstrates that affective empathy and certain moral tendencies are associated with higher COVID-19 risk perception, greater perceived personal and societal impact, and greater acceptance of quarantine measures. Prosocially oriented traits, especially empathic concern and personal distress, emerged as stronger predictors than cognitive empathy or ToM, while moral judgments regarding instrumental harm and impartial beneficence also contributed. These results highlight the potential of leveraging prosociality to design behavioral interventions that promote compliance with public health guidelines in pandemic contexts, particularly in the Global South. Future research should employ larger, probabilistic, and gender-balanced samples across the lifespan; integrate subjective, objective, and real-life behavioral measures; and expand outcomes to include vaccine acceptance, adherence to medical requirements, and post-infection behavior change, while considering mediated impacts of COVID-19.

- Single-country sample from Colombia limits generalizability to other cultural and socioeconomic contexts. - Convenience, online sampling skewed toward educated participants with internet access reduces representativeness and may introduce participation bias (including female overrepresentation). - Sample primarily comprised younger adults, potentially affecting observed relations between prosociality measures and pandemic behaviors. - Reliance on self-report instruments with an unbalanced number of items across measures introduces common-method biases. - Limited set of pandemic-related outcomes; did not assess behaviors such as vaccine uptake, medical adherence, or behavior changes following negative COVID-19 outcomes. - Potential unmeasured confounding and situational variability in moral choices; cross-sectional design precludes causal inference.