Twenty-eight moderately depressed outpatients were randomly assigned to 12 weeks of cognitive therapy (N = 14) or pharmacotherapy (N = 14). Significant changes in mood, cognitive processes, and content were similar to those found in previous studies demonstrating effectiveness of cognitive therapy. Patients treated with medication, however, demonstrated nearly identical change on all measures, including cognitive measures, despite the absence of direct focus on cognitive activity. Further analyses disclosed that cognitive change may be an important feature of overall clinical improvement, as patients whose conditions did not improve (regardless of treatment modality) showed significantly less change on cognitive measures. These findings suggest that cognitive change may be more accurately seen as a part of improvement rather than the primary cause of improvement. This suggests a more complex conceptualization of the role of cognitions in the change secured by cognitive therapy.