This study analyzes the effect of ethnicity, classified as African-American and Anglo-European, on diagnosis, symptomatology, and response to inpatient treatment in a sample of geriatric patients of a research unit of a tertiary care facility. Variables that might confound the effects of ethnicity, such as age, gender, education, and associated diagnoses, were statistically controlled. Psychotic diagnoses were significantly associated with ethnicity, being higher in African-Americans, weakening the claim that such an association stems from the confounding effects of social class. Variability in the ratings on the Brief Psychiatric Rating Scale and Hamilton Depression Scale was significantly associated with ethnicity in several instances. African-Americans appeared to obtain comparatively higher therapeutic benefits from hospitalization. Results are discussed using knowledge of the field of cultural and social psychiatry.