Background: The authors sought to evaluate the associations between depressive symptoms and social support in a sample drawn from a relatively understudied population — depressed elderly patients. The present study also used a multi-measure approach to assess both depressive symptomatology and social support. Methods: In this prospective study of 115 patients we examined: (1) the baseline relations among a self-report measure of depressive symptoms, two clinical assessments of depressive symptoms, and subjective and received social support, and (2) the ability of social support to predict changes in clinical assessments of depressive symptoms at 6 months and 1 year. Education level, financial concerns, activities of daily living ratings, and gender were controlled for. Results: Baseline subjective support was negatively related to self-reports of depressive symptoms, but unrelated to clinical assessments at baseline or follow-up. Conversely, received support was unrelated to self-reported depressive symptoms, but positively related to both clinical assessments at baseline. However, higher ratings of received support at baseline predicted decreases in clinical ratings of depressive symptoms at 6 months and 1 year. Limitations: These data were gathered in a primarily Caucasian sample, thus the findings may not generalize to more diverse ethnic populations. Potential confounding due to treatment mode and setting was not controlled in the present analyses. Conclusions: These results have important implications for interpreting clinical data in elderly depressed patients. Specifically, when depressive symptoms are assessed using clinician ratings, the most informative aspect of social support with respect to future clinical status appears to be received, rather than perceived, support measures.