In-Group Bias

Examples

The unsettling discovery from decades of “minimal group” experiments: people sorted into groups by a coin flip — told they’re “group A” — immediately allocate more resources to fellow As, rate As as nicer, and trust As more. No history, no shared interests, no rivalry. A label alone is enough.

Outside the lab, the labels are richer and the effect stronger:

Marek: “Someone from our forum wrote a critique of the study. Sharp stuff.” Sofia: “A critique of the same quality from that other forum — would you have called it sharp or ‘motivated nitpicking’?”

The same argument arrives with a team jersey on, and the jersey gets judged first.

Why it happens

Groups run on trust, and expecting cooperation from “one of us” is usually a decent bet — in-group members really are more likely to reciprocate. The bias is this useful shortcut running beyond its evidence: the positive tilt applies to people you’ve never met, on questions (like who’s right) that group membership can’t answer. Notably, the research finds the effect is mostly in-group love rather than out-group hate — you don’t have to despise outsiders to consistently shortchange them.

In arguments, it pairs dangerously with the halo effect: “one of us” reads as more honest and more competent, so their claims skip the scrutiny queue. Dismissing an argument because of the arguer’s group is the genetic fallacy; the bias is what makes that fallacy feel reasonable.

How to counter it

• Blind the jersey when you can: evaluate the writing sample, the code, the claim before learning which side produced it.
• Apply the swap test: “If someone from the other group said exactly this, how would I receive it?”
• Argue against your side’s weakest claims occasionally — it recalibrates your sense of where the errors live.
• Widen the label: people cooperate more when reminded of a group that contains both sides (“we’re all trying to get this right”).