Convergence versus Divergence of Social and Developmental Measures of Adult Attachment: Testing Jay Belsky’s Proposals

The current study assessed the empirical relationship between two social and developmental attachment measures in a sample of female adults, half of whom were diagnosed with BPD. Following Belsky (2002)’s conceptualization of the possible relationship between these two attachment traditions, the current study assessed two mutually exclusive propositions regarding the Adult attachment interview (AAI) and the experiences in close relationships (ECR) questionnaire. First, it is possible the AAI and ECR assess the same mental representations of attachment, but empirical correspondence does not emerge unless accommodations for method variance are made. Or second, AAI and ECR are not related to each other directly, but converge on another attachment-related construct.

Correspondence analyses suggested a significant positive correlation between AAI preoccupied/unresolved and ECR anxiety dimensions, but not between AAI dismissing and ECR avoidance or AAI 4-way classifications and ECR cluster-based categories. This partially supports Belsky’s first proposition and suggests the AAI and ECR are not assessing entirely the same attachment representations but do assess one important aspect. Convergence analyses found both ECR anxiety and the AAI preoccupied/unresolved dimension were significantly correlated with borderline features. Because they were significantly related to each other, too, this did not fully support Belsky’s second proposition of independent contributions to an attachment related construct. However, the correlation was moderate and the ECR provided more variance in borderline features than did the AAI, suggesting partial support for Belsky’s second proposition

As a whole findings suggest limited support for Belsky (2002)’s first proposition that if methodological differences are removed, there is some association between the two sets of measures. Moreover, there is some suggestion that AAI and ECR’s provide some independent contribution to borderline features, in support for Belsky (2002)’s second proposition Future comparison research would benefit from using research paradigms that are adaptive and assess social and developmental attachment in a range of contexts using a variety of methodologies. This could enhance our understanding of how these traditions relate and identify key points of convergence and divergence.