This introductory paper discusses three issues: theory, validity of measurement and benefits of assessment. It is argued that the conceptual entity 'quality of life' remains ill-defined, although researchers commonly agree on its multi-dimensional nature. The selectivity of dimensions under study is illustrated by referring to recent work on the quality of life in cancer and in cardiovascular disease. Despite conceptual weaknesses, promising developments are apparent in research methodology. The issue is discussed with special reference to the convergent and discriminant validity of some widely used research methods. Improvements in therapy and patient-physician relationships are considered among the major benefits of assessing the quality of life in clinical practice. In conclusion, this field of research, if based on appropriate theory, opens important opportunities for cross-fertilization between advanced medicine and the social sciences.