Intergroup bias and correspondence bias: People engage in situational correction when it suits them

This study tests the hypothesis that the correspondence bias is reduced when the correspondent inference of a behaviour runs counter to the goal of forming a more favourable impression of in-group than of out-group members. Participants were leaders and subordinates in an organization and read favourable or unfavourable behaviour descriptions of either a leader or a subordinate. The descriptions were accompanied by information indicating either that the behaviour was produced by an underlying disposition of the actor or by situational causes. Supporting the hypothesis, unfavourable behaviours with situational causes produced moderate inferences only when the actor was an in-group member (indicating that the situational causes were taken into account); conversely, when the actor was an out-group member, favourable behaviours with situational causes produced moderate inferences.