OBJECTIVE This study aimed to evaluate the diagnostic validity of the Beck Depression Inventory (BDI) in the elderly and to suggest an optimal cut-off score in order to screen major depressive disorder. METHODS The BDI and an elderly health questionnaire were administered to 2729 subjects over the age of 60 chosen by stratified random sampling in a Ansan City, South Korea. The BDI and geriatric depression scale (GDS) were examined at about a two-year interval. A reliability and validity test, a factor analysis and an ROC curve analysis were performed. RESULTS Eighty-four subject had depression and 2645 subjects were rated as normal. The BDI showed significant positive internal consistency (r = 0.88) and test-retest reliability (r = 0.60). Convergent validity with GDS was significantly positive (r = 0.59), and an exploratory factor analysis revealed four factors. We suggest a score of 16 as the optimal cut-off point for the BDI when screening for major depression. CONCLUSIONS The results of this study showed that the Korean version of the BDI is appropriate for screening for depression and 16 is the optimal cut-off score for the Korean elderly. Screening of elderly depression with BDI in the community would be valuable when comparing with younger adults and with their former BDI data which were taken when they were young.