Women with breast cancer are thought to be vulnerable to depression for reasons associated with impact of diagnosis, treatment, and metabolic/endocrine changes. While the literature shows that most of these women do not become clinically depressed, 15% to 30% report elevated depressive symptoms that may be clinically important. The purpose was to identify and determine the relative importance of predictors of depressive symptoms in women treated for early-stage breast cancer. A total of 2,595 women (< or = 4 years following completion of initial treatment for early-stage breast cancer) provided data on cancer-related variables, personal characteristics, health behaviors, physical functioning/symptoms, and psychosocial variables. Participants were divided into high or low depressive groups using the Center for Epidemiologic Studies Depression Scale screening form. Results of the binary logistic regression analysis were significant (overall R2 = 32.4%). Before entry of psychosocial variables, younger age, being unmarried, poorer physical functioning, and more vasomotor and gastrointestinal symptoms were significant risk factors for elevated depressive symptoms (R2 = 16.1%), but objective cancer-related variables were not. After inclusion of psychosocial variables in the model (DeltaR2 = 16.3%), none of the preceding variables remained significant. Greater risk for depressive symptoms was associated with stressful life events, less optimism, ambivalence over expressing negative emotions, sleep disturbance, and poorer social functioning. Depressive symptoms in women treated for early-stage breast cancer are not associated with objective cancer-related factors. Rather, they are most strongly linked with many subjective psychosocial variables.