Some have proposed that all psychotherapies for depression, as well as for other common mental disorders, are equally efficacious and that they all work through common processes, especially a relationship with a therapist. One reason it may be difficult to discern differential efficacy among treatments, as well as how they exert their effects, is that depression and other disorders are heterogeneous in both presentation and prognosis. The studies presented in the dissertation aimed to explore how heterogeneity in depression may moderate treatment effects and process-outcome relations. In study 1, a prognostic index (PI) was developed and treatment differences along the PI were explored in a sample of patients (N = 622) randomized to treatment as usual (TAU) or stepped care starting with brief therapy (BT) or with cognitive-behavioral therapy (CBT). The PI comprised five variables: unemployment status, depression severity, hostility, sleep problems, and lower positive emotionality, all of which predicted a lower likelihood of recovery. For patients whose PI indicated a high likelihood of recovery (73% of the sample), recovery rates were similarly high across the treatments. Among patients whose PI indicated a lower likelihood of recovery, patients in the CBT condition experienced a substantially higher recovery rate (65%) than patients in TAU (40%) or BT (44%). In study 2, variability in the predictive relationship between the therapeutic alliance and depressive symptom change was explored in a sample of patients receiving cognitive therapy (CT) for depression (N = 60). The alliance predicted outcome in the subgroup of clients with 0–2 prior episodes (r = .52), but not in those with 3 or more prior episodes (r = -.02). In study 3, these findings were replicated in an independent sample of patients receiving CBT for depression, but they did not extend to patients in a psychodynamic therapy condition. Taken together, these findings suggest that there may be identifiable subgroups of patients for whom factors common to all treatments will promote symptom change. By contrast, complementary subgroups, such as those with poorer prognoses or more recurrent histories of depression, may reveal differences in the efficacies of treatments and their active mechanisms.