Instant favoritism from minimal group assignment
Psychological experiments demonstrate that even arbitrary group assignments produce favoritism within minutes. Divide people randomly into Group A and Group B, announce the split, and ask them to distribute resources. Both groups consistently allocate more to their own group, rate their group as more competent, and remember their group members more favorably. No history, no interaction, no shared identity precedes this. Just knowing you are labeled the same is enough.
This is not about explicit prejudice. Subjects report no conscious bias and do not believe they are being unfair. Yet their behavior shifts systematically. The effect emerges from several mechanisms: people process information about their group more deeply, attribute group members' failures to circumstance and successes to ability, and selectively remember positive examples from their group.
Bias at scale in organizations and society
At scale, minimal group effects become the foundation of tribalism. A company divided into departments, a city divided into neighborhoods, a nation divided into regions or parties, all activate the same basic mechanism. Given a group label and minimal contact with the out-group, people naturally favor the in-group and view the out-group with suspicion or disdain.
Reducing in-group bias requires conscious design. Mixing groups for collaborative tasks, highlighting shared superordinate goals, and regularly exposing people to individual out-group members (not as representatives but as people) can reduce but not eliminate the effect. The bias is not malice; it is automatic. The only reliable defense is awareness plus structural arrangements that force integration.