National parochialism is ubiquitous across 42 nations around the world

The study investigates whether national parochialism—greater cooperation with national ingroup versus outgroup members—is ubiquitous or varies across nations, and what factors explain its prevalence. In an increasingly interdependent world, parochialism may hinder global public goods provision (e.g., pandemic control, environmental conservation). Competing theories make different predictions: reputation-based or indirect reciprocity accounts suggest parochial cooperation arises from expectations of ingroup cooperation and reputational concerns, potentially appearing mainly when behavior is observable. Cultural evolution accounts posit that norms and institutions (e.g., rule of law, effective government, world religions) and ecological conditions (e.g., pathogen stress, relational mobility) shape group boundaries and should produce cross-national variation in national parochialism. The authors test preregistered hypotheses about the ubiquity, variability, and determinants of national parochialism across 42 nations using an incentivized prisoner’s dilemma paradigm adapted for cross-national comparability.

Prior work documents ingroup favoritism in cooperation and suggests parochial tendencies can coevolve with intergroup conflict and social norms. Much empirical evidence comes from WEIRD populations or self-reports, limiting generalizability. Meta-analytic findings indicate parochialism occurs similarly with hypothetical and real incentives, and reputation cues can increase cooperation. Theories link lower parochialism to institutions enabling safe interactions among strangers (rule of law, government effectiveness) and to prosocial norms spread by world religions. Ecological theories predict links with pathogen stress and relational mobility. However, comprehensive cross-national experimental evidence testing these determinants of national parochialism has been sparse, motivating the current large-scale study.

Design and preregistration: Hypotheses and design were preregistered (OSF: https://osf.io/68wds/). Data were collected online in December 2018 via Qualtrics. Participants and sampling: N = 18,411 adults from 42 nations, stratified by age, gender, and income, recruited via Nielsen/Harris panels. Languages were professionally translated using back-translation/committee methods. Incentives: Decisions were hypothetical in most nations; in Brazil (n≈832), India (n≈834), and Poland (n≈776), participants were randomly assigned to real monetary incentives vs hypothetical outcomes for robustness. Payoffs: In a one-shot, continuous prisoner’s dilemma (12 independent decisions), participants received 10 Monetary Units (MUs) per trial and chose 0–10 to send to a partner; each MU sent was doubled for the recipient. MU value was standardized as 2.5 minutes of the national average wage. Treatments: Within-subject manipulation of partner’s nationality across three categories—Ingroup (same nation), Outgroup (one of 16 specified foreign nations; two outgroup sets counterbalanced), and Stranger (no nationality info). Observability: Two blocks (each six trials) varied observability—Private (choices not published) vs Public (choices posted online under a self-chosen alphanumeric pseudonym; what-did-people-do.com). A pilot (MTurk, N=369) confirmed that public observability increased cooperation. Expectations: After each decision, participants reported expected partner cooperation. Additional measures: Perceived relational mobility scale (7 items). Demographics: age, gender, and education. Data quality: An attention check (do-not-respond item) was used; comprehension was assessed with a payoff question. Analytic strategy: Multilevel mixed-effects models with participant (level 2) nested within nation (level 3) random intercepts; partner nationality entered as a random slope. Two orthogonal contrasts tested: (1) ingroup vs (outgroup + stranger) for national parochialism; (2) outgroup vs stranger. Observability entered as a level-1 predictor; age and gender as level-2 covariates; nation-level indicators as level-3 predictors in cross-national analyses. Meta-analyses used the metafor R package. Effect sizes reported as unstandardized coefficients (b) and standardized mean differences (Cohen’s d).

• Incentives robustness: Across Brazil, India, and Poland, incentivized vs hypothetical treatments did not affect mean cooperation (b = −0.14, p = 0.12), national parochialism (b = 0.01, p = 0.74), or the observability effect (b = 0.04, p = 0.34).
• Ubiquity of national parochialism: Participants cooperated more with ingroup than with outgroup/strangers (mixed-effects b = 0.29, p < 0.001). National parochialism was significant in 39/42 nations; it was positive but non-significant in the remaining 3. Overall effect size for national parochialism was small-to-medium (d ≈ 0.22).
• Cultural distance and parochialism: Bilateral cultural distance did not predict differences in national parochialism between nations (b = −0.022, p = 0.70).
• Observability: Cooperation was higher in public vs private blocks (b = 0.12, p < 0.001), with a pooled observability effect size RE model d = 0.09 [0.08, 0.11]. National parochialism occurred in both public and private conditions and did not differ by observability (p = 0.59).
• Expectations: Participants expected more cooperation from ingroup members; these expectations were positively associated with national parochialism (indirect effect b = 0.06, p < 0.001).
• Cross-national predictors of parochialism: No significant associations with rule of law (p = 0.56), government effectiveness (p = 0.77), kinship norms and historical exposure to western church (p = 0.52, p = 0.70), religiosity (p = 0.44), church attendance (p = 0.11), pathogen stress (p = 0.43), or relational mobility (p = 0.66).
• Variance decomposition and moderators: Greater within- than between-nation variance in parochialism (SDwithin = 1.25; SDbetween = 0.13). Parochialism–cooperation association was stronger among men than women (b = 0.03, p = 0.04), and weaker among higher-educated individuals (b = −0.02, p = 0.02).
• Impersonal cooperation (overall giving irrespective of partner’s nationality) varied substantially across nations and tracked cultural and ecological indicators: greater cultural distance predicted larger cross-national differences in cooperation (b = 0.59, p < 0.001); higher cooperation was associated with lower historical pathogen prevalence (b = −0.65, p < 0.001), greater historical exposure to western church (b = −0.53, p = 0.003), more egalitarian values (lower power distance b = −0.49, p = 0.006; lower hierarchy b = −0.43, p = 0.04), more indulgent/loose norms (indulgence b = 0.49, p = 0.007; tightness–looseness b = 0.40, p = 0.06), more individualism/self-expression (b = 0.39, p = 0.03), and higher relational mobility (national b = 0.64, p = 0.009). At the individual level, perceived relational mobility correlated with higher cooperation (b = 0.35, p < 0.001).

Findings demonstrate that national parochialism is pervasive with remarkably little cross-national variability, challenging theories predicting substantial ecological or institutional modulation of parochialism. The presence of parochialism in both private and public contexts and the minimal moderating role of observability provide limited support for accounts that emphasize reputational incentives as the primary driver of national parochialism, though ingroup-positive expectations do play a role. Contrary patterns reported in small-scale or less industrialized societies may reflect differences in populations, tasks (e.g., allocation vs prisoner’s dilemma), or design features. Importantly, while parochial bias is stable, the baseline level of impersonal cooperation differs substantially between nations and aligns with institutional, normative, and ecological factors, suggesting distinct pathways for parochial bias versus generalized cooperation. Addressing global challenges that require cross-border cooperation may thus benefit from interventions that elevate impersonal cooperation across group boundaries, rather than solely targeting reductions in parochial preferences.

The study contributes large-scale experimental evidence that national parochialism is a stable and widespread feature across 42 nations, largely invariant to cultural distance, institutional quality, religiosity, pathogen stress, or relational mobility. However, overall cooperation with strangers varies meaningfully across nations and is associated with culture, norms, and ecology, implying that cultural evolution mechanisms likely shape impersonal cooperation more than parochial bias. Future research should develop and test strategies that expand cooperation across group boundaries, use comparable paradigms across diverse populations (including small-scale societies), and incorporate tasks that allow measurement of potential outgroup derogation or harm to better parse ingroup favoritism from outgroup hostility.

• Demand characteristics: Within-subject manipulation of partner nationality and observability might cue hypotheses; authors did not elicit participants’ stated reasons. Comparable results in hypothetical vs incentivized conditions and past literature reduce this concern.
• Incentive structure: Many nations used hypothetical payoffs; robustness checks in five nations and prior meta-analyses suggest minimal impact on parochialism.
• Uncertainty explanation: Similar variability in expected cooperation for ingroup vs outgroup/stranger partners suggests uncertainty does not explain parochialism, but residual confounds cannot be fully ruled out.
• Task generalizability: The prisoner’s dilemma may not capture outgroup derogation or aggression; alternative paradigms allowing costly outgroup harm could reveal different patterns.
• Effect sizes: Effects are small-to-medium (overall d ≈ 0.22), typical for large social science samples; small effects may still have cumulative societal impact.