Knowledge About Individuals’ Interracial Friendships Is Systematically Associated With Mental Representations of Race, Traits, and Group Solidarity

The paper, in plain language

Painting faces from friendship lists

With Ivuoma Onyeador and John Dovidio, I used reverse correlation — hundreds of noisy-image choices averaged into a portrait of what participants imagine — to test whether knowledge of a person's interracial friendships reshapes their mental representation. Black and White Americans generated images of targets described as having mostly Black, racially mixed, or mostly White friends.

Racial assimilation in the mind's eye

The effect was symmetric and large: Black targets with White friends were pictured with lighter skin and more Eurocentric features; White targets with Black friends, darker and more Afrocentric — and independent raters confirmed it in the features, not just the shading. Friendships also wrote group politics onto faces: same-race-friend targets were judged likelier to support their own group's causes. The sharpest result is the defector penalty — White participants' image of a White person with mostly Black friends was rated markedly less trustworthy, less warm and more threatening.

Why it matters

Intergroup-contact research celebrates cross-race friendship as prejudice's remedy; these data show the same friendships carry a social price tag levied by one's own group, which helps explain why segregated networks are so stable. The caveats: US Black–White relations only, deliberately extreme friendship descriptions, and composite images that cannot yet say which facial ingredient does the work. The method links this paper to my reverse-correlation work on phenotypic prototypicality and dehumanized faces.