Free riding occurs when people benefit from a resource or collective effort without contributing to its upkeep, thereby undermining cooperation and leading to suboptimal group outcomes (Olson, 1965). It is most studied within traditional rational-choice approaches which emphasize strategic deterrents such as punishment and reputational sanctions as key mechanisms for curbing opportunism (Fehr & Gächter, 2000). However, research in moral psychology suggests that we should consider not only external incentives but also deeply rooted moral intuitions that determine whether individuals perceive opportunism as a minor infraction or a serious violation.

Moral Foundations Theory (MFT, Graham et al., 2013) proposes that moral judgments are guided by distinct foundational domains some of which may be especially relevant in case of shaping attitudes towards free riding. For one, Fairness/Cheating emphasizes proportionality and the expectation that all individuals contribute equally. From this standpoint, free riding represents a clear violation of reciprocity norms. When individuals fail to reciprocate, they disrupt the implicit social contract of mutual exchange, which almost universally leads to moral disapproval (Vauclair & Fischer, 2011).

However, this default disapproval can be further strengthened or weakened by making other foundations salient through moral suasion. Loyalty/Betrayal highlights the importance of group solidarity, reinforcing moral expectations that ingroup members contribute to shared goals. Individuals who strongly identify with a social group may therefore perceive free riding not only as a fairness violation but also as a betrayal of collective duty. Similarly, Authority/Subversion contributes to moral opposition to free riding by stressing the importance of rule maintenance, institutional legitimacy, and social hierarchy. Rule following is viewed as a moral obligation grounded in respect for legitimate authority and tradition.

Yet not all moral intuitions inherently oppose free-riding and self-serving behavior (Fiske & Rai, 2014). For instance, Liberty/Oppression, introduced in subsequent research (Iyer et al., 2012), captures concerns about coercion, personal freedom, and resistance to institutional constraint. Prioritizing autonomy may lead to a view of free riding as a rational assertion of independence, not a moral failure. Although MFT also includes the Care/Harm and Sanctity/Degradation foundations, these are less directly related to free riding. They focus instead on concrete harm to specific individuals, or on issues of purity and contamination, respectively.

To explain how salience of these distinct moral foundations influence attitudes toward free riding, we propose collective orientation – a sense of moral responsibility toward the group welfare – as a key psychological mechanism. This construct builds on research in social psychology emphasizing internal commitments to cooperation (Bicchieri, 2006; Tyler, 2006; Van Lange et al., 2013) and reflects the degree to which individuals prioritize group welfare over personal advantage.

The logic for this mediating pathway is clarified by the traditional MFT classification of foundations. MFT divides moral domain into two categories: binding foundations, which emphasize social cohesion and group harmony, and individualizing foundations, which prioritize individual well-being and personal rights (Haidt, 2012). Loyalty/Betrayal and Authority/Subversion belong to the binding type. Liberty/Oppression, in turn, while not officially a part of individualizing foundations, is seen as being in tension with the Authority/Subversion foundation and placing individual and their freedom first (Haidt, 2012; Iyer et al., 2012).

This distinction allows us to suggest how salience of these foundations should affect attitudes towards free riding through collective orientation. The Authority/Subversion foundation should heighten collective orientation by emphasizing duty to uphold social stability. The Loyalty/Betrayal foundation should similarly increase collective orientation by stressing solidarity and shared goals. In contrast, Liberty/Oppression foundation may lessen collective orientation by legitimizing self-interest and downplaying obligations to others.

In sum, we posit that collective orientation is the mediating pathway through which moral framings influence opposition to free riding. These propositions are consistent with research showing that stronger internalized group obligations increase the likelihood of condemning norm violations (Bicchieri, 2006; Cialdini & Goldstein, 2004).

Institutional trust and political ideology both shape how strongly moral narratives that emphasize particular foundations influence attitudes toward free riding. Institutional trust reflects confidence in the fairness, efficiency, and legitimacy of public institutions (Newton, 2001; Rothstein & Uslaner, 2005). Thus, individuals with high trust in institutions tend to support the norms upheld by these institutions, which implies that additional moral cues may have only marginal effects due to ceiling constraints. In contrast, low trust individuals, who are skeptical of institutional motives, may be more responsive to moral persuasion when the content aligns with their concerns, such as in the case of personal autonomy framing emphasizing resistance from oppression. In turn, binding appeals can substitute for missing institutional trust by appealing to group solidarity and social stability.

Political ideology further shapes responses to moral cues (Hibbing & Theiss-Morse, 2001; Jost et al., 2003). Left-leaning individuals typically endorse equality and redistribution, predisposing them toward self-sacrifice for collective benefit. Because their baseline opposition to self-serving behavior is already high, the influence of moral cues is likely to be limited and primarily reinforcing (Carmines & D’Amico, 2015). Right-leaning individuals, while generally opposing opportunism that undermines reciprocity and fairness (Weeden & Kurzban, 2014), may have more ambivalent reactions. On the one hand, they tend to give greater weight to authority and loyalty intuitions (Haidt & Graham, 2007), which makes them particularly receptive to messages emphasizing rule maintenance, deterrence, and group commitment. At the same time, liberty concerns are also more salient on the right (Iyer et al., 2012), meaning that autonomy-based appeals can resonate with this public when institutional control is viewed as excessive.

Taken together, levels of institutional trust and ideological concerns about government overreach condition how moral narratives are interpreted. As a result, the effects of moral suasion aimed at making specific foundation salient are unlikely to be uniform; instead, depend on the predispositions individuals bring to the interpretive task.

This study employs a between-participants experimental design embedded within a broader survey conducted by the Centre d’Estudis d’Opinió (Center for Public Opinion Studies), the official polling institute of the Catalan Government in Spain. The survey, designed for methodological research, included various experimental components, with the experiment analyzed here as the first in the sequence. It uses a moral framing approach which investigates how presenting information through the lens of moral values influences people’s attitudes, judgments, and behaviors (Brugman, 2024; Feinberg & Willer, 2019; Kodapanakkal et al., 2022).

The total sample consisted of 2,023 internet users aged 16 and older residing in Catalonia, recruited from the Netquest’s internet access panel and stratified by sex, age, and education level. Data collection was conducted via self-administered online interviews, ensuring consistency in the delivery of experimental stimuli¹1

Comprehensive details about the survey, including access to the dataset can be found at: https://osf.io/7jez8/ and also in the Supplementary Information link: https://osf.io/t36rf/files/tgnx8

After ascertaining the scale’s coherence, we examined how each moral framing shaped collective orientation by running a regression analysis with control group as a reference. As Figure 1 shows, consistent with expectations, binding framings (social order and pro-community altruism) were associated with higher levels of collective orientation, although the magnitude varied. The social order treatment produced the most robust effect (b = 1.066, p < .001)³3

Standardized effect sizes for all treatments relative to the control are reported in the online SI, Table A1.

Surprisingly, the personal autonomy framing did not lessen collective orientation. Its effect was nonsignificant (b = 0.107, p = .451) and, if anything, trended in the opposite direction. This suggests that the exposure to an autonomy-based appeal may have triggered a counterreaction, thus reinforcing collective orientation rather than weakening it.

We next examined whether these treatments changed participants’ justifications for free riding. As shown in Figure 2, the social order framing and the pro-community altruism framing led to statistically significant decreases in the justification of free riding (b = -0.287, p = .025 and b = -0.273; p = .032, respectively), suggesting that exposure to binding moral narratives strengthens moral opposition to opportunistic behavior⁴4

The same pattern emerges when using the latent factor scores from the CFA; predicted values are reported in the SI, Figure A3.

However, again, the personal autonomy framing does not produce the expected increase in justification of free riding. Instead of making participants more accepting of opportunism, the effect appears directionally aligned with the binding treatments (b = -0.248, p = .052).

These results support H1 and offer tentative backing for H2, with the pro-community altruism framing producing a weaker effect. Overall, the binding moral framings tended to reduce justification for free riding, though the strength of the effect varied across treatments. However, H3 must be rejected, as the personal autonomy framing failed to increase justification for free riding and may even trigger a counterreaction.

Additional analyses (Figure A4, SI) show that when collective orientation was included as a covariate, much of the direct effect of binding treatments on free-riding justification diminished⁵5

Equivalent attenuation is observed when using the latent factor score (Figure A5, SI).

We next examined whether moral framing exerts a direct impact on free-riding attitudes or operates indirectly via collective orientation, understood as the internalized prioritization of group welfare. To estimate these mediation models, we created three dichotomous treatment variables (each comparing a specific framing to the control) and applied structural equation modeling (SEM) using Stata’s ‘medsem’ command (Mehmetoglu, 2018). Each model was estimated separately with bootstrapped standard errors based on 500 Monte Carlo replications to compute indirect effects. This approach clarifies how binding and autonomy-based framings differentially foster or diminish collective orientation, which then predicts stricter or more lenient attitudes towards free riding. The findings reinforce that internalized cooperative dispositions, not momentary responses to external prompts, primarily drive judgments about free riding.

Figure 3 shows that exposure to the social order framing (X₁) lowered free-riding justification (Y) via collective orientation (M). Specifically, the indirect path through collective orientation was statistically significant (b = 0.094, p < .001), whereas the direct path was not (b = 0.167, p = .115). According to Zhao et al.’s criteria (2010) the pattern is consistent with indirect-only (full) mediation. This framework identifies the indirect effect as the dominant pathway, though the absence of a significant direct effect should not be interpreted as evidence that the direct path is zero. We, therefore, interpret this with caution given the correlational nature of the design.

Around 35.9% of the total effect is mediated by collective orientation, with an indirect effect about 0.56 times the size of the direct effect. These results support the view that reinforcing the priority of group welfare – via shifts in collective orientation – may plausibly explain how moral framing reduces justification of free riding. While this analysis cannot establish causality definitively, the observed pattern is consistent with our theoretical model and provides support for this part of H4.

Figure A6 in the SI examines whether emphasizing altruistic volunteers pursuing a shared societal goal (X₂) reduces free-riding justification (Y) through collective orientation (M). The estimated indirect effect is small and not statistically significant (b = 0.017, p = .103 by Sobel test; p = .110 by Monte Carlo), whereas the direct effect is marginally nonsignificant (b = 0.191, p = .051). According to Zhao et al.’s (2010) framework, this pattern provides no evidence of mediation.

The underlying path coefficients are directionally consistent with the theoretical model: the pro-community altruism framing increased collective orientation (b = 0.276, p = .044), and collective orientation was associated with lower justification of free riding (b = -0.063, p = .005). However, because the indirect effect was not statistically significant, there is no evidence that collective orientation mediates this relationship.

By contrast, the personal autonomy treatment exhibited no statistically significant direct or indirect effects (see Figure A7, in the SI), thus providing no evidence of an effect collective orientation.

To assess whether institutional trust or political ideology modifies the indirect effect of social order framing on free-riding attitudes via collective orientation, we conducted a first-stage moderated mediation analysis (Crowson, 2021). Figure 6 presents a conceptual representation of this analysis for institutional trust. As shown in Table A2 in the SI, contrary to H6, there is no evidence that institutional trust conditions the indirect pathway through collective orientation. The confidence intervals for the moderated mediation index include zero, indicating that trust does not significantly condition the strength of the indirect effect. Thus, although the direct moderation results are broadly consistent with theoretical expectation about variation by institutional trust, the data do not support a moderated mediation process in which trust conditions the indirect pathway through collective orientation.

An equivalent moderated mediation test for left-right ideology also failed to reveal significant moderation of the mediator of the mediating pathway (IMM = -0.016, CI = [-0.038, 0.001]; see Table A3, SI). Although right- and left-leaning individuals differed in their direct responses to the social order frames, both groups displayed similar mediation patterns through collective orientation. The indirect effects were significant across ideological levels but did not vary significantly in strength, suggesting that while right-leaning individuals may be somewhat more responsive to social-order arguments. This difference did not alter the underlying mediating process. These findings indicate that political ideology, like institutional trust, does not substantially modify whether moral framings operate through internalized priority for group welfare.

These outcomes carry significant implications for public messaging and institutional strategies aimed at promoting civic cooperation and discouraging free riding. Social order appeals, such as stressing the societal benefits of following tax laws or traffic regulations, appear more potent than purely altruistic frames. Public agencies, including tax authorities, traffic regulators, and health services, could apply this logic by framing rule-following as a moral contribution to societal stability (e.g., “When everyone pays their share, society runs smoothly”).

Further, our findings suggest that to be efficient, such messages should be tailored to distinct ideological groups and groups with different levels of institutional trust. Left-leaning audiences, though generally favoring redistribution, may still find fairness-based arguments compelling, while right-leaning groups respondents appear more receptive to messages emphasizing stability and institutional order. Likewise, trust in institutions cannot be assumed to magnify moral persuasion; among low trust individuals, moral justifications may partially offset skepticism about institutional authority.

Taken together, these findings extend moral-framing research in several directions. By applying Moral Foundations Theory to the problem of free riding, the study demonstrates that moral narratives influence attitudes toward everyday collective obligations, not only abstract ethical issues such as environmental protection. The results also highlight the importance of collective orientation as an attitudinal mechanism through which moral messages shape moral judgment, clarifying how moral cues become internalized rather than merely shifting opinions. At the same time, the surprisingly similar patterns observed across binding and liberty-based appeals reveal the scope conditions of moral persuasion, showing that moral narratives may produce the opposite effect if misaligned with cultural norms. Finally, the observed differences across institutional trust and political ideology indicate that moral framing operates within a broader social context, where values and beliefs shape responsiveness to moral appeals. Together, these insights offer a more comprehensive understanding of when and why moral communication fosters collective responsibility.

Despite these contributions, certain caveats warrant attention. First, self-reported attitudes may not translate seamlessly into real-world behavior. Future research should include behavioral measures (e.g., actual tax compliance or participatory tasks) to assess if moral framing extends to concrete actions. Second, the study did not include an embedded manipulation check, a limitation that future work could address through pilot testing or post-exposure validation. In our design, however, the mediation through collective orientation provides an indirect indication of construct validity: the framings first influenced collective orientation and only then altered judgments of free riding, suggesting activation of the intended moral intuitions rather than surface-level alignment. This sequential pattern helps mitigate concerns about conceptual overlap between the social-order framing and the outcome measure, as the indirect effect through collective orientation indicates shifts in perceived group-related duty rather than mere wording correspondence. Nevertheless, given the correlational nature of the mediator, these indirect pathways should be interpreted as suggestive rather than definitive causal mechanisms.

A further limitation concerns the restricted set of moral foundations included in the experiment. Our design intentionally focused on Authority/Subversion, Loyalty/Betrayal, and Liberty/Oppression because these domains most directly speak to the tension between individual and collective interests inherent in free-riding dilemmas. However, we did not examine the effects of Fairness/Cheating, Care/Harm and Sanctity/Degradation foundations. In the first case, we reasoned that the mere content of foundation predetermines its salience when making moral judgments about free riding behavior, while we were more interested in how evoking other foundations may shift the general disapproval caused by opportunistic behavior. In contrast, the latter two foundations, emphasizing the immorality of causing direct harm to another person and impurity, respectively, in our view, may be linked to such acts as tax evasion or claiming governmental benefits only indirectly. However, we recognize that including these additional foundations could reveal different or stronger patterns of moral persuasion, particularly among left-leaning individuals. Future studies may explore Equity vs Proportionality domains which Fairness was lately divided into (Atari et al., 2023) or test whether moral framing is effective if it relies on more indirectly linked representations (e.g., recovery of trauma survivors or cleanliness of public spaces obtained through collective efforts).

An additional limitation concerns the use of a single-item measure to capture collective orientation⁶6

This limitation also applies to the moderator variables, which are measured using single-item indicators and may therefore be more susceptible to measurement error.

The study was conducted in Spain, a context characterized by relatively high institutional centralization, moderate civic trust, and comparatively strong relational collectivist orientations in everyday social life (Hofstede et al., 2010; Triandis, 1995) – features that may shape how moral narratives about authority, duty, and group solidarity are interpreted. The cultural specificity of moral foundations could therefore influence both the strength and direction of moral-framing effects. Replications in more diverse sociopolitical environments, particularly those differing in institutional trust, political polarization, or prevailing forms of collectivism, would help clarify the generalizability and cultural situatedness of these findings⁷7

Given the cultural and contextual specificity of the experimental materials, alternative interpretations of the moral framings cannot be ruled out. The pro-community altruism vignette, centered on the 1992 Barcelona Olympics, may also have evoked national pride or collective identity, while the personal autonomy vignette might have triggered associations with social class or elite privilege. Although these possibilities could partly shape responses, they are consistent with the broader view that moral language in public discourse activates multiple, culturally embedded meanings. The sequential pattern observed – moral framing influencing collective orientation, which in turn predicts attitudes toward free riding – suggests that, despite these possibilities, the manipulations engaged the intended moral intuitions rather than simply reflecting contextual associations.

Additionally, while collective orientation emerged as a central mechanism, other psychological processes, such as moral emotions or social identity, could interact with moral narratives. Lastly, longitudinal designs would clarify whether repeated exposure to social order or pro-community altruism framings produces durable transformations in cooperative behavior. Addressing these possibilities would further refine our understanding of how moral persuasion and institutional contexts intersect to shape public disapproval of free riding.

Our findings demonstrate the potential of moral framing to shape attitudes toward free riding, with social order narratives producing the most consistent and substantively larger effects. Importantly, collective orientation functions as a central mediating pathway linking moral framing to attitudes toward free riding in the case of social order framing. Still, individual differences – particularly institutional trust and political ideology – structure patterns of response to this framing.

These findings advance research on moral framing, when and how individuals can be prompted to interpret cooperation through distinct moral lenses. Acknowledging these variations can inform the design of interventions aimed at fostering civic cooperation and discouraging free riding, particularly in contexts where trust and civic engagement are strained.