Elsevier

Economics Letters

Volume 219, October 2022, 110794
Economics Letters

In-group favoritism in natural and minimal groups

https://doi.org/10.1016/j.econlet.2022.110794Get rights and content

Highlights

  • We compare within-subject in-group favoritism in natural and minimal groups.

  • There are three conditions: Weak and Strong Real Groups, and the Minimal Group.

  • Real groups are based on random assignment to university residential colleges.

  • Weak and Strong decisions are: prior to matriculation; and several weeks later.

  • The Weak and Strong Group favoritism bracket the Minimal Group result.

Abstract

We examine giving to an in-group member relative to an out-group member in the third party (other-other) dictator game. Individuals are randomly assigned to real groups under a “weak” and a “strong” condition, and also assigned to artificial groups using the minimal group paradigm. We compare the effect of the type of group on subject’s allocation of an amount of money between an in-group member and a person who is not a member of the in-group, using a within-subject design with repeated measures. We find that the Weak and Strong real group conditions bracket the Minimal group condition.

Introduction

We examine how different group identities affect favoritism toward in-group members. Numerous studies document the positive effect of in-group identity on economic interactions (see Charness and Chen, 2020). However, little is known about how different group types affect the strength of in-group favoritism. Two meta-analyses directly addressing the question use different sets of games, look at different characteristics of the subjects, and find mixed results. Balliet et al. (2014) find no significant difference for in-group favoritism across natural and artificial identities. By contrast, Lane (2016) suggests that in-group favoritism may be stronger or weaker for natural groups, depending on the context. For ethnic or gender-based natural groups, favoritism is weaker than for minimal groups, while for social and geographical groupings where social interaction is high, favoritism is stronger than for minimal groups. However, group composition and populations change across treatments in these studies, and therefore neither of the meta-analyses is able to say, for a given set of individuals, which type of group identity most strongly affects favoritism. This is our focus.

We exploit a unique panel that enables us to study within-subject differences in in-group favoritism. We recruit a sample of freshmen at Rice University who participate in a series of third-party dictator game (or other-other) allocation decisions, where the dictator divides a fixed amount of money between two other people, one of whom is a member of their in-group. This game allows us to measure in-group bias free from the dictator’s self-regarding motives. In his meta-analysis, Lane (2016) finds that, among all the games he considers, the third-party dictator game is the most likely to produce in-group favoritism.

Subjects complete three such tasks at different times and under different group conditions. The first is the Weak Real Group condition. All incoming freshmen at Rice University are randomly assigned to one of eleven residential colleges: this assignment serves as a naturally-occurring group identity. Before arriving on campus, and before any social interactions, all subjects completed the allocation task, dividing $20 between someone from their residential college and someone from a different college. Twelve weeks after arrival, and once subjects were immersed in their college environment, they faced the same decision task, now the Strong Real Group condition. These subjects also completed the same task for the Minimal Group condition, where they divided $20 between someone from their minimal group and someone not from their group (Tajfel and Turner, 1979).

We find that subjects allocate the most to their in-group in the Strong Real Group condition, while they allocate the least amount in the Weak Real Group condition. The Minimal Group Condition falls between the two real group conditions. Thus our result supports and strengthens the conclusion of Lane (2016), that real groups can lead to stronger or weaker in-group favoritism than minimal groups, depending on the specific context. We thus provide evidence that real and minimal groups can affect behavior differently, even for the same set of subjects.

Section snippets

Experimental design and procedure

A random sample of two-thirds of Rice’s 2016 entering freshman class (661 of 992) was recruited prior to arriving on campus (July 14–August 12, 2016): 553 completed Phase 1 of the study online. Three months after starting classes, those who completed Phase 1 were invited to participate in Phase 2 of the study; 521 subjects participated either in the lab or online, in November 2016. Each phase of the study had multiple tasks (see the Supporting Information — SI): We focus on three third-party

Experimental results

Fig. 1 illustrates that the average amount of in-group giving is the greatest in the Strong Real Group condition ($13.11) and the least in the Weak Real Group condition ($11.70). The difference is statistically significant (p<0.01). The only difference between the two real group conditions has to do with the subjects’ experience with their college. The higher amount of in-group giving in the Strong Real Group condition reflects a strengthened college identity.

We also find that the Weak and

Conclusion

We find that as subjects form a greater attachment to their naturally-occurring groups, they show greater in-group favoritism. Rice University’s residential college system fosters strong group identity, and subjects favor fellow college residents even though the initial assignment to colleges is explicitly random. While prior research on in-group favoritism has been divided in its assessment of minimal as compared to naturally occurring groups, our results are consistent with the findings of 

References (4)

There are more references available in the full text version of this article.

We are grateful for comments from Erin Krupka, Tanya Rosenblat, Oluwagbemiga Ojumu, and Hanna Hoover, as well as the Texas A&M Economic Research Lab Meeting attendees, and participants in the Economic Science Association International Meeting. The lab manager at the Behavioral Research Lab at Rice University was Michelle Harris, who managed the data collection for this project. Research assistants at the BRL were Yu Jin Chang, Sierra Jenkins, Yun Min Oh, and Rakesh Vijayakumar. This study was funded by the National Science Foundation, USA (for Wilson: SES-1534403 and SES-2027556; for Eckel: SES- 2027548 and SES-1534411), and has received IRB approval at Texas A&M, USA and Rice University, USA . Protocols, data and Supporting Information is available at: DOI 10.17605/OSF.IO/DPYBW. Pre-registered hypotheses are available at: DOI 10.17605/OSF.IO/WA6H6.

View full text