Learn it → Apply it → Save as Skill
Part of who we are comes from the groups we belong to. The moment people sort themselves into "us" and "them" — a team, a company, a department, a nationality, a fandom — they start favoring their own group and judging the other one more harshly, often with little or no real difference between the groups to justify it.
The Dr. Wallace's Favorite Mental Models is designed to turn useful ways of thinking into reusable tools. Each model gives you a fast visual overview, deeper interactive learning, practical application exercises and an AI-ready Markdown skill you can save for future use.
Instead of merely reading about mental models, you can gradually build your own AI skills library: a collection of portable reasoning skills that can help ChatGPT, Claude, Grok and other AI assistants apply the mental models you value.
People do not just evaluate others as individuals. They first notice which group someone belongs to, then borrow part of their own self-esteem from how well that group compares to others.
Important: the famous 1971 experiments assigned people to groups using a literal coin flip. The favoritism still appeared. The group did not need to be real or meaningful — it only needed to exist.
The model is most useful whenever a conflict, a hiring decision, a merger, or a piece of marketing involves more than one group identity.
Henri Tajfel and John Turner described three ordinary mental steps that, together, quietly produce group bias.
People sort themselves and everyone around them into groups — team, department, company, nationality, alma mater.
They adopt the norms, values and status of the group they belong to as part of their own self-concept.
They compare their in-group to relevant out-groups, usually on dimensions where their own group looks better.
A favorable comparison boosts self-esteem, which motivates favoring the in-group again next time — the loop repeats.
Click below to be randomly sorted into a meaningless group, then watch the default point allocation.
Open a use case to see what the model changes.
After an acquisition, employees often keep sorting colleagues by their old company rather than by role or skill. Decisions, trust and blame tend to follow the old lines. The fix usually starts with a shared superordinate identity (a combined team name, mixed project pods) rather than more memos about "culture fit."
Engineering distrusts sales input, support distrusts product roadmaps. Check whether objections track the actual argument or just which department raised it. If it is the latter, the obstacle is identity, not evidence.
Strong brand communities (a sneaker label, a sports club, an open-source project) work partly because belonging to them raises self-esteem. This explains loyalty that outruns product quality, and also why attacking a competitor's brand can backfire by hardening the rival group's identity.
Shared school, shared hobby, or shared background can quietly read as competence. Structured, criteria-first scoring reduces the chance that in-group resemblance gets mistaken for merit.
Once a negotiation becomes "our team versus their team," concessions start to feel like status losses rather than trade-offs. Reframing both sides as members of one larger group (one supply chain, one industry) can lower the temperature.
Political parties, sports rivalries and online fandoms are classic, intense versions of the same mechanism. The strength of the reaction is a clue: if it is bigger than the stakes, group identity is probably doing some of the work.
This is an illustrative example designed to teach the method.
Expand for the research this model actually rests on.
In a 1971 paper, "Social categorization and intergroup behaviour" (European Journal of Social Psychology), Henri Tajfel, Michael Billig, R.P. Bundy and Claude Flament sorted schoolboy participants into groups on a trivial, almost arbitrary basis (a stated preference between two abstract painters), with no face-to-face contact and no personal stake in the outcome. Even so, participants reliably allocated more reward points to anonymous members of their own group than to the other group. This became known as the "minimal group paradigm." Tajfel, later working with John Turner, formalized Social Identity Theory through the late 1970s and 1980s to explain why.
Social categorization (sorting self and others into groups), social identification (taking on that group's norms and status as part of one's own identity), and social comparison (comparing the in-group favorably against relevant out-groups to protect or raise self-esteem). The theory's central claim is that favorable in-group comparison is partly a self-esteem mechanism, not only a judgment about the groups themselves.
No. Groupthink describes how pressure to preserve harmony inside one group can degrade that group's own decisions. Social Identity Theory describes how simply belonging to a group changes how members see outsiders. The two frequently appear together but explain different failures.
The minimal group studies show bias appears even with a coin-flip group assignment. Do not wait for "real" differences before taking the dynamic seriously.
In-group favoritism is a near-universal cognitive default, not a sign that specific people are unusually biased or unusually virtuous.
Bias intensifies when a group's status or legitimacy feels threatened. A merger, a reorg, or a budget cut often makes group lines harder, not softer.
Banning team names or forcing "one culture" overnight tends to backfire. A shared superordinate identity usually works better than deleting the old ones.
Not every dispute is identity-driven. Check whether the reaction tracks the argument or tracks which group raised it before reaching for this model.
Group lines can re-harden after a new threat, a leadership change, or a new comparison target. Treat the intervention as ongoing maintenance, not a one-time cure.
the conflict is really about scarce budget, time, or credit rather than group status — identity framing can obscure a straightforward allocation problem.
one specific person's performance or conduct is the issue. Relabeling an individual problem as a "group" problem can let a real performance issue hide.
the groups genuinely differ in training, tools, or incentives. Then the gap may be real and solvable with resources, not just reframing.
This browser tool structures your thinking. It does not send your information anywhere or perform AI analysis.
Save as Skill converts the model into portable Markdown instructions that teach an AI when and how to apply Social Identity Theory, what questions to ask, what evidence to seek, what mistakes to avoid, and how to structure its analysis.
Save the skills you find useful and organize them into your own reusable library. The Markdown is intentionally model-neutral, so it can be used wherever an AI system accepts uploaded files, project knowledge, custom instructions or reusable text context. Platform capabilities differ, so this is a portable skill file rather than a claim of one-click installation.
No. Tajfel and Turner's own framework implies the opposite: because the bias runs through self-esteem and comparison, changing what people compare (a shared superordinate goal, a new shared identity, cooperative contact) changes the outcome. The theory explains the mechanism so it can be interrupted, not to excuse it.
Classic prejudice research often focused on long-standing, real-world group histories (race, religion, nationality). The minimal group paradigm's contribution was showing the same favoritism pattern from a group with no history, no face-to-face contact and no real stakes — suggesting the mechanism is a general feature of social cognition, not only a product of specific histories of conflict.
No. Team identity and pride can be a genuine source of motivation and cooperation within the group. The risk this model flags is specifically what happens at the boundary with an out-group, not the existence of group identity itself.