Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Religion, Prosociality, and Moral Cooperation

Abstract. Religion and prosociality are linked through multiple interacting psychological processes: beliefs about supernatural monitoring, ritual commitments, reputation, community identity, and institutional opportunities for cooperation. This fundamental research monograph synthesizes original experiments, cultural-evolutionary theories, historical and ethnographic comparisons, experimental economic games, cross-cultural field studies, and meta-analytic evidence. It critically distinguishes generosity toward strangers from religious ingroup assistance, charitable self-presentation from observable behavior, and short-term conceptual priming from sustained institutional participation. The 2024 quantitative synthesis reports a modest average correlation between religiosity and prosociality, weaker for directly observed behavior than for self-report. Divergent evidence is evaluated through construct validity, selection effects, replication, and cross-cultural inference. The monograph also examines corrections and retractions in developmental and historical studies, including debates over whether moralizing gods preceded sociopolitical complexity. Its integrative conclusion specifies a multidimensional research program for studying religious cooperation, norm enforcement, social inclusion, and moral boundaries. The method is a critical narrative synthesis rather than a preregistered systematic review, an original experiment, or a new meta-analysis.

Keywords: religious prosociality; moral cooperation; religiosity; supernatural monitoring; moralizing gods; religious priming; costly signaling; credibility-enhancing displays; altruism; trustworthiness; religious identity; outgroup generosity; cross-cultural economic games; religious institutions; ritual; prosocial behavior

Introduction: The Psychological Problem of Religious Cooperation

Religion, prosociality, and moral cooperation form a research problem at the intersection of social psychology, cultural evolution, behavioral economics, and anthropology. Prosociality refers to actions or preferences that benefit others, including sharing, helping, generosity, trust, and costly contributions to common goods. Cooperation concerns the coordinated provision of mutual or collective benefit under conditions in which individuals might profit from withholding a contribution. Moral cooperation adds beliefs about obligation, fairness, sanctions, legitimate recipients, and collective enforcement. Religious participation can influence each component through interpersonal networks, moral teachings, imagined observers, rituals, identities, and institutions. This dossier examines those mechanisms separately because a charitable self-description, a donation to a coreligionist, an anonymous transfer to a stranger, and compliance with a punitive norm are different observations. The core question is which dimensions of religion predict which dimensions of cooperation, among whom, and in what conditions Norenzayan & Shariff (2008) Galen (2012) Kelly et al. (2024).

The proper explanatory unit varies across studies. Individual religiosity includes belief, salience, identification, attendance, private practice, and institutional affiliation; communities supply repeated interaction, reciprocal obligations, and opportunities for collective action; cultural ecologies regulate how strangers interpret moral conduct. A supernatural punishment model emphasizes expectations about unseen accountability, whereas social-functional models emphasize belonging and institutional embeddedness. These explanations may coexist within a population, and a single observed correlation seldom identifies their relative contributions. The monograph therefore distinguishes individual disposition, momentary psychological activation, dyadic recipient choice, community practice, and population-level social organization. It also treats religion as culturally diverse, including traditions whose cooperation systems do not center on an omniscient moralizing deity Graham & Haidt (2010) Purzycki et al. (2016).

The research method is a critical narrative synthesis of identifiable original experiments, cross-cultural behavioral games, observational and longitudinal studies, historical comparisons, theoretical contributions, systematic assessments, and meta-analyses. Claims are ordered by design strength and accompanied by explicit discussion of selection, measurement, and causal interpretation. The bibliography records authentic publisher or indexing links; accessible abstracts support only claims stated in those abstracts, while fuller original material was examined where accessible. No new participants were recruited, no data were reanalyzed, and the present narrative synthesis should not be interpreted as a registered systematic review. Its purpose is to reconstruct a complex explanatory literature and state what different forms of evidence warrant Shariff et al. (2016) Kelly et al. (2024).

Conceptual Foundations: Religion, Prosociality, and Moral Norms

Religious belief is a proposition concerning the sacred, supernatural beings, or a religious cosmology; ritual practice is structured action, often performed in the presence of others; religious affiliation is a social relationship; and religiosity is an umbrella variable that instruments operationalize in different ways. Treating these as interchangeable conceals the causal mechanism. For example, service attendance may predict volunteering because local organizations recruit volunteers, while beliefs about supernatural punishment may predict rule-following in an anonymous task. A self-rating of spiritual importance can have still another association. Theories of religious cooperation require explicit predictions for each indicator and must account for their covariance with education, socioeconomic status, family socialization, political identity, and national institutions Galen (2012) Graham & Haidt (2010).

Prosociality likewise has several operational meanings. An allocator's gift in a dictator game measures willingness to sacrifice an assigned sum under task-specific expectations. A public-goods contribution measures participation in a social dilemma, while cooperation in a prisoner's dilemma additionally depends on expectations about a partner. Trust games introduce beliefs about others and responses to vulnerability. Volunteering measures behavior sustained by schedules, organizations, opportunity, and identity; charitable giving depends on resources and recipient selection. Moral judgment, empathy, reputation, perceived trustworthiness, and observed reliability belong in distinct columns of an evidence map. No one instrument is a universal index of moral character Norenzayan & Shariff (2008) Kelly et al. (2024) Valencia Caicedo et al. (2023).

Religious morality may encompass fairness, compassion, duty, loyalty, authority, purity, forgiveness, and obligations toward strangers. A norm that increases assistance within one community may strengthen group continuity while simultaneously marking boundaries against outsiders. The concept of cooperation must therefore specify beneficiaries and excluded parties, the voluntary or coercive nature of contribution, the intensity of sanction, and the distribution of costs. In a dyadic experiment, generosity toward a religious outgroup has a different theoretical meaning from generosity within one's congregation. Preston and Ritter's experimental comparison of God-related and religion-related primes illustrates the importance of distinguishing concept content and recipient category rather than treating all reminders of religion as equivalent Preston & Ritter (2013).

Secular cooperation offers a central comparison, since reputational incentives, repeated interaction, moral norms, monitoring, punishment, and effective institutions can coordinate behavior regardless of theology. Religious and secular moral cues can partially overlap in the psychological functions they activate. Shariff and Norenzayan reported experimentally that secular moral-institution primes also increased giving relative to neutral conditions. This finding is compatible with a more general accountability mechanism, although a shared behavioral effect does not establish that the underlying thoughts were identical. Religious-specific effects should be evaluated against appropriately matched secular alternatives rather than against the absence of social organization Shariff & Norenzayan (2007) Bloom (2012).

Historical Development of the Research Programs

Early social-scientific approaches to religion and morality addressed the collective organization of rituals, social obligations, and community authority. Contemporary behavioral research translated these broad problems into testable predictions involving costly signals, economic decisions, and representations of supernatural observers. The 2003 historical analysis of nineteenth-century American communes by Sosis and Bressler treated community survival and demanding rules as observable indicators in a theory of commitment. In the same year, Sosis and Ruffle compared economic-game cooperation in religious and secular Israeli kibbutzim. These investigations made membership, costly participation, and collective action empirically examinable, although the social settings were too distinctive to warrant conclusions about all religions or all cooperative behavior Sosis & Bressler (2003) Sosis & Ruffle (2003).

A second program developed experimental manipulation of accessible religious ideas. Shariff and Norenzayan's 2007 dictator-game studies tested the effect of implicit God-related concepts on anonymous allocation. Pichon, Boccato, and Saroglou likewise examined primed religious concepts and accessible prosocial associations. Norenzayan and Shariff's 2008 synthesis subsequently proposed that reputation and imagined supervision might help bridge the social-scale problem of cooperation among people who lack repeated personal contact. By 2010, Graham and Haidt had articulated a social-functional account of religions as moral communities, broadening the explanatory emphasis from private propositions to group processes. These lines of research coexist because laboratory accessibility, community reputation, and institutional coordination occupy different levels of analysis Shariff & Norenzayan (2007) Pichon et al. (2007) Norenzayan & Shariff (2008) Graham & Haidt (2010).

An important phase of criticism refined the central hypothesis. Galen's 2012 review asked whether conclusions derived from self-report, stereotypes, and selected social settings exceeded the evidence for changed behavior. The 2016 priming meta-analysis by Shariff and colleagues evaluated aggregated experiments and publication-bias sensitivity. Purzycki and colleagues brought the moralizing-gods account into diverse field settings, using interviews and allocation games. Subsequent work directly examined religious outgroups and the effects of experimental prompts, while Kelly, Kramer, and Shariff's 2024 meta-analysis consolidated decades of individual-level studies. The chronology displays scientific maturation: a striking result generates hypotheses, design-sensitive criticism, broader cross-cultural testing, and more precise aggregate estimates Galen (2012) Shariff et al. (2016) Purzycki et al. (2016) Pasek et al. (2023) Kelly et al. (2024).

Causal Questions and Levels of Explanation

The relationship between religious participation and helping behavior can reflect selection into groups, effects of participation, mutually reinforcing processes, or a third cause. People with high social motivation might join congregations and volunteer, thereby producing a correlation even without an independent theological effect. Congregations can also create repeated opportunities for helping that strengthen habits, social recognition, and access to charitable programs. Experimental priming addresses momentary activation of meanings under random assignment, but cannot by itself estimate the lifelong effects of living within religious networks. Cross-national comparisons test structural relationships, but national institutional history, demographic composition, and measurement differences complicate causal attribution Galen (2012) Valencia Caicedo et al. (2023).

Evidence becomes more informative when a study separates the treatment, psychological mediator, observed outcome, and population boundary. If religious cues increase generosity, researchers might test whether the cue changes perceived monitoring, anticipated sanction, accessible norms, social identity, empathy, or the interpretation of an anonymous recipient. These mediators are conceptually distinct and can be experimentally varied or measured at successive times. Even well-designed randomization of a prompt establishes a causal effect of the prompt under that setting, not a unique causal effect of religion as a whole. A theoretically disciplined synthesis therefore seeks convergent mechanisms without treating one experimental manipulation as a proxy for an entire civilization Shariff & Norenzayan (2007) Preston & Ritter (2013) Pasek et al. (2023).

Cooperation is multiply determined. The repeated-game problem, conformity to norms, credible commitments, and governance arrangements provide general background conditions within which religious communication can operate. Mechanisms may substitute for or amplify one another: explicit secular sanctions may reduce the marginal effect of supernatural accountability, whereas weak formal enforcement may make communal reputation more salient. The globally comparative analysis by Valencia Caicedo, Dohmen, and Pondorfer reports positive associations between reported religious belonging and prosocial preferences, with institutional context as a moderator. Such patterns motivate interactions between institutional quality and religious sociality, while their observational nature leaves important confounding explanations open Valencia Caicedo et al. (2023).

Supernatural Monitoring and Moral Accountability

Supernatural monitoring theories start with a recognizable cooperation problem: an individual may benefit from violating a rule when other people cannot observe the violation. The belief that a morally concerned supernatural agent can detect concealed misconduct might change the actor's expectation of detection or moral sanction. That prediction differs from the broader claim that a religion teaches kindness. A god conceived as knowledgeable about intentions, punitive toward wrongdoing, and invested in interpersonal norms supplies an especially specific monitoring representation. Cultural distributions of these representations vary, and individuals may attribute distinct moral concerns to the same religious figure Norenzayan & Shariff (2008) Purzycki et al. (2016).

The two 2007 studies by Shariff and Norenzayan manipulated implicit accessibility of God-related ideas before a dictator-game decision. Participants exposed to religious concepts allocated more to strangers than comparison groups. Similarity to the effect of secular moral-authority prompts suggested that socially learned concepts of accountability could be important. The design offered random assignment of the cue and a costly choice, an important advance over correlations between questionnaire religiosity and charitable self-description. Nevertheless, the measured dependent variable was a short economic allocation under laboratory instructions, and the design did not conclusively separate social surveillance from generosity schemas, self-presentation, or expectations about the experiment Shariff & Norenzayan (2007).

Purzycki and colleagues examined moralistic gods through field interviews and economic games among 591 participants in eight culturally diverse communities, yielding 35,400 allocation observations. They reported that greater attribution of punitive and knowledgeable qualities to gods predicted more impartial allocations to geographically distant coreligionists in their game. This pattern connects the monitoring model to distribution across social distance, and the research deliberately included both world religions and local traditions. Interpretation requires preserving the recipient boundary: a distant coreligionist is an expanded religious ingroup member and does not automatically stand for an unrestricted stranger. Furthermore, naturally held beliefs were measured rather than randomly assigned, so the findings primarily support a cross-cultural association consistent with the proposed mechanism Purzycki et al. (2016).

An adequate monitoring test would independently vary whether observers are supernatural or human, whether they know private actions or public behavior, and whether they punish, reward, or merely care about moral outcomes. It would also measure participants' beliefs about the agent rather than assuming that a shared label denotes a shared concept. The predicted effect on cooperation should be most clearly visible when detection is otherwise unlikely, yet institutional punishments and generalized moral commitments can generate similar patterns. This is a reason to specify alternative pathways in the study design rather than interpret every increase in giving as evidence of fear of divine punishment Bloom (2012) Pasek et al. (2023).

Religious Priming and the Reproducibility Problem

Religious priming attempts to manipulate immediate accessibility of religious representations without changing a participant's long-standing religious affiliation. Investigators have used scrambled-sentence tasks, words related to religion, subtle representations, and explicit prompts to think about God. The methods differ in visibility and content. A word associated with institutional religion may evoke group participation, while a concept of a watching deity may activate an expectation of moral accountability. An explicit request to contemplate a god may also communicate expectations about the experiment. Causal attribution is strongest for the particular manipulation and outcome tested, and more ambitious interpretations require mediation and generalization studies Pichon et al. (2007) Shariff & Norenzayan (2007) Preston & Ritter (2013).

Shariff, Willard, Andersen, and Norenzayan's meta-analysis, published in the 2016 volume of Personality and Social Psychology Review, covered 93 studies with 11,653 participants. The authors combined effect-size analyses, p-curve examination, and publication-bias adjustments. They reported effects across several outcomes, including prosocial measures, while finding that religious primes did not reliably influence nonreligious participants. This boundary is psychologically informative: an activation mechanism depends on accessible learned beliefs and their significance for the person. It also means experimental samples with varying religious literacy should not be merged conceptually without testing moderation Shariff et al. (2016).

The existence of an aggregate positive estimate does not settle whether every priming paradigm replicates, whether published studies cover the full distribution of null findings, or whether brief task outcomes predict meaningful cooperative behavior in ordinary life. Small samples, analytic flexibility, variation in prime content, incomplete reporting, and weak blinding can distort effects. Design comparison should distinguish explicit religious instruction from implicit cues, simple word accessibility from social-norm salience, and outcomes such as generosity, honesty, helping intention, and punitive judgment. Future preregistered multisite experiments should report effects by religiosity, familiarity, task expectancy, local norms, and recipient identity, including meaningful null outcomes Shariff et al. (2016) Galen (2012).

Preston and Ritter's three-study investigation found different patterns for religious and God-related primes: the former favored ingroup-oriented prosocial responses, whereas the latter favored outgroup-oriented responses in the reported tasks. The result does not show an immutable feature of theology; it shows that conceptual distinctions matter for experimental construct validity. The later cross-cultural experiments of Pasek and collaborators used explicit attention-to-God manipulations among believers and provide a separate, broader test of generosity beyond a single local experiment. The results encourage an analysis of what participants believe God wants, whom they consider eligible for aid, and how those expectations change with political and cultural context Preston & Ritter (2013) Pasek et al. (2023).

Ritual Commitment, Costly Signaling, and Credibility

Costly signaling models propose that difficult, visible, or resource-intensive rituals can communicate commitment to a group's expectations. If feigning commitment requires an appreciable sacrifice, other members may use that action when choosing collaborators and allocating trust. The psychological logic concerns observers' inferences and participants' incentives; it does not require that painful or expensive practices are intrinsically morally valuable. A signal might stabilize shared expectations, reduce uncertainty about future participation, or screen prospective members. Yet a practice can be costly without being reliable evidence of actual kindness, and some signals may be valuable primarily as badges of social position Sosis & Bressler (2003) Henrich (2009).

Historical data on 83 nineteenth-century American communes allowed Sosis and Bressler to test relationships among religiously demanding restrictions, communal commitment, and organizational longevity. Longer survival is an institution-level outcome: it may reflect cohesion, resource pooling, leadership, demographic conditions, or exclusion of dissent, as well as cooperative work. The study supports aspects of the signaling approach but cannot establish that the same relationship would hold in congregations without communal property or in modern voluntary associations. Its contribution is a distinct methodological strategy: comparing institutional histories where costly commitment is recorded independently of retrospective self-evaluations of moral character Sosis & Bressler (2003).

Sosis and Ruffle's study of religious and secular Israeli kibbutzim used a behavioral cooperation setting to examine ritual participation and cooperative conduct within organizations that shared some institutional features. Attendance at collective worship was theoretically relevant because it indicated repeated group participation, potentially visible reliability, and common ritual obligation. However, distinguishing ritual from group selection, social monitoring, kin ties, leadership, and local economic structures remains difficult. The central scientific lesson is that contextual comparison can make religious practice observable but does not automatically randomize belonging or isolate an individual psychological mediator Sosis & Ruffle (2003).

Henrich's account of credibility-enhancing displays adds a learning mechanism. Cultural learners may trust professions of belief more when teachers' conduct makes deception appear costly. A leader's costly devotion might thereby influence observers' beliefs and their commitment to common practice even before explicit calculation of mutual benefit. This process is compatible with ordinary social learning and can be applied to secular ethical commitments. Its explanatory scope differs from punishment models: credible action affects learning about a norm and trust in its transmitters, while supernatural monitoring affects anticipated accountability when behavior is hidden Henrich (2009).

Collective Ritual, Synchrony, and Shared Action

Religious collective action includes recurring worship, coordinated movement, shared recitation, calendrical observance, pilgrimage, and extraordinary high-arousal ceremonies. These activities organize attention, bodily timing, emotion, and interpersonal expectations. Synchrony and shared arousal have been proposed as psychological routes to attachment and collective identification, although a ritual's specific religious meanings and rules also determine the social outcome. A procession may create visible joint commitment, while private worship provides a different set of social cues. Analytical comparison must therefore attend to bodily co-presence, audience composition, frequency, expense, emotional intensity, and perceived sacred significance Graham & Haidt (2010) Xygalatas et al. (2013).

Xygalatas and colleagues' 2013 study, Extreme Rituals Promote Prosociality, investigated participants and observers in an intense devotional ritual and used donations as a behavioral outcome. It contributes evidence that unusual, costly ceremonies may be associated with giving beyond the performer alone, consistent with emotional engagement and community signaling. However, the immediate beneficiary, the particular religious institution, and the special occasion matter: donations toward a temple cannot be translated directly into evidence of universal generosity. The field design has ecological strengths but also requires careful treatment of prior commitment and social affiliation Xygalatas et al. (2013).

Shared practices may organize cooperation through recurrent contact, mutual recognition, expectations of reciprocity, and opportunities for norm enforcement. Some elements can persist in secular volunteer organizations, sports teams, and civic ceremonies, allowing mechanism-focused comparisons. At the same time, practices that strengthen cohesion may increase demands for conformity or reduce trust in outsiders; the relevant outcome is thus a vector of benefits, costs, and beneficiaries rather than a single moral score. These interactions link ritual research with social identity theory and institutional psychology but are best tested through designs that separate group exposure from doctrinal content Graham & Haidt (2010) Bloom (2012).

Ingroup Prosociality, Outgroup Boundaries, and Moral Scope

Cooperation must be indexed to whom it benefits. Religious institutions establish membership categories through shared narratives, sacred obligations, rites of passage, and communal boundaries. Such categories can direct care to local vulnerable people, connect geographically distant coreligionists, and shape expectations about strangers. These effects can increase prosocial action within a group while leaving other beneficiaries unchanged. The term parochial prosociality describes assistance conditional on collective identity, and intergroup cooperation describes costly assistance across salient group boundaries. Neither outcome can be inferred solely from the aggregate amount donated because distribution among recipients carries distinct moral and psychological implications Preston & Ritter (2013) Purzycki et al. (2016).

In experimental comparisons, a participant's identity and a recipient's category should be crossed rather than collapsed. A study comparing allocation to a local coreligionist, a distant coreligionist, a different-religion recipient, and an explicitly nonreligious recipient can distinguish local obligation, extended religious solidarity, and broader moral inclusion. Purzycki and colleagues found associations involving distant coreligionists, an important enlargement beyond face-to-face membership. Preston and Ritter reported that different primes directed prosocial responses toward different groups. These studies together recommend a multidimensional analysis of recipient identity and religious representations rather than a binary religious-versus-nonreligious goodness comparison Purzycki et al. (2016) Preston & Ritter (2013).

Cross-cultural experimental research by Pasek and colleagues in 2023 enrolled 4,753 Christian, Muslim, Hindu, and Jewish adults in settings across the Middle East, Fiji, and the United States. Asking participants to think about God increased monetary giving by 11 percent relative to the comparison condition, corresponding to 4.17 percent of the total stake, and the reported increase extended comparably to religious ingroup and outgroup recipients. This design supplies more diverse evidence than a single Western student sample and directly addresses a central boundary hypothesis. The result remains the effect of a specific prompt on assigned monetary decisions among sampled believers; it should be interpreted alongside baseline differences, cultural diversity, and possible variation in attributed divine preferences Pasek et al. (2023).

Religious outgroup relations are contingent on norms and history. Some contexts encourage hospitality and common humanity, while others make purity, loyalty, or perceived threat especially salient. Group membership can foster valuable mutual aid and generate duties that compete with universal distributive ideals. The outcome of interest must therefore specify whether the research concerns everyday cooperation, fairness between unfamiliar people, protection of vulnerable outsiders, or punishment of perceived defectors. Theory should predict a distribution of behavior across targets, including the possibility of simultaneous generosity and exclusion, rather than assert that one collective trait explains all conduct Bloom (2012) Graham & Haidt (2010).

Trust, Reputation, and Perceived Religious Commitment

Trust is a belief about another person's likely conduct under vulnerability; trustworthiness is the person's actual tendency to honor obligations when entrusted with something valuable. Cooperation may depend on both, but the two measures are not equivalent. Religious identity markers can affect how trustworthy someone appears before any interaction, while actual cooperative performance may not follow the same pattern. Theoretical models of reputation predict that public ritual participation, shared moral vocabulary, and organizational membership alter social expectations, especially when independent evidence about the partner's reliability is scarce Norenzayan & Shariff (2008) Henrich (2009).

Reputation-based hypotheses can be tested through partner-selection tasks and trust games in which researchers independently manipulate identity badges, previous cooperative behavior, and the visibility of past conduct. If a religious badge boosts trust even after information about behavior is held constant, the effect concerns categorical expectations. If documented ritual attendance predicts later returns or actual cooperation under appropriate controls, an additional behavioral account becomes plausible. Yet performance may reflect social monitoring and opportunity rather than private moral conviction. Observers' favorable stereotypes can also affect the opportunities available to religious participants, generating self-reinforcing social advantage Galen (2012) Norenzayan & Shariff (2008).

Religious signaling research reviewed in 2024 emphasizes the difference between being perceived as especially trustworthy and demonstrating superior trustworthiness. Its synthesis notes mixed evidence for actual cooperative conduct and indications that routine religious cues may be associated with ingroup preferences. High-cost, highly social practices may sometimes enlarge social identity, but findings are heterogeneous across practices and settings. Accordingly, a scientific model should measure belief in another person's reliability separately from their decisions, and include a secular civic-association control where possible Northover et al. (2024) Henrich (2009) Sosis & Bressler (2003).

One important question is whether reputational advantages and the actual capacity to sustain collective goods can diverge. Social systems may reward visible conformity while overlooking unobserved caregiving or anonymous contributions. Institutional incentives may encourage symbolic demonstrations rather than costly benefits to outsiders. Cooperative quality therefore requires evaluation of what was supplied, who received it, which obligations were honored, whether participation was free, and whether the benefits persisted. A community's positive reputation is informative about members' perceptions but does not substitute for direct behavioral measurement Galen (2012) Kelly et al. (2024).

Charity, Volunteering, and Institutional Opportunities

Surveys frequently associate religious involvement with voluntary service or charitable participation, yet the causal pathways require decomposition. Congregations create recurring recruitment channels, moral language about duties, personal requests for assistance, and reliable schedules for collective projects. Attendance can increase exposure to opportunities that would otherwise be absent. Some gifts circulate within religious organizations, maintaining facilities or supporting fellow members; other gifts reach beneficiaries across group boundaries. A sound study records the destination, opportunity structure, amount or time invested, and the social conditions of the decision rather than treating all transfers as identical Galen (2012) Graham & Haidt (2010).

Volunteering may express altruism, reciprocity, identity, role commitment, or a combination. Contributors sometimes receive companionship, recognition, professional skills, or personal meaning. These benefits do not nullify the prosocial value of their work; they instead show why motivational categories should not be inferred from behavior alone. Studies should distinguish stated reasons for helping from the real costs and consequences of assistance. Religious and secular organizations can organize similar collective activities, enabling matched comparisons of opportunity, support, governance, and social reward Bloom (2012) Graham & Haidt (2010).

Longitudinal panels improve identification by measuring religious participation and cooperative outcomes at several time points. Yet even person fixed effects cannot fully resolve time-varying confounding, measurement error, reverse causation, or shared changes in family and community life. When a participant increases attendance and later volunteers more, it remains necessary to ask whether the same life transition brought new relationships or altered priorities. Natural experiments that change access to local religious organizations or civic alternatives can provide additional leverage, but careful exclusion restrictions and pretrend checks are needed Aksoy & Wiertz (2024) Valencia Caicedo et al. (2023).

A comprehensive account should also assess distributional inequalities. Volunteer recruitment depends on free time, health, transportation, caregiving duties, social class, and institutional inclusion. Measures of religious service attendance may understate the practical contribution of isolated caregivers who lack opportunity to join public collective activities. Treating participation as a simple virtue scale risks confusing capacity with moral preference. Better evidence records the accessibility of helping contexts, the beneficiary's actual receipt of support, and the longer-term sustainability of the contribution Kelly et al. (2024).

Cross-Cultural Variation and the Institutional Ecology of Cooperation

The same religious label can identify different practices, cultural histories, institutions, and expectations across societies. Comparative research must avoid treating Christianity, Islam, Buddhism, Hinduism, Judaism, or local traditions as internally uniform psychological treatments. A person's response to a punitive deity question may reflect whether moral surveillance is emphasized locally, how terms are translated, and whether a scale captures the relevant indigenous concept. Similarly, trust decisions depend on local norms of stranger exchange and the economic meaning of the transferred stakes. Cross-cultural evidence should test interpretive and measurement equivalence before attributing a statistical difference to doctrine Purzycki et al. (2016) Pasek et al. (2023).

Valencia Caicedo, Dohmen, and Pondorfer analyzed globally representative social-preference data across major religious traditions. They reported that religious respondents expressed higher levels of positive reciprocity, altruism, and trust, while negative reciprocity and punishment patterns varied among traditions. Their analysis also identified contextual moderation: the association was stronger in more populous societies and weaker where secular institutions functioned more effectively. Such findings motivate a functional-substitution hypothesis, in which alternative institutions can address cooperation problems. They remain observational associations, so institution quality, selection into religious categories, and local political history need careful treatment Valencia Caicedo et al. (2023).

Cross-cultural cooperation results also caution against assuming large and stable differences in behavior between national cultural groups. A 2022 preregistered meta-analysis of impersonal cooperation across 70 societies and many experimental dilemmas reported limited cross-societal variance after adjustment for study characteristics. This finding does not refute context-sensitive religious mechanisms, since country and religious affiliation are imperfectly aligned and many outcomes differ. It does constrain sweeping claims that a single broad cultural factor straightforwardly determines cooperative behavior across the globe. Generalization must be established through comparably measured actions and explicit moderators rather than broad historical narratives Spadaro et al. (2022) Valencia Caicedo et al. (2023).

The study of cultural ecologies benefits from multilevel modeling that distinguishes personal religiosity, community religious concentration, legal institutions, and the norms of actual interaction partners. An individual may identify weakly with a tradition within a highly religious society; another may be devout within a largely secular environment. Both face incentives shaped by public expectations. Strong causal claims require attention to migration, historical selection, religion-state relationships, inequality, and secular civic infrastructure. Religion and cooperation can be causally connected in multiple directions across this ecological system Norenzayan & Shariff (2008) Valencia Caicedo et al. (2023).

Measurement: Reports, Behavior, and Moral Judgments

Self-reported religiosity and self-reported prosociality can share response styles and idealized self-views. Questions asking whether one is compassionate or generous may be interpreted as statements of moral identity rather than accurate records of conduct. A behavioral economic game provides a more constrained observation, but its hypothetical or small material stakes, anonymity, instructions, and short duration also limit interpretation. Administrative records of volunteering, audited charitable donations, and independent recipient reports can improve ecological relevance while introducing unequal opportunities, selection, and privacy constraints. Evidence assessment must preserve these differences rather than average every measure into an unqualified moral outcome Galen (2012) Kelly et al. (2024).

Kelly, Kramer, and Shariff's 2024 Psychological Bulletin meta-analysis pooled 701 effects across 237 samples and 811,663 people. The overall association between religiosity and prosociality was r = .13, with substantial heterogeneity. Studies using self-reports yielded a stronger association (r = .15) than studies using directly measured behavior (r = .06). This precise moderator is central to interpretation: both forms of evidence are informative, and their divergence can reflect biased self-enhancement, limitations of short behavioral tasks, or processes common to the two self-report scales. The meta-analysis does not decide uniquely between these accounts, and the small behavior-level effect should not be inflated into a prediction of large differences in individual moral character Kelly et al. (2024).

Behavioral games must be read in terms of incentives and social information. A dictator-game allocation is an asymmetric transfer; public-goods cooperation usually entails contribution with benefits spread across players; a trust game includes the risk of another participant's response. Participants' beliefs about whether the experimenter watches them, whether the recipient is real, and whether their choice has future consequences can affect results. Anonymous in the investigator's description does not guarantee psychologically experienced anonymity. Experimental studies of religion should therefore measure or manipulate perceived observability and clarify economic stakes, actual payment, deception, partner characteristics, and the nature of random assignment Shariff & Norenzayan (2007) Purzycki et al. (2016).

Construct-valid measurement also requires attention to morality's multiple domains. An individual may strongly condemn dishonesty yet offer little financial assistance, contribute within a congregation while excluding outsiders, or help strangers despite rejecting supernatural propositions. Such patterns are possible because rule compliance, distributive generosity, fairness, empathy, and moral concern are partially distinct. A multivariate research program will measure them separately and test whether religious predictors retain incremental validity after relevant social affiliation, personality, institutional exposure, and cultural norms are modeled Graham & Haidt (2010) Decety et al. (2019) Galen (2012).

Evidence Hierarchies and Confounding

Randomized experiments identify the average effect of an assigned manipulation under the trial's conditions, subject to protocol adherence, sufficient power, and valid measurement. Religious-priming studies therefore offer evidence about the accessibility of learned concepts; they do not simulate years of theological formation or eliminate long-term selection. Correlational surveys capture a broader social population but cannot by themselves distinguish religion's influence from the characteristics that lead people into religious participation. Ethnographic and historical studies disclose institutional processes missed by laboratory games yet face difficulties of replication and causal isolation. Triangulation is valuable when each design is assigned a specific inferential role Shariff et al. (2016) Sosis & Bressler (2003) Purzycki et al. (2016).

The causal structure can be described as an interlocking system: family upbringing and local institutions influence religious involvement; involvement shapes network membership, recurrent rituals, belief accessibility, and opportunities for service; these pathways influence expectations of others and actions; previous actions may in turn strengthen community identification. Socioeconomic resources, personality, cohort, national regulation, local social conflict, and selection into studies influence several links simultaneously. Treating all these influences as controls can be just as misleading as ignoring them, because a variable may be a mediator rather than a confounder. Careful studies should specify causal diagrams before estimation and conduct sensitivity analyses for missing variables Galen (2012) Valencia Caicedo et al. (2023).

Research quality should include preregistered hypotheses, defensible sample-size planning, exact reporting of exclusions, outcome transparency, data sharing where ethically feasible, and tests of measurement comparability. Distinct levels of clustering matter in congregation or village studies: many observations from few communities do not constitute independent community-level replications. Field games should document exchange rates and the material significance of stakes. Cross-cultural projects should report translation procedures, local research partnerships, and ethical responses to potentially sensitive religious differences. Strong designs state which inference is supported and where extension remains conjectural Purzycki et al. (2016) Pasek et al. (2023).

Quantitative synthesis requires attention to heterogeneity, dependence among effects, research-design moderators, and publication-selection processes. One sample can generate many outcomes, and multiple studies may reuse participants, so naive pooling can overstate precision. Reliance on statistically significant primes can produce biased publication records. The 2016 priming review used several analytic methods to address these problems, while the 2024 broad meta-analysis clarified the difference between attitudinal and behavioral measurement. Neither finding licenses indiscriminate aggregation across religiosity, spiritual belief, ritual frequency, and organizational participation Shariff et al. (2016) Kelly et al. (2024).

Historical Causation and the Moralizing-Gods Debate

An evolutionary-historical hypothesis proposes that morally concerned supernatural agents helped populations coordinate cooperation as human settlements and institutions expanded. The mechanism is plausible at a functional level, but demonstrating its historical primacy requires dating changes in beliefs, social complexity, and institutions with sufficient precision. Ethnographic associations between currently held beliefs and allocations to strangers do not independently establish which feature emerged first in ancient history. Religious traditions themselves can evolve in response to expanding states, markets, administrative records, and law. Distinct hypotheses therefore concern present-day psychological effects, cultural selection of doctrines, and historical temporal precedence Norenzayan & Shariff (2008) Purzycki et al. (2016).

The 2019 Nature paper Complex Societies Precede Moralizing Gods Throughout World History became a prominent attempt to reconstruct the sequence using Seshat historical data. Its original report should not be used as a valid empirical estimate because the paper was formally retracted in 2021 following a dispute over the treatment of missing values. Beheim and colleagues argued that recoding unknown information as absence shaped the inference. The authors acknowledged problems with the original analysis and issued a retraction while defending aspects of their interpretation. A responsible scientific history identifies the retraction and then evaluates newer corrected analyses on their own terms rather than silently treating the original result as intact Beheim et al. (2021) Whitehouse et al. (2021) Whitehouse et al. (2023).

Whitehouse and collaborators published a corrected and expanded analysis in 2022, Testing the Big Gods Hypothesis With Global Historical Data: A Review and Retake. This paper argued that large increases in social complexity generally preceded the emergence of moralizing big gods in the data and that its revised methods supported the central temporal claim of the earlier investigation. Other analysts have emphasized the sensitivity of results to coding, missing observations, definitions, and time series assumptions. The revised work is a distinct publication; the retraction status of the 2019 paper remains. The debate illustrates why historical causal theories require transparent data provenance, alternatives for unobserved cases, and reproducible model specification Whitehouse et al. (2023) Beheim et al. (2021).

Even a firmly established chronological ordering would leave additional questions. A supernatural monitoring system could maintain cooperation after large institutions appear without having caused their initial formation. Conversely, other collective institutions may facilitate the spread of religious doctrines after they become administratively or socially advantageous. Cross-cultural histories need to distinguish presence of a moralizing god, perceived supernatural punishment, general ritual obligation, and actual enforcement practices. Each can have a different trajectory. Psychological experiments and archaeological reconstructions should be integrated through clearly labeled hypotheses rather than mutual substitution of evidence Purzycki et al. (2016) Whitehouse et al. (2023).

Revisions, Retractions, and Scientific Integrity

The literature on religion and cooperation provides unusually clear examples of how correction should shape synthesis. A high-profile investigation of religiousness and children's altruism published in Current Biology in 2015 was retracted by its authors after problems involving the treatment of country of origin. It cannot serve as independent established proof that children from religious households are less altruistic across the world. The article's retraction notice explains that country origin drove several reported relationships after reanalysis and that the investigators withdrew the paper to correct the record. Research on developmental religious prosociality must therefore be evaluated from other validated designs Galen (2012).

The 2019 moralizing-gods historical paper was also formally retracted, and a subsequent corrected analysis was published separately. These events show that transparent scientific revision can change the evidential standing of a claim without ending investigation of its underlying question. A monograph must distinguish evidence reported in a retracted source from evidence available through surviving independent experiments or later analyses. The original disputed articles can be discussed as episodes in the methodological history, while core substantive conclusions should rest on unretracted evidence with identifiable methods Beheim et al. (2021) Whitehouse et al. (2023).

Preregistered replication, publication of null findings, open code, and active study of moderators are particularly consequential because religion is connected to moral identity and social expectations. Investigators and participants can infer the expected direction of findings, sometimes altering responses through demand characteristics or interpretive flexibility. Research teams should separate questions about the public value of religion from testable claims about behavior and avoid treating statistically significant associations as moral verdicts on populations. Scientific appraisal concerns mechanisms, measured outcomes, and uncertainty, with the same standard applied to religious and nonreligious comparisons Shariff et al. (2016) Kelly et al. (2024).

Longitudinal Religious Participation and Prosocial Outcomes

A central limitation of the classic religious-prosociality literature is the difference between a point-in-time association and a change associated with an individual's changing religious involvement. Aksoy and Wiertz applied cross-lagged panel models with individual fixed effects to two substantial British longitudinal surveys. The British Household Panel Survey covered 1991–2009 with approximately 26,000 respondents; the UK Household Longitudinal Study covered 2009–2021 with approximately 80,000 respondents. Their measures included religious service attendance, generalized trust, reported volunteering, and an interviewer assessment of the respondent's cooperativeness. The cross-wave design and person-level fixed effects address stable personal differences and some reverse-causation patterns that ordinarily frustrate one-time survey comparisons Aksoy & Wiertz (2024).

The authors reported that more frequent attendance predicted higher trust, volunteering, and interviewer-rated cooperation on average. Their results imply that the social practice of attendance has greater explanatory value in the analyzed setting than a broad religious category alone. Their pooled estimates for a difference between never and weekly attendance corresponded approximately to nine percentage points in generalized trust, three percentage points in volunteering, and two percentage points in perceived cooperativeness. These values quantify associations under the study's longitudinal modeling assumptions rather than direct results of randomly assigned worship attendance. The interviewer rating measures visible cooperativeness within an interview; it should not be reinterpreted as an objective census of helping strangers Aksoy & Wiertz (2024).

The analysis also found heterogeneity between religious affiliations, including stronger positive attendance-related outcomes for Anglicans and other Protestants and less consistent estimates for other groups. This makes the British social and historical environment integral to interpretation. An explanation based on recruitment channels and majority-community integration is plausible, but the estimates cannot independently isolate doctrinal content, network embeddedness, and social opportunity. Fixed-effects controls help with stable traits, while changes in health, family formation, social ties, residential location, or motivation can remain time-varying confounders. The strongest inference is a longitudinal relationship consistent with effects of religious involvement under specified assumptions, complemented rather than replaced by experimental tests Aksoy & Wiertz (2024).

Combining the British panels with short-term priming experiments offers an instructive multilevel comparison. Experimental tasks manipulate a momentary representation and record an immediate allocation, whereas surveys track repeated participation and more enduring life activities. Different effect sizes or even different directions would not create a logical contradiction because the treatments and outcomes differ. A more integrated program would randomize opportunities to take part in religion-affiliated or secular volunteer networks, measure participants' motivations and belonging, and independently collect service outcomes after several months. Such an approach could distinguish organizational exposure from specific theological persuasion, provided participation remained voluntary and research ethics protected religious choice Aksoy & Wiertz (2024) Shariff & Norenzayan (2007).

Cooperation Across Societies: A Direct Test of Strong Cultural Predictions

Broad cultural accounts occasionally suggest that religious traditions generate large national differences in cooperative behavior. Spadaro and colleagues' 2022 preregistered meta-analysis supplies an unusually broad benchmark for evaluating such claims. They assembled 1,506 experiments of impersonal cooperation across 70 societies, yielding 2,271 effect estimates, and examined 17 preregistered explanations. Their outcome was costly cooperation in experimental social dilemmas, not charitable self-report or social-service participation. After adjusting for ten features of study design, the authors found very little cross-societal variation in impersonal cooperation, and cultural group labels explained virtually none of the remaining variance Spadaro et al. (2022).

The distinction between a societal effect and a within-society religious association is essential. A global analysis can report weak variation between societies even if religiously affiliated participants differ somewhat from unaffiliated participants within many societies. Conversely, a positive correlation within societies does not prove a religious explanation for economic development or historical state formation. Spadaro and colleagues cautioned against exaggerating large differences in anonymous cooperation between contemporary societies; Valencia Caicedo and colleagues found individual religious affiliation correlated with particular prosocial preferences and that the institutional environment moderated the correlation. Read together, these results define different levels of the same empirical landscape rather than mutually exclusive conclusions Spadaro et al. (2022) Valencia Caicedo et al. (2023).

Design-level adjustment is necessary because social dilemmas vary in the multiplier for shared contributions, numbers of players, information about previous choices, laboratory incentives, framing, participant recruitment, and anonymity. Any of these can produce a sizable change in measured cooperation. Cross-cultural comparisons in which a country difference is smaller than a task-format difference should therefore avoid cultural essentialism. Researchers should preregister whether religion is treated as an individual-level predictor, a community practice, or a country-level institutional feature, and avoid moving opportunistically among levels after inspecting the results Spadaro et al. (2022).

Cross-cultural experiments can also enhance measurement by local qualitative work. Before translating a word such as generosity, trust, sin, or punishment, researchers need to establish what participants mean by the term. The same transfer may express compassion in one setting and obligation to kin, elders, or religious institutions in another. A procedure that obscures local beneficiaries risks confusing substantive differences with linguistic differences. Combination of comparable economic choices, ethnographic interviews, culturally validated belief measures, and transparent contextual records provides a more defensible route to generalizing research on religious cooperation Purzycki et al. (2016) Spadaro et al. (2022).

Religious Signaling: Perceived Trustworthiness and Cooperative Performance

Religious costly-signaling theories have become sufficiently established to require clear separation of the observer's judgment from the actor's performance. A participant might judge someone who adheres to demanding rituals as reliable, cooperative, or committed to a moral community. The same participant might later observe the person's behavior in a situation involving a genuine conflict of interest. These are separate outcomes, and a theoretical claim about signaling should specify whether the primary effect is on beliefs, partner choice, actual cooperation, community longevity, or some combination. Costly performance can create a reputational advantage without proving superiority in every moral domain Henrich (2009) Sosis & Bressler (2003).

The 2024 review Religious Signaling and Prosociality synthesizes this distinction. Its authors report evidence that religious actors are often perceived as more trustworthy, while findings concerning actual trustworthiness and prosocial conduct are less consistent. Routine religious behavior can coincide with ingroup preference, and unusual high-cost collective practices may sometimes expand social identity and tolerance. The evidence places boundaries around theories that move directly from costly ritual to generalized benevolence. Context, practice type, beneficiary, and how cooperation is measured remain decisive for interpretation Northover et al. (2024) Henrich (2009) Preston & Ritter (2013).

A targeted research design can test the gap. First, randomly assign observers information about a hypothetical partner's costly religious ritual, low-cost religious affiliation, equivalent costly secular volunteering, or no identity signal. Next, record expected trustworthiness and willingness to collaborate. Finally, provide genuine behavioral histories that sometimes confirm and sometimes contradict the cues, and measure how observers update their beliefs. A companion design would record actual choices by signaling participants and matched controls. This structure separates prior stereotypes, informational cues, strategic incentives, and the true predictive validity of reputation. It also makes the secular comparison intellectually necessary rather than ceremonial Henrich (2009) Bloom (2012).

The institutional function of an effective signal is to coordinate expectations when reliable behavioral information is scarce. In societies with strong independent reputation systems, the marginal advantage of a religious identity cue may diminish. Conversely, in closed communities with recurrent interaction and enforceable norms, a signal can carry information about obligations to known group members. The hypothesized mechanism therefore varies with social distance, monitoring institutions, and shared moral language. Future comparative investigations should quantify whether signal effects arise through increased belief in the actor's honesty, fear of punishment, expectation of shared norms, or social penalties for declining cooperation Norenzayan & Shariff (2008) Valencia Caicedo et al. (2023).

Integrative Methods and Testable Research Questions

A cumulative research program requires a common vocabulary without collapsing distinctive variables. For religious exposure, measurement should record belief centrality, representation of supernatural punishment, organized attendance, practice frequency, social embeddedness, and recognized identity. For cooperating behavior, independent outcomes should include stranger generosity, public-goods contribution, trust-game return, verified sustained volunteering, distributional fairness, and outgroup-directed assistance. For context, variables should include institutional enforcement, recipient identity, opportunities to help, visibility of choices, economic stakes, and voluntary membership. The analytic purpose of this matrix is to make competing explanations empirically discriminable across research designs Kelly et al. (2024) Purzycki et al. (2016).

Several propositions follow from the reviewed literature. If supernatural monitoring operates through anticipated detection, experimentally increasing perceived anonymity should moderate its influence when secular controls are held constant. If ritual primarily signals commitment, public information about ritual performance should shift partner expectations more strongly than private knowledge of an equivalent unobserved practice. If congregations chiefly create opportunities, the effect of attendance on verified volunteering should attenuate when equivalent opportunities are experimentally available through secular institutions. If broad moral inclusion is the relevant mechanism, experimentally induced changes in generosity should persist toward religiously unfamiliar recipients rather than merely shift resources within the ingroup. Each proposition names a potential falsifying result and therefore improves the clarity of subsequent inquiry Shariff & Norenzayan (2007) Henrich (2009) Aksoy & Wiertz (2024) Pasek et al. (2023).

The strongest evidentiary strategy combines preregistered experimental manipulations with longitudinal observation, transparent behavioral outcomes, cultural interviews, and multilevel statistical models. Primary outcomes and contrasts should be declared before data collection, as should exclusion rules and treatment of missing values. Multi-institution teams can test whether effects reproduce among populations with different beliefs, socioeconomic structures, and religious traditions. Hierarchical models should acknowledge that participants nested within a small number of communities do not provide independent community-level evidence. Identifiable human subjects should retain privacy protections when their religious beliefs and cooperation choices are recorded Spadaro et al. (2022) Beheim et al. (2021).

The resulting framework also clarifies scientific limitation. Psychological monitoring experiments cannot establish the chronological origin of historical state complexity; historical coding cannot replace experiments on current moral judgment; a survey measure of identity cannot stand in for a behavioral measure of generosity. The study of religion and prosociality grows more rigorous when each level contributes its own piece of evidence and when large interpretive claims are traced back to concrete designs. This standard enables evidence-based comparisons across sacred traditions and secular systems while preserving a central empirical question: by what observable processes do people learn, expect, enforce, and sustain cooperation under different institutional conditions? Whitehouse et al. (2023) Kelly et al. (2024) Valencia Caicedo et al. (2023).

Developmental and Socialization Pathways

Religious participation often begins through family, school, neighborhood, peer groups, and community rituals. Children can learn expectations of helping and punishment from both explicit instruction and observation of adults' actions. A developmental analysis must distinguish learning to articulate a rule from choosing to follow it when enforcement is absent. It must also distinguish internalization of concern for others from obedience to authority and willingness to contribute to a familiar group. Because children inhabit distinct institutional and family contexts, comparisons by household affiliation alone are vulnerable to cultural and socioeconomic confounding Graham & Haidt (2010) Henrich (2009).

Social learning accounts predict that credible adult models, repeated institutional practices, and consistent feedback help establish normative expectations. Henrich's theory of credibility-enhancing displays is useful here: conduct perceived as costly for a merely strategic actor may make proclaimed beliefs or norms appear trustworthy. Ritual familiarity may then influence both moral reasoning and the social networks in which assistance occurs. Such pathways can operate without a child's detailed endorsement of theological propositions and can generalize to secular institutions with visibly committed mentors. Observational and longitudinal designs should measure exposure to particular practices rather than infer it from inherited affiliation Henrich (2009).

The developmental evidence base also requires explicit correction of a highly publicized error. The 2015 Current Biology article reporting a negative relationship between household religion and children's sharing across countries was retracted. Because its analysis of country context affected interpretation, the original report should be used only as a cautionary case in research methodology. Future work needs within-community comparisons, cultural measurement alignment, age-specific tasks, measures of observed generosity outside laboratory allocation, and designs that can separate household selection from socialization Decety et al. (2019) Galen (2012).

Adolescent and adult changes in religious identity are also relevant to causal ordering. People can join religious communities when they seek belonging, engage in volunteering before conversion, or withdraw from institutional affiliation while retaining moral commitments and trusted social ties. A person who loses affiliation may lose recruitment opportunities without changing willingness to help. Conversely, new membership may supply a network through which existing cooperative dispositions become visible. Longitudinal attention to these transitions could reveal whether networks, beliefs, reputational concerns, and actual actions change together or on different timescales Graham & Haidt (2010) Galen (2012).

Normative Conflict, Coercion, and Institutional Harm

Moral cooperation is not exhausted by group survival or quantity of coordinated action. Organizations may coordinate effectively while requiring costly conformity, restricting autonomy, or allocating benefits unequally. The psychological processes that support cooperation—shared identity, obedience, reputation, and enforcement—can also support exclusion or punitive treatment. For that reason, empirical research needs separate outcome measures for voluntary aid, fairness, protection of vulnerable members, minority inclusion, and coercive compliance. A community's efficiency at mobilizing members is an institutional achievement, but its ethical quality requires analysis of the ends and means of mobilization Bloom (2012) Graham & Haidt (2010).

Some religious moral norms support hospitality, mutual care, and reconciliation, while others create conflicts between loyalty to insiders and equal treatment of outsiders. Researchers must identify the specific norm participants perceive as binding and test how it interacts with target identity, personal conscience, and social sanction. Experimental scripts describing a generic good deed cannot represent all possible moral obligations within a tradition. Contextual comparisons should preserve moral plurality and investigate how members negotiate conflicting duties in family, economic, and civic settings Preston & Ritter (2013) Pasek et al. (2023).

The distinction between moral motivation and surveillance deserves particular attention in coercive environments. If a person contributes primarily because of anticipated punishment, the behavior may sustain a public good but also produce stress and hidden resentment. If members share resources because they trust one another and endorse the goal, similar behavior can arise from different psychological conditions. A research program that records contributions but overlooks perceived voluntariness, punishment threat, and exit options will miss consequential distinctions. Studies of supernatural monitoring should compare sanction expectations with positive internalization and social care when the design permits Purzycki et al. (2016) Norenzayan & Shariff (2008).

Future Research: Integrating Mechanisms and Institutions

The next generation of research should replace global propositions about whether religion makes people good with precise predictions linking particular practices and cognitions to particular decisions. A useful factorial study would manipulate the salience of a knowing deity, a secular observer, and a supportive community norm, while independently varying anonymity and the recipient's group membership. Multiple measures would separate allocation, beliefs about supervision, perceived duty, and emotion. Recruitment should include religious and nonreligious participants from comparable local settings, with preregistered tests of treatment-by-belief interactions and careful measurement of demand effects Shariff et al. (2016) Preston & Ritter (2013).

Cross-national field studies should employ shared but locally validated procedures and explicitly model congregation, neighborhood, and country levels. Multiwave designs can measure changes in attendance, belief centrality, peer networks, and volunteer activity while tracking time-varying life circumstances. Where legal and ethical, anonymized administrative records could complement interviews and behavioral tasks, increasing coverage of sustained service. Historical studies need independent dating, reproducible coding, and sensitivity analysis of missingness, especially where early religious practices left sparse records. These improvements directly address the limitations visible in present evidence and in the moralizing-gods chronology dispute Whitehouse et al. (2023) Beheim et al. (2021).

The Artificial Era introduces further research possibilities concerning cooperation mediated by artificial agents. Digital systems can recommend volunteer opportunities, allocate collective resources, simulate moral observers, or influence which religious and nonreligious identities are salient. These are prospective extensions, not established effects of Artificial Sapiens or empirical conclusions of the historical literature. Their study should distinguish human perceptions of automated monitoring, trust calibration, algorithmic fairness, and genuine patterns of collective contribution. Comparisons with human oversight can illuminate which parts of reputation and accountability depend on shared belief and which respond to institutionally credible observation. The research direction connects the dossier to Psychology of Human–Artificial Relations while keeping the evidential claims grounded in existing human studies Shariff & Norenzayan (2007) Norenzayan & Shariff (2008).

More productive syntheses will treat cooperation as an outcome vector rather than a trait. A study can report anonymous stranger giving, long-term volunteer work, fairness to outsiders, retaliation toward norm violators, and the voluntary character of contribution separately. Theoretical interpretation should then state which mechanisms are supported for each component and population. This preserves meaningful differences between social cohesion, interpersonal kindness, moral evaluation, and maintenance of public goods. It also creates clearer opportunities for replication and policy-relevant comparison with secular civic institutions Kelly et al. (2024) Valencia Caicedo et al. (2023).

Conclusion

Religion can contribute to prosociality and moral cooperation through several interacting pathways: religious representations may heighten attention to accountability; rituals and costly commitments may communicate reliability; groups organize repeated interaction, norms, and opportunities for assistance; and beliefs about the scope of moral obligation can shape generosity toward insiders or outsiders. No single pathway explains all types of religious action across populations. The fundamental scientific conclusion is therefore differentiated: the effects of religiosity depend on its operational meaning, the outcome being measured, social setting, and the identity of beneficiaries Norenzayan & Shariff (2008) Graham & Haidt (2010) Purzycki et al. (2016).

The most comprehensive behavioral synthesis available among the sources reviewed reports a modest positive average relationship between religiosity and prosociality, with a weaker association for directly observed conduct than for self-reports. Experiments establish that activating some God-related concepts can change monetary allocations in certain conditions, including in cross-cultural designs involving religious outgroups. Historical and field research provides plausible accounts of shared norms and collective commitment. Important limitations remain in generalizing short tasks to sustained civic action, separating selection from influence, identifying psychological mediators, and inferring ancient historical precedence Kelly et al. (2024) Pasek et al. (2023) Sosis & Ruffle (2003).

The literature's controversies are part of its evidentiary architecture. Correct interpretation requires attention to the explicit retractions of prominent developmental and historical papers, the corrected historical studies that followed, and distinctions between observed associations and causal effects. Scientific knowledge advances when the mechanisms of cooperation are specified precisely and alternative explanations remain testable. Religion's psychological significance is best understood through the interaction of belief, practice, community, monitoring, identity, and institutional structure, examined with transparent methods and with equal attention to beneficiaries, outsiders, and the costs of collective coordination Galen (2012) Shariff et al. (2016) Whitehouse et al. (2023).

References

  • Aksoy, O., & Wiertz, D. (2024; published online 2023). The impact of religious involvement on trust, volunteering, and perceived cooperativeness: Evidence from two British panels. European Sociological Review, 40(1), 143–159. https://doi.org/10.1093/esr/jcad024
  • Beheim, B. A., Atkinson, Q. D., Bulbulia, J., Gervais, W., Gray, R. D., Henrich, J., Lang, M., Monroe, M. W., Muthukrishna, M., Norenzayan, A., Purzycki, B. G., Shariff, A. F., Slingerland, E., Spicer, R., & Willard, A. K. (2021). Treatment of missing data determined conclusions regarding moralizing gods. Nature, 595, E29–E34. https://doi.org/10.1038/s41586-021-03655-4
  • Bloom, P. (2012). Religion, morality, evolution. Annual Review of Psychology, 63, 179–199. https://doi.org/10.1146/annurev-psych-120710-100334
  • Decety, J., Cowell, J. M., Lee, K., Mahasneh, R., Malcolm-Smith, S., Selcuk, B., & Zhou, X. (2019). Retraction notice to: The negative association between religiousness and children's altruism across the world. Current Biology, 29(15), 2595. https://doi.org/10.1016/j.cub.2019.07.030
  • Galen, L. W. (2012). Does religious belief promote prosociality? A critical examination. Psychological Bulletin, 138(5), 876–906. https://doi.org/10.1037/a0028251
  • Graham, J., & Haidt, J. (2010). Beyond beliefs: Religions bind individuals into moral communities. Personality and Social Psychology Review, 14(1), 140–150. https://doi.org/10.1177/1088868309353415
  • Henrich, J. (2009). The evolution of costly displays, cooperation and religion: Credibility enhancing displays and their implications for cultural evolution. Evolution and Human Behavior, 30(4), 244–260. https://doi.org/10.1016/j.evolhumbehav.2009.03.005
  • Kelly, J. M., Kramer, S. R., & Shariff, A. F. (2024). Religiosity predicts prosociality, especially when measured by self-report: A meta-analysis of almost 60 years of research. Psychological Bulletin, 150(3), 284–318. https://doi.org/10.1037/bul0000413
  • Norenzayan, A., & Shariff, A. F. (2008). The origin and evolution of religious prosociality. Science, 322(5898), 58–62. https://doi.org/10.1126/science.1158757
  • Northover, S. B., Quillien, T., Conroy-Beam, D., & Cohen, A. B. (2024). Religious signaling and prosociality: A review of the literature. Evolution and Human Behavior, 45(5), Article 106593. https://doi.org/10.1016/j.evolhumbehav.2024.06.002
  • Pasek, M. H., Kelly, J. M., Shackleford, C., White, C. J. M., Vishkin, A., Smith, J. M., Norenzayan, A., Shariff, A. F., & Ginges, J. (2023). Thinking about God encourages prosociality toward religious outgroups: A cross-cultural investigation. Psychological Science, 34(6), 657–669. https://doi.org/10.1177/09567976231158576
  • Pichon, I., Boccato, G., & Saroglou, V. (2007). Nonconscious influences of religion on prosociality: A priming study. European Journal of Social Psychology, 37(5), 1032–1045. https://doi.org/10.1002/ejsp.416
  • Preston, J. L., & Ritter, R. S. (2013). Different effects of religion and God on prosociality with the ingroup and outgroup. Personality and Social Psychology Bulletin, 39(11), 1471–1483. https://doi.org/10.1177/0146167213499937
  • Purzycki, B. G., Apicella, C., Atkinson, Q. D., Cohen, E., McNamara, R. A., Willard, A. K., Xygalatas, D., Norenzayan, A., & Henrich, J. (2016). Moralistic gods, supernatural punishment and the expansion of human sociality. Nature, 530, 327–330. https://doi.org/10.1038/nature16980
  • Shariff, A. F., & Norenzayan, A. (2007). God is watching you: Priming God concepts increases prosocial behavior in an anonymous economic game. Psychological Science, 18(9), 803–809. https://doi.org/10.1111/j.1467-9280.2007.01983.x
  • Shariff, A. F., Willard, A. K., Andersen, T., & Norenzayan, A. (2016). Religious priming: A meta-analysis with a focus on prosociality. Personality and Social Psychology Review, 20(1), 27–48. https://doi.org/10.1177/1088868314568811
  • Sosis, R., & Bressler, E. R. (2003). Cooperation and commune longevity: A test of the costly signaling theory of religion. Cross-Cultural Research, 37(2), 211–239. https://doi.org/10.1177/1069397103037002003
  • Sosis, R., & Ruffle, B. J. (2003). Religious ritual and cooperation: Testing for a relationship on Israeli religious and secular kibbutzim. Current Anthropology, 44(5), 713–722. https://doi.org/10.1086/379260
  • Spadaro, G., Graf, C., Jin, S., Arai, S., Inoue, Y., Lieberman, E., Rinderu, M. I., Yuan, M., van Lissa, C. J., & Balliet, D. (2022). Cross-cultural variation in cooperation: A meta-analysis. Journal of Personality and Social Psychology, 123(5), 1024–1088. https://doi.org/10.1037/pspi0000389
  • Valencia Caicedo, F., Dohmen, T., & Pondorfer, A. (2023). Religion and cooperation across the globe. Journal of Economic Behavior & Organization, 215, 479–489. https://doi.org/10.1016/j.jebo.2023.09.025
  • Whitehouse, H., François, P., Savage, P. E., Hoyer, D., Feeney, K. C., Cioni, E., Purcell, R., Larson, J., Baines, J., ter Haar, B., Covey, A., & Turchin, P. (2021). Retraction note: Complex societies precede moralizing gods throughout world history. Nature, 595, 320. https://doi.org/10.1038/s41586-021-03656-3
  • Whitehouse, H., François, P., Savage, P. E., Hoyer, D., Feeney, K. C., Cioni, E., Purcell, R., Larson, J., Baines, J., ter Haar, B., Covey, A., & Turchin, P. (2023; published online 2022). Testing the Big Gods hypothesis with global historical data: A review and “retake.” Religion, Brain & Behavior, 13(2), 124–166. https://doi.org/10.1080/2153599X.2022.2074085
  • Xygalatas, D., Mitkidis, P., Fischer, R., Reddish, P., Skewes, J., Geertz, A. W., Roepstorff, A., & Bulbulia, J. (2013). Extreme rituals promote prosociality. Psychological Science, 24(8), 1602–1605. https://doi.org/10.1177/0956797612472910