This paper describes the development, validation and use of a computerized assessment for minor psychiatric disorder based on the Clinical Interview Schedule (CIS; Goldberg et al. 1970). There was good agreement between the computerized assessment and the CIS administered by psychiatrists, both in assessing overall severity and in defining 'cases' of psychiatric disorder. Individual symptoms elicited by the computer and the CIS were compared, and the levels of agreement found were similar to those from inter-observer studies of standardized interviews. Subjects from a variety of non-psychiatric settings regarded the assessment as acceptable, accurate and easy to use. It is concluded that this computerized assessment of neurotic symptoms is valid and reliable. It eliminates observer bias, it is an efficient use of research resources and it may have clinical applications in primary care.