This article describes a study of the degree of concordance between assessments made by patients and informants (n = 96) and between assessments made by informants (n = 92) on the basis of the Gerontological Personality disorder Scale (gps). The patientinformant concordance was weak (κ = 0.24) to fair (κ = 0.51) and the concordance between two informants assessing the same patient was moderate (κ = 0.40) to fair (κ = 0.45). The results of the study indicate that if it is better to ask the informant questions that relate solely to observable and interpersonal characteristics. The assessment of intrapersonal characteristics such as mood and selfperception is less important as far as the diagnosis is concerned.