Assessing attachment: convergent validity of the adult attachment interview and the parental bonding instrument

Objective: The aim of this study was to determine whether or not the Parental Bonding Instrument (PBI) can provide information about parent–child attachment that is comparable to information obtained from the Adult Attachment Interview (AAI), a more complex measure of attachment. Method: One hundred and thirty emotionally and/or behaviourally disturbed adolescents (73 male, 57 female; ages 13–19 years, x = 15.3 ± 1.47 years) participating in a study of attachment and suicidality completed the PBI and the AAI. Data from these measures were compared within participants. Results: Maternal care and overprotection on the PBI differed significantly by AAI attachment classification (F3,122 = 2.79, p = 0.012), with autonomous participants showing the most optimal and unresolved participants the least optimal PBI results. Maternal love and maternal involvement/role reversal on the AAI were significant predictors of maternal care and maternal overprotection, respectively, on the PBI (R2 = 0.15; R2 = 0.16). These predictions improved when AAI scales measuring idealisation and involving anger towards the mother were included in the regression analyses (R2 = 0.35; R2 = 0.20). Autonomous participants on AAI showed the highest scale correlations across instruments. Conclusions: Attachment information obtained from the PBI and the AAI is comparable in participants with optimal attachment histories, but not in participants showing idealisation or anger towards their mothers. Caution is, therefore, advisable when using the PBI to obtain attachment information in clinical samples where suboptimal attachment histories are likely.