Interview codings of attachment style:using profile analysis to understand the patterns involved.

Abstract

Attachment style is frequently discussed in terms of profiles of early childhood risk factors. Those using attachment interview methods use their ratings of these risk factors in developing an attachment style rating. In spite of this, profile analysis has yet to be used to model specific attachment styles. By using a multiple regression profile analysis to model attachment style in terms of coder's ratings of early risk factors, we can test empirically whether individual elements are relevant and how. The study of attachments began with Freud in the middle of the last century. Since that time attachment style has been studied first by John Bowlby and Mary Ainsworth, and since that time by many others. Early views of attachment, including the identification of specific attachment styles, and the investigation of its stability are discussed, as well as the limitations of the existing research. Specifically, the paper addresses the need for additional research to support or refute the theoretical models of attachment structure. Many methods have been developed to assess attachment style, most of which are closely tied to one particular theoretical view of attachment structure. Because the data for this paper are drawn from a study which utilized a four-prototype model of attachment as assessed for a coded semi-structured interview, the best way to understand the resultant codes is through a profile analysis. By using a two-step multiple regression profile analysis procedure, we can assess the unique contributions of both the level of risk, and the pattern of risk factors. The multiple regression methodology has the additional benefit of allowing for both continuous predictors and criterion variables; something that is not possible with other profile analysis methodologies. This allowed me to run the regressions with both dichotomous and semi-continuous criterion variables which enable the detection of different patterns. The results indicate that both the profile patterns and level can predict the criteria. The pattern component however is significantly more predictive of the criteria. While the derived patterns differ from the predicted patterns, they remain consistent with theory. Overall, environmental risk factors such as abuse, neglect, and parental rejection were not predictive of attachment style or score, while individual risk factors such as anger at parents and rebellion, and the interactive factor of role reversal were highly predictive. This leads us to conclude that profiles are a viable method of understanding attachment styles, and that it is the individual's responses to the risk factors present in the childhood environment rather than those factors themselves which determine attachment style.