A group of yound adults completed the Attributional Style Questionnaire and measures of depressive affect and hopelessness on two occasions separated by a time interval of three years. Attributional style was demonstrated to be relatively stable over this time period, and was correlated with the measures of psychological well-being. Specifically, those who scored highest on depressive affect and hopelessness attributed good outcomes more externally and less stably, and attributed bad outcomes to relatively more stable and global causes. In contrast to the hopelessness model of depression, however, multiple regression analyses showed that depressive attributions were not antecedent to increased psychological distress; nor were they a consequence. Furthermore, negative life-events did not contribute to the prediction. It was concluded that the data are most consonant with Brewin's (1985) symptom model, in which depressive attributions are a concomitant or symptom of depression but have no causal impact on the onset or course of the disorder.