Patients with right and left hemisphere stroke (RH, LH) and a control group (CG) were administered a structured clinical interview (SCID-R) and rating scales for anxiety and depression (Center for Epidemiologic Studies Depression Scale, Beck Anxiety Inventory, and Hamilton Depression Rating Scale). The three groups did not differ when compared for the mean level of distress on the rating scales or frequency of mood or anxiety disorder diagnosis using the SCID-R. By contrast, participants in the LH group were more likely to be classified as distressed using rating scales cutoff scores and on measures of severity when compared only to the RH group. The rating scales were sensitive to psychiatric disorders, but lacked specificity in all groups. There were significant correlations between scales suggesting the rating scales measure a common factor in the controls, while in the stroke groups associations were weaker and less likely to be significant. Together, these results suggest the need for caution in using rating scales of depression and anxiety in neurologic patients and support the notion that these scales are sensitive to distress rather than specific for identifying depressive and anxiety disorders.