In a case-control study, two potential interpersonal risk factors of non-endogenous depression, namely a patient's perception of their current intimate partner as dysfunctional and a patient's recall of exposure to previous deprivational parenting, were quantified. The interpersonal characteristics of the partner were assessed principally by a brief self-report questionnaire, the Intimate Bond Measure (IBM). By cross-sectional and longitudinal comparison of this instrument with other interview-derived and self-report measures, the convergent, discriminant and predictive validity of the IBM in depressed patients was established. Further, little evidence of any distorting effect of depressed mood or neuroticism was detected. The perception of the current intimate partner as dysfunctional imparted a risk to non-endogenous depression of at least five times, while reported exposure to parental 'affectionless control' was quantified as a four times' risk. Importantly, IBM care scores predicted the course of the depressive disorder over a six-month period.