In-group Favouritism
People favour their own group even when the group was assigned at random moments ago.
In-group favouritism is the tendency to allocate more trust, credit, and resources to members of a group one belongs to. The group does not need history or stakes; an arbitrary label is enough to start it. Favouritism toward the in-group is more reliable than hostility toward the out-group.
How it shows up in software
Products create groups constantly: workspaces, servers, teams, cohort: A group of users bucketed by when they joined, tracked over time rather than blended together., subreddits, fandoms, plan tiers. Membership changes how content is read, since the same post lands differently inside a community than in a public feed. Group boundaries raise trust and contribution inside them and raise hostility across them when the product puts groups in competition.
Using it well
- Give new users a small, specific group to join rather than dropping them into a global feed where they belong to nothing.
- Make membership visible in ways members choose, such as a team name or a role badge they can set.
- Design shared goals that require cross-group cooperation when you need the boundary softened.
- Keep moderation and norms local to the group, since people accept rules from their own group more readily than from a platform.
Where it turns manipulative
- Ranking groups against each other for engagement, through cross-team or cross-server leaderboards, manufactures hostility that the platform then has to moderate.
- Using tribal identity as a growth lever, by recommending content that sharpens the boundary because conflict drives sessions, profits from a harm the product created.
- Framing a pricing tier as membership in a superior class encourages users to treat other customers as lesser, which poisons shared spaces.
Where you have seen it
Discord
Scopes identity, roles, and nicknames to each server, so the same account carries different standing in different groups.
Strava
Organises activity around clubs and segment leaderboards, which makes contribution to a named group a reason to log a workout.
Slack
Bounds membership at the workspace, so channel norms and trust are set inside an organisation rather than by the platform.
What the research says
- Tajfel, Billig, Bundy and Flament, 1971Mixed evidence
Boys assigned to groups on a trivial basis allocated more money to anonymous in-group members, and sometimes chose options that maximised the gap rather than in-group gain.
Small samples of schoolboys, and critics have long argued the allocation matrices signalled to participants what was expected. The core finding has been reproduced many times, but the original design carries real demand-characteristic problems.
- Balliet, Wu and De Dreu, 2014Well evidenced
Meta-analysis of over two hundred studies found consistent in-group favouritism in cooperation, larger when group membership was mutually known than when it was one-sided.
The best available summary. It supports the effect while showing it depends on reputational visibility, not on group labels alone.
- Sherif et al., 1961Contested
The Robbers Cave study reported that competition between assigned boys' groups produced hostility, and shared goals reduced it.
Archival work, including Gina Perry's, shows the researchers intervened to provoke conflict and that an earlier failed run of the study went unreported. Use it as illustration, not as evidence.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.