Twenty‐one clients were assigned to three conditions in a four‐month trial of once per week psychodynamic therapy. Clients in an imagery condition worked with subpersonalities elicited by psychosynthesis guided visual imagery. Clients in a non‐imagery condition worked in a similar way, but without guided imagery. The third condition was a minimal contact control group in which clients had all assessments (at baseline, mid‐point, outcome and follow‐up) but no active therapy. Mean outcomes in imagery and non‐imagery groups were significantly better than for the minimal contact group on target aims of therapy, and on self‐acceptance, but there were no significant differences between the imagery and non‐imagery groups. Repertory grids enabled the trajectory of each individual in therapy to be tracked in a conceptual space in which subpersonalities were related to people in the subjects' real lives. A case history of an imagery client who improved during the course of therapy is discussed to show these relationships and the utility of a combined nomothetic and idiographic research design. 1990 The British Psychological Society
