Why do Catholics have lower COVID-19 vaccine intentions than atheists? The mediating role of belief in science and moral foundations

The study investigates why Catholics report lower COVID-19 vaccination intentions than atheists, focusing on belief in science and moral foundations as explanatory mechanisms. In Poland, vaccine hesitancy remains substantial, threatening public health goals. Prior work links higher religiosity to lower vaccination uptake, especially among Christians, but mechanisms are unclear and atheists are often underrepresented or conflated with low religiosity. The research tests whether lower belief in science among Catholics and stronger endorsement of binding moral foundations (purity, loyalty, authority) explain reduced vaccine intention. It also examines whether identification with one’s denominational group moderates these relationships. Understanding these mechanisms is important to design effective, culturally sensitive interventions to increase vaccination uptake.

- Religion and COVID-19 vaccination: Highly religious individuals often show greater vaccine hesitancy; cross-national data link Christianity (but not other major religions) to lower COVID-19 vaccination rates. Religious orientations differ: intrinsic orientation (living faith as a guiding worldview) versus extrinsic (instrumental, social uses of religion). Prior findings on orientations and vaccination are mixed across contexts. - Belief in science: For some, scientific and religious worldviews are perceived as competing. Higher belief/trust in science predicts greater vaccination intention and lower hesitancy. Religiosity is generally associated with more skepticism toward science, though relations vary by culture. Prior studies suggest trust in the scientific community or scientific credibility mediates religiosity–vaccination links, but direct comparisons of believers and atheists with belief in science as a mediator are lacking. - Moral foundations theory: Binding foundations (purity, loyalty, authority) are more strongly endorsed by religious individuals, while individualizing foundations (care, fairness) are similarly endorsed by believers and atheists. Prior studies (mostly UK/USA) show mixed associations between specific foundations and vaccine hesitancy; purity often predicts greater hesitancy. Whether binding foundations mediate believer–atheist differences in vaccine intention outside Anglo-American contexts remains underexplored.

Design: Three cross-sectional online survey studies conducted in Poland (Aug–Nov 2023). All variables standardized to Z-scores after confirmatory factor analyses (CFA) to retain items with factor loadings ≥0.30. Mediation and moderated mediation models estimated using SPSS PROCESS with 5000 bootstrap resamples; covariates included age, sex (0=female, 1=male), and education. Study 1 (Catholics vs. atheists): - Sample: National online panel; Catholics and atheists only; quota sampling by sex and age (18–65; M=39.43, SD=12.90). N=691 (345 females); two non-binary excluded from models; analytic n≈689; 366 Catholics (53%), 325 atheists (47%). - Measures: COVID-19 vaccination intention (single item, 1–7). Group identification with Catholic/atheist group (Identity subscale of Collective Self-Esteem; 3 items after removal; α=0.71). Belief in Science Scale (10 items; α=0.92; all loadings >0.63). Demographics; prior vaccination history. - Analysis: Moderated mediation (PROCESS Model 8). Predictor: believing (0=atheist, 1=Catholic). Mediator: belief in science. Moderator: group identification. Covariates: age, sex, education. 5000 bootstrap samples; 95% CIs. Study 2 (Catholics only): - Sample: National online panel of Catholics; quota by sex and age (18–75; M=40.58, SD=14.47). N=754 (345 females); two non-binary excluded; analytic n≈752. - Measures: Intrinsic and extrinsic religious orientations (I/E-Revised; intrinsic 7 items α=0.86; extrinsic 6 items α=0.79; one intrinsic item removed for low loading 0.23). Belief in Science Scale (α=0.84; loadings >0.62). Vaccination intention single item (1–7). Demographics. - Analysis: Mediation (PROCESS Model 4). Predictor: intrinsic religious orientation. Mediator: belief in science. Covariates: extrinsic religious orientation, age, sex, education. 5000 bootstrap samples; 95% CIs. Study 3 (Catholics vs. atheists): - Sample: National online panel; Catholics and atheists only; quota by sex and age (18–65; M=40.34, SD=13.40). Initial N=474 (249 females); after exclusions for missing vaccination variables, analytic n≈469; 258 Catholics (54.8%), 213 atheists (45.2%). - Measures: Moral Foundations Questionnaire (Polish MFQ; 30 items across care α=0.80, fairness α=0.83, loyalty α=0.74, authority α=0.75, purity α=0.75; one fairness item removed). Vaccination intention single item (1–7). Demographics. - Analysis: Parallel multiple mediation (PROCESS Model 4). Predictor: believing (0=atheist, 1=Catholic). Mediators: care, fairness, loyalty, authority, purity. Covariates: age, sex, education. 5000 bootstrap samples; 95% CIs. Power: Sensitivity analyses via G*Power indicated adequate power to detect small effects in Study 1 (f²=0.02, n=689) and Study 2 (f²<0.02, n=752); Study 3 nearly adequate for small effects (f≈0.03, n=469). Data and analysis software: CFA using Jamovi/lavaan; regression-based mediation via PROCESS in SPSS v29. Datasets available on OSF (link provided in paper).

Study 1 (n≈689): - Catholics reported lower COVID-19 vaccination intention than atheists; this difference was fully mediated by lower belief in science among Catholics. - Paths: Believing (Catholic) → Belief in science: b=-0.39, SE=0.03, t=-11.98, p<0.001, 95% CI [-0.458, -0.329]. Belief in science → Vaccination intention: b=0.21, SE=0.04, t=4.96, p<0.001, 95% CI [0.129, 0.297]. Direct believing → vaccination intention became non-significant when mediator included (b=-0.04, p=0.275), but was significant without mediator (b=-0.12, p<0.01). - Moderation: Group identification moderated believing → belief in science (interaction b=-0.27, p<0.001). Conditional effects stronger at higher identification: low b=-0.18; medium b=-0.37; high b=-0.64 (all p<0.001). Index of moderated mediation for vaccination intention = -0.057, SE=0.01, 95% CI [-0.087, -0.032]. Group identification did not significantly moderate believing → vaccination intention. - Covariates: Age (b=0.11, p=0.005) and education (b=0.11, p=0.004) positively predicted vaccination intention. Study 2 (n≈752; Catholics): - Intrinsic religious orientation negatively related to belief in science (b=-0.34, p<0.001) and, via belief in science, to vaccination intention (indirect b=-0.03, 95% CI [-0.059, -0.006]). The direct intrinsic → vaccination path was non-significant with mediator (b=-0.06, p=0.160) but significant without mediator (b=-0.09, p<0.05), indicating full mediation by belief in science. - Belief in science positively predicted vaccination intention (b=0.09, p=0.013). Extrinsic religious orientation positively predicted vaccination intention (b=0.13, p=0.001). Age (b=0.32, p<0.001) and sex (male b=0.08, p=0.020) were positive predictors. Study 3 (n≈469): - Catholics scored higher than atheists on binding foundations: loyalty (b=0.31, p<0.001), authority (b=0.33, p<0.001), purity (b=0.48, p<0.001). Believing did not significantly predict care; fairness showed a small negative association. - Vaccination intention predicted by loyalty (b=-0.18, p=0.006) and purity (b=-0.17, p=0.01) negatively; care showed a positive zero-order relation but was non-significant in the multivariate model; authority and fairness were non-significant. - Indirect effects from believing to vaccination intention via loyalty: b=-0.06, 95% CI [-0.104, -0.015], and via purity: b=-0.08, 95% CI [-0.145, -0.018] (both significant). Indirects via care, fairness, authority were non-significant. Total effect of believing on vaccination intention: b=-0.13, SE=0.04, t=-2.97, p<0.01; direct effect non-significant when mediators included. Across studies, Catholics had lower vaccination intentions than atheists; lower belief in science (Study 1) and higher endorsement of purity and loyalty (Study 3) explained these differences. Within Catholics, intrinsic religiosity lowered vaccination intention via reduced belief in science, whereas extrinsic religiosity increased vaccination intention (Study 2).

The findings address the central question by demonstrating that Catholics’ lower COVID-19 vaccination intentions relative to atheists are explained by both cognitive (lower belief in science) and moral (higher purity and loyalty) mechanisms. Group identification strengthened the inverse link between being Catholic and belief in science, suggesting identity-based reinforcement of worldview tensions between faith and science. Within Catholics, intrinsic religiosity reduced belief in science, fully accounting for its negative association with vaccination intention, whereas extrinsic religiosity was positively related to intention, possibly reflecting conformity and social integration effects. Moral foundations further clarified differences: stronger purity and loyalty values among Catholics were associated with lower intention, consistent with concerns about bodily sanctity and in-group moral norms that may discourage vaccination in the Polish context. These results highlight culturally contingent pathways by which religiosity influences health decisions and underscore the importance of tailoring interventions to belief in science and salient moral values.

This work advances understanding of why Catholics report lower COVID-19 vaccination intentions than atheists by identifying two key mediating pathways: diminished belief in science and greater endorsement of purity and loyalty moral foundations. It further differentiates religious orientations, showing intrinsic religiosity lowers intention via belief in science, while extrinsic religiosity elevates intention. Practically, the study suggests targeting belief-in-science interventions, especially among highly identified Catholics, and designing persuasive messages framed around moral purity (e.g., cleanliness, safety, protecting bodily integrity) and loyalty (e.g., safeguarding the community) to promote vaccination. Future research should test these strategies experimentally, examine longitudinal causality, assess generalizability beyond Poland and beyond Catholicism, and explore how religious leadership and community norms shape the observed effects.

- Sampling restricted to Catholics and atheists; other faiths excluded, limiting generalizability. - Conducted in Poland, a predominantly Catholic context; religiosity–science–vaccination relations may be culture- and religion-specific. - Multiple mediation models with bootstrap inference increase risk of Type I error; Bonferroni-adjusted checks indicated some effects (e.g., Study 1 group identity → vaccination; Study 2 belief in science → vaccination; Study 3 purity/loyalty → vaccination) may not remain significant under strict correction. - Cross-sectional design precludes causal inference; temporal precedence and causality of mediation paths require experimental or longitudinal designs. - Self-report measures and single-item vaccination intention outcome may introduce measurement limitations.