The Psychiatric Assessment Schedule for Adults with Developmental Disability (PAS-ADD) is a semi-structured interview for use with respondents who have learning disability and for key informants. This report investigates the ability of the instrument to detect symptoms that had been found to exist during routine clinical assessment of the patients. Field trials involved 95 referred patients with learning disability and a key informant for each sample member. Clinical opinions of the referring psychiatrists were sought using a symptom checklist. Referrer checklist symptoms and PAS-ADD data were both factor analysed. Validity testing involved (a) computation of correlations between PAS-ADD factors and checklist data and (b) comparison of PAS-ADD and referrers' diagnoses. Results indicated good validity for the PAS-ADD in relation to psychotic symptoms and depressive symptoms. Anxiety symptom identification was not well validated, probably due to small numbers. Expansive mood identified by the referrers was not detected by the PAS-ADD because there is currently no corresponding section in the interview. Where the PAS-ADD produced a diagnosis (in 58 members of the sample), 44 were in agreement with the referrer. Probability of diagnosis by PAS-ADD increased with the number of relevant active symptoms identified by the referrer. The PAS-ADD has been shown in a previous report to have the sensitivity to detect mental disorders not known to psychiatric services. For psychotic and depressive conditions, our results showed that symptom detection was in good agreement with the information provided by the referring psychiatrists on their patients. The PAS-ADD needs a section on hypomania and further investigation of its detection of anxiety disorders.