Submission 103
Ingroup Favoritism under Reputation Systems
PS2-G09-01
Presented by: Anna Sokolova
A large body of literature shows the existence of ingroup bias — a preference for members of one's own group. This bias can manifest in various ways, such as choosing members of the same group for interactions, trusting them more, and displaying greater cooperative behavior toward them. Conversely, strangers and outgroup members often face a less favorable disposition.
In situations where trustworthiness is difficult to assess (e.g., large marketplaces with frequent interactions between strangers), an exogenously established reputation system can offer a solution to the cooperation problem. Such a system reflects an actor’s past behavior and informs potential partners about it, compensating for the lack of interpersonal interaction history. A high reputation serves as a reliable signal of trustworthiness and fosters positive expectations about cooperative intentions.
Within the context of such a system, a reputation score should be the primary driver of partner selection and trust decisions. However, can ingroup bias disrupt rational motivation and lead to favoritism toward ingroup members, even those with lower reputations? We investigate ingroup favoritism within a reputation system and argue that it varies based on past experiences with one's own group.
This study explores the trade-off between ingroup and reputation-based preferences in decision-making using a two-stage experimental design. Participants interact with bots programmed by the researchers but believe they are playing with real people.
In Stage 1, each subject acts as a Trustor in several rounds of a one-shot Trust Game. The between-subjects manipulation involves three conditions representing varying experiences with ingroup interaction during this stage: negative, neutral, and positive. For example, in the "negative experience" condition, a subject is frequently exposed to breaches of trust from the Trustee. Participants believe they have been grouped with other subjects based on similar preferences, reinforcing their group identity.
In Stage 2, participants place their trust in one of two potential Trustees over a series of rounds. One potential partner is from the ingroup, while the other is from the outgroup. Both partners are assigned reputation scores meant to reflect their behavior in Stage 1. Reputation score combinations vary based on both the level of the scores and the magnitude of the difference between them.
We hypothesize that individuals exposed to positive or neutral ingroup interactions in Stage 1 will exhibit ingroup favoritism (i.e., a higher likelihood of choosing an ingroup member over an outgroup member when the reputation scores are equal or when the ingroup member's score is lower). Conversely, those with negative experiences in Stage 1 will develop ingroup aversion. Additional hypotheses examine how these preferences are influenced by the difference in reputation scores and the overall reputation score level.
This study provides insights into the interplay between social identity and reputation, with implications for understanding cooperation, trust in diverse social contexts, and the emergence of inequality.
In situations where trustworthiness is difficult to assess (e.g., large marketplaces with frequent interactions between strangers), an exogenously established reputation system can offer a solution to the cooperation problem. Such a system reflects an actor’s past behavior and informs potential partners about it, compensating for the lack of interpersonal interaction history. A high reputation serves as a reliable signal of trustworthiness and fosters positive expectations about cooperative intentions.
Within the context of such a system, a reputation score should be the primary driver of partner selection and trust decisions. However, can ingroup bias disrupt rational motivation and lead to favoritism toward ingroup members, even those with lower reputations? We investigate ingroup favoritism within a reputation system and argue that it varies based on past experiences with one's own group.
This study explores the trade-off between ingroup and reputation-based preferences in decision-making using a two-stage experimental design. Participants interact with bots programmed by the researchers but believe they are playing with real people.
In Stage 1, each subject acts as a Trustor in several rounds of a one-shot Trust Game. The between-subjects manipulation involves three conditions representing varying experiences with ingroup interaction during this stage: negative, neutral, and positive. For example, in the "negative experience" condition, a subject is frequently exposed to breaches of trust from the Trustee. Participants believe they have been grouped with other subjects based on similar preferences, reinforcing their group identity.
In Stage 2, participants place their trust in one of two potential Trustees over a series of rounds. One potential partner is from the ingroup, while the other is from the outgroup. Both partners are assigned reputation scores meant to reflect their behavior in Stage 1. Reputation score combinations vary based on both the level of the scores and the magnitude of the difference between them.
We hypothesize that individuals exposed to positive or neutral ingroup interactions in Stage 1 will exhibit ingroup favoritism (i.e., a higher likelihood of choosing an ingroup member over an outgroup member when the reputation scores are equal or when the ingroup member's score is lower). Conversely, those with negative experiences in Stage 1 will develop ingroup aversion. Additional hypotheses examine how these preferences are influenced by the difference in reputation scores and the overall reputation score level.
This study provides insights into the interplay between social identity and reputation, with implications for understanding cooperation, trust in diverse social contexts, and the emergence of inequality.