Behavioral Consequences of Attributional Bias

Abstract

Two experiments were conducted to examine the behavioral consequences of the “fundamental attribution error.” In the first experiment, subjects viewed a general-knowledge quiz, shown by Ross, Amabile, and Stemmetz (1977) to produce biased attributions of expertise: Questioners are perceived as more knowledgeable than contestants. Subjects were assigned to a partner, either the questioner or the contestant in the quiz. As predicted, (1) subjects tended to conform more to questioners than to contestants, and (2) subjects predicted for themselves a lower score on a general-knowledge test when a score of 23 correct out of 30 was ascribed to questioners than when the same score was ascribed to contestants. In the second experiment, subjects interviewed a target they believed was a questioner from a quiz game. Interviewers who were moderately confident about their attributions of knowledgeablity were more likely to elicit behavior from the target that was consistent with the attribution than were interviewers who were highly confident. These results are analyzed here in terms of hypothesis-testing behavior.