A psychometric and theoretical investigation of the measurement of attachment relationships in adolescence

Abstract

Adolescence is marked by rapid changes in biological, cognitive, psychological and social domains. This thesis examines the nature and importance of close dyadic bonds in adolescence. Using an attachment theory perspective and focussing on those relationships which provide physical and psychological security, the composition of these attachment relationships will be examined. Despite the well-documented importance of attachment relationships for psychological adjustment in adolescence, there remains a paucity of theoretically informed and validated measurement tools for this age group. Controversy over the self-report assessment of attachment abounds with several recent reviews of measurement instruments for adults and children. However, there is a lacuna with regard to assessing attachment in adolescence. Theoretical and methodological considerations regarding the measurement of adolescent attachment are discussed in this thesis. It is argued that theoretical and empirical developments in the area of adolescent attachment are limited by the dearth of measurement tools specifically developed for this age group. The key objectives of this thesis are as follows: i) to review existing measures of attachment in adolescence in order to organise and clarify the body of literature and delineate the necessary conditions for a new measure of adolescent attachment relationships; ii) to examine the psychometric properties of the new measure of adolescent attachment relationships in order to provide initial content and construct validation; iii) to examine the relationship between the new measure of adolescent attachment relationships and a number of existing measures of attachment relationships in order to provide convergent and discriminant validation; iv) to further validate the new measure of adolescent attachment relationships via multi-method validation; v) to examine the relationship between the new measure of adolescent attachment relationships and a range of measures of psychological health in order to provide discriminant validation and information regarding clinical utility. In addressing the first objective, results of a systematic literature review are reported. Following this, the most commonly used measures of attachment for adolescents are reviewed and critiqued. Several necessary conditions for a psychometrically sound, theoretically coherent measure of adolescent attachment are formulated. A pool of items was developed based on existing scales and theoretical conceptualisations of attachment. The Domains of Adolescent Attachment Scale (DAAS) includes four independent sections: general attachment orientation; and attachment to mother, father and best friend. The scale was administered as a self-report composite questionnaire to two samples of Australian high-school aged participants (N=720). Studies One and Two investigated the psychometric characteristics and factor structure of each section of the DAAS. Findings provide support for the preliminary validity and reliability of the DAAS. The latent structure of the DAAS sections demonstrates the uniqueness of each attachment bond. Study Three validated the DAAS using a categorical measure of attachment style, the Relationships Questionnaire; and two measures of the attachment network, the Attachment Networks and Functions Questionnaire and the Bull's eye hierarchical mapping technique. Results of this study indicate that the DAAS has good convergent validity with the RQ and demonstrates the ability to discriminate between individuals of different attachment styles. Scores on the DAAS relate effectively to attachment network characteristics including network size, strength of attachment and quality of attachment. Study Four presents a multi-informant validation design utilising parent reports of their adolescent's attachment style as well as the parent's attachment style and psychological health. Analysis of the relationship between the DAAS sections and measures of psychological adjustment are undertaken in Studies Five and Six. The DAAS is thus suggested to have potential for the measurement of adolescent attachment relationships in a range of contexts. The current research provides evidence for the distinctiveness of general and specific attachment relationships. Adolescence is a unique developmental period where the individual is faced with developmental, educational, and relational challenges. Research needs to employ refined and sensitive instruments that offer the opportunity to assess individual differences in attachment relationships, in all their complexity, in order for us to understand their true significance.