Outgroup Homogeneity

Outgroup Homogeneity is the tendency to see "them" as interchangeable - same views, same motives, same type - while seeing "us" as varied individuals with nuance. The outgroup looks like one block; the ingroup looks like people.

People tend to encode and recall more differentiating detail for ingroup members than for outgroup members. You remember that your colleagues disagree about tactics, humor, and taste. The rival team, the other party, or "those users" collapses into a single cartoon. Less contact and less motivation to individuate make the flattening easier.

The pattern feeds stereotyping: if they are all alike, one story fits everyone. It also feeds hostile politics. A single loud outgroup voice becomes "what they all think." Ingroup dissent stays visible; outgroup dissent disappears into the blob.

Ingroup bias favors your side. Essentialism freezes group traits as deep natures. Ultimate attribution error skews explanations by group. Outgroup homogeneity is the perceptual step - "they are all the same" - that makes those other bugs cheaper to run.

Groups can share real averages and norms. The bug is erasing variance on the other side while preserving it on yours - then making decisions as if the cartoon were a census.

• "They're all like that."
• "The other party is a hive mind."
• "Users always want the same dumb thing."
• "That school produces one kind of person."
• "Fans of that team are all toxic."
• "Remote workers are all checked out."
• "Every vendor from that country is the same."
• "If one of them said it, they all believe it."

Rival team blob: your department contains mentors, skeptics, and clowns. The neighboring department is "all political." Park and Rothbart's memory asymmetry shows up in who gets individuating stories.

Feed as census: three outgroup influencers rant. You conclude the whole side agrees. Ingroup dissent still feels like healthy debate; outgroup dissent never enters the sample.

Customer monolith: a product team treats "the user" as one persona after hearing two loud tickets. Internal stakeholders still get six nuanced roles.

Campus sorting: students describe their own club as diverse in motives and the opposing club as a single ideology. Contact is high inside, thin across.

Hiring cartoon: candidates from a rival company are assumed to share one culture. Candidates from your alma mater get individual reads.

Outgroup homogeneity makes prejudice easier. One anecdote becomes a group verdict.

Negotiation and politics worsen when each side bargains with a cartoon instead of a coalition of factions. Deals that could peel allies never get tried because "they" are assumed locked.

Products and policies miss real segments because "they" were never differentiated. Design serves a stereotype and then blames the audience for not fitting it.

Conflict escalates: punishing the blob feels fair when you cannot see dissenters inside it. Collective blame gets cheaper than targeted accountability.

Hiring and admissions grow noisier when category membership substitutes for individuating evidence.

Over time, reduced contact and flattened memory reinforce each other - less curiosity, fewer individuating facts, harder stereotypes.

We have more exposure to ingroup variation. Familiarity supplies differentiating detail; outgroups are sampled thinly.

Category labels invite prototype thinking. Once someone is filed as "one of them," subordinate traits get less encoding and less retrieval.

Motivation matters. Individuating outsiders takes effort; in conflict, effort goes to defending the ingroup story instead.

Media and feeds often present outgroups through a few loud exemplars. The sample looks unanimous because the camera never pans to their internal arguments.

Park and Rothbart's 1982 Journal of Personality and Social Psychology paper on perception of out-group homogeneity showed that people remember more differentiating, subordinate attributes for ingroup members than for outgroup members.

The effect appears across many social categories and is stronger when groups are in competition or when contact is limited. It is a perceptual and memory bias, not only a mean remark.

The practical lesson is to treat "they" as a hypothesis about variance - then collect individuating cases before you let one voice stand for a population.

• You can name factions on your side and only a stereotype on theirs.
• One outgroup example updates your whole model of the group.
• You say "they all" about people you barely know.
• Ingroup disagreement feels normal; outgroup disagreement feels fake.
• Your stories about them are interchangeable.
• Contact is low, confidence about their sameness is high.

Individuate on purpose. Sample variance before you summarize.

• Name three outgroup members who disagree with each other.
• Replace "they" with a specific person in arguments.
• Seek outgroup sources that conflict with each other.
• In research, segment users; do not worship one persona.
• Increase contact that reveals subordinate detail, not only category labels.
• When angry, assume you are flattening - then look for a counterexample.

Against Outgroup Homogeneity, keep group patterns if the data support them - then refuse to erase the people inside the label.

All like that

Hive mind

One voice

Practice this pattern in the Reframing App - capture the trigger, label it (like Outgroup Homogeneity), check evidence, and write a more balanced thought.