Attachment and the Self at Age Six

There is wide-spread agreement among psychologists that the self develops not in a vacuum but rather through social interaction, particularly interaction with the child's primary caregivers. Bowlby has emphasized the attachment component of the child-parent relationships as being of central importance. This study investigated the relation between quality of attachment to mother and self-esteem in six-year-olds.

Subjects were 52 white, middle-class, six-year-olds, and all but two were from two-parent families (26 boys). Children and their mothers came to the laboratory for two 1 1/4 hour sessions, one month apart. Quality of attachment was assessed in each session based on the child's behavior on reunion with mother following a separation of approximately one hour using a procedure devised by Main and Cassidy (1985). Reunion behavior was first rated on a security-insecurity scale and on an avoidance scale, and was then classified as indicating one of four patterns of child-mother attachment: secure, insecure/avoidant, insecure/ambivalent, and insecure/controlling. Self-esteem was assessed in four ways: with a puppet interview, a series of incomplete doll stories, the global self-esteem subscale from Harter's Scale of Perceived Competence, and a measure devised for this study that correlates the rank ordering of what a child thinks "it's most important for kids to be good at" with the rank ordering of "what he's best at." A fifth measure, Harter's Scale of Perceived Competence and Social Acceptance for Young Children was also administered. The self as assessed in one session was analyzed in connection with attachment of the other session.

Quality of attachment, the puppet interview, the ranked domains assessment, and the one doll story that was repeated were all stable across sessions. The puppet interview, the assessment using the incomplete doll stories, and Harter's global self-esteem subscale were all somewhat related to attachment, as were three of four subscales on Harter's Scale for Young Children. In addition, there was a tendency for particular patterns of responses from the puppet interview and the incomplete doll stories assessment to relate to particular patterns of attachment. The measure that correlates the two rank orders was not related to attachment.

The connections between attachment and self-esteem found in this study were moderate, and there was much of the variance that could not be explained. Methodological issues that may relate to this pattern of results are discussed. Additional areas of possible influence on the child's self-esteem are considered.