Summary:
When public reputation and direct personal experience conflict, people tend to trust their own interactions more than social gossip when judging refusals to help. A new study from Rissho University describes an asymmetric moral calculus: acts of helping are praised broadly, regardless of background signals, while refusals to assist are evaluated largely through the lens of prior personal encounters. Even refusals that appear justified do not receive positive moral credit from observers.
Key Facts:
- Asymmetric moral evaluation: Cooperative behavior is judged positively across contexts, but judgments of non-cooperation depend heavily on direct prior experience with the requester.
- Personal experience outweighs reputation: When a person’s reputation conflicts with how they treated the observer, observers rely primarily on firsthand experience to evaluate refusals to help.
- Justified refusal is neutral, not praised: Even when refusal is supported by both poor reputation and prior mistreatment, third-party observers rate the refusal as neutral rather than virtuous.
Source: Rissho University
Overview: Human cooperation often relies on indirect reciprocity, where reputational information guides decisions about whether to help others. Yet everyday social life also provides direct personal experiences that can conflict with a person’s public reputation. To understand how people resolve these conflicts, Professor Hitoshi Yamamoto and colleagues at Rissho University ran experiments testing how observers weigh reputation versus personal history when evaluating helping behavior.
Published in Evolution and Human Behavior, the study shows that people use distinct cognitive rules for assessing cooperation and defection. While helping behavior is largely credited regardless of reputation or prior interactions, refusals are judged through a different filter: the observer’s own prior treatment by the requester.
“In daily life we combine hearsay and firsthand experience when judging others,” said Professor Yamamoto. “Our results reveal one way people prioritize those information sources when they disagree.”
Experiment design and everyday settings
The research team ran two scenario-based experiments using realistic workplace and neighborhood contexts—specifically scenes involving restaurant coworkers and local community members. Participants read short vignettes in which an actor either agreed to help or refused a direct request. The researchers experimentally varied two main factors:
- Public reputation: whether the recipient had a positive or negative standing in the wider group;
- Direct personal history: whether that recipient had previously helped or refused to help the participant in an earlier interaction.
Participants then rated the moral standing, social approval, and perceived fairness of the person who answered the request.
Results revealed a robust double standard in moral judgment. When someone offered help, observers rated that behavior positively across the board—the recipient’s prior reputation or past interactions had little effect on praise for cooperative action. But when someone refused to help, the observer’s own personal history with the requester dominated evaluations.
Two clear patterns emerged:
- Harsh treatment for a once-helpful but poorly reputed person: If the requester had a bad public reputation but had previously helped the participant, observers judged a refusal to help that requester much more negatively.
- Lenient judgment for a well-regarded but previously rude person: If the requester had a strong public reputation but had personally slighted the participant, observers were more forgiving of a refusal to help.
Neutral view of justified refusal
A notable finding is that observers do not reward punitive or retaliatory non-cooperation. Even when a refusal to help was clearly supported by both negative reputation and prior mistreatment—conditions that might justify withholding assistance—participants rated the refusal neutrally rather than positively. This suggests social norms favor restraint: withholding cooperation may be acceptable, but it is not celebrated as morally good.
Implications for reputation systems and AI trust metrics
These findings have practical implications for online reputation systems, decentralized networks, and algorithmic trust metrics. Digital environments frequently rely on aggregate ratings, upvotes, or automated scores, but the study indicates that human social trust remains anchored to direct personal engagement. Automated systems that depend exclusively on public reputational data may miss how people actually combine firsthand experience with social signals when making moral and cooperative judgments.
“Future work should explore environments where direct experience is scarce—such as anonymous online marketplaces or AI-mediated interactions—to see how people adapt their reliance on reputation,” Professor Yamamoto noted.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by staff.
About this social neuroscience and psychology research:
- Media contact: YOSHIMITSU Omachi
- Source: Rissho University
- Image credit: Image credited to Neuroscience News
- Original research (open access): Evolution and Human Behavior (August 31, 2026). Title: “Direct experience overrides reputation in evaluations of defection in indirect reciprocity.” Authors: Yamamoto, H., Suzuki, T., & Okada, I.
- DOI: 10.1016/j.evolhumbehav.2026.106980
Abstract
Direct experience overrides reputation in evaluations of defection in indirect reciprocity
Cooperation in human societies is often sustained through indirect reciprocity, where reputational signals influence who receives help. But real-world decisions also include direct personal experience, and when firsthand behavior conflicts with shared reputation, it is unclear which source observers prioritize. This research experimentally examines how third-party observers evaluate donor actions when a recipient’s public reputation conflicts with that recipient’s past behavior toward the observer.
Across two scenario experiments with more than 2,500 participants recruited through online crowdsourcing in Japan, the study manipulated whether a donor cooperated or defected, the recipient’s reputation, and the recipient’s prior behavior toward the observer. Findings reveal a pronounced asymmetry: cooperative acts are consistently rated positively, whereas judgments of defection are highly sensitive to the observer’s direct experience. In many cases, firsthand experience overrides reputational information. Yet even justified defection—backed by both poor reputation and prior mistreatment—fails to be endorsed as morally good, receiving neutral appraisal instead.
These results point to a social reluctance to celebrate punitive withdrawal of help and suggest norms of indirect reciprocity may prioritize restraint over retribution.