Objective - The aim of this study was to verify the construct validity of the Italian version of Parental Bonding Instrument (PBI) a questionnaire which estimates the parental style as reported by the son or daughter. Method - The questionnaire was administered to a group of 102 students (62 males and 40 females) attending University of L'Aquila and to a sample of 128 patients (76 males and 52 females) consecutively admitted to a psychiatric unit for an index episode. We compared the means of the two factors (care, protection) separately for each parent in the two groups using a t-test for indipendent samples. After having estimated the internal consistency of items of each scale by calculating Cronbach's coefficient α, a factor analysis was performed for students and patients to find the structural factors of the questionnaire; then, we conducted a confirmatory factor analysis of the PBI items, for the students only, to evaluate the fit of the real items to models proposed in the literature. Results - The Italian version of the Parental Bonding Instrument, demonstrated the ability to discriminate between patients and controls; it showed an high internal consistency. The factor analysis identified a two factors solution which accounted for 44.6% and 44.3% of the variance of the mother's and father's PBI scores respectively for the group of students and it identified two factors which accounted for 49.3% and 46.6% of the variance in the group of patients. Conclusions - The psychiatric patients showed a low «care»-high «protection » confirming an association between the «affectionless control» pattern and psychiatric disorders.