A theory of the structure of perturbations in observer ratings and questionnaire data in personality research

Oliver Wendell Holmes once outlined the six personalities involved in a dialogue between two people, here John and Thomas: three Johns the real John, John's ideal John, and Thomas's ideal John; and correspondingly, three Thomases. The psychologist may be concerned with the real John, but he has access only to Thomas's ideal John (observer‐reports) and John's ideal John (self‐report). How can he extract John's true behavior from the reports of his behavior? Is there any systematic way in which he can account for the distortions which must arise from internal and external influences on the perception of the observer? The present paper attempts to answer these questions by outlining a theory of perturbations and describing an experimental design for weighing and evaluating the perturbing influences which affect personality data.