The impaired performance of Alzheimer's disease (AD) patients on the clock drawing test (CDT) relative to age-matched normal controls is a well-documented finding in the literature. On the other hand, there is sparse information regarding the use of this test in schizophrenia. We examined three groups of subjects matched for gender and education: institutionalized patients with schizophrenia (n = 32), patients with AD (n = 32), and normal controls (n = 36). The CDT ("free-drawn", "pre-drawn", and three "examiner" conditions) and Mini-Mental State Examination (MMSE) were administered to all participants. In patients with schizophrenia, symptom severity was assessed with the Brief Psychiatric Rating Scale (BPRS). Patients with schizophrenia were significantly younger than AD patients and normal controls (56.78 versus 71.41 and 66.25, respectively), and normal controls had significantly higher MMSE scores than patients with schizophrenia and AD (27.58 versus 20.75 and 18.44, respectively). In all of the clock conditions, the two patient groups performed significantly poorer than the normal controls, with the exception of the "pre-drawn" clock in which AD patients also performed worse than patients with schizophrenia. Age and duration of illness did not correlate significantly with CDT scores. When MMSE scores were used as a covariate, all significant differences on the CDT among the three groups disappeared, with the exception of the "pre-drawn" clock (AD patients had lower scores than both control and schizophrenia groups). In patients with schizophrenia, scores on the BPRS were not related with any CDT variable. Institutionalized patients with schizophrenia and AD patients showed similar deficits on a neuropsychological test sensitive to changes in visual-analytic function, attention, receptive language, and executive functions such as planning, organization, and simultaneous processing.