The dynamics of measuring attachment.

Shaver and Mikulincer’s review makes a strong case that self-reports of adult attachment are associated with attachment-related processes that are unconscious or at least that occur automatically and outside of awareness. Therefore, the common criticism that self-reports cannot meaningfully assess attachment orientations in a manner that is reective of dynamic processes is clearly false. The belief that interview measures are the only valid way to assess adult attachment is based in fundamental assumptions about what can and cannot be assessed in self-reports, and what needs to be assessed in determining adult attachment orientation. The AAI approach to assessing adult attachment is based on the assumption that assessment of defensive strategies, which operate at least partially outside of conscious awareness, is a necessary component in the evaluation of ‘states of mind’ with respect to attachment. Moreover, it is assumed that these strategies can be accurately scored (given appropriate training of coders) from the transcripts of interviews in which participants discuss the relationships they had with their parents in childhood. The attachment strategies identiŽ ed in the AAI are assumed to re ect processes that operate outside the interview setting, especially in attachment-relevant contexts such as parenting. By deŽnition, individuals with particular attachment strategies deny some types of psychological experiences and/or distort their responses to questions tapping these experiences. AAI advocates would therefore tend to conclude that it makes little sense to ask individuals with insecure attachment patterns direct questions about processes that are assumed to be defensively distorted and not open to conscious access. In contrast, they would argue that interview procedures are the most appropriate way to assess adult attachment because defensive strategies can be assessed and provide important clues to underlying attachment issues. Thus, it is common in the attachment literature to see comments such as: ‘Because we are interested in differences in processing attachment-related thoughts and feelings, we deal only with Ž ndings linking attachment states of mind [as assessed by the AAI] with psychopathology’ (Dozier, Stovall, & Albus, 1999, p. 498). How can self-reports be useful if this is true? One possibility is that self-reports about attachment are valid assessments of dynamic processes. This seems unlikely since these reports are not designed to assess such processes. Another possibility, suggested by Hesse (1999), is that states of mind with respect to early attachment experiences are not necessarily available to conscious awareness (therefore requiring interview measures to assess), but attachment orientations in the romantic domain are (therefore allowing for self-report assessments). However, we cannot think of any good reasons why the process of recalling and reporting on childhood experiences would be qualitatively different from the process of recalling and reporting on Department of Psychology, Simon Fraser University, Burnaby, BC Canada.