The present paper argues that a need will remain for data which have been collected retrospectively. Recent developments in oral history and sociology are described to suggest a method of collecting retrospective data which may minimize recall bias. Pilot work is reported on the adaptation of the method for a study of chronic respiratory disease. The results of the pilot study proved promising. The technical properties of the measures appeared adequate. The results of the substantive analysis were consistent with existing knowledge and went beyond existing knowledge to suggest new areas of research. Ways of further validating the method are identified and its wider application discussed.