Does getting to know several people of another race reduce bias?
Chinese preschoolers who learned the names of and moved in sync with four different Black adults lost their implicit anti-Black bias, while children who did the same with only one Black adult did not.
Source
Individuating multiple (not one) persons reduces implicit racial bias
Study at a glance
- Design
- Human experiment — Two-arm randomized pretest-posttest experiment: a 2-minute video-based synchronous movement game with four named Black instructors (differentiation) versus one named Black instructor (no-differentiation), with no same-race control arm
- N
- N=61 · 66 children were randomized (33 per condition); 5 were excluded from the implicit-bias analysis by speed/accuracy criteria, leaving 61
- Population
- Han Chinese preschool children aged about 4-6 years (mean age 5.90 years) tested at school in China
- Outcome
- Implicit racial bias (D score on a child-friendly Implicit Racial Bias Test) and explicit bias (own-race choices across four everyday scenarios), before and after the manipulation
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
66 Han Chinese children aged around 5-6 were randomly assigned to one of two versions of a 2-minute video game in which they copied simple movements such as clapping and jumping jacks. In the differentiation condition each movement was led by a different named Black man, and children repeatedly recalled all four names; in the no-differentiation condition one named Black man led every movement. Implicit bias was measured before and after with a tablet reaction-time test pairing Chinese or Black faces with smiles or frowns, and explicit bias with forced choices between a Chinese and a Black doctor, teacher, coach or guide.
What they found
At pretest the children showed strong implicit anti-Black bias. Only the differentiation group's implicit bias fell, from a D score of about 0.55 to 0.06 (no longer different from zero), whereas the one-instructor group's change was not significant, giving a significant condition-by-time interaction. Explicit own-race preference dropped in both groups (the proportion of own-race choices fell from 0.79 to 0.60 overall), and at posttest the four-instructor group chose own-race people at chance (a proportion of 0.52) while the one-instructor group still showed a preference (0.68).
The limits
What it doesn't show
Bias was measured only immediately after a 2-minute session, so the study says nothing about how long the effect lasts or whether it changes behaviour toward real people. There was no same-race control group here (the authors rely on earlier studies), so general practice or retest effects are only partly ruled out, and the explicit bias drop in both groups could partly reflect that. Contact was via video with Black men only, children were one age group in one culture, and the sample was modest; the reported statistics also contain some inconsistencies (for example group sizes and one repeated t value), so exact numbers should be read with caution.
Key terms
- Implicit racial bias
- Automatic, less consciously accessible associations between a racial group and good or bad, measured through reaction times rather than self-report.
- Individuation
- Treating members of a group as distinct individuals, for example by learning to tell their faces and names apart.
- Perceptual-Social Linkage hypothesis
- The idea that poor ability to tell other-race faces apart leads children to lump the group together and apply attitudes to the whole category.
- Outgroup homogeneity effect
- The tendency to see members of other groups as more similar to one another than members of one's own group.
- D score
- The standard IAT bias index: the difference in reaction times between incongruent and congruent pairings divided by their pooled standard deviation.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
What was the key manipulated variable in this experiment?
Common questions
Why compare four instructors with one instead of with no instructors?
Both conditions gave children names and a friendly synchronous activity with Black people; the only difference was whether they had to tell several individuals apart. This isolates differentiation as the active ingredient.
How can you measure implicit bias in 5-year-olds?
With a simplified IAT on a tablet: children tap a smile or frown for faces, and are slower when the pairing (for example Black face with smile) conflicts with their automatic associations.
Did the one-instructor condition do nothing?
It did not significantly reduce implicit bias, but explicit own-race preference did fall in that group too, just less than in the four-instructor group.
More on Prejudice & bias