Prejudiced Liberals?

Ninety white female freshmen very much in favor of integration played a Prisoner's Dilemma game in which information about the cooperativeness or competitiveness of the other player (significant, p <.04), the race of other player (p < .09), the native region of the United States of the subject (nonsignificant), and experience over three trial blocks of ten trials each (nonsignificant) were manipu lated. Subanalyses in the cooperative information condition showed cooperation was significantly less with the Negro other player than it was with the white other player (p <.02). Attitude and personality measures did not correlate significantly with total cooperation in the game, but high cooperators had low anxiety scores. The Negro other player was rated consistently lower than the white other player, especially on sociometric and ability traits. Further analyses of attitude and personality measures, ratings, and submeasures of cooperation showed the subjects were reacting differently to the Negro and white other players; consequently, an unexpected high rate of cooperation with the competitively described Negro other player, which resulted in a weak main effect for race, was itself interpreted as evidence of prejudiced behavior.