Evidence locker · triangulate the cluster
Out-Group Homogeneity Bias
The tendency to perceive members of groups one does not belong to ("them") as more similar to one another than members of one's own group ("us").
Quattrone & Jones (1980) — rival universities
— · n?
Park & Rothbart (1982) — sororities
— · n≈90
Boldry, Gaertner & Quinn (2007) — meta-analysis of measures
— · n≈12078
Mullen & Hu (1989) — meta-analysis on group size
— · meta
★ Out-group homogeneity bias in competi…
field incident · 2020
★ Political out-group homogeneity and a…
field incident · 2025
★ Developmental out-group homogeneity i…
field incident · 2019
The claim
First experimentally established by Quattrone and Jones (1980), who found Princeton and Rutgers students judged a single rival-school student's choice as more predictive of his whole university than they did for their own school.
Reading the scatter. ★ chits are real-world incidents (vivid, but n = 1, so they sit low on strength). The cluster is split: the best-powered field evidence lands upper-right, while several lab paradigms sit lower and weaker. Read the construct paradigm by paradigm.
Exhibit index
Quattrone & Jones (1980) — rival universities. Participants generalized one out-group member's choice to the whole out-group more than they did for an in-group member, treating the out-group as more internally uniform.
George A. Quattrone, Edward E. Jones, 1980 · University students (Princeton/Rutgers); multiple studies in the paper
Park & Rothbart (1982) — sororities. Participants judged their own sorority's members as more dissimilar/variable than out-group raters judged them, and rated out-group sororities as more internally similar — the out-group homogeneity pattern.
Bernadette Park, Myron Rothbart, 1982 · n = 90 (about 30 per sorority)
Boldry, Gaertner & Quinn (2007) — meta-analysis of measures. A small but reliable overall tendency to perceive more variability in the in-group than the out-group; the effect was stronger for real (non-minimal) than for minimal/experimentally-created groups; different measures produced systematically different — sometimes opposite — patterns, with some measures yielding in-group homogeneity.
Jennifer G. Boldry, Lowell Gaertner, Jeff Quinn, 2007 · 177 effect sizes from 173 independent samples; 12,078 participants
Mullen & Hu (1989) — meta-analysis on group size. Majority-group members showed relative out-group homogeneity, whereas minority-group members tended to show the reverse — relative in-group homogeneity — indicating the effect is moderated by, and can reverse with, group size/status.
Brian Mullen, Li-tze Hu, 1989 · meta-analytic (multiple studies)
Why the shortcut forms
Three main accounts compete. (1) Differential familiarity / exemplar storage: people encounter far more in-group than out-group members, so memory holds more distinct in-group exemplars, making the in-group look more varied (Linville, Fischer & Salovey, 1989). (2) Differential representation: in-groups are stored as both individual exemplars and abstract prototypes, while out-groups are stored mainly as prototypes (Judd & Park, 1988). (3) Self-categorization / intergroup-context accounts: homogeneity perceptions are context-dependent comparative judgments, not fixed memory artifacts (Haslam, Oakes et al.), which explains why context manipulations and minority status can eliminate or reverse the effect.
The familiarity and dual-representation accounts struggle to explain that the effect is context-sensitive, can be made to disappear, reverses for minority-group members (Mullen & Hu, 1989), and does not appear reliably for some categories such as gender. This is the core mechanism debate: a stable cognitive bias vs. a flexible, comparison-driven judgment.
Replication
Mixed
The basic direction replicates across many paradigms and is meta-analytically reliable, but the effect is small and highly conditional. Boldry, Gaertner & Quinn (2007) showed its appearance depends on the measure used, with some measures producing the opposite (in-group homogeneity) pattern; Mullen & Hu (1989) showed it reverses for minority-group members; and it is unreliable for certain categories such as gender. It is best described as a real but context-dependent and moderator-laden effect rather than a uniformly robust law.
The seminal Quattrone & Jones (1980) and Park & Rothbart (1982) studies predate standardized effect-size reporting, so their magnitudes are described qualitatively here ("not reported"). The granular per-measure d-values circulating in secondary summaries (e.g., d≈0.45 for dispersion vs. d≈0.20 for range) appeared in a search-engine synthesis and could not be verified against the Boldry et al. (2007) full text (publisher returned 403), so only the verified characterization "small but reliable" with 177 effect sizes / 12,078 participants is asserted. The Mullen & Hu (1989) reversal and the meta-analytic counts were confirmed via secondary/encyclopedic sources rather than the original PDFs. The most important honesty point: this is a real but small and heavily moderated effect that reverses for minority groups and depends on the measurement instrument — not a universal law.
What to log instead
Force individuation and seek disconfirming variety: deliberately list several specific, distinct out-group members or sub-segments before making any judgment, and increase genuine contact/familiarity with the out-group, since the effect shrinks or reverses with greater exposure (the minority-group reversal in Mullen & Hu, 1989) and is sensitive to how variability is assessed.
Mullen & Hu (1989) found minority members, who have more exposure to the majority out-group, show reduced or reversed out-group homogeneity; Boldry et al. (2007) showed the effect is measure-dependent, implying judgments can be reframed.
Catch it in the act
Watch for "they're all the same" reasoning — generalizing from one member of an out-group (a rival company, a foreign market, the opposing political party, another profession) to the entire group, while readily acknowledging your own group's internal diversity. The asymmetry (rich variety for "us," a single type for "them") is the tell.
Adjacent lockers
- Cross-race effect (own-race bias / other-race effect)mechanistically-linkedThe difficulty recognizing and individuating out-race faces is often treated as a perceptual manifestation of out-group homogeneity.
- In-group homogeneity effectoppositeThe reverse pattern, reliably found among minority-group members (Mullen & Hu, 1989) and with certain measures.
- In-group favoritism / in-group biassiblingA distinct intergroup bias about evaluation/liking rather than perceived variability; the two often co-occur but are conceptually separate.
- Law of small numbersparentQuattrone & Jones framed out-group homogeneity as an application of Tversky & Kahneman's tendency to over-generalize from small samples.
- Ultimate attribution erroreasily-confusedConcerns attributing out-group behavior to dispositional causes, not perceived within-group similarity.