The Balloon Analogue Risk Task (BART) is a computerized decision-making task that provides a test of behavioral risk taking. The task is increasingly used in laboratory settings and in the field with young adults and adolescents. However, there are currently no published data about the test-retest characteristics of the task when it is administered on separate days. The current paper addresses this gap. Risky behavior on the BART (adjusted average pumps) showed acceptable test-retest reliability across days (r = +.77, p < .001). The data indicate that risk behavior on the BART has adequate test-retest stability and therefore performance on the task on a single occasion is likely to be representative of an individual's performance on other occasions.