OBJECTIVE: It is frequently reported that patients with psychotic disorders have poor insight into their illness. Previous research has suggested that poor insight may have considerable power in predicting the long-term course of chronic mental disorders and an impact on patients' compliance with treatment plans. The authors, proposing that insight is best viewed as a multidimensional phenomenon, developed the Scale to Assess Unawareness of Mental Disorder, which samples discrete and global aspects of insight across a variety of manifestations of illness. This article reports on a reliability and validity study of the scale. METHOD: The study subjects were 43 patients with schizophrenia and schizoaffective disorder. Various aspects of insight into illness were evaluated with the scale. In addition, ratings of psychopathology, course of illness, and compliance with treatment were made. RESULTS: Item variability was high and normally distributed, supporting the authors' contention that insight can be rated on a continuous rather than dichotomous scale. Results of the analyses examining the relations between the various dimensions of insight assessed and the psychopathology, course, and compliance variables were generally as hypothesized. Convergent validity with other global measures of insight was found, and aspects of poor insight were correlated with poorer compliance and course of illness. Examination of the interrelations among the four insight subscales revealed that these subscales sample independent phenomena. CONCLUSIONS: The Scale to Assess Unawareness of Mental Disorder has good reliability and validity and has certain advantages over previous measures of insight, suggesting the usefulness of a multidimensional view of this complex concept.