Public ability attributions of public task performance were investigated as a finetion of test performance (high or low), task performance (high or low), and the availability of information about test performance to the audience (audience informed or audience not informed). The results were largely in agreement with self-presentation expectations, Ability attributions concerning public task performance addressed to an audience which was not in formed about test results, were found to be strongly self-serving, although they were also influenced by consistency of test and task performance. In ability attributions directed at an in formed audience, self-serving biases were completely absent, the attributions being in accordance with the consistency or inconsistency of test and task performances.