Psychometric properties and latent classes of the Attachment Style Questionnaire in mothers and their offspring

Background: The Attachment Style Questionnaire (ASQ) is used to assess an individual’s inner working model of interpersonal relationships and attachment style. The study evaluated the factor structure, sex invariance, and convergent validity of the ASQ among middle-aged women and their young adult offspring. Methods: Middle aged mothers (n = 3,605, M age =46.5 years) and their 3,712 offspring were administered the ASQ when their offspring were young adults ( M age =21.6-years). Symptom measures of anxiety and depression were concurrently administered to mothers (Delusions Symptoms States Inventory) and offspring (Young Adult Self-Report). Confirmatory factor analyses were conducted to test the structural validity in both groups as well as sex invariance in the offspring cohort. Convergent and divergent validity were assessed using depression/anxiety subscales. Latent class analyses (LCA) were performed to examine patterns of responses across ASQ items. Results: Confirmatory analyses supported the utility of a shortened 33-item ASQ for both cohorts, and the structure was invariant between male and female offspring. Strong convergence was found between anxious subscales on the ASQ and depression/anxiety symptoms in both cohorts. Measures of anxiety and depression correlated positively with ASQ anxiety and avoidance subscales. LCA identified four latent classes across a continuum of secure and insecure attachment rather than four distinct attachment styles (i.e., Bartholomew & Horowitz, 1991). Conclusions: The ASQ-33 is a reliable and valid instrument that measures internal working models of relationships in both male and female young adults, and women in middle-age. Findings support a dimensional view of attachment security/insecurity over the four-part categorisation of attachment styles.