An experiment was conducted to investigate factors that may affect people's tendency to underestimate situational determinants of behaviours when forming impressions of self and others. Subjects were randomly designated as questioners or contestants in a quiz game, which had been used to demonstrate a ‘questioner superiority effect’ as a result of situational underestimation (Ross, Amabile&Steinmetz, 1977). While the questioners were composing challenging questions for use in the quiz game, contestants either prepared questions of the same sort as a ‘warm‐up’ exercise, or did something unrelated to the game. After the quiz one‐third of the contestants were told that the questions used were actually not composed by the questioner but were supplied by the experimenter. Another third were given false feedback about themselves and their partner on a written general knowledge test. The rest were not given any information. Results showed that the question preparation task did not reduce the questioner superiority effect. It was drastically undercut, however, when false feedbac