The objective of this:study was to assess how children's racial evalua
tions were affected by talking about these with a friend whose level o
f prejudice was different from their own. We compared the kinds of eva
luative statements and explanatory strategies used by the high- and lo
w-prejudice partners of a dyad, as well as the change in attitude that
followed from the discussion. White children from the third and fourt
h grades were identified as above or below the median for their class
on the Multi-response Racial Attitude (MRA) measure, which assessed po
sitive and negative evaluations of White, Black, and Chinese children.
They were then paired with a friend who differed in level of prejudic
e, and asked to discuss one positive and one negative item from the at
titude measure. Instructions were to Salk about how the three races sh
ould be evaluated and why. After the discussion, each child was privat
ely reassessed on the MRA. Analyses of the discussion variables:indica
ted that low-prejudice children stated significantly more negative eva
luations and examples of Whites and more cross-race similarity than hi
gh-prejudice children. High-prejudice children became significantly le
ss prejudiced in their evaluations after the discussion. Changes were
greater in children whose low-prejudice partner made more statements a
bout cross-race similarity (e.g. ''everyone can be mean sometime''), a
long with more positive Black and negative White evaluations, Low-prej
udice partners remained unprejudiced. Dyadic discussion may be a usefu
l context for reducing prejudice.