This study examines the psychometric properties of two new abbreviated versions of standard measures of depression, a revised eight-item Center for Epidemiological Studies--Depression Scale (CES-D) and a short-form Composite International Diagnostic Interview (short-form CIDI). A sample of 6,133 elders, age 70 years or older, completed both measures as part of the Asset and Health Dynamics Study of the Oldest Old. The revised CES-D had an internal consistency and factor structure comparable to that of prior versions of the CES-D. The sources of discordance between the two measures were examined and the two measures were compared on self-report of four clinical variables: medical illness, physician diagnosis, psychiatric treatment, and antidepressant or tranquilizer use. Both measures were associated with self-report of physician diagnosis and psychiatric treatment. Respondents positive for depression on the CES-D reported higher rates of antidepressant use. Respondents positive on the short-form CIDI only did not report more antidepressant use than nondepressed respondents.